International Journal On Multicultural Societies (IJMS)

International Journal
on Multicultural Societies
(IJMS)
Vol. 4, No. 2, 2002
“Protecting Endangered
Minority Languages:
Sociolinguistic Perspectives”
International Journal on Multicultural Societies (IJMS)
Vol. 4, No. 2, 2002
The Impact of Government Policies on Territorially Based

Ethnic or Nationalist Movements
DIRECTOR OF PUBLICATION: Paul de Guchteneire
EDITOR-IN-CHIEF: Matthias Koenig
GUEST EDITOR: Eda Derhemi
MATTHIAS KOENIG, Editorial 148
EDA DERHEMI, “Thematic Introduction: Protecting Endangered Minority Languages: 150

Sociolinguistic Perspectives”
SALIKOKO MUFWENE, “Colonisation, Globalisation, and the Future of Languages in 162

the Twenty-First Century”
SUZANNE ROMAINE, “The Impact of Language Policy on Endangered 194

Languages”
RAJESHWARI PANDHARIPANDE, “Minority Matters: Issues in Minority Languages 213

in India”
PIETER MUYSKEN, “Uchumataqu: Research in Progress on the Bolivian Altiplano” 235
EDA DERHEMI, “The Endangered Arbresh Language and the Importance of 248
Standardised Writing for its Survival: The Case of Piana degli Albanesi”
JOSHUA FISHMAN, “Endangered Minority Languages: Prospects for Sociolinguistic Research” 270
Editorial
MATTHIAS KOENIG
University of Marburg
he present issue of UNESCO’s International Journal on Multicultural

T Societies takes up a previously established thematic thread by addressing
questions of linguistic diversity and language policy. Here, our focus is on
problems of endangered minority languages and their protection through public
policy-making. Whereas previous issues have investigated the legal framework of
language policies at international, regional, and national level (see Vol. 3, No. 1
and Vol. 3, No. 2), the current issue, edited by Eda Derhemi, approaches the topic
from a sociolinguistic perspective. It attempts to analyse the causes, circumstances
and results of language endangerment as well as the social conditions and effects of
political intervention in favour of the survival of endangered minority languages.
As noted in the Thematic Introduction, the debate over the preservation of

endangered minority languages has gained momentum in recent years. With
international organisations and policy-makers being engaged in recording
endangered languages and promoting the linguistic rights of persons belonging to
linguistic minorities in various regions, sociolinguists have not only renewed their
interest in problems of language loss and language maintenance but, further, have
situated their own work within an increasingly politicised discourse. This has led to
new controversies over the desirability and the actual impact of language policies
designed to preserve or even to revitalise endangered languages. What is ultimately
at stake in theses controversies, is the professional role of sociolinguistics or, more
precisely, the linkages between sociolinguistic research, the interests of minority
speech communities, and the dynamics of public policy-making.
The contributions to this issue take different positions vis-à-vis these questions and
thereby provide an overview over the current state of discussion, both in theoretical
and empirical sociolinguistics. At the same time, they collectively emphasise the
importance of sound scientific knowledge about the linguistic characteristics of
endangered languages, about the social conditions of the respective speech
communities, and about the intended and unintended consequences of political
intervention for devising viable language policies. Joshua Fishman, commenting on
the articles collected in this issue, underlines this point when he calls for more
systematic meta-analyses of case studies on languages in demise, so as to arrive at
complex and parsimonious theories that would ultimately lead to more precise
evaluations of language policies.
International Journal on Multicultural Societies (IJMS), Vol. 4, No.2, 2002: 148 - 149
ISSN 1817-4574, www.unesco.org/shs/ijms/vol4/issue2/ed © UNESCO
149 Matthias Koenig
The thematic thread of linguistic diversity will be pursued in a subsequent issue

which addresses patterns of language use on the Internet. It will include discussion
of relevant theoretical debates and an evaluation of current policies promoting
multilingualism on the Internet. Furthermore, it will report on new original
research carried out in the framework of UNESCO's B@bel project. Other issues in
preparation will establish another new theme of the Journal, by addressing patterns
of international migration, links between social science research and migration
policies in various countries, and the relation of migration and integration policies
in industrial countries. As always, readers are invited to contribute to this debate by
sending their comments to the mailing list attached to this Journal.
Protecting Endangered Minority Languages:
Sociolinguistic Perspectives -
Thematic Introduction
EDA DERHEMI
University of Illinois at Urbana-Champaign
1. Main Foci of this Issue

This issue is devoted to problems of endangered languages, particularly
endangered languages spoken by minorities, focusing on the sociolinguistic study
of the causes, circumstances and results of endangerment, and other structural and
social processes related to endangered languages and to their survival. The papers
analyse aspects of the loss of minority languages, either indigenous or migrant, in
the presence of a dominant second language. They investigate the relationship of
these languages to previous and current policies, emphasising the positive or
negative impact of institutional intervention for the survival of endangered
minority languages (EMLs). As shown by the different authors, EMLs are often
stigmatised and communities of speakers are often marginalised and dominated
populations. Matters of power and control of the speech communities are also
discussed, as their role is important in accounting for language endangerment.
This issue therefore focuses first on the characteristics of EMLs, on their level of
attrition and the specific characteristics of attrition from a sociolinguistic
perspective. A second focus is the evaluation and assessment of the possibilities for
linguistic preservation and revitalisation of EMLs. Taking as a starting point the
empirical analysis of the languages discussed, the authors attempt to draw some
theoretical conclusions about the chances for survival of EMLs in general, and
about the specific circumstances that would facilitate their maintenance. A third
focus is the empirical effects of language policies and institutional action on EMLs
and on communities of EML speakers. The papers also discuss whether
institutional intervention is necessary and the importance of preliminary
sociolinguistic research for an effective language policy.
The authors call attention to the causal relations between the characteristics of
languages and speech communities, on the one hand, and language policies and
other institutional action, on the other. They examine how policies and the process
of implementing them recognise or neglect the needs of the speech communities
and the state of the languages, and how they contribute, in turn, to changes that
occur in the language and in the linguistic behaviour of its speakers. In dealing
International Journal on Multicultural Societies (IJMS), Vol. 4, No. 2, 2002: 150 - 161
ISSN 1817-4574, www.unesco.org/shs/ijms/vol4/issue2/art1 © UNESCO
151 Eda Derhemi
with language policies affecting EMLs, the speech communities involved are given
particular emphasis; their efforts to preserve their native languages are seen as an
active factor in the existence and implementation of language policies and for the
fate of EMLs.
2. Relevant Sociolinguistic Research on First-Language Attrition

An endangered language1 is a language that may soon vanish, ceasing to be used as
a vehicle of communication, perhaps even disappearing completely from human
history. As the contributions to this issue show, an endangered language is not
necessarily a minority language, and not every minority language is necessarily
endangered. But there is a high probability that with time a neglected minority
language will become endangered.
At a time when it is estimated that “80% of the world’s 6,000 or so living

languages will die within the next century” (Crystal 1997, 17), when language
endangerment is increasingly seen as a topic that primarily concerns linguists
(Nettle and Romaine 2000, 23), and with the increase in the last decade of the
number of international organisations formed to record endangered languages
(Crystal 1997, 18) and to regulate and promote the linguistic rights of language
communities, it is natural that there is a great interest in the field by sociolinguists,
policy-makers and analysts.
Research on language endangerment shows that the social status and prestige of
EMLs, one of the most important forces in the process of attrition and
maintenance, depend on a complex set of economic and cultural factors, reflecting
the power relations among the communities of speakers involved. In order to assess
symbolic indicators of dominance and control, it is important to investigate the
functional aspects of the language and its use in different registers and domains.
The tendencies for change in functional roles in bilingual and diglossic situations
are very important in the process of reversing language shift and revitalisation of
an endangered language. The linguistic attitudes of the community members are
also an important parameter. These attitudes are historical and cultural construc-
tions and relate directly to the prestige of EMLs. The sociocultural and ethnic
context, as well as sociolinguistic indicators of language use and attitudes of the
speakers, have been at the centre of research on endangered languages in the last
twenty years.
There are important landmarks in the field of endangered languages that embody
the directions of research and methodological approaches outlined above. Lambert
and Freed (1982) present a volume on endangered languages without separating
sociolinguistic from psycholinguistic and educational approaches and language
acquisition from loss. This work is important for its theoretical contributions,
especially the attempt to come up with methods of measuring maintenance and
1. Grinevald Craig (1997) presents a general discussion of language endangerment.

Thematic Introduction 152
loss. Dorian (1989) assembles a large number of sociolinguistic contributions,

mainly case studies on endangered languages, and she clearly divides the two foci
of the publication: the social context of the endangered languages on the one hand,
and the structural studies on grammar on the other. Robins and Uhlenbeck (1991)
edit a collection of articles about endangered languages from all the continents of
the world, not only in order to study the causes of language death, but also to call
international attention to this phenomenon. Brenzinger (1992) publishes a number
of case studies and surveys on African endangered languages, introduced by four
articles pertaining to theoretical discussions on language death. Grenoble and
Whaley (1998), besides the study of the mechanisms of language loss that have
been traditionally investigated in language attrition studies, focus attention on the
community response and the importance of linguistic diversity for human society.
Besides communicative functions and other sociolinguistic parameters of EMLs,

structural indicators such as lexical and grammatical loss and marked grammatical
asymmetries are also important factors in language attrition. Sociolinguists have
shown a constant interest in the formal changes shown by endangered languages.
Some of the main publications on attrition and death mentioned above have
reserved a separate chapter for contributions on structural grammatical changes and
loss. There are even a few monographs focused on the grammatical system of the
endangered languages. Sociolinguists agree that structural degeneration and loss
are caused by changes in the communicative domains of a language undergoing
decay: “structural restrictions in grammar have been convincingly correlated with
reduction in speech genres” (Tsitsipis, in Dorian 1989, 117). But, by itself,
structural disintegration or structural recovery is an important factor that influences
the use of the language in the community and the attitude of the speakers towards
their language.
Thus research on linguistic endangerment is characterised by a holistic approach

that integrates multiple methodologies and directions of inquiry. Sasse (1992)
maintains that an approach that reduces the study of endangered languages to one
separate domain is “unrealistic and counterintuitive”. Serious studies on attrition
should include the analysis of external settings, such as the history and ethnicity of
the community, cultural and religious features, economic status, sociopolitical and
constitutional structures. They should also include sociolinguistic analysis of
speech behaviour related to the communicative functions pertaining to diglossic or
just bilingual patterns, domains of use, prestige of the endangered language, and
structural phenomena such as erosion of linguistic categories and forms, deviation
and other changes of internal systems of endangered languages. The pivotal
longitudinal work of Dorian (1981) on the Scottish Gaelic dialects and on language
death in general is a good example of this approach. The interest in processes of
linguistic endangerment shown by sociolinguists, ethnolinguists, psycholinguists,
anthropological linguists, language policy analysts, scholars in education and even
in fields not directly related to linguistics, are an indicator of the multifaceted
problems involved in the study of language endangerment. Crystal, in his
introduction to Language Death, admits that the issue of language endangerment
153 Eda Derhemi
“is now so challenging in its unprecedented enormity that we need all hands –
scholars, journalists, politicians, fundraisers, artists, actors …” (2000, ix).
Analysis of language policy is another important trend of sociolinguistic research

on endangered languages. Each of the aspects of sociolinguistic studies analysed
above should be seriously investigated, evaluated and reflected in the process of
language planning pertaining to endangered minority languages. In return, the
impact and results of language planning should be analysed from the perspective of
the changes that they have stimulated in the sociolinguistic levels mentioned above.
The relationship between sociolinguistic research and institutional and community
action to encourage processes of maintenance and revitalisation has not been a
central concern of sociolinguists working on linguistic attrition. Language policies
have not been analysed in relation to the functional and structural characteristics of
endangered languages. 2 An exception to this is the two volumes by Fishman,
published in 1991 and 2001, in which efforts to reverse language shift have been
analysed from a theoretical perspective and on the basis of a large number of
longitudinal case studies. On the other hand, language policies have been
thoroughly questioned and evaluated in relation to educational settings (Skutnabb-
Kangas 2000) and linguistic culture (Schiffman 1996).
Concerning language policies and threatened languages, a “role dilemma” is

always faced by sociolinguists: Are they researchers who objectively study
endangered languages as neutral analytic machines, or do they also get involved
and see their responsibilities as rescuers? Often the position of the researcher is
made clear in the title of the book. It is harder to understand the role of the
researcher in a volume entitled Endangered Languages, but it is much easier in
titles such as Vanishing Voices, Reversing Language Shift, or Linguistic Genocide
in Education. Nettle and Romaine (2000) dedicate a whole volume very rich in
facts and examples from all over the world to the importance of language
maintenance and the factors that threaten languages today. Their work strongly
criticises the lack of care and support for endangered languages and the devastating
effects of such indifference, intentional or not, on linguistic diversity. The authors
consider it “a strategic error that will be regretted as time goes on” (14).
Skutnabb-Kangas (2000, 249), being herself a sociolinguist who supports

assistance and revitalisation processes, makes a thorough presentation of the
arguments for and against the preservation of linguistic diversity. The main
argument for language maintenance in all sociolinguistic works is that culture and
language “stand for each other” (Fishman 1991, 22) and the loss of language is in
fact a permanent loss of human culture and knowledge. The weakness of this
argument is very well described by Fishman himself (1991, 15): “The ‘whole truth’
about the relationship between language and culture may be too complex and too
subtle, as well as too subjective and self-fulfilling to be fully told.” On the other
hand, efforts to maintain endangered languages are considered by some politicians
2 Most research of this kind consists of single papers in volumes dealing with general phenomena of
endangerment.
and even researchers as “a primordialist dream, creating employment for the

world’s linguists” (Skutnabb-Kangas 2000, 249). Their main argument is that the
preservation of endangered languages is very costly and even if one decides to
spend what is necessary, the result is not cost-effective. Fishman describes the
proponents of this group as “reductionists”: their ‘realism’ “reduces human values,
emotions, loyalties and philosophies to little more than hard cash and brute force”
(Fishman 1991, 19). In Can Threatened Languages Be Saved (2001), Fishman
maintains his position in favour of linguistic salvation of endangered languages but
with less optimism. More than anything else, he stresses the complexity and
difficulties of this “struggle”, and argues that it is very important to construct
strategies to address the functional diversification of the languages involved,
strategies that are very focused and very carefully chosen.
The dynamics of this field of research are defined by the need to document
endangered languages, and the need for new case studies to consolidate theoretical
findings on endangerment. In 1992 Sasse stated, “there are huge geographical areas
for which no comprehensive studies are available” and “theoretical or model-
establishing approaches are scarce” (9). This is still true ten years later. There is
still a need for the participation of sociolinguists in longitudinal research and
projects connected to single communities for the study and revitalisation of
endangered languages, when it is desirable and feasible.
3. Theoretical Studies Presented in this Issue

The first three papers, by Salikoko Mufwene, Suzanne Romaine and Rajeshwari
Pandharipande, deal with theoretical questions concerning language policies in
cases of endangered minority languages. The contributions constitute parts of a
whole picture, ranging from a more general view towards a more detailed focus.
Mufwene discusses whether language policy is necessary in cases of endangered
languages; Romaine assesses past policies around the world and analyses their
problems, focusing on ways to raise their effectiveness; and Pandharipande links
the factors of endangerment and low status of minority languages in India with the
way language policy is designed and implemented in that country.
Mufwene attempts to answer the question: Is it possible and necessary to interfere

with the natural development of languages through linguistic policies and change
the direction of language shift in order to stop language death? He recognises
sociohistorical factors that influence, accelerate and slow down language change,
but supports the natural path a language takes – endangered or not – whatever the
final result turns out to be. He argues that changes occurring in the linguistic
behaviour of speakers are simply an adaptation to changes in the socioeconomic
conditions of the speech communities, driven by interests related to costs and
profits that come along with language use, and therefore they must be respected.
The underlying message of Mufwene’s paper is to let languages naturally compete,
even eliminate each other, as they have done for thousands of years. Mufwene is
sceptical about the wisdom and even the possibility of deliberate efforts to prevent
155 Eda Derhemi
language shift. In arguing for this, he takes a historical and comparative approach,
comparing languages and situations in the past and the present, their development
and eventual death. Language vitality has historically been affected by contacts of
languages and of populations, changes in power, and other sociopolitical factors.
According to Mufwene, the process of globalisation today does not bring any
special new element to the competition among languages. Therefore, in general, it
does not constitute a threat for the small and less powerful languages, nor a reason
for institutional intervention through language policies.
Mufwene also presents a review and analysis of the terminology and approaches in
recent literature with respect to language change and colonialism, as well as
language change and globalisation. As he notes in his analysis of relevant
sociolinguistic literature, his view is not shared by many sociolinguistics scholars
today. Mufwene sees language endangerment as an engine that produces a natural
diversity, while most sociolinguists see it as a death machine that needs to be
stopped. Although the author reinforces that lack or presence of practice are the
real source of attrition or revival of a language (see also Romaine and Fishman in
this issue), he does not account for the fact that language policy, when realistic and
based in community will, can change the direction of language use and practice.
Mufwene’s paper, like others in this issue, emphasises the need for the description
of endangered languages and for the recording of chunks of discourse from these
languages before it is too late.
Romaine’s paper is a response to some of the issues raised by Mufwene. Unlike

Mufwene, Romaine opposes the idea of free and natural competition among
languages because a “no language policy is, in reality, an anti-minority-languages
policy”. Along with this position, she analyses the impact of language policies,
addressing two main questions: Why do languages policies often fail?; In what
ways can they be more effective? Romaine bases her analysis of language policies,
their implementation and their impact on minority endangered languages, on her
conviction that endangered languages must be saved while it is still possible and
still desired by their speakers, and that it is worth doing whatever it takes to
preserve this cultural heritage. She emphatically supports and has confidence in the
effectiveness of institutional action, but underlines how difficult it is to have a
realistic and effective linguistic policy, and how complex the road is from the
approval of a legal statement to its realisation in languages.
Romaine brings a large number of examples of problematic understanding and

impact of language policies concerning EML in Canada, the United States, New
Zealand, Nigeria, Senegal, Norway, Spain, Greece, Ireland, Turkey and Australia,
among others. In the analysis of these cases, she focuses on the factors that prevent
a linguistic policy from reaching the endangered language or the community that
speaks it. She draws attention to two main factors. First, what makes a policy
ineffective is the lack of congruence between what the policy states and the actual
sociolinguistic conditions of the community. The outcome of this is the absence of
realistic planning, which isolates the policy in itself and destroys the purpose of its
existence. Second, language policies are often available only when it is too late and
the languages are practically extinct. The cultural and economic context and the
attitudes of speakers towards their languages are the main variables that must be
studied and considered in the design and implementation of any policy for
language maintenance.
Romaine is aware of the fact that “effective language policies will and must affect
all aspects of national life and will have to be sustained for decades, if not forever”.
In fact, the main problem that scholars less enthusiastic about language mainten-
ance see in the institutional efforts to save endangered languages is that the costs of
the maintenance and revitalisation can be great. But this issue is not often
addressed by sociolinguists.
In her paper, Pandharipande analyses the main sociolinguistic factors that

determine the high or low status of minority languages in India. She examines the
low social status of minority languages in relation to linguistic endangerment and
offers an explanatory basis for both phenomena. The main focus among the
variables that determine and measure social status and endangerment is on the
concept of “functional load”. She uses the term “functional load of a language” as a
synonym for “language dominance” and “language power” in the social domains of
language use. She explains that she prefers this term as a diagnostic tool, because it
allows a quantification of the concept of “power” of a language. She uses the
concept of functional load to explain the position of a language in the power
hierarchy among languages, therefore the low and endangered status of a minority
language is explained by its low functional load or low ability to function in the
public sphere. Functional load in this sense also accounts for the fact that a
minority language is not necessarily endangered, and for the fact that one minority
language can be well maintained while another is undergoing attrition. A vertical
axis of power or functional load from the upper levels of technology and state
languages, to the lower levels of tribal languages, symbolises the scale in which the
status of any language in a linguistic repertoire can be measured and compared.
Other factors, such as number of speakers, linguistic attitude, and usage in different
functional domains are also considered as important variables in the process of
language attrition and linguistic salvation, but in order for them to have an impact
on the language, the functional load of this language must change first. In other
words, the status of EMLs is connected indirectly to all the factors mentioned
above, but directly only to the functional load of the languages.
Pandharipande emphasises that these sociolinguistic factors must be included in the

basis of any language policy dealing with EMLs. Although Indian states have
approved a central language policy intended to promote the development of
indigenous vernaculars, its implementation has often not been successful.
Sometimes the output of such policies, instead of raising the social status and
prestige of a minority language, has lowered it even further
157 Eda Derhemi
As Pandharipande’s paper deals with the complex situation of EMLs in India,

where the current norm of linguistic behaviour in any community is
multilingualism, it helps to clear up some of the confusion surrounding the
meaning and definitions of minority and majority Indian languages. In the same
spirit as the other authors of this issue, she underlines the lack of linkage between
the higher (constitutional) levels of language policy and the way the policy is
transformed into action to stop or change language shift. In general, a well-
intentioned language policy cannot succeed if the sociolinguistic situation in a
speech community and other idiosyncrasies of the languages are ignored.
4. Empirical Studies
The two case studies of this issue reflect recent fieldwork conducted in Bolivia on
the endangered language Uchumataqu, by Pieter Muysken, and in Sicily on the
endangered Arbresh, by Eda Derhemi. Both papers focus on the characteristics of
the languages, from a functional and structural linguistic perspective, in order to
assess the viability of these EMLs and the possibilities for language shift and
revitalisation through effective language policies. The issues discussed in the
previous papers about the need for realistic policies based on a detailed and
complex description of the sociolinguistic situation of the endangered languages
are analysed in the specific settings of Uchumataqu and Arbresh.
Muysken discusses Uchumataqu, the language of the Uchuma in Bolivia, from an

anthropological, ethnographic and sociolinguistic perspective, demonstrating how
the prestige and state of health of this language and other languages of the Uru
people have changed over time. He analyses Uchumataqu in relation to other
languages used by the same community and by others nearby, such as Aymara,
Quechua and Spanish. Linguistic data on functional restriction and other changes
Uchumataqu has undergone, influenced by the dominant languages of the region,
indicate its state of attrition. Together with the historic, ethnic and cultural features
of the speech community, and its economic and political status, Muysken integrates
the linguistic loyalties of the speakers and the phonological and morphological
analysis of linguistic data collected by the very few semi-speakers that are left.
Muysken recognises the need for documenting Uchumataqu before it is too late, as
an important source of information on the early linguistic history of Bolivia and the
whole continent. But he does not give up hope for a linguistic revitalisation of
Uchumataqu, based first of all on political changes in the community that have
greatly improved the linguistic attitudes of the speakers towards their language.
Although he stresses the capability of rural communities in “language planning”,
he calls for a complex and difficult range of economic and political changes that
are necessary for the planning process to be successful.
Unlike Uchumataqu, which has lost most of its speakers and has no fluent speakers
today, the Arbresh of Piana degli Albanesi, Sicily, is in a less-advanced stage of
endangerment. Derhemi presents sociolinguistic arguments from the functional
domains of Arbresh and from structural data showing linguistic corrosion that
demonstrate a clear language shift from Arbresh to Italian and the state of
endangerment of Arbresh. She focuses on the sociolinguistic analysis of those
features of language attrition that specifically show the importance today of a
normative written form of the language, arguing that if this codified written
language spreads among young speakers through the schools, it will have a strong
corrective effect on the aberrant uses of forms and on the unusual free linguistic
variation that is characteristic of the Arbresh speech community at present.
Considering that Arbresh is still used at different degrees and levels of competence
as an informal means of communication, and considering that linguistic loyalty
towards Arbresh is relatively high, there is a strong possibility that linguistic
policies aiming for wider use of the language can still succeed. But the paper
underlines that, in any process of language planning in Piana, linguistic
codification and the use of the written form in schools are the main factors stopping
the process of linguistic disintegration of Arbresh. Derhemi analyses recent
implementation of linguistic policy in Piana, its effectiveness at every step, and the
degree of involvement of the community elite and grassroots in such efforts. She
also presents some prescriptive observations on how efforts for revitalisation of the
language could be more successful, based on the specific sociolinguistic conditions
of Piana degli Albanesi and its linguistic repertoire.
5. Fishman’s Criticism and Some Conclusions for This Issue

In the final paper in this issue, Joshua Fishman presents some critical opinions on
problems raised by the other authors and offers a few suggestions on future
directions for research on endangered languages. As in recent works discussed in
this thematic introduction, Fishman again stresses the need for new case studies in
this field, not partial ones, but detailed and complete research that put together
language as social behaviour, as well as other social phenomena related to
language in a community. Language is not a variable separable from culture and
other forms of social behaviour. Once a considerable amount of such case studies is
available, a collective methodological and conceptual analysis that aims for further
scientific parsimony and clarity is needed.
Fishman also focuses on the importance of language use and language practice in a
situation of language decay. He argues that sociolinguists and other scholars
dealing with EMLs should always keep in mind that language planning is only a
means for change, and as long as the policy does not demonstrate any impact on
language use, it has been unsuccessful. The “unplanned” and “spontaneous” use of
a language by the speech community is the real arena where a language changes,
and the only means of measuring the efficacy of a language policy.
Below I briefly summarise the main conclusions of sociolinguistic research on

endangered languages and the impact of language policies as discussed in this
issue:
159 Eda Derhemi
• Sociolinguists are divided on whether they should press for intervention

and institutional protection in cases of endangered languages or should just
be passive recorders of the dying languages. Many are sceptical about
possible revitalisation but consider it an important contribution to the
maintenance of linguistic and cultural diversity. However, all the
researchers agree that the most decisive factor in the future of these
languages is the will and the attitude of the speech communities. Without
the interest of the speech community in revitalisation, any effort to
promote institutional protection would be egoistic and meaningless. The
sociolinguists also agree on the complexity, immense efforts and high costs
of language revitalisation.
• Language policies will have a stronger impact on EMLs if they are

constructed in a realistic and effective way, on the basis of sufficient
empirical and integrative sociolinguistic studies on endangered languages.
With such studies, the state of endangerment will be easier to control and
the process of analysing language shift and revitalisation more feasible.
Sociolinguistic studies are very important in the first stages of designing a
language policy, but also in the aftermath of the policy, in analysing its
impact and efficacy, and in eventual undoing or redoing corpus planning.
• In order to grasp the nature of the phenomenon of language attrition and to

provide a strong sociolinguistic basis for a sustainable linguistic policy,
sociolinguistic studies have to be holistic in character. Extra-linguistic
phenomena such as the ethnocultural settings of minorities with
endangered languages, the historical, economic and political developments
of these communities, the speakers’ loyalties, attitudes and interests, are as
important as the analysis of language use and language structures. Only
extensive fieldwork by researchers in the speech communities can result in
such studies.
• The main goal of any language policy should be a change that would result
in spontaneous language use by a large community of speakers. This is the
only indicator that can measure the efficacy of a language policy: factors
such as legislative changes, the amount of money spent, the number of
conferences and meetings, and the active participation of the elite are not
indicators of language shift or of the impact of a language policy on
endangered minority languages.
• In the process of research on linguistic endangerment and language policy,

one has to face up to the complexity of the phenomenon of language
attrition and death: not only the many functional, structural and cognitive
aspects of language itself, and the strong influence of extra-linguistic
processes all deeply different from each other, but also the implication of
values, ethical stands, political advantages and disadvantages, social and
civil responsibilities, sentimental positions, even the ability to follow a

dream.
References
BRENZINGER, M., ed. 1992. Language Death. GRENOBLE, L. and L. WHALEY, eds. 1998.
Factual and Theoretical Explorations with Endangered Languages: Current Issues and
Special Reference to East Africa. Berlin/New Future Prospects. Cambridge: Cambridge
York: Mouton de Gruyter. University Press.
COULMAS, F., ed. 1997. The Handbook of GRINEVALD CRAIG, C. 1997. “Language contact
Sociolinguistics. Oxford: Blackwell Pub- and language degeneration”. In: Coulmas,
lishers. ed., op. cit.
CRYSTAL, D. 1997. English as a Global LAMBERT, R. D. and FREED, B. F., eds. 1982.
Language. Cambridge: Cambridge University The Loss of Language Skills. Rowley, Mass.:
Press. Newbury House.
CRYSTAL, D. 2000. Language Death. NETTLE, D. and ROMAINE, S. 2000. Vanishing
Cambridge: Cambridge University Press. Voices: The Extinction of the World’s
Languages. Oxford: Oxford University Press.
DORIAN, N. 1981. Language Death: The Life
and Cycle of a Scottish Gaelic Dialect. ROBINS, R. and E. M.UHLENBECK, eds. 1991.
Philadelphia: University of Pennsylvania Endangered Languages. New York: Berg.
Press.
SASSE, H.-J. 1992. “Theory of language death”.
DORIAN, N., ed. 1989. Investigating In: Brenzinger, ed., op. cit.
Obsolescence. Studies in Language Con-
traction and Death. Cambridge: Cambridge SCHIFFMAN, H. 1996. Linguistic Culture and
University Press. Language Policy. London: Routledge.
FISHMAN, J. A. 1991. Reversing Language SKUTNABB-KANGAS, T. 2000. Linguistic

Shift: Theoretical and Empirical Foundations Genocide in Education or Worldwide
of Assistance to Threatened Languages. Diversity and Human Rights? New Jersey:
Bristol, Pa.: Multilingual Matters. Lawrence Erlbaum Associates.
FISHMAN, J. A., ed. 2001. Can Threatened TSITSIPIS, L. 1989. “Skewed performance and
Languages Be Saved? Reversing Language full performance in language obsolescence:
Shift, Revisited: A 21st Century Perspective. the case of an Albanian variety”. In: Dorian,
Clevedon, UK: Multilingual Matters. ed., op cit.
About the Author

Eda Derhemi is a doctoral student in the Institute of Communications Research at
the University of Illinois at Urbana-Champaign, United States. She did her
undergraduate work in language and literature at the University of Tirana, Albania,
and was a member of the faculty at the same university from 1985 to 1990. From
1990 to 1995 she lived in Sicily, and worked as a journalist for Deutsche Welle. In
1995 she entered the graduate programme in linguistics at the University of
Illinois, receiving her Master’s degree in 1997. In 2000 she worked in Paris as an
intern on UNESCO’s Management of Social Transformations (MOST)
161 Eda Derhemi
Programme. In 2001 she did extensive fieldwork in Sicily on the endangered

Arbresh language. Her interests are in language contact, language change, and
endangered languages and cultures. Recently she has also been working on issues
of Albanian migration after the fall of communism and the cultural integration of
Albanian immigrants.
Address: Institute of Communications Research, University of Illinois at Urbana-

Champaign, United States; email: [email protected]
Colonisation, Globalisation, and the Future of
Languages in the Twenty-first Century*
SALIKOKO S. MUFWENE
University of Chicago
The typical academic discourse on language endangerment has

presented languages as anthropomorphic organisms with lives
independent of their speakers and capable of negotiating on their own
the terms of their coexistence. Not surprisingly it has become
commonplace to read about killer languages in the same vein as
language wars, language murders and linguicides. I argue below that
languages are parasitic species whose vitality depends on the
communicative behaviours of their speakers, who in turn respond
adaptively to changes in their socio-economic ecologies. Language
shift, attrition, endangerment and death are all consequences of these
adaptations. We must develop a better understanding of the ways in
which one ecology differs from another and how these dissimilarities
can account for variation in the vitality of individual languages.
Globalisation is discussed as part of the relevant language ecology. I
submit that only local globalisation has endangered or driven most
languages to extinction.
T his article is a general critique of the literature of the past decade on language
endangerment, including the following recent major works, which are
typically not cited individually here except for peculiarities that warrant singling
out any one of them: Mühlhäusler (1996), Dixon (1997), Brenzinger (1998),
Grenoble and Whaley (1998), Calvet (1998), Crystal (2000), Fishman (2000),
Hagège (2000), Nettle and Romaine (2000), Maffi (2001) and Renard (2001). I
exhort linguists to embed the subject matter in a historical perspective longer than
European colonisation of the past 400 years, to highlight the competition and
*
This article has largely developed from my contribution to a debate with Professor Claude Hagège,
under the title Quel avenir pour les langues?, at UNESCO Headquarters in Paris on 19 September
2001 (part of the series Entretiens sur le XXIe siècle). The original French title was Colonisation,
mondialisation, globalisation et l’avenir des langues au XXIe siècle, from which mondialisation has
now been omitted, for reasons that soon become obvious in the text. The essay has also benefited
from lectures I gave on 7 November and 3 December 2001 at, respectively, the National University
of Singapore and Hong Kong University entitled “Colonization, globalization, and language
endangerment”. I am equally indebted to Michel DeGraff, Claude Hagège, Alison Irvine, Paul
Newman and my anonymous referees for comments on earlier drafts of this publication.
ISSN 1817-4574, www.unesco.org/shs/ijms © UNESCO
163 Salikoko S. Mufwene
selection (Mufwene 2001) that has characterised the coexistence of languages since
probably the beginnings of agriculture (Nettle and Romaine 2000), and thus to shed
better light than hitherto on natural trends of language shift and loss.1 Such an
approach would make the linguistic enterprise comparable to that of
environmentalists concerned with endangered species, who have first sought to
understand the conditions that sustain or affect biodiversity in the same econiche.
I submit that the subject matter of language endangerment will be better understood
if discussed in the broader context of language vitality, with more attention paid to
factors that have favoured particular languages at the expense of others, factors
which lie in the changing socio-economic conditions to which speakers respond
adaptively for their survival. Linguists have typically bemoaned the loss of
ancestral languages and cultures especially among populations colonised by
Europeans, arguing that relevant languages and cultures must be revitalised or
preserved by all means. Missing from the same literature are assessments of the
costs and benefits that the affected populations have derived from language shift in
their particular socio-economic ecologies. Also worth addressing is the question of
what actions, if any, can realistically be taken on the relevant ecologies to prevent
shift from the ancestral languages. I start by articulating the senses of the notions of
“colonisation” and “globalisation” (as in global/globalised economy) that have
figured prominently in the relevant literature, highlighting how they bear on
language vitality.
1. Terminology Matters
Outside population genetics, colonisation conjures up political and economic
domination of one population by another. This form of control is often associated
with military power, which, based on human history, is the means typically used to
effect such domination. This has been made more obvious by the European
colonisation of the world over the past four centuries, at least until the
independence of African and Asian countries in the mid-twentieth century. Often in
alternation with (neo)-colonialism, the term has also been used to describe the
economic relations of less industrialised countries (LICs) with their former colonial
1
The dominant trend in the literature has been to discuss languages as agents with lives somewhat
autonomous from their speakers. This has led to unfortunate titles such as Language Wars (Calvet
1998), which suggest something contrary to the history of language loss. Barring cases of absolute
genocide, languages have typically been endangered or driven to extinction under peaceful
conditions, through an insidious process of assimilation. Wars and political conflicts have fostered
ethnic or national distinctiveness, which has revitalized languages as identity markers. Languages
are also parasitic species whose vitality depends on the communicative behaviours of their speakers.
Although I speak of them as competing with each other in a multilingual community, the notion of
“competition” in this discourse, as in population genetics, means no more than a coexistence set up
in which alternate entities are not equally valued. In the same vein, I also use the notion of
“selection” to refer to the resolution of the competition in favour of one of the alternatives, with the
agency attributed to the “ecology” of the relevant languages. This consists of speakers and the
socio-economic systems in which they evolve. Much of the discussion that follows is framed by
these concepts (for details on this approach, see Mufwene 2001, especially Chapters 1 and 6.).
Colonisation, Globalisation and Languages in the 21st century 164
metropoles, in which the latter have continued to determine the terms and language
of economic exchange. This interpretation of colonisation is present in the current
debate on language endangerment, in which European languages have been
depicted as “killer languages” about to replace all other languages (see for example
Crystal 2000; Nettle and Romaine 2000; Hagège 2000). Thus, power has usually
been invoked as an important factor that has favoured the language of the powerful
over those of the dominated, hence less powerful, populations.
Exceptions to the above observations include volumes such as Brenzinger (1998),

which, by focusing on the competition among languages of the colonised,
oppressed or powerless rural populations of Africa, highlight the fact that the
vitality of a language often depends on factors other than power. They show that if
power has any role to play, basic cost-and-benefit considerations having to do with
what a speaker needs a particular language for, or to what extent a particular
language facilitates survival in a changing socio-economic ecology, determine
what particular languages are given up and doomed to attrition and eventual
extinction. Many African languages have recently lost the competition not to
languages of economic and/or political power but to peers that have guaranteed a
surer economic survival. What such literature shows is that, like the emergence of
new language varieties, language endangerment is one of the outcomes of language
contact and is also subject to patterns of interaction among the populations in
contact.
In order to understand the above view, it helps to also think of colonisation in its
population-genetics interpretation, when a population relocates in a new territory,
regardless of whether the latter is or is not inhabited by an indigenous population.
Thus the eighteenth-century settlement of French colonists on Réunion and
Mauritius, then uninhabited, was as much a form of colonisation as the settlement
of several Caribbean islands by Europeans during the sixteenth to eighteenth
centuries, the establishments of trade forts on the African and Asian coasts in the
same period, or the political and economic domination of several African and
Asian countries from the nineteenth to the mid-twentieth centuries. Bearing in
mind that even the spread of Indo-European populations in Europe involved as
much of settlement colonisation as the domination of North America and Australia
by the English, history tells us that colonisation as understood in population
genetics has assumed many styles involving different patterns of interaction. The
more common, political notion of colonisation rests largely on the more neutral,
population-genetics notion.
From the point of view of language contact, the consequences of colonisation have
not been uniform. Although several languages have died in the process (e.g. Celtic
languages in Western Europe and several Native American languages), new ones
have also emerged (e.g. English out of the contact of Germanic languages among
themselves and with Celtic languages, the Romance languages out of the contact of
Vulgar Latin with continental south-western European Celtic languages, and
today’s pidgins and creoles out of contacts typically of Western European with
non-European languages in some extra-European colonies during the seventeenth

to nineteenth centuries. It is not always the colonised populations that have lost
their languages. Sometimes it is the colonists and colonisers, as in the case of the
Norman French in England, or the Tutsi (formerly speakers of Nilotic languages)
in Rwanda and Burundi, or the Peranakan Chinese in the Strait of Malacca.2 There
are also interesting cases where the old and new languages have coexisted. What is
now interpreted as a threat to the more indigenous language (e.g. Basque vis-à-vis
Spanish) is only a recent development in a long history.
It is thus difficult to produce a general and uniform formula of what happens when
one population colonises another, no more regarding language vitality than
regarding the development of new language varieties. As argued in Mufwene
(2001), the ecology of every case of language contact is somewhat unique. Despite
similarities among them, what happens in one setting is not necessarily replicated
in another. To be sure, we cannot overlook similarities, such as the fact that
language loss has been the most catastrophic in settlement colonies and new
language varieties have emerged additively in trade colonies (i.e. without replacing
some extant languages). On the other hand, we must still note differences from one
colony to another, regardless of whether the members of the relevant subset can all
be identified as plantation or non-plantation settlement colonies, or as trade or
exploitation colonies. Settlement colonies of North America still differ from those
of Latin America, plantation colonies of the Atlantic and Indian oceans were not
quite the same as those of the Pacific, and exploitation colonies of Africa were not
quite the same as those of Asia.
Like colonisation, the terms globalisation in English and mondialisation in French

have figured prominently in the literature on language endangerment. 3
2
Heeding Hoeningswald (1989), I invoke here an often-neglected aspect of language loss especially
among immigrants (invaders, colonists, slavers, or otherwise), who have often lost their languages
while resettling in the new land. This loss, which is partial in that only some of the diaspora
population is affected, is quite relevant, because it is informative about the impact of ecological
changes on the vitality of a language. Just like biological species, language may die in one setting
and yet thrive in another (see also Mufwene 2001, Chapter 6). Their fates are not uniform across
populations of their speakers, especially when the communities are discontinuous (on the model of
what macroecologists identify as metapopulations).
The Peranakans are descendants of male Chinese traders who settled in the Strait of Malacca in the
fifteenth century, married local women, and gave up Chinese while preserving some aspects of their
Chinese cultural background. Their children, who spoke nothing but Baba Malay, are the
Peranakans. (Literally, Baba Malay means Malay of the male Peranakans, based on the fact that
these Chinese men were instrumental in the divergence of this variety from the local varieties.) They
have formed a culturally mixed group distinct from traditional Chinese (who have only reproduced
among themselves) and the local Malay and Javanese populations. Today many of them speak
English as their first language and learn Chinese in school. Their cuisine, characterized as nonya (as
female Peranakans are referred to), reflects local Malay influence. Their communities are to be
found in cities such as Penang, Melaka, Singapore and Jakarta, the original Chinese trade colonies. I
explain the different types of colony below.
3
Rare are books on globalization that bother to define the term and lead the reader to some
understanding of the different ways in which it can be interpreted, depending on context. Yeung
Yue-man (2000) is rather exceptional in providing a discussion that makes it possible for the alert
Globalisation and mondialisation have typically been assumed to be cross-

linguistic equivalents and therefore synonymous. Actually, they do not express the
same meanings. They reflect different perspectives on the present socio-economic
state of the world, which do not bear equally on language vitality. A more adequate
English translation of mondialisation seems to be universalisation, having to do
with worldwide distribution of some institutions such as McDonald’s (hence the
terms McDonaldisation and macdonaldisation in both languages), of cultural
products such as Hollywood movies, American toys and pop music, and the spread
of English in several parts of the world. I show below that this interpretation,
related to the spread of English and other colonial European languages around the
world, does not help to articulate the differential ways in which language shift and
loss have proceeded around the world.
Globalisation need not be universal, as in global warming, or regional, as in global

war (involving several countries but not necessarily the whole world). The most
relevant interpretation on which I wish to capitalise is local (as in global taxation),
meaning “comprehensive” and having to do with interconnectedness of parts of a
complex system, as is more common in local or regional uses of the phrase global
economy in North America or Western Europe. In fact, in the debate on language
vitality, it becomes critical to address the question of whether the worldwide
interpretation of global economy (économie mondiale in French) bears on the life
of a language in the same way as does the local interpretation of the same phrase. I
show below that the phenomena are not the same and therefore do not have the
same linguistic effects. (For an informative discussion of these distinctions, see
Yue-man Yeung 2000.)
Not all countries have developed (significant) local global economies. Not all of
them participate equally in the worldwide global economic system. Although
places like Singapore and Hong Kong depend largely on worldwide globalisation,
many LICs in especially Africa participate only marginally in this networking.
When a particular common language, such as English or French, is required for
communication among the different branches of multinational companies that
foster worldwide globalisation, not all employees of these companies are expected
to be fluent in the lingua franca, especially where most of the labour is involved in
the production of raw materials to be processed outside the country, or a large
proportion of the adult population is unemployed and thus seriously disfranchised
from the economic system. In such places, the vast majority of the populations
continue to function in their ancestral or other local vernaculars, which they in fact
adopt as their identity marker to distinguish themselves from the affluent minority.
To my knowledge, Caribbean territories reflect some of the earliest experiences of

loss of ancestral languages by the enslaved Africans and by the Arawakans and
Caribs in European settlement colonies since the sixteenth century. In most of
reader to identify the wide range of interpretations of the term globalization. Another author coming
close to this is Friedman (1999).
them, the creole vernaculars that later replaced these languages (through shifts to
European colonial vernaculars) have become identity markers for the present mass
of disfranchised proletarians who function only in the local and low sectors of their
economies. 4 They stand in contrast with the acrolectal varieties spoken by
minorities of the more affluent members of their societies. Creole speakers have
either resisted shifting to the acrolects, or have seldom faced opportunities and real
pressure to do so, despite a long history of stigmatisation of their own vernaculars.
Things are not necessarily so different in economically more affluent former

colonies where English or other Western European languages appear to play an
important function and have been claimed to endanger the indigenous languages.
For example, as much as the participation of Hong Kong, Malaysia, Singapore and
Taiwan in such multinational production networks depends on usage of English as
a worldwide lingua franca, the proportion of employees that must be fluent in it is
quite small. The English used by many not highly educated local people has often
been described as pidgin. The reason is that locally, or nationally, the low sector of
the economy is run in a local language (Cantonese, Malay or Putonghua) and
English is only an interface among countries that use different vernaculars or local
lingua francas. While in most parts of the United States and Canada, it would be
difficult to travel and communicate with the local population without speaking
English, knowledge of only English can be frustrating while travelling in Taiwan,
Malaysia and Hong Kong. A visitor often comes across locals who speak no
English at all, especially in less-affluent neighbourhoods. Anyone who claims that
the spread of English around the world endangers indigenous languages should
explain how this is possible in countries where it is only a lingua franca of an elite
minority but is barely spoken by the vast majority, or a large proportion, of the
population.
The above observations do not of course demonstrate that these territories have not
suffered any language loss, nor that local globalisation has played no role in this
process. In becoming the major business language of Taiwan, Chinese has
seriously endangered the more indigenous, Formosan languages in much the same
way that Japanese has caused the attrition of Ainu – just as English and the
Romance languages have driven to extinction most of the Celtic languages that
preceded them in Europe. The prevalence of Malay as the vernacular of Malaysia
has certainly been at the expense of several other indigenous languages. Usage of
these equally indigenous languages in wide and diverse sectors of the national
economies has nurtured their vitality by providing them some raison d’être in what
Bourdieu (1991) identifies as the “language market”. In terms of costs and benefits
relative to English as a global language, their association with lucrative functions
4
While it is obvious that the Arawakan and Carib languages were lost because their speakers were
killed or driven out, it is an oversimplification to assume that the African languages were lost
because of the pressures exerted on their speakers by slavery. As explained towards the end of
Section 2, it is the particular form of assimilation exerted on the slaves of the homestead phase that,
by the founder principle, doomed the vitality of African languages early in the history of settlement
colonies.
in local, national and/or regional economies has limited the need for English for
most Asian populations, and it has thus been confined to the role of elite supra-
regional lingua franca. The division of labour is such that the threat of English to
indigenous languages in Asia, as in other former European exploitation colonies, is
exaggerated.
2. Importance of Distinguishing Different Colonisation Styles

It is helpful to start this section with my observation that European colonial
languages have endangered other languages, or driven them to extinction, typically
in settlement colonies, not in exploitation nor in trade colonies. It is also important
to bear in mind that globalisation is not as recent a phenomenon as may be
assumed. It is in some ways as old as colonisation in its population-genetics
interpretation, to the extent that when a population relocates and/or dominates
another, it more or less imposes a form of geographical globalisation by connecting
the political and economic structure of the colony to that of the homeland. The
colonists may import into the new territory production techniques that are more
typical of the metropole, they may make the colony part of the same industrial
network, and they often adopt the same business language at least for some level of
the socio-economic and political system. So, even the use of European languages
as the official varieties in some former colonies is a form of globalisation, to the
extent that they represent some uniformity or unity (as partial as it is) in the way
that business is conducted in the metropole and the colony. Thus, today’s
globalisation differs from its earliest ancestors, say of the time of the Roman
Empire, particularly in complexity and speed of communication rather than in the
fact of interconnectedness and uniformity of economic systems, technology and
production of goods.
In the context of this article, in which socio-economic ecology is invoked to

explain variation in the vitality of languages, the distinction between different
colonisation styles sheds some light on why local globalisation is not equally
extensive or integrated everywhere. Each colonisation style has determined
particular patterns of interaction between the colonisers and the indigenous
populations as well as the particular kind of economic structure that is now in
place. The categorisation is far from being clear-cut, and there are mixed cases.
However, this rough distinction, which needs refining in future work, will help to
make more sense than has been suggested in most of the literature of how
languages have been vanishing over the past 400 years of Western European
hegemony.
Mufwene (2001) distinguishes between trade, settlement and exploitation colonies.

Trade colonies (such on the west coast of Africa from the fifteenth to the
eighteenth centuries) were the first to develop. This typically happened soon after
Europeans explored new territories and established trade relations with the local
people on more or less egalitarian terms, although the terms of interaction changed
later, at the expense of indigenous populations (see below). The relationships were
sporadic and generally led to the development of new language varieties called
pidgins, typically lexified by a European language on the west coast of Africa but
by a Native American language in the Americas.
In the latter part of the world, the trade colonisation was concurrent with settlement
colonisation. Europeans settled to build new homes, or better Europes than what
they had left behind (Crosby 1986). The nature of regular interactions among
different populations in these new colonies often led to protracted competition and
selection among the languages and dialects they brought with them, leading to
shifts from some to others and to the loss of several of them, as well as to the
emergence of new language varieties typically lexified by European languages.
Some of these have been identified as creoles (typically in plantation settlement
colonies), but others have been identified as new, colonial dialects of the European
lexifiers, such as American English(es) and Québécois French (in non-plantation
colonies). 5 No significant language loss has so far been associated with trade
colonisation, even when trade was abused to enslave and deport some of the
indigenous populations.
Especially noteworthy about settlement colonies is the fact that they gradually
produced local or regional monolingualism, favouring the language of the
colonising nation but dooming to extinction the languages brought by the Africans
(who were first to lose theirs, as explained below) and Europeans originating from
countries other than the colonising one (the case of Gaelic/Irish, German, Italian,
French, Dutch and Swedish in North America, except in Quebec and Ontario).
Native Americans lost their languages either because they were decimated by
diseases and wars, or because they were forced to relocate to places where they
could not continue to speak their languages, or because they eventually got to
function in the new, European-style economic world order which imposed a new
language of business and industry. Unlike trade colonies, settlement colonies
everywhere gradually evolved to some form of economic (and social) integration
that has endangered languages other than those of the colonising European nation,
or one adopted by it.6
The balance sheet has of course involved more losses than gains, but we must
always remember that the outcome of the contacts of population and of languages
in settlement colonies anywhere, including Australia and New Zealand, has not
consisted of losses only. This is especially important because we do not know what
5
As explained in Mufwene (2000, 2001), the criteria for the distinction are social, not structural. The
geographical or socio-economic distinction simply serves to identify places that coincide with the
spurious opposition widely accepted to date in linguistics between creole and non-creole languages.
We need not discuss this question here. Suffice it to note the emergence of new language varieties,
regardless of whether they are considered as new dialects of the same European colonial languages
or as separate languages.
6
The latter was the case for English in Suriname, which evolved into creoles such as Saramaccan,
Sranan and Ndjuka. Dutch serves as the language of the elite in this former plantation settlement
colony, not as a vernacular. Almost the same is true of the Netherlands Antilles, where Papiamentu,
a creole largely lexified by Portuguese, functions as a vernacular.
the future of creoles is, nor whether American and Australian Englishes will be
considered as new dialects of English or as separate languages a couple of
centuries from now, if nothing changes in the present world order and in the
dynamics of the coexistence of languages.
The question of the future of creoles is relevant, because the former plantation
settlement colonies in which they developed have had an economic history
different from those of non-plantation settlement colonies, which are more
industrialised. After the abolition of slavery, plantation settlement colonies evolved
economically on a hybrid model between the non-plantation settlement colonies
and the exploitation colonies (explained below). With the exception of those that
have become French overseas départements, most of the former plantation
settlement colonies have not industrialised and belong in the LIC bloc of nations,
marginally engaged in the recent trend of world or regional global economy. The
mass of their populations is under hardly any pressure to speak a language (variety)
other than Creole. Jamaica is a good example, with Patois gaining in vitality.
The above considerations are simply a reminder that, just as colonisation has not
been uniform worldwide, the vitality of languages has not been uniformly affected
everywhere, not even in former settlement colonies. In future research, it will help
to examine the social structures of these former colonies in terms of which have
majority European populations and which do not, whether this has some correlation
with economic development, and to what extent particular patterns of interaction
across language or dialect boundaries are linked to the process of language
endangerment.
It is also worth determining the extent to which settlement is advanced in a

particular territory and what can be learned about the factors that bring about
language endangerment. If the documentation provided by Nettle and Romaine
(2000) is accurate, why are there proportionally more Native American languages
surviving in Canada than in the USA, and why are there more indigenous
languages still spoken in Latin America than in North America? Are these
differences a consequence of variation in colonisation patterns within the
settlement style (including patterns of interaction with the indigenous populations),
are they a consequence of variation in the physical ecologies of the settlement
colonies, or do they reflect a combination of both factors? For example, can the
size and density of the Amazon forest be overlooked as a factor in the survival of
indigenous languages in a large part of South America – any more than the role of
rain forests in the preservation of linguistic diversity elsewhere? Is this
phenomenon entirely different from Nettle’s (1999) and Nettle and Romaine’s
(2000) observation that the greatest linguistic diversity obtains along the equatorial
forest, in a worldwide belt between the tropics?
We cannot be shocked by the fact that indigenous languages have survived the
most in exploitation colonies, which have typically replaced and expanded former
trade colonies of Africa and Asia since the mid- or late-nineteenth century. Even
those languages that have died or are moribund in these territories have suffered
not from European colonial languages but from other indigenous languages that
have been favoured by the new socio-economic ecologies implemented by
European colonisers (e.g. Swahili in East Africa, Wolof in Senegal, and Town
Bemba in Zambia).
Although both settlement and exploitation colonies developed from trade colonies,
in part as the consequence of European commercial greed in wanting to control the
sources of raw materials and other products needed in Europe, very few colonisers
planned or decided to build new homes in the exploitation colonies. As the term
exploitation colony suggests, these colonies were intended to be exploited for the
enrichment of the European metropole. The colonisers were generally civil
servants or companies’ employees who served limited terms and had to retire back
in Europe. With the help of missionaries and their schools, they generally
developed an intermediary class of indigenous bureaucrats or low-level adminis-
trators through which they communicated with the local populations or they
themselves learned the most important of the local languages, but they encouraged
no more than this local colonial elite to learn scholastic varieties of their languages
(Brutt-Griffler 2002).
Instituting economic systems that generally reaped raw materials to be processed in

metropolitan industry, the colonisers fostered a two-tiered economic system in
which the overwhelming mass of the population continues to communicate in the
ethnic languages or in the (new) locally-based lingua francas, such as Lingala in
the Congo Basin, Sango along the Ubangi River, Swahili in East Africa, Wolof in
Senegal, Songhay in parts of West Africa east of Senegal (along Arab north-south
trade routes), Hausa in Nigeria, Fanagalo in the Copper Belt extending from South
Africa to Zambia, and Bazaar Malay in South-East Asia. In a few places, such as
Nigeria, Cameroon and Papua New Guinea, pidgins based on European languages
were developing from naturalistic, trial-and-error attempts to communicate in these
languages (without a teacher) by the mass of the population who participated in the
lower sector of the colonial economy. The expansion of these pidgins into major
lingua francas sometimes competed with, but did not eliminate (the development
of), other indigenous-based lingua francas, such as Pidgin Ewondo in Cameroon or
Police Motu in Papua New Guinea.
Overall, as in the case of trade colonisation, these colonial languages were just
additions to local repertoires of languages and constituted little threat to the more
indigenous ones, which were protected by clear divisions of labour in their
functions – with the more indigenous languages functioning as vernaculars and the
colonial languages, including the few indigenous ones favoured by the colonial
regimes, used as lingua francas. Socio-economic changes of the late colonial and
post-colonial periods, with many of the new lingua francas becoming urban
vernaculars and with relatively more lucrative jobs based in urban centres and
operating in them gave a competitive edge to the new indigenous lingua francas.
Ethnic vernaculars fell into attrition in the cities, and the trend is expanding to
some rural areas. The collapse of LIC economies and the increasing relative
economic importance and lure of urban centres, which led to rural exodus,
compounded to further erode the beneficial significance of rural indigenous
languages. Still, these have been eroded not by the European languages but by the
indigenous lingua francas be they traditional (such as Swahili, according to Nurse
and Spear 1985) or new (such as Lingala).
We really must remember that in the evolution of languages, the balance sheets
from European contact with other countries look very different in settlement
colonies than in their exploitation counterparts. An important reason is that the
colonial agents were less socially and psychologically invested in the exploitation
colonies than were the colonists in settlement colonies. The latter considered their
colonies as their homes (Crosby 1986) and the patterns of their interactions with
the indigenous populations gradually moved from sporadic to regular, with the
involvement of the indigenous populations in the local economy growing from
marginal to engaged. Also, unlike in exploitation colonies, where the European
colonisers remained a small, though powerful, minority, the colonists in non-
plantation settlement colonies (the continental Americas, Australia and New
Zealand) became the overwhelming majorities and instituted socio-economic
systems that function totally in their own dominant language.
Once demarginalised and now relatively absorbed minorities, the indigenous

populations in former settlement colonies have felt more and more pressure on
them to also speak the majority languages for their economic survival, especially
after the transformation of their physical ecologies made it impossible for them to
continue their traditional economic systems. Their gradual assimilation to the
mainstream made it less and less necessary for their children to learn their ancestral
language or even stick to their traditions. Demographics have played a more
important role in language loss than has been highlighted in the relevant literature.
In most former exploitation colonies, the local people did not even feel the same
pressure to shift, because they remained the overwhelming majorities who in the
rural areas have barely been affected by the economic and political transformations
undergone by their territories, including the formation of nation-states. Most of
them have not even had options other than to continue operating in their traditional
world or, at best, to work in the low-cost colonial and post-colonial labour system
that does not require a European language.
In fact, the new world order in former exploitation colonies is such that even the
elite participating in the interfacing sector of the economy have had no pressure,
except from their own personal attitudes, to give up their indigenous languages. If
anything, unless they decided to sever links with their ancestral customs, the
pressure has been just the opposite: to preserve competence in the ancestral
languages in order to continue interacting with relatives in the rural areas.
The closest approximation of European values is evident in the development of

urban societies, in which traditional and colonial ways have mixed and the new
indigenous lingua francas (such as Wolof, Swahili and Lingala) have gained
economic power and prestige, and have gradually displaced (other) ancestral ethnic
languages. It is these that can be said to have endangered indigenous languages, to
the extent that some rural populations have been shifting to the urban vernaculars,
abandoning some of their traditional cultural values for those practised in the city.
On the other hand, the city has also been perceived as the source of some negative
transformations and the main beneficiary of economic progress at the expense of
the rural environment. Negative attitudes towards it have often been concurrent
with resistance to its language, thus providing the ethnic languages an identity
function that has slowed down their demise.
In the same vein, unemployment in cities and the ever-growing size of the
proletariat in African and other LICs have also disfavoured the usage of European
languages. There are fewer and fewer incentives for speaking these languages
which have sometimes been interpreted as a means of exploitation by indigenous
rulers. Even in more prosperous former exploitation colonies such as Singapore
and Malaysia, European languages have continued to function primarily as bridges
with the world outside the home, or outside the ethnic group or neighbourhood, or
outside the country.7 Otherwise, it remains natural to communicate with members
of an inner group in an indigenous, or non-European, language.
We should thus not overrate the importance of European languages regarding

language endangerment. The experience in former exploitation colonies has
certainly not been the same as in former settlement colonies, although European
colonisation has undeniably spread European languages to territories where they
were not spoken 400 years ago. Moreover, former plantation settlement colonies
reveal features of both exploitation and settlement colonies. They are like the latter
in that the indigenous languages have generally disappeared, due to the rapid and
dramatic deaths of their speakers or to the relocations of indigenous populations to
places where they discontinued speaking their languages.
The settlement colonies are also similar in that several immigrants lost the
languages of their homeland. The homestead period in these settlement colonies
must have exerted a serious negative founder effect on the languages of the
enslaved Africans. They were originally integrated as small minorities in the
homesteads, which were isolated from each other. They had nobody with whom to
speak their languages within the homestead, and in the rare events that they
happened to know somebody on another homestead who spoke the same language,
there was not enough regular interaction to have permitted the active retention of
7
As noted above, my categories of colonization styles are not perfect and need refining. Singapore is
definitely not a typical former exploitation colony. To date, the Malays, the most indigenous of its
current almost fully Asian population, represent less than 15 per cent of the total, as opposed to
more than 75 per cent of Chinese. However, neither is Singapore a European-dominant settlement
colony and it developed its present socio-economic structure after independence. Time will
determine whether its ethnolinguistic diversity will survive the promotion of English as their
common language by its political leaders.
that common language. Attrition and loss were simply caused by lack of
opportunities to interact in the African languages.8 Their Creole children learned to
speak the colonial languages as their vernaculars and they would in fact become
the models emulated by the mass of bozal slaves of the plantation period, those
slaves who had recently arrived from Africa and were most likely to work in the
field.
While the colonies were growing from homestead societies to plantation societies,
Creole slaves were typically preferred to bozal slaves, as they were generally more
familiar with the local customs and vernaculars (see, for example, Berlin 1998).
They were often spared the hardship of working as field hands, and they thought of
themselves as superior to the bozal slaves, whom they had the responsibility of
seasoning. This process entailed acculturating the bozal slaves to the local
vernacular. The constant decrease in opportunities to speak African languages,
especially in socio-economic settings marked by high societal multilingualism,
fostered more and more erosion of the African languages, and eventually their loss.
The situation is somewhat reminiscent of how rural populations have been
absorbed over the past century in sub-Saharan African cities, except that here the
existence of ethnic neighbourhoods has slowed down the process of language shift.
As in sub-Saharan African cities, the African slaves formed the overwhelming

majority of the plantation societies. People of European descent have been small
minorities, with small subsets of them emerging as affluent. Yet, the countries that
evolved from such plantation societies still contain large proletarian majorities that
speak Creole and identify socio-economically with it. Because of lack of incentives
in an economic system depending on foreign markets and industry, participating
only marginally in the world’s global economy, and becoming poorer and poorer,
Creole has gained more vitality in relation to the acrolectal language varieties
spoken by the upper class. In places such as Jamaica and Haiti, it is also clear that
the overt prestige of a language does not necessarily guarantee its vitality. The
underprivileged do not necessarily aspire to the varieties spoken by the more
affluent members of their societies, especially if these varieties will not improve
their conditions. The fact of being economically disfranchised is often a good
reason for despising supposedly prestigious varieties.
8
This does not mean that the African languages died soon after their speakers arrived in the colonies.
In Haiti, some African languages were apparently used as secret codes during the Revolution wars
(Ans 1996, Manessy 1996). The fact that Voodoo and Kumina rituals contain remnants of African
languages is also evidence that some African languages continued to be spoken up to the nineteenth,
or perhaps the early twentieth, century, although they did not function as vernaculars. The few
languages that seem to have assumed this function were reintroduced with the importation of
indentured labourers from specific ethnic groups that remained segregated from the mainstream of
slave descendants, who speak European-based languages. This was the case of Trinidad Yoruba,
which was spoken up to the mid-twentieth century (Warner-Lewis 1996). However, the gradual
integration of speakers of such languages eventually led to their demise. Usage of some African
languages in nineteenth-century Haiti can certainly be associated with the bozal slaves who arrived
soon before the Haitian Revolution.
3. Why Speakers Shift Languages: What Linguists Should Not Ignore

As argued in Mufwene (in press), prestige alone will not favour a particular
language (variety) over others. Shifting to a particular language is typically
associated with particular benefits to be derived from its usage, especially
economic benefits. Otherwise, speakers stick to the languages they have
traditionally spoken, although they may learn another one for interaction with
outsiders. However, even this behaviour is benefit-driven. Most LIC populations
will not shift to European languages, because the alternatives are not likely to
improve their conditions. In the first place, the division of labour that relies on
indigenous lingua francas in the lower sectors of the economy (in which most of
the workforce are engaged) even makes it unnecessary to target a European
language, because the jobs associated with them are very few.
Immigrants to the New World and Australia shifted to the dominant languages
because they had emerged as the only languages of the colonies’ economic systems
and they had something to gain from the shift, or at least they avoided the danger
of not being able to compete at all in the new labour markets. Although slaves gave
up their languages because they often had nobody else to speak them with, an
important reason why their children never bothered to learn their parents’
languages (just like children in African cities) is that they had everything to gain in
speaking the colonial languages as fluently as they could.
Now the question arises of whether linguists can help some languages to thrive by
encouraging their speakers to have pride in their ancestral heritage, even if they
lack control over situations that have led them to give up their languages. Over the
past decade language endangerment has become a major preoccupation among
linguists. In a seminal article (1992), Michael Krauss instilled a certain amount of
guilt among linguists, accusing them of negligence to the vitality of the subject
matter of their own research: languages. The number of publications has increased
since then. They have typically blamed the European colonisation of the past 400
years and today’s global economy for this state of affairs. Some linguists have even
spoken of “killer languages”, which are held guilty of linguicide (by analogy with
homicide) as if languages had independent lives and weapons of their own.
The issues have sometimes become confusing, especially when language

preservation and language maintenance are confused as one and the same (see
below), and the very linguists whose party line is that language is primarily oral
and spoken have privileged the school system and the written medium as ways of
saving the endangered languages. Very little scholarship has been invested in
understanding the ecology of language and what it takes to sustain the vitality of a
language, especially in territories where several languages have coexisted
apparently happily with one another under an efficient division of labour in the
repertoires that contain them. As explained in note 1, languages have no lives that
are independent of their speakers. Therefore, languages do not kill languages; their
own speakers do, in giving them up, although they themselves are victims of
changes in the socio-economic ecologies in which they evolve. Solutions that focus
on the victims rather than on the causes of their plights are just as bad as
environmental solutions that would focus on affected species rather than on the
ecologies that affect the species.
European colonisation of the past four centuries has certainly contributed to the
predicament of languages around the world, as it has introduced new socio-
economic world orders that have pre-empted the usefulness of some languages.
However, it is helpful to put things in historical perspective too. Language shift and
language loss are neither new nor recent phenomena, as evidenced by the curious
fact that only 3 per cent of the world’s languages are spoken in Europe (Mayor and
Bindé 2001), although it is one of the most densely populated parts of the world.
Today’s prevalence of English (a Germanic language) in the United Kingdom and
of Romance languages in south-western Europe has been accomplished at the
expense of Celtic languages, only a handful of which are still spoken today.
Likewise, the Indo-European languages have spread and prevailed in territories
where other languages, survived today by Basque and Finnish, for example, used to
be spoken.
The Stammbaums ('family trees') of genetic linguistics, which illustrate language

diversification and therefore an increase in the number of languages, have masked
the concomitant loss of indigenous languages replaced by the new, Indo-European
languages. Things seem to have proceeded in the same way as they have recently,
with some languages prevailing at the expense of others and being transformed in
the process, becoming new varieties and eventually being recognised as separate
languages. It would be informative to learn why and how Basque and Finnish
survived the dispersal of Indo-European languages, while the majority of others
vanished. We could then investigate similarities and differences between what
happened then and what is happening now, and why some populations just cannot
preserve their languages against the invaders while some invaders (e.g. the Norse
and Norman French in England and the Tutsi in Rwanda and Burundi) have
actually given up their own languages.
Linguists have typically bemoaned loss of linguistic, especially typological,

diversity. Rarely have they focused on speakers themselves in terms of motivation
and costs and benefits to them in giving up their languages. Seldom have they
addressed the question of whether the survival of a language would entail more
adequate adaptations of its speakers to the changing socio-economic ecologies.
They have decried the loss of ancestral cultures as if cultures were static systems
and the emergence of new ones in response to changing ecologies was necessarily
maladaptive. The following questions arise from this particular approach to
change: Are the ancestral cultures more adaptive to the current world order than the
new ones? Are the peculiarities of the lost or endangered languages more
informative about the nature of universal grammar as a biological endowment for
language than are those that have survived and the new ones that have emerged?
None of the treatises cited at the outset of this article addresses these questions.
It should help to recall that much of the concern for language endangerment has
been modelled on environmentalists’ concern about the degradation of our physical
ecology due to modern industry. Like linguists, environmentalists are ecologists,
scholars who have specialised in the co-evolution of species and their
environments. We would really be their counterparts if there were a research area
in linguistics specialising in the coevolution of speakers, their socio-economic
ecologies, and their languages. The concern for language endangerment seems to
have caught linguists off guard and we have been prescribing remedies without the
requisite understanding of the socio-economic dynamics that have affected the
vitality of languages negatively or positively in different parts of the world
throughout human history.
There is another important point of difference. Environmentalists are concerned

with the environment relative to humans, with the way we have coexisted with
other species, and with how we have been affected by what affects them and vice
versa. Their case for the preservation of biodiversity has been less for the benefit of
their discipline than for various residents of our planet. However, things are not so
similar in the literature on language endangerment. If languages are there to serve
their speakers, it is strange that the costs and benefits to the latter have been
overlooked for so long! Because languages do not have independent lives from
their speakers, it is bizarre that the hosts, whose socio-economic behaviours affect
them, have been ignored.
Such literature could likewise have bemoaned language change, as this process
substitutes one kind of (sub)system for another. The literature has ignored the fact
that speakers make their languages as they speak; and cultures are being shaped as
members of particular communities behave in specific ways. These are dynamic
systems that keep evolving as people behave linguistically and otherwise and as
they keep adapting these systems to new situations. That is, languages co-evolve
with their speakers. Language shift, which is the main cause of language
endangerment and death, is part of this adaptive co-evolution, as speakers
endeavour to meet their day-to-day communicative needs. It is not so much that
linguistic changes are bringing about cultural changes, but that linguistic changes
echo cultural changes. That is, language shift is no more than an adaptive response
to changes in a particular culture, most of which I have identified as a socio-
economic ecology. Arguments for language maintenance without arguments for
concurrent changes in the present socio-economic ecologies of speakers seem to
ignore the centrality of native speakers to the whole situation.
To suggest that native speakers will maintain or preserve their cultures if they
continue speaking their language is to ignore the fact that in the first place they
would not stop speaking it if they valued its association with their ancestral culture
over their necessary adaptation to the current world order – a simple matter of
prioritising things in their struggle for survival. The position in the average
literature on the subject is also tantamount to assuming that language and culture
go hand in hand, that only one language can best mirror or convey a particular
culture, and that another language cannot be adapted to convey it. Sapir (1921)
argues convincingly for decoupling language and culture as separate systems.9 The
literature of indigenised English and African French, for example, have made it
quite obvious that a language can be adapted to a different culture – which gives
more meaning to the notion of “language appropriation” (so much preferred by
Chaudenson (2001) over those of “language learning” or “language acquisition”).
So populations shifting to another language have always had the option of adapting
the new language to their ancestral culture. After all, it is generally influenced by
their substrate systems and typically develops into a new variety.
We can perhaps argue that a language mirrors a culture because it is itself part of a
culture. Changes affecting it reflect changes in a particular culture. Arguing for its
maintenance when the population of its speakers behaves differently reflects a
value judgement on the part of the linguist, who rates the ancestral culture more
highly than the one that is being fashioned by the speakers’ linguistic behaviour. A
problem then arises when nothing is being done or advocated to change the
ecology, to which speakers adapt. Linguists are thus different from
environmentalists, who have realised that the survival of a particular species
depends largely on restoring the ecology in which it thrives. Curiously, linguists’
proposal for rescuing endangered languages (as articulated in, for example, Crystal
2000; Nettle and Romaine 2000) suggests that speakers must continue their
traditional communicative behaviours regardless of changing socio-economic
ecologies. Somebody should explain how adaptive such resistance to changing
ecologies is or how a language can continue to be spoken as a vernacular when the
ecological structures that used to support it barely survive.
As there are countries such as Taiwan which have succeeded in appropriating the
Western capitalist economic system without losing much of the Chinese culture
and language, it is obviously clear that other countries could have taken that path. It
should help to know why they did not choose to do so. And the following question
also remains: Can the process be reversed in nations whose cultural and linguistic
experiences have been different, and under what realistic conditions?
In this context, it becomes important to distinguish between language maintenance

(sustaining an ecology in which a population can continue to speak their language)
and language preservation (recording texts from a particular language graphically
or mechanically). If the current ecology cannot be changed, should not linguists be
more realistic and focus on language preservation (Paul Newman 1998 and
forthcoming) rather than on maintenance? Such a response would of course also
9
See especially Chapter 8: “Language, culture, and race” (207–20 in the 1949 printing). In the
particular case of Athabaskan, Sapir states: “The cultural adaptability of the Athabaskan-speaking
peoples is in the strangest contrast to the inaccessibility to foreign influences of the languages
themselves” (214). Invoking factors that are subsumed by what I have identified here as socio-
economic ecology, he writes: “A common language cannot indefinitely set the seal on a common
culture when the geographical, political, and economic determinants of the culture are no longer the
same throughout its area” (215). Much of my discussion capitalizes on this view.
entail investing more time into understanding natural laws that since the beginnings
of humanity, and through colonisation, have regulated language shift, the loss of
some languages, the emergence of new ones, and the balance sheets of losses and
gains at different states in history. Then we would be able to deal with language
endangerment with justifications other than benefits and costs to linguistics. My
position remains that costs and benefits to speakers as individuals adapting to
socio-economic changes that affect them should have played a more central role
than is evident from the literature to date. Even from an environmentalist
perspective, in which all members of an econiche matter, speakers are far more
important to our planet than their languages, which are being lost through their own
communicative practices.
Scholars such as Nettle and Romaine (2000, cited here because they have the most
explicit discussion of all publications on this subject in 2000 and 2001) argue that a
certain amount of traditional folk knowledge of their environments is lost with
dying languages. The observation is undeniably true, but it fails to note that the
environment itself is changing and this particular knowledge may be becoming
quite irrelevant to it. Moreover, the culture and this specific knowledge must have
been eroding concurrently with the language itself, if not before it; otherwise they
would be transferred to the new language. One way or another, insisting on the
utility of the endangered language and on bilingualism, when the socio-economic
ecology can no longer sustain them, suggests that a language can be sustained
regardless of whether or not it really contributes to the socialisation of the young
into new realities. Yet experience everywhere suggests that linguistic behaviour is
profit-driven (Bourdieu 1991). Speakers would like to invest not only in forms and
structures that maximise their linguistic capital but also in a language that is
beneficial to them. Individual multilingualism is possible typically when it is
advantageous to the speaker. It is perhaps not by accident that in highly stratified
societies multilinguals seem to be the most numerous in the lower classes. In
societies that are typically monolingual, multilingualism is practised by those who
can travel outside their communities and interact with outsiders. Not everyone has
a vested interest in speaking more than one language. A profile of individuals or
communities that give up their languages in favour of others should be informative
in future research.
4. Colonisation and Globalisation: Not Such New Phenomena

The current literature on language endangerment has presented the phenomenon
primarily as one of the negative side-effects of European expansion and
colonisation of most of the non-European world over the past half millennium. It is
true that the geographical and political extent of European expansion has been
unprecedented, for example when the size of the British Commonwealth, as
discontinuous as it has been, is compared with that of the Roman Empire a
millennium earlier. However, putting things in perspective, the difference in size is
also seen to be a function of differences in modes of communication. About 1,500
years ago, the size of the Roman Empire was certainly also unprecedented, in fact
it was too large to have central control over, at least under the conditions of the
time. Easier and faster transportation systems since the fifteenth century have
allowed the European conquest of territories much farther away from the
metropole. Easier and faster means of communication (especially with the
invention of the telegraph and telephone, of the radio and television, and now of
the Internet) have facilitated the political, military and economic controls of larger
and larger colonies, making the world look even smaller. Improvements in control
techniques have also facilitated the control of more and more aspects of the
colonies.
However, today’s colonisation differs from that of earlier times more in size and
complexity than in kind. It is not so common to refer to the dispersal of the Bantu
populations from the southern Nigeria and western Cameroon area into central and
southern Africa as colonisation. The same applies to the spread of Indo-Europeans
from Asia Minor to Europe. In reality, these are instances of colonisation, at least
in the population-genetics sense of relocation to a new territory. As Nettle and
Romaine (2000) point out, agriculturalists generally colonised hunter-gatherers and
imposed their economic systems on them. Thus the Bantu populations have largely
assimilated or decimated the Pygmies and Khoisans in central and southern Africa,
and only a few of these latter populations remain today as distinct minorities in a
wide area considered to be Bantu. Of the non-Indo-European languages that
preceded the European languages, Basque, Finnish, and Lap are notorious
exceptions whose survival conditions need uncovering. Basque is an especially
interesting case, because it has survived both the Indo-European and Roman
colonisations, although it has lost a lot of its geographical space. Much of the
present linguistic map of Western Europe represents consequences of language
shift, under colonisation, for Roman or Germanic languages. Celtic languages have
become moribund minorities in a wide territory, from Germany to the British Isles,
that used to be dominated by the Celts (Green 1998).
We stand to learn a lot by trying to understand similarities and differences between

those earlier forms of colonisation, and between them and the recent European
phenomenon of the past 400 years. For example, both the British Isles and the
southern part of Western Europe were colonised by the Romans. In both places
Latin was the colonial language, but the Romance languages have developed only
in the latter. The subsequent colonisation of the British Isles by the Germanics can
perhaps be invoked to explain this difference. However, we cannot ignore the fact
that following Roman colonisation, Iberia was dominated first by the Arabs and
then by the Visigoths, and France was dominated by the Frankish. Also noteworthy
in this context is the fact that the colonisation of England by the Norman French
caused no language shift of the kind that would produce a new language from that
of the colonisers. Its main consequence was the development of the ancestor of
today’s Standard English varieties.
Also significant is the fact that, as in former exploitation colonies of Asia and
Africa, it was after the colonisers had left that the important proportions of the
indigenous populations adopted the coloniser’s language in today’s Romance

Europe. Can we assume that if the Germanics had not settled permanently in the
British Isles, these territories would have become Romance too? Or should other
factors be taken into account? Why did the Arab, Visigoth, Frankish and Norman
colonisations of Iberia, France and England not have the same effects regarding the
vitality of indigenous and colonial languages as did the Roman and Germanic
colonisations of the same territories? Did all these cases involve colonisation of the
same style, such as settlement or exploitation? If so, how did they vary?
There are nevertheless similarities between England and North America in the
styles of their settlement colonisation by outsiders and in the fates of their
indigenous languages. When the Germanics settled in England, they drove the
Celts westwards and later they assimilated the survivors. So did the Europeans in
North America, obtaining concessions on the eastern coast of North America and
driving the indigenous populations westwards. Eventually, they assimilated the
survivors, after the American Revolution (which was primarily the independence
of European colonists from England) and the present United States had been
formed.
Native Americans were really not brought into American politics and recognised as
American citizens until late in the nineteenth century, and this assimilation process
in itself was quite reminiscent of the gradual absorption of the Celts in the British
Isles by the Germanic invaders. Colonised since the fifth century, some Celts such
as the Irish did not become subjects of the United Kingdom and have to speak
English as a vernacular until the nineteenth century, long after Oliver Cromwell
had initiated the settlement colonisation of Ireland in the seventeenth century and
potato plantations had become one of its major industries. In both cases, the loss of
indigenous languages did not start until the assimilation of the local people to the
current socio-economic system.10
Noteworthy in all such cases is the fact that absorption of the indigenous
population by the colonisers has generally led to the loss of indigenous languages,
especially when the colonised are kept in a subordinate position. The critical factor
is their involvement in an economic system in which they must use the language of
the new ruler in order to compete in the labour force and function adaptively. This
is an aspect of globalisation as homogenisation, requiring that things work more or
less the same way in the colony as in the metropole, especially in the exercise of
power and control of the working class. Here similarities may be seen between the
Germanicisation of England and the rest of the British Isles, the Islamicisation of
North Africa and Iberia, and the Romanisation of south-western Europe. To the
questions asked above about differential impacts of colonisation, the following can
be added: Why did the eastern Roman Empire, which was colonised for longer, not
undergo the same kind of language shift as did the western empire?
10
The case of Scotland is different because this was more a merger of kingdoms than regular
colonisation. English was not imposed by the English (thus Germanic) refugees but adopted by an
enthusiastic Scottish monarch who loved both an English princess and her language.
Did the Romans colonise territories of their empire on the exploitation model and
is their departure comparable to the recent independence of European exploitation
colonies? If so, what are the specific ecological factors that account for language
shift in their western empire? Why has a similar evolution not taken place in sub-
Saharan Africa, where any serious danger to minor indigenous ethnic languages
arises more from the expansion of the indigenous lingua francas than from the
European colonial languages (Mufwene 2001)?11
One noteworthy social ecological factor here is that Roman soldiers and
administrators married into the local communities and obviously transmitted their
language to their children. The latter, who shared power with their parents, also
used their Romance languages (i.e. Celticised Vulgar Latin, such as today’s
Africanised French) in ruling their countries, continuing basically the same Roman
administrative style. In sub-Saharan Africa, segregation was the rule and cross-race
unions were relatively rare. Most such unions occurred between the European
merchants with African women, but the merchants had no political or
administrative power and were more disposed to speaking indigenous languages.
Their children had barely more advantages than the more indigenous colonial elite,
who had the same kind of colonial education and, as noted above, have not given
up the indigenous languages.
Overall, as auxiliaries to colonial rule, the African elite were just intermediaries
between, on the one hand, the indigenous populations and, on the other, the
European colonisers. They worked for the latter but socialised more with the less-
privileged indigenous mass than with their rulers. Thus, their usage of European
colonial languages was highly circumscribed, despite their additional function as
lingua francas between those from different ethnolinguistic backgrounds who did
not share an indigenous lingua franca. Even the few mulattoes that were to be
found were still under pressure to speak African languages in order to be integrated
in the majority populations.
While running post-independence Africa, the elite have generally tried to maintain
the socio-economic structure of colonial sub-Saharan Africa, although they have
had more success in maintaining the linguistic division of labour than in sustaining
the colonial economic (infra)structure. The decline of their nations’ economies has
in fact favoured the indigenous lingua francas over the European official
languages. In the United Republic of Tanzania, Swahili has been promoted at the
expense of English (although it is debatable how successful the policy has been),
and in cities such as Kinshasa (Democratic Republic of the Congo) Lingala has
gained more prestige than French in modern popular culture, where French is often
derided.
Former plantation settlement colonies are somewhat like sub-Saharan African

11
Exceptions to this observation appear to be Gabon, where French is spoken by the urban population,
and Mozambique, where the protracted liberation war promoted Portuguese as the lingua franca of
the liberation fighters and the major lingua franca of the post-independence state.
countries in that language varieties of the proletarian masses are far from being
endangered by the acrolects that were privileged by the colonial systems. As a
matter of fact, former English and French plantation settlement colonies were, to
all intents and purposes, converted into exploitation colonies after emancipation.
They were assigned administrators from the colonial metropoles. The economic
systems of all these territories, which are still in the LIC group, have remained
generally the same as those of sub-Saharan African countries, with the exception of
French overseas départements whose economic discrepancies from the metropole
are just being addressed now. Haiti, which became independent in 1804, before
emancipation in the remaining colonies, shows perhaps the highest proportion of
creole speakers. As Dejean (1993) points out, the only vernacular of the
overwhelming majority of the Haitian population is far from being threatened by
French.12
Yet students of language endangerment cannot continue to dodge interesting

questions that arise from variation in colonisation styles. These linguistic
developments are like natural evolution in population genetics, where it is
absolutely imperative to understand what ecological factors bring about particular
consequences for varying species in an econiche. The non-uniform linguistic
consequences of colonisation over the world makes it compelling for linguists to
have to investigate and better understand the socio-economic factors that affect
language vitality, favouring colonial languages in some settings but indigenous
ones in others.
It is also obvious that many of the developments today have antecedents in earlier
history, especially in the colonisation of England by the Germanics and of south-
western Europe by the Romans. Adequate interpretations of those earlier cases
depend partly on how well we understand recent developments and what parallels
we find between the latter and the former. In turn, our understanding of the past
will shed new light on different aspects of what we thought we already understood
about the present.
5. Imperial Languages and Language Endangerment: A Myth that

Cannot Go On
As noted in Section 1, globalisation also applies to “the emergence of international
12
According to Dejean, 95 per cent of Haitians are monolingual in creole (77), many of them do not
interact with French speakers (78), and members of the French-speaking elite also speak creole (76).
The latter situation is similar to that of the African elite explained above. Moreover, the proletarian
mass of creole speakers does not even aspire to speaking French (79). Although the size of the
proletariat is apparently much greater in Haiti than other Caribbean islands, the situation described
by Dejean has counterparts in them. It may in fact develop in the direction of the same Haitian
extreme if their economies do not improve. In places such as Jamaica, patois seems to have gained
more vitality over the past few decades, or perhaps acrolectal speakers have become more
uncomfortable with speaking their variety in domains where patois is becoming the norm and where
the acrolect carries no particular prestige, such as in music and local cuisine.
and regional economic networks with blurred national boundaries” as well as to

“the economic monopoly that highly industrialised countries have exercised over
LICs for raw materials and as outlets of their technology”. It has thus led world
languages such as English and French to compete with each other as imperial or
hegemonic languages. These are languages that need not serve as lingua francas
among the elite of the indigenous populations (although they often do) but are
primarily needed to interface local economies (regardless of how globalised they
are) with foreign and more globalised systems. For example, French in Haiti is
needed to maintain some economic ties with France, although the elite also use it to
isolate themselves from the proletarian masses (DeGraff 2002). Taiwan and Hong
Kong could apparently manage locally with their Chinese varieties and without
English, but they use this language to maintain their global associations with the
United States and the United Kingdom. Malaysia and Singapore could probably
also do without English and use only Malay as their national lingua franca if their
economies did not depend so largely on American and British markets. More and
more LICs, especially in Africa, have become arenas where English and French are
competing with each other for monopoly.
To be sure, French as an imperial language (not as a vernacular!) has been losing

ground to English in many places around the world. Works such as Hagège (2000),
joined by Crystal (2000), Nettle and Romaine (2000), and Renard (2001), decry
this English expansionism. They connect it with the McDonaldisation of the world
or the worldwide spread of American movies and other cultural products. In fact,
as if to trivialise the language endangerment “problem”, La Francophonie claims
that French is endangered by English. In a March 2002 posting in The Linguist, the
British Professor Geoffrey Sampson made a similar absurd observation, apparently
confusing the French population’s now better disposition to use English as a lingua
franca with an unfounded fear of seeing it used as a vernacular in France or
francophone Belgium. There is no evidence of such an evolution yet in these
strongholds of French as a vernacular, not even in Quebec, where the economic
pressure for such a development is stronger.
Interestingly, McDonald’s outlets around the world operate in the local lingua
francas, if not their vernaculars (as in the case of France and Germany). Hollywood
films are often translated into local lingua francas/vernaculars, although the music
lyrics are not. Those who learn in English to partake in American pop culture do
not even dream of using it as a vernacular — which is true of many parts of the
world, including France, Germany, Latin America and Russia. What we learn here
is that exportation of desirable technology often carries along the language and
culture of the powerful manufacturer. However, in the vast majority of places
where the imperial languages were not already adopted during the colonial period,
the languages are being learned as international lingua francas. An older imperial
language may become less attractive if it becomes globally less advantageous to
speak it. The competition in such cases is resolved on the basis of costs and
benefits to the local population. It makes little sense to characterise the losing
imperial language as endangered.
Practical considerations prevail a great deal more than linguists have

acknowledged. Proximity to North America has made English more attractive than
French to many Haitians today. Economic or technological aid from the United
States (even if only symbolic), rather than ideological drives on the part of France
to propagate French culture, has made English more attractive to several LICs.
Economic and professional incentives have made English an asset, albeit as a
second lingua franca, even to local francophone professionals. In any case, we
cannot lose sight of the fact that imperial languages are far from becoming
vernaculars in those places where the elite still use their indigenous languages in
domains associated with their local cultures. It is contradictory on the part of
linguists to advocate multilingualism as a possible solution for the survival of
languages around the world and yet discourage people from appropriating
international languages that should allow them to satisfy personal economic and
other cultural interests. There would perhaps be a cause to worry if the hegemonic
languages were becoming vernaculars, but they are not, except in former settlement
colonies, where it is too late to reverse the course of events. Even in places such as
Singapore and Hong Kong, where English is widely spoken among Asians of
different ethnolinguistic backgrounds, my impression is that the indigenous
languages are far from being threatened by it.
There is an exaggerated view of language endangerment as a uniform problem,

based only on numbers of speakers without consideration of history. This is best
illustrated with Nettle and Romaine’s (2000) inventory of the world’s most widely
spoken languages, which includes Chinese varieties, Bengali, Hindi, Japanese,
Javanese, Korean, Vietnamese and Telugu alongside colonial/imperial languages,
viz. English, French, Portuguese, Spanish, Russian and Arabic. It is partly
corroborated by the following list of “eight most widely spoken languages”
produced by Mayor and Bindé (2001, 334): Chinese, English, Hindi, Spanish,
Russian, Arabic, Portuguese and French.
There is no doubt that colonisation of one style or another in the distant past
accounts for the fact that all these languages are so widely spoken. The history of
the world is marked by regular waves of population movement on small and large
scales, with the stronger people assimilating or displacing those they did not kill.
This is as true of the current distribution of the Bantu languages as it is of Indo-
European languages. 13 Asia is no exception, and the current movement for the
independence of Tibet from China is but an evolution from that old expansionist
colonisation which brought together populations speaking different languages.
To be sure, with the exception of Arabic, all the non-European languages in the
above lists function today primarily as vernaculars rather than lingua francas. They
are also dispersed worldwide, with diasporic communities that are largely a
consequence of European colonisation and its demand for labour. Even when they
13
See, for example, James Newman (1995), Mazrui and Mazrui (1998) and Mufwene (2001)
regarding the present linguistic landscape of Africa, Martinet (1986) and Renfew (1988) for the
dispersal of Indo-European, and Cavalli-Sforza (2000) regarding the world overall.
are spoken outside their homelands, Chinese, Hindi, Bengali, Japanese, Javanese,
Korean, Vietnamese and Telugu function primarily as vernaculars among
transplanted people from the same ethnolinguistic background. Thus, in North
America and Europe, Chinese is spoken typically in Chinatowns (although we
cannot even take it for granted that the younger generation is acquiring it in these
neighbourhoods).
The other languages (English, French, Arabic, etc.) are recent hegemonic
languages that owe their large numbers of speakers mainly to their lingua franca
function. English and French in particular have more non-native than native
speakers. While Chinese vernaculars may be a real threat to some Tibetan
languages, they hardly compete with English in North America, the United
Kingdom or the Caribbean. As noted above, English, French and Arabic are
certainly no danger to many languages in the LICs, where they are spoken as
second-language varieties, for highly circumscribed functions, and only by small
fractions of the indigenous population. Likewise, despite France’s present
commitment to the economic development of its overseas départements (mainly by
supporting their infrastructures for tourism), there is no indication that French is a
threat to créole in these territories. Similar doubt can be cast about all territories
where creoles have coexisted with their lexifiers and have derived much vitality
from association with the cultures of the disfranchised proletarian majorities.14
It is also noteworthy that Spanish and Portuguese are widely spoken today largely
thanks to the settlement colonisation of several parts of the world by their
European speakers since the fifteenth century. Portugal and Spain have no
economic or military hegemonies today that would make them threats to other
languages outside those same settlement colonies. In more or less the same vein,
note that Arabic has become so much associated with Islam that it can hardly stand
up to the competition of English and French for the function of international lingua
franca, even in those territories of North Africa and the Middle East where Arabic
vernaculars are spoken. Reading Nettle and Romaine’s statistics (2000) at their
face value leads to a misinterpretation of the dynamics of competition and selection
among the world’s languages.
Also, as noted above, the lingua franca function is scarcely a threat to indigenous
languages in those territories where the hegemonic languages do not function as
vernaculars. In fact, the best lesson here comes from the fact that standard varieties
of the same languages have generally not displaced their nonstandard vernaculars,
14
It is fundamentally inaccurate to count Nigeria and India as anglophone countries in the same way
as the United States is; or the Republic of the Congo, the Democratic Republic of the Congo, the
island of Dominica and Viet Nam as francophone in the same way as France, Quebec and Belgium
are. Dejean (1993) also finds it problematic to count Haiti, with its overwhelming majority of
monolingual creole speakers, as francophone. The only reason for doing so would be in considering
Haitian Creole as a French dialect – a position that is defensible diachronically but is likely to be
disputed politically, especially by creolists. The same seems to be true of all territories where
creoles lexified by European languages have developed.
just as acrolectal varieties have not displaced basilectal and mesolectal ones in
creole-speaking territories. In the now-celebrated case of Ocracoke Brogue as an
endangered dialect (Wolfram and Schilling-Estes 1995), the dialect has actually
been endangered by other vernacular varieties, not by Standard English. Not even
highly stigmatised varieties such as African-American English and Appalachian
English are at all threatened by Standard English. French patois, as either
traditional Celtic languages or rural non-standard French dialects (français
populaires), have been threatened by urban colloquial French, not by Standard
French. Perhaps one of the very reasons why hegemonic languages are a false
perceived threat to indigenous languages in several places around the world is that
they are not vernaculars in the first place.
6. Conclusions
Language endangerment is a much more complex subject than most of the
literature has led us to think. The process is far from being new in human history. It
has been a concomitant of language diversification, which is itself a little-acknow-
ledged by-product of language contact, in which a language is influenced by others
into whose territory it has been taken or which have been brought into its territory.
Such contacts have sometimes caused language shift (instead of sustained bi- or
multilingualism). This process is directly related to language loss. The effects of
language contact are far from being uniform from one territory to another, being in
part correlated with variation in different colonisation styles and in the
communicative functions that the new languages have assumed in various
territories relative to their indigenous counterparts. They are largely a function of
the new economic systems that have replaced the indigenous ones and of the extent
to which local people have been absorbed, assimilated or integrated in the current
systems.
Integration happens when populations coexist in some sort of peace. This state of
affairs makes it ironical and inadequate to speak of language wars, rather than of
competition as a coexistence relation in which alternatives have different ethno-
graphic values to speakers, such that they often must select one or another
alternative during their verbal interactions. It also reveals an interesting point about
how language loss occurs, viz. the more highly valued language stealthily
endangers the less-highly valued one(s) while speakers, unaware of the long-term
effects of their repeated selections, are happy simply to be able to communicate
(successfully) with others. The procedure is the same even during periods of
enslavement, including the most oppressive, from the sixteenth to the nineteenth
centuries in the New World and the Indian Ocean.
Adding significance to the strength of the founder effect, the homestead societies
inflicted a devastating blow on the languages of the enslaved Africans, with the
Africans of the homestead phase being forced by circumstances to operate only, or
most often, in colonial European vernaculars and the plantation-phase slaves being
seasoned by Creole slaves into the colonial vernaculars (Chaudenson 2001;
Mufwene 2001). If a few African languages survived until the mid-twentieth

century in places such as Trinidad and Brazil, it is largely thanks to the indentured
labour system which replaced slavery and brought Africans from ethno-
linguistically more homogeneous areas, keeping them in relative segregation from
former slaves. These languages would gradually die, concurrently with the
integration of (descendants of) these indentured labourers within the populations
that preceded them, by the same process that likewise gradually absorbed Asian
indentured labourers in the same plantation communities also at the expense of
their indigenous languages. This was the same process that absorbed most later
immigrants in the dominant socio-economy of the host countries and led them to
lose their languages.
These relatively recent incidents of language loss also have precedents in older
history. Like the enslaved Africans, the Jews enslaved in Babylon and Egypt had
lost their language (Hagège 2000), through absorption in the local socio-economic
infrastructure, although they had a low social status and were not integrated. On the
other hand, as has been made obvious by the linguistic experience of countries with
large indigenous or indigenised proletarian populations, economic marginalisation
can produce just the opposite effect. The disfranchised proletarians stick to their
indigenous or nonstandard vernaculars as markers of their identity and are forced
by circumstances to avoid the language associated with their economic
exploitation.
In the big picture of competition and selection among languages, cases of language
extinction by genocide remain exceptional. Those due to absorption of
demographically or economically less-powerful groups are more typical. The
distinction between different colonisation styles was proposed in part to distinguish
those territories where peaceful coexistence resulted in language loss from those
where it did not.
Language loss is indeed one of the outcomes of competition and selection among
languages sharing the same econiche. Competition and selection among languages,
not just between indigenous and non-indigenous ones, is similar to that which
obtains among structural features in language evolution (Mufwene 2001). Like
structural features, languages or dialects can be a threat to each other only when
they compete for the same functions. Languages or dialects that have separate
communicative or social functions can coexist quite happily, which has typically
been the case with European and indigenous languages in former exploitation
colonies. Overall, it is when a language is adopted as a vernacular that it becomes a
threat to the speaker’s previous vernacular. European languages have been such
threats to indigenous languages in former settlement colonies because they have
become vernaculars, albeit in new, restructured forms. On the other hand, their
status as lingua francas in exploitation colonies has made them primarily economic
assets for a chosen few, the educated elite, and of rather marginal significance to
the proletarian masses. No colonisation style has proceeded uniformly everywhere
and more factors that distinguish one ecology from another need to be understood,
in the way advocated by Fishman (2000).
There is an advantage that follows from the distinction I have proposed between,
on the one hand, plantation settlement colonies, where descendants of non-
Europeans have constituted demographic majorities (as in the Caribbean and Indian
Ocean islands), and, or the other, non-plantation settlement colonies, where descen-
dants of Europeans have become majorities (as in the American mainland and
Australia), viz. it becomes possible to explain why creoles are not as endangered as
has been suggested by decreolisation hypotheses since DeCamp (1971). In
plantation settlement colonies, creoles have functioned as vernaculars of large
proletarian masses, assuming an ethnographic function that has not competed with
the acrolectal variety spoken by the local elite and required in the white-collar
sector of the economy. They have acquired a status similar to that of indigenous
vernaculars in former exploitation colonies, also serving as identity markers for
their speakers against their economic exploitation by the ruling elite. They are not
at all threatened by the acrolectal varieties. Their ethnographic status is also similar
to that of new, likewise restructured vernaculars that have emerged in other
settlement colonies but have been identified as nonstandard dialects of the same
European languages. These too serve as identity markers for their low-class and
rural speakers and are also used in the blue-collar sectors of their economic
systems. All these new vernaculars (creole and non-creole) are those that have
actually driven to extinction other indigenous and non-indigenous vernaculars.
Globalisation has been a useful consideration in this essay because it sheds

interesting light on the role of socio-economic structure in language vitality. There
is at least a partial correlation between, on the one hand, the type and extent of
globalisation in a setting and, on the other, whether or not the primary language of
the economy is endangering other languages. Generally, language endangerment is
most serious where local globalisation is the most advanced and inclusive of
virtually all economic sectors. However, we must remember that globalisation is
not necessarily implemented in a European language, and the latter may be used
only for international interfacing, as in the case of Japan and Taiwan. Thus, major
European languages are not necessarily threatening non-European languages
everywhere. For this reason, I found it relevant to distinguish between hegemonic
lingua-franca status of a European language and its vernacular function in a
different territory. English is certainly a threat to other languages in polities where
it functions as a vernacular, but not at all in countries where it has been adopted
only to help the local economy interface with the worldwide economy. Thus it is
not a threat to Japanese nor to Putonghua in Taiwan, although it seems to be a
threat to French in francophone African countries, where French also has a
hegemonic status.
The future of languages in the twenty-first century obviously depends on how

individual nations will evolve socio-economically during that time. In some parts
of the world, globalisation is progressing without any serious obstacles that can
stop its effects on disfranchised languages. Then again, the economic future is
already so uncertain in some other parts of the world that no indigenous languages
and cultures are being affected by the present course of events, except somewhat
by the indigenous lingua francas. In most such polities, numbers matter little in
determining whether or not a particular population will carry on their ancestral
language, as long as the speakers remain isolated from developments outside their
communities, as well noted by Mühlhäusler (1996).
While we linguists are so concerned with linguistic diversity as a dimension of

biodiversity to be maintained (Maffi 2000, 2001; Nettle and Romaine 2000), we
cannot ignore a moral dilemma that arises. The socio-economic ecologies of most
populations around the world have changed since the recent European colonisation
of the world started four centuries ago (especially during the past century), and so
have their aspirations for decent living. The changes in these socio-economic
ecologies have often included the emergence of new languages in which both the
indigenous people and immigrants are expected to develop some competence in
order to compete for jobs. Despite their attachment to their pre-globalisation
traditions, the pressures of the new socio-economic systems have made it
increasingly difficult to practise their ancestral languages and cultures. Lack of
practice has stealthily led to attrition and eventually death of the languages. In
other words, the loss of ancestral traditions is a consequence of changes in the
socio-economic ecologies of speakers. Can linguists advocate the maintenance of
cultural heritage without restoring the older ecology? Can it be restored and at what
costs and benefits to the relevant populations? I have not seen these issues
discussed in the literature.
Much of the literature on language endangerment has also promoted linguistic

rights. To the list presented at the outset of this article may be added, among others,
International Journal on Multicultural Societies, vol. 3, no. 2 (2001). As suggested
above, linguistic and human rights are not necessarily congruous. It is certainly not
unnecessary to echo Ladefoged (1992) with the following questions: Can we
linguists work against the aspirations of the affected populations and exhort them
to hold on to their languages and cultures only in the interest of a kind of diversity
that should benefit our disciplines? Note that despite Nettle and Romaine’s (2000)
characterisation of such questions as the “benign neglect” position, languages and
cultures are nurtured by practice. 15 Practice is fostered by various ecological
factors. Energy may be wasted when the prescriptions to loss of traditions pays no,
or little, attention to these factors.
Typically, as suggested above, speakers do not consciously give up their languages.
15
As Nettle and Romaine formulate it (153), the “benign neglect” position amounts to the following:
“there have always been massive extinctions, so why should we be concerned about the prospect of
another?” Speakers of endangered languages “quite reasonably have more pressing concerns, such
as improving their economic prospects” than worrying about the fates of their languages (153). This
is not of course the position I advocate. We should be concerned with whether linguistically a
particular population is adapting adequately to the changed, or changing, socio-economic ecology
that affects them.
Languages die gradually and inconspicuously as a consequence of the communi-

cative practices of the relevant population, in ecologies where the speakers
themselves can be considered as victims, as they themselves have adapted to
change. We cannot just encourage them to maintain their ancestral languages even
if only as home varieties without providing the ecologies that can support our
prescriptions.
From a purely academic perspective, language shift, endangerment and death are
all part of language evolution. In order to work on them, linguists should, like
environmentalists, better understand the ecology of language evolution and focus
on the real factors that have brought the demise of some languages. The work
should be on those factors and focus should be on the kind of socio-economic
world that can be promoted. In order to convince the parties involved in all these
processes to change their behaviours, we must convince them of the benefits that
humanity, especially the affected populations, can derive by changing their
behaviours. As both languages and cultures are dynamic and constantly (re)shaping
themselves through the behaviours of the populations with which they are
associated, bemoaning ancestral traditions alone will not do the job. Nor does it
sound humanitarian to decry loss of linguistic diversity in the interest of research
on the architecture of universal grammar, about which any kind of variation, old or
new, is likely to be informative.
References
ANS, A.-M. D’. 1996. “Essai de socio- Linguistic Politics. Oxford: Oxford
linguistique historique à partir d’un University Press.
témoignage inédit sur l’emploi des langues,
notamment africaines, en Haiti, au cours de la CAVALLI-SFORZA, L. L. 2000. Genes, Peoples,
guerre de libération et des premières années and Languages. New York: North Point
de l’indépendance”. Etudes Créoles 19: 110– Press.
22. CHAUDENSON, R. 2001. Creolization of
BERLIN, I. 1998. Many Thousands Gone: The Language and Culture, revised in
First Two Centuries of Slavery in North collaboration with Salikoko S. Mufwene.
America. Cambridge, Mass.: Harvard London: Routledge.
University Press. CROSBY, A. W. 1986. Ecological Imperialism:
BOURDIEU, P. 1991. Language and Symbolic The Biological Expansion of Europe, 900–
Power. Cambridge, Mass.: Harvard 1900. Cambridge: Cambridge University
University Press. Press.
BRENZINGER, M., ed. 1998. Endangered CRYSTAL, D. 2000. Language Death.

Languages in Africa. Cologne: Rüdiger Cambridge: Cambridge University Press.
Köper Verlag. DECAMP, D. 1971. “Toward a generative
BRUTT-GRIFFLER, J. 2002. World English: A analysis of a post-creole speech continuum”.
Study of its Development. Clevedon, UK: In: Dell Hymes, ed., Pidginization and
Multilingual Matters. Creolization of Languages, 349–70.
Cambridge: Cambridge University Press.
CALVET, L.-J. 1998. Language Wars and
DEGRAFF, M. 2002. Linguists’ Most Dangerous
Myth: The Fallacy of Creole Exceptionalism. MAFFI, L., ed. 2001. On Biocultural Diversity:
Ms. Linking Language, Knowledge, and the
Environment. Washington, D.C.: Smithsonian
DEJEAN, Y. 1993. “An overview of the Institution Press.
language situation in Haiti”. International
Journal of the Sociology of Language 102: MANESSY, G. 1996. “Réflexions sur les
73–83. contraintes anthropologiques de la
créolisation: de l’improbabilité du métissage
DIXON, R. M. W. 1997. The Rise and Fall of linguistique dans les créoles atlantiques
Languages. Cambridge: Cambridge exogènes”. Etudes Créoles 19: 61–71.
University Press.
MARTINET, A. 1986. Des steppes aux océans:
FISHMAN, J. A. 2000. “Why is it so hard to save l’indo-européen et les “Indo-Européens”.
a threatened language? (A perspective on the Paris: Payot.
cases that follow)”. In: Joshua A. Fishman,
ed., Can Threatened Languages Be Saved?, MAYOR, F. and J. BINDÉ. 2001. The World
1–22. Clevedon, UK: Multilingual Matters. Ahead: Our Future in the Making. Paris:
UNESCO Publishing; London/New York:
FRIEDMAN, TH. L. 1999. The Lexus and the Zed Books.
Olive Tree. New York: Farrar, Straus and
Giroux. MAZRUI, A. and A. MAZRUI. 1998. The Power
of Babel: Language in the African
GREEN, D. H. 1998. Language and History in Experience. Oxford: James Currey; Chicago:
the Germanic World. Cambridge: Cambridge University of Chicago Press.
University Press.
MUFWENE, S. S. 2000. “Creolization is a social,
GRENOBLE, L. A. and L. J. WHALEY. 1998. not a structural, process”. In: Ingrid
“Toward a typology of language Neumann-Holzschuh and Edgar Schneider,
endangerment”. In: Lenore A. Grenoble and eds., Degrees of Restructuring in Creole
Lindsay J. Whaley, eds., Endangered Languages, 65–84. Amsterdam: John
Languages: Current Issues and Future Benjamins.
Prospects, 22–54. Cambridge: Cambridge
University Press. MUFWENE, S. S. 2001. The Ecology of
Language Evolution. Cambridge: Cambridge
HAGÈGE, C. 2000. Halte à la mort des langues. University Press.
Paris: Editions Odile Jacob.
MUFWENE, S. S. In press. “Language
HOENINGSWALD, H. M. 1989. “Language endangerment: What have pride and prestige
obsolescence and language history: Matters got to do with it?” In: Brian Joseph, Johanna
of linearity, leveling, loss, and the like”. In: DeStefano, Neil G. Jacobs and Ilse Lehiste,
Nancy C. Dorian, ed., Investigating eds., When Languages Collide. Columbus,
Obsolescence: Studies in Language Ohio State University Press.
Contraction and Death, 347–54. Cambridge:
Cambridge University Press. MÜHLHÄUSLER, P. 1996. Linguistic Ecology:
Language Change and Linguistic Imperialism
International Journal on Multicultural Societies. in the Pacific Region. London: Routledge.
2001. The Human Rights of Linguistic
Minorities and Language Policies. 3 (2). NETTLE, D. 1999. Linguistic Diversity. Oxford:
Oxford University Press.
KRAUSS, M. 1992. “The world’s languages in
crisis”. Language 68: 4–10. NETTLEE, D. and S. ROMAINE. 2000. Vanishing
Voices: The Extinction of the World’s
LADEFOGED, P. 1992. “Another view of Languages. Oxford: Oxford University Press.
endangered languages”. Language 68: 809–
11. NEWMAN, J. L. 1995. The Peopling of Africa: A
Geographic Interpretation. New
MAFFI, L. 2000. “Linguistic and biological Haven/London: Yale University Press.
diversity: The inextricable link”. In: Robert
Phillipson, ed., Rights to Language: Equity, NEWMAN, P. 1998. “We has seen the enemy and
Power, and Education, 17-22. London: it is us: The endangered languages issue as a
Lawrence Erlbaum Associations. hopeless cause”. Studies in the Linguistic
Sciences 28 (2): 11–20. Press.

NEWMAN, P. Forthcoming .“The endangered SAPIR, E. 1921. Language: An Introduction to
language issue”. In: Mark Janse and Sijmen the Study of Speech. New York: Harcourt,
Tol, eds., Language Death and Language Brace & World.
Maintenance. Amsterdam: John Benjamins.
WARNER-LEWIS, M. 1996. Trinidad Yoruba:
NURSE, D. and TH. SPEAR, . 1985. The Swahili: From Mother Tongue to Memory.
Reconstructing the History and Language of Tuscaloosa: University of Alabama Press.
an African Society, 800–1500. Philadelphia:
University of Pennsylvania Press. WOLFRAM, W. and N. SCHILLING-ESTES. 1995.
“Moribund dialects and the endangerment
RENARD, R. 2001. Une éthique pour la canon: The case of the Ocracoke Brogue”.
francophonie. Questions de politique Language 71: 696–721.
linguistique, 2nd edition, revised, 405 p.
Mons, CIPA. YUE-MAN YEUNG. 2000. Globalization and
Networked Societies: Urban-Regional
RENFEW, A. C. 1988. Archaeology and Change in Pacific Asia. Honolulu: University
Language: The Puzzle of Indo-European of Hawai’i Press.
Origin. Cambridge: Cambridge University
About the Author

Salikoko S. Mufwene is a Distinguished Service Professor of Linguistics at the
University of Chicago. He is the author of The Ecology of Language Evolution
(Cambridge University Press 2001) and is editor of the series “Cambridge
Approaches to Language Contact”. His current primary research area is the
development of creoles, which he approaches from a population-genetics
perspective – which he has extrapolated to the evolution of other languages. He has
(co)-edited several books and authored numerous essays on various linguistic
topics. For more information, visit his website at:
http://humanities.uchicago.edu/faculty/mufwene/.
Address: Department of Linguistics, University of Chicago, 1010 East 59th Street,

Chicago, IL 60637, USA; email: [email protected]
The Impact of Language Policy on Endangered
Languages
SUZANNE ROMAINE
Merton College, University of Oxford
Evaluation of the potential and actual impact of language policy on

endangered languages is complicated by lack of straightforward
causal connections between types of policy and language maintenance
and shift, as well as by confusion of policy and planning. Language
policy is not an autonomous factor and what appears to be ostensibly
the “same” policy may lead to different outcomes, depending on the
situation in which it operates. Weak linkages between policy and
planning render many policies ineffective. Conventions and treaties
adopted by international organisations and agencies recommending
the use of minority languages in education usually lack power to
reinforce them. Furthermore, policies have negligible impact on home
use, which is essential for continued natural transmission of
endangered languages. Although survival cannot depend on
legislation as its main support, legal provisions may allow speakers of
endangered languages to claim some public space for their languages
and cultures.
F ewer than 4 per cent of the world’s languages have any kind of official status
in the countries where they are spoken. The fact that most languages are
unwritten, not recognised officially, restricted to local community and home
functions, and spoken by very small groups of people reflects the balance of power
in the global linguistic market place. Campaigns for official status and other forms
of legislation supporting minority languages often figure prominently in language
revitalisation efforts, despite the generally negative advice offered by experts on
their efficacy. As Fishman (1997, 194) has pointed out, endangered languages
become such because they lack informal intergenerational transmission and
informal daily life support, not because they are not being taught in schools or lack
official status. Nevertheless, because official policies banning or restricting the use
of certain languages have been seen as agents of assimilation, if not also by some
such as Skutnabb-Kangas (2000) as tantamount to acts of genocide, it is no wonder
that hopes of reversing language shift have so regularly been pinned on them.
Skutnabb-Kangas (2000, 312), for example, maintains that “unsupported
coexistence mostly ... leads to minority languages dying”.
195 Suzanne Romaine
Nevertheless, we have here a good example of unwarranted and simplistic

conclusions being drawn about causal relationships between language and policy,
if not outright confusion of policy and planning. As Benton (1999, 23) so aptly puts
it, “there is a difference between permission to speak, and actually speaking”.
Basque speakers in Spain’s Basque Autonomous Community (BAC) have been
hesitant to use their language in relations with the administration not because they
are not allowed to, but because they have difficulty in doing so. A long history of
dealing with officialdom in Spanish and lack of education in Basque leaves most
ordinary people unfamiliar with the newly coined terminology used in this domain
(Gardner 1999).
Likewise, McCarty and Watahomigie (1998, 321) observe that “in practice,
language rights have not guaranteed language maintenance, which ultimately
depends on the home language choices of native speakers. Such decisions are
notoriously difficult for extra-familial institutions to control, even when those
institutions are community controlled”. Nettle and Romaine (2000, 39–40) warn in
a similar vein that “conferring status on the language of a group relatively lacking
in power doesn’t necessarily ensure the reproduction of a language unless other
measures are in place to ensure intergenerational transmission at home. ...
conferring power on the people would be much more likely to do the trick”.
Looking to schools and declarations of official status to assist endangered

languages is much like looking for one’s lost keys under the lamp-post because that
is where the most light appears to shine rather than because that is where they have
been lost. Just as it is easier to see under the lamp-post, it is far easier to establish
schools and declare a language official than to get families to speak a threatened
language to their children. Yet only the latter will guarantee transmission. This
points to the negligible impact of official language policies on home use. Strubell
(2001, 268) notes that “the way people bring up their families – including the
language they choose – is not for the authorities to decide”. In any case, these acts
fall short of what is required in practical terms if the language is to survive in
spoken everyday use.
Many language-policy statements are reactive ad hoc declarations lacking a

planning element. The Native American Languages Act (NALA) of 1990 is one of
the most explicit statements on language ever issued by the United States
Congress, yet it is a classic example of a policy with no planning dimension.
Among other things, NALA states that “the United States has the responsibility to
act together with Native Americans to ensure the survival of these unique cultures
and languages” and “to preserve, protect, and promote the rights and freedom of
Native Americans to use, practice, and develop Native American languages”. As
Schiffman (1996, 246) observes, now that the languages are practically extinct and
pose no threat to anyone, we can grant them special status. Those who think that
NALA is a pro-active policy rather than a recommendation lacking means of
enforcement just because it is written and carries the grand name of “act” deceive
themselves. However, this does not mean that policy is totally useless. As Lucas
The Impact of Language Policy on Endangered Languages 196
(2000) points out in a quite different context (that of assessing the legal status of
Hawaiian), the 1978 state constitutional amendments declaring Hawaiian and
English as the state’s official languages may provide language advocates with the
tools to compel the state to take various measures to support Hawaiian, but they
must be tested in court. No state courts have yet interpreted the legal implications
of these provisions.
In this article I examine some of the obstacles faced in evaluating language policies
and some examples of weak linkages between policy and planning which render
ineffective most policies aimed at assisting endangered languages.
1. Evaluating Policies and the Fallacy of Autonomy

The ideal way to evaluate language policies in a systematic fashion would be to
control all the independent variables but one, and examine the consequences.
Needless to say, in practice, things are otherwise. Evaluation of the efficacy of
policies is made difficult, if not impossible, by the existence of almost as many
variables as there are polities and policies as well as the lack of congruence
between the sociolinguistic condition of the group in question and the language
policy (see Schiffman 1996, 26). A plethora of interlocking factors make it difficult
to discern any direct relationship. Bourhis (2001, 114), for example, says that
“cause and effect relationships are difficult to establish when evaluating the impact
of language policies on language behaviour and language shift”.
At first glance, a number of typologies of language policies appear to offer some

guidance through the entangled thicket (see e.g. Cobarrubias 1983, Skutnabb-
Kangas 2000), but upon closer examination we are forced to conclude that
language policy is not an autonomous factor. As Conversi (1997, 1) puts it in a
different context, “no country’s politics exists independently of its culture”. What
is ostensibly the “same” policy may lead to different outcomes, depending on the
situation in which it operates. Strubell’s (1999, 27–8) comparison of the status of
Catalan in Catalonia and Valencia is an insightful case in point. He concludes that
“the same degree of devolution granted to Catalans and Valencians ... has not led to
the same increase – or rather recovery – in the use of the (same) language” (1999,
26).
Carrington (1997, 88) furthermore notes how change of status can be used as a
political instrument to neutralise those pressing for recognition of their language by
reducing the rallying power of their cause. Amery (2000, 231) suggests that
Australia’s adoption of a “softer approach to language and culture by the federal
government may be a trade-off for their hardline stance on land matters – a partial
compromise which directs some additional resources to those areas which do not
pose a direct threat to the economic interests of the rich and powerful”. After years
of suppressing the indigenous languages of New Caledonia, France provided
financial support to encourage their use in education. This was clearly part of an
attempt to promote peace with militant Kanaks who have long struggled against
197 Suzanne Romaine
French control, and to mitigate anti-French sentiment in advance of a referendum

on independence.
Elevation in status of a previously unrecognised or unsupported minority language

or efforts to extend its use to new domains may also trigger backlash from speakers
of the dominant language, as in Spain where Spanish nationalists have protested
against legislation in Catalonia requiring knowledge of Catalan for certain jobs. In
the Basque Autonomous Community, similar efforts to “normalise” the use of
Basque in education and government through legal measures prompted battles over
the rights of individuals. The 1982 Basic Law for Normalising Basque Language
Use made the right to use Basque an individual rather than a territorial right. The
declaration of officiality, however, was challenged by the Spanish Constitutional
Court, which declared that it could not affect bodies of the Spanish Government
operating in the BAC. More recently in December 2000, the Navarre Government
passed an Autonomous Decree regulating the use of Basque in public
administrative bodies. One result is that knowledge of Basque has ceased to be a
requirement for many public-service positions. The government has justified the
decree as a corrective measure in face of discrimination suffered by Spanish
speakers. Meanwhile, bilingual road signs, advertisements and other public notices
are being replaced with Spanish monolingual ones (Peña 2001, 9). Yet another
example comes from Peru, where Quechua was made co-official with Spanish in
1975, with provision made for Quechua to be taught at all levels from 1976, and
from 1977 for it to be used in court actions involving Quechua speakers. Again,
resistance from the Spanish-speaking majority made implementation difficult, and
it has fallen far short of its ambitions.
Magga and Skutnabb-Kangas (2001, 26) underline similar difficulties in

implementing the provisions of the Saami Language Act passed in 1992 in
Norway, which designated certain areas as Saami administrative districts. Many of
the municipalities outside these districts withdrew services in Saami, claiming that
the law did not require them. Even in traditional Saami areas, where there may be
one Norwegian speaker in a class, it is assumed that all teaching must be done in
Norwegian. When teachers have used Saami in such contexts, allegations of
discrimination against Norwegians ensued. Magga and Skutnabb-Kangas attribute
such actions to a culture clash between the Saami community’s collective right to
develop their language and the right of individual Norwegian speakers. The choice
to use Saami is thus politicised and restricted territorially.
Fishman (1991, 84) writes of the damage, both locally and beyond, done by
previously disadvantaged language activists who become “cultural imperialists”
themselves within their newly dominated networks. When Quebec francophones
adopted various legislative measures designed to protect French, in particular a
requirement for newcomers to learn it and direct financial incentives to increase the
birth rate, anglophones felt threatened. Bill 101 mobilised anglophones to mount
legal challenges and to boycott Montreal stores with French monolingual signs; by
1988 the Canadian Supreme Court ruled that the legal requirement for French-only
signs contravened both the Quebec and Canadian Charter of Rights and Freedoms.
Quebec’s linguistic laws also stirred up much negative feeling among anglophones
outside the province as well as outside Canada. Bourhis (2001, 133) observes how
the English Only movement in the US regularly uses controversial features of
Quebec language laws to justify its campaigns against minority language
maintenance. Nevertheless, he sees democratically adopted language laws as
necessary tools allowing modern states to harmonise class and ethnic conflicts (see
also Kymlicka 1995, 2000).
In proposing such measures, Quebec’s francophones sought no more than to

guarantee for themselves similar “rights” to control their own reproduction that
anglophone Canadians have felt unnecessary to state as policy because they were
implicit in practice anyway. Quebec anglophones, in particular, benefited from the
provision of a state-financed English-medium education system ranging from pre-
kindergarten through university, to an extent rarely granted a linguistic minority
elsewhere (except perhaps to the Swedish-speaking minority in Finland). What
amounted to an affirmative action plan for Quebec anglophones passes unnoticed
because it is regarded as “normal”. In this way, all nations unavoidably promote
and support the languages sanctioned for use in education, at the same time as they
marginalise other languages denied the same public space.
This reminds us not to overlook the fact that policy is implicit even if no specific
mention is made of language. Probably most majority languages dominate in many
domains where they have only de facto and no legal status. As Fishman (2001a,
454) comments, “even the much vaunted ‘no language policy’ of many
democracies is, in reality, an anti-minority-languages policy, because it delegiti-
mizes such languages by studiously ignoring them, and thereby, not allowing them
to be placed on the agenda of supportable general values”. Proponents of what is
sometimes called “benign neglect” ignore the fact that minorities experience
disadvantage that majority members do not face.
Advocates of minority languages have repeatedly stressed that demographically

weak languages need firm pro-active policies in order to survive and thrive (see
e.g. Strubell 1999, Skutnabb-Kangas 2000). Yet the legal approach to reconciling
status differences in languages with equality in a world where majority rights are
implicit, and minority rights are seen as “special” and in need of justification, is
fraught with difficulty. Magga and Skutnabb-Kangas (2001, 31) emphasise that
“equality is misunderstood if it leads to an equal division of time and resources
between a minority and a majority language”. As Hickey (2001, 466–7) has
observed in connection with Irish immersion pre-schools, “equal treatment of
different children does not necessarily mean the same treatment is given to each
child”. Thus, there is an important distinction between legislating equal use of
languages and guaranteeing equitable treatment of their speakers, a point to which I
return below in my discussion of South Africa’s post-apartheid language policy.
In assessing the impact of Quebec’s protective legislation, Bourhis (2001, 115)

199 Suzanne Romaine
says that the 1996 census suggests increasing intergenerational shift towards
French since 1971, although the change can be largely attributed to allophones (i.e.
those whose native language is neither French or English) adopting French as their
home language. In addition, unfavourable reaction to Bill 101 led to anglophone
out-migration.
Schiffman (1996) says that we cannot assess the chances of success of policies
without reference to culture, belief systems, and attitudes about language. The idea
that linguistic rights need protection has never been part of American culture, and
so they have not been seen as central to American courts unless allied with more
fundamental rights such as educational equity, etc. (Schiffman 1996, 216, 246).
Elsewhere, however, even international courts have opined that there is no basic
human right to education in one’s own language. UNESCO’s (1953, 6) much-cited
axiom “that the best medium for teaching is the mother tongue of the pupil ...” did
not lead to any widespread adoption and development of vernacular languages as
media of education. In most parts of the world schooling is still virtually
synonymous with learning a second language.
Although a basic right to education cannot function equitably unless the child
understands the language of instruction, this is of little use to groups whose
nationalities and languages do not “officially” exist (as is the case with the Kurds
and Kurdish in Turkey) or to groups whose language has been so eroded by shift
that their children do not speak it. A case in point is that of Hawaiian, where the
Board of Education’s official position is that Hawaiian immersion schools
constitute a programme of choice and not of right within the public school system.
Hence it has refused to recognise an affirmative duty to provide adequate funding
for Hawaiian-medium schools for children desiring education through the medium
of Hawaiian.
2. Weak Linkages
Although Fishman (2001a, 478) admits that conclusive evidence is lacking at both
the state and international level to evaluate the efficacy of policies, he believes that
“there is no reason to be overly optimistic in either case, because a lack of priorities
and linkages seems to characterise the entire legalistic approach”. He does,
however, advocate monitoring certain “litigious climates” surrounding languages
such as Maori and Frisian in order to gauge the likelihood and the circumstances
needed for legislation, and various other legal measures to be able to make a
practical difference in language revitalisation efforts. Even if such actions do make
a difference, Fishman warns that they must still be distinguished from the possible
effects of the conventions and treaties adopted by international agencies and
organisations lacking the power to enforce their resolutions.
A good example of weak linkages is the European Charter for Regional or

Minority Languages, created to provide a legal instrument for the protection of
languages. Although it specifies no list of actual languages, the languages
concerned must belong to the European cultural tradition (which excludes

“migrant” languages), have a territorial base, and be separate languages identifiable
as such. The terms of reference are deliberately vague in order to leave open to
each member state how to define cultural heritage and territory. Thus, each state is
free to name the languages which it accepts as being within the scope of the charter
(see the issue of International Journal on Multicultural Societies on “Lesser Used
Languages and the Law in Europe” 2001a; and Ó Riagáin 1998, 2001, for
information in the status of languages in the European Union and a summary of
legislation relating to minority languages). The UK, for example, which ratified the
treaty in March 2001, does not include Manx and Cornish. The effectiveness of any
initiatives on the supranational level can always be undermined by individual states
unless there is some way of guaranteeing the implementation of language-related
measures on a supranational level. The only institutions with authority to regulate
language policies exist within the political bodies of individual states, and the
European Union has generally avoided taking any action that would interfere with
national laws or policies concerning linguistic minorities, or for that matter with
laws concerning its national languages. Moreover, the charter does not grant rights
to speakers or minority language groups, but to languages.
Despite the fact that Greece is signatory to many international covenants and
treaties on human rights, as well as a member of the European Union, it voted
against the European Charter for Regional or Minority Languages in 1992. In July
1995 Sotiris Bletsas, a member of the minority Aroumanian (Vlach) community,
was arrested after he distributed publications of the European Bureau for Lesser
Used Languages which mentioned the existence of the Aroumanian language and
four other minority languages in Greece (Arvanitika, Macedonian, Turkish and
Pomak). The police obliged him to make a statement saying that he was Greek. As
a result of charges brought by Mr Haitidis, a right-wing Member of Parliament of
the New Democracy party, Bletsas was convicted under Article 191 of the Greek
Penal Code which states that dissemination of false information could create fear
and unrest among Greek citizens and damage the country’s international relations.
The European Court of Human Rights had already ruled that this article was in
violation of the European Convention on Human Rights, but an Athens court gave
Bletsas a 15-month sentence (suspended) and a fine. After several postponements
of his appeal, much international pressure, and concern expressed to the Greek
Government by the EU Commissioner for education and culture, among many
others, Bletsas was finally acquitted in December 2001 by unanimous decision of
the Athens Three-Member Appeal Court (see http://www.eurolang.net for coverage
of this case). Meanwhile, Turkey, an aspiring member of the European Union, still
maintains that it has no minorities.
Most European nation-states still apply one set of rules to the national language
and another to minority languages within their boundaries, and often in addition
apply differing standards to indigenous and non-indigenous minorities (see
Romaine 1998). Similarly, New Zealand has progressed in its treatment of Maori
language issues, while it has lagged behind in recognition of the rights of migrant
201 Suzanne Romaine
Pacific-islander communities.
Differing practices within different regions of the same country, and with respect to
different minority groups, add a further dimension to the vexed problems of
evaluation and implementation. The effects of policy proposed at the national level
can be complex, depending on political structures. In Australia, for example, the
1990 National Language Policy did not really challenge the dominance of white
anglophone society after centuries of assimilation and restrictive immigration
practices (see Romaine 1991, 1994). Fishman (2001a, 479) offers a more recent,
but equally pessimistic assessment, and Lo Bianco and Rhydwen (2001, 417) say
that community language maintenance has been relegated to a subordinate status
with insufficient resources to sustain the few token acclamations remaining in the
policy.
Clyne (2001, 386) points out how individual states subsequently developed vastly
different policies, and chose different priority languages. Lo Bianco and Rhydwen
(2001, 404) explain how the “second languages policy” of the Northern Territory
Government’s Department of Education serves only as a recommendation to
schools and does not cover the specific needs of Aboriginal communities. Neither
do its Social and Cultural Education guidelines cover the kinds of programmes that
Aboriginal people want to implement. The lack of strong policy support has meant
that Aboriginal language and culture programmes have not achieved a secure place
in the schools. In 1998 the Northern Territory abandoned public funding for
indigenous bilingual education, which had originally been established by the
Commonwealth Government when education in the Northern Territory was under
its jurisdiction.
Lo Bianco and Rhydwen (2001, 418–19) conclude that policy can lead to change in
the ongoing trend of attrition and extinction if control of resources and the means
for decision-making, as well as the institutional domains where language sociali-
sation occurs, are in the hands of those affected. They doubt whether Aboriginal
Australians will be given the space for self-determination and regulation to a
sufficient degree.
Benton and Benton (2001) contrast the Kura Kaupapa Maori (a special category of
New Zealand state schools with a Maori language and culture orientation) with the
Ataarangi movement aimed at the Maori language needs of whole families, which
works through homes rather than schools. Because the latter receives no govern-
ment support, it is not subject to government controls. Attempts to manage the
Kura Kaupapa Maori at government level have been divisive. The Council
governing these schools lobbied the House of Representatives in 1998 for a bill to
require schools seeking designation as Kura Kaupapa Maori to subscribe to a
particular set of philosophical principles. Not all communities favoured this move,
prompting Benton and Benton (2001, 436) to comment that it remains to be seen
whether what were originally independent schools will come under the ideological
control of a group selected by the state to enforce a “legislatively defined” Maori
world view.
Returning to NALA as an instance of weak linkage between policy and implemen-

tation, we can see that it also illustrates how lack of federally mandated language
planning has led to a hodgepodge of policies potentially in conflict with state, local
or other federal rules. Legislation in Arizona and Hawaii provides at least two
examples.
Hawaii is the only state with an official language in addition to English. Article
XV, Section 4, states that “English and Hawaiian shall be the official languages of
Hawaii”. A second amendment (Article X, Section 4) contains a provision “to
revive the Hawaiian language, which is essential to preservation and perpetuation
of Hawaiian culture”. Lucas (2000, 13) mentions two cases involving legal claims
brought under the auspices of NALA, both initiated by native Hawaiians. In
Tagupa v. Odo (1994), attorney William Tagupa refused to give his deposition in
English, despite his fluency in the language, on the grounds that Article XV,
Section 4, of the state Constitution and NALA prohibit federal courts from
mandating that deposition testimony be made in English.
In rejecting Tagupa’s claim, the federal district judge argued that the intention of
NALA was directed at increasing the use of Native American languages in
education and not at judicial proceedings in federal courts. He also quoted
President Bush’s remarks on signing NALA into law to the effect that it was
construed as a “statement of general policy” and should not be understood as
conferring “a private right of action on any individual or group” (Lucas 2000,
26fn.76). Moreover, the judge opined that allowing deposition in Hawaiian would
be contrary to the Federal Rules of Civic Procedure, which mandate the “just,
speedy and inexpensive determination of every action”, because additional costs
and delays would be needed to appoint an interpreter.
In 1996 the Office of Hawaiian Affairs (OHA) brought a case against the
Department of Education claiming that the department’s failure to provide
sufficient financial and technical support for the Hawaiian immersion programme
was a violation of both state law and NALA. The state removed the suit from state
court to federal court, where the same federal district judge who ruled against
Tagupa also ruled against OHA. He said that NALA does not create affirmative
duties on the states but merely evinced a federal policy to encourage states to
support Native American languages. In 2000 the Department of Education and
OHA reached an out-of-court settlement which provided an additional US$7.5
million to the immersion programmes under a 2:1 funding partnership, with the
state to spend up to a million dollars a year for the next five years.
In November 2000, 63 per cent of Arizona voters passed Proposition 203 to end
bilingual education and replace it with one year of untested English immersion
marketed with the slogan “English for the children”. The proposition was
spearheaded by Ron Unz, who portrays himself as a strong believer in assimilation,
203 Suzanne Romaine
and backed a similar successful initiative in California in 1998 (Proposition 227).

Seeing Proposition 203 as an attack on the languages spoken by Arizona’s Indian
tribes, Arizona State Senator Jack Jackson, a member of the Navajo nation,
requested an Attorney General’s opinion about whether the proposition applied to
the Navajo. In February 2001, Janet Napolitano gave her opinion that it did not
apply to any Arizona Indians living on or off reservations. She invoked “principles
of tribal sovereignty” and NALA’s provision that “the right of Native Americans to
express themselves through the use of Native American languages shall not be
restricted in any public proceeding, including publicly-supported education
programs” (Reyhner 2001, 23).
In Nigeria, also, weak linkages prevent most schools from implementing the
National Policy on Education, which stipulates that pupils’ mother tongues be used
in the lower levels of public education. More importantly, no government sanctions
are applied to schools that do not follow the policy. Indeed, 80 per cent of African
languages lack orthographies (Adegbija 2001) making it difficult to contemplate
their effective use in schools. In Senegal, six African languages (Mandingo, Diola,
Peul/Poular, Serer, Soninke and Wolof) have been declared official, but little effort
has been made to use them in education. Various factors inhibit implementation,
such as lack of funding for materials development, teacher training, parental
anxiety about their children’s acquisition of the dominant language, along with fear
among the elite of losing their status gained through education in the colonial
language. Brenzinger (1998, 95) estimates that fewer than 10 per cent of African
languages are included in bilingual education programmes, with the result that
more than 1,000 African languages receive no consideration in the education
sector.
As Bamgbose notes (1991, 100–1) “the paradox of mother tongue education in

many African countries is that while it is negligible at the primary level, it seems to
flourish at university level”. It is possible in Nigeria, for example, to take a degree
in Hausa, Yoruba and Igbo. International aid agencies, colonial regimes, former
and current, often tie aid packages to economic, social and educational policies that
support and maintain the colonial language. One result is that many minority
languages have more status outside their territories than within them, as is
evidenced in the fact that Quechua is taught in universities in the United States and
elsewhere.
The new democratic regime in South Africa has recognised the linguistic reality of
multilingualism that had been ignored under apartheid. Henrard (2001) points to a
difference between the 1993 Interim Constitution containing a proclamation
promoting the state’s “equal use” of eleven official languages (among them nine
indigenous languages plus the colonial languages Afrikaans and English) and the
1996 Constitution aiming at “equitable treatment” and “parity of esteem” of the
official languages. The need for differential and preferential treatment of the
indigenous languages, given the past history of denigration and discrimination, was
recognised in the stipulation that the state must take practical and positive measures
to elevate their status and advance their use. More specifically, the national
government and provincial governments must use at least two official languages.
Nevertheless, this case also shows the difficulty in attempting to enforce equality
of use and status legislatively among a number of languages unequal in social
practice. Despite the Constitutional Court’s proclamation to the contrary, the
elevation in status of nine previously unrecognised indigenous languages has had
the practical effect of diminishing the status of Afrikaans, just as the National Party
feared. In practice, the public life of the country has actually become more
monolingual (Webb 1998). Afrikaans, which no longer enjoys legal and political
protection as a language co-official with English, has experienced dramatic losses,
one of the most visible in the area of television, where it formerly shared equal
time with English. The new broadcasting time is now more than 50 per cent for
English, while Afrikaans, Zulu and Xhosa get just over 5 per cent each. Although
greater emphasis is to be given to languages heretofore marginalised and more than
20 per cent of broadcasting time is supposed to be multilingual, in practice this
time has been taken up mostly by English. Similarly, the South African National
Defence Force, which formerly used Afrikaans, declared in 1996 that English
would be the only official language for all training and daily communication. The
demand for English among pupils and parents also works against implementing
multilingualism in education (Kamwangamalu 1998).
These examples show that without additional measures to support teacher training,
materials development, and a variety of other enabling factors, policy statements
which merely permit, encourage, or recommend the use of a language in education
or in other domains of public life cannot be very effective. Political ideology drives
policy in particular directions, creating various divergences between stated policy
and actual practice. Lo Bianco and Rhydwen (2001, 416–17) point out how in
Australia the low achievement of Aboriginal children in English literacy is used to
justify eliminating bilingual education, just as it is in the United States.
Gardner-Chloros (1997, 217) writes that lawyers agree that “the only way to
guarantee fundamental rights effectively is to restrict declarations as to what these
rights consist in to the most basic and incontrovertible one”. In other words, it is
pointless to think that “grand declarations of policy ... would be effective if they
are not tied to a – preferably existing – legal instrument with an effective
machinery for reinforcement”. An interesting case of a grand declaration with no
such ties is Eritrea’s 1995 declaration not to recognise an official language. Thus,
President Isayas Afewerki (Brenzinger 1998, 94):
“When we come to the question of language as a means of instruction in schools,

our principle is that the child should use its mother tongue or a language chosen
by its parents in the early years of its education, irrespective of the level of
development of the language.
Our policy is clear and we cannot enter into bargaining. Everyone is free to learn
in the language he or she prefers, and no one is going to be coerced into using this
or that ‘official’ language.”
205 Suzanne Romaine
In the case of Hawaiian, however, Lucas (2000, 17–19) suggests that a strategic
opportunity lies in Article XII, Section 7, of the state Constitution, which enjoins
the state to “protect all rights, customarily and traditionally exercised for
subsistence, cultural, and religious purposes”. The Hawaii State Supreme Court has
already held that this imposes an affirmative duty to protect and perpetuate
traditional and customary practices. Although Lucas is doubtful whether this article
would create a means of forcing increased funding of immersion schools, language
activists might get the court to recognise the speaking of Hawaiian as a traditional
and customary practice.
The case made for Maori under the provisions of the Treaty of Waitangi is
instructive. In 1974 a largely decorative amendment to the Maori Affairs Act
“officially” recognised the Maori language as “the ancestral language of the
population of Maori descent”. While it allowed the Minister of Maori Affairs to
take such steps as were considered appropriate to the encouragement of the
learning of the language, it had no practical effect until five years later, when it
became clear that this statement meant nothing in the courts when an appellant
claimed the right to address the District Court in Maori and was refused. The High
Court upheld the ruling on the basis of the Pleadings in English Act of 1362, which
became part of New Zealand law by virtue of the English Laws Act of 1858 when
the New Zealand legislature adopted all the laws of England in force on 14 January
1840. Ironically, the 1362 statute was passed at a time when the official language
of court proceedings in England was French. The High Court’s decision came as a
disappointment to those activists who had seen legislation as a way of
strengthening the position of Maori. It was not until 1987 that an act made Maori
an official language of New Zealand and established Te Taura Whiri i Te Reo
Maori (the Maori Language Commission).
Maori activists have seen the efficacy of linking the struggle for language rights
with natural resource management and preservation provisions guaranteed to them
in the Treaty of Waitangi of 1840 signed by Maori chiefs and the British. In 1975
the Waitangi Tribunal was created to consider Maori grievances over breaches of
the treaty. Although the British regard Maori assent to the treaty as the basis for
their sovereignty over New Zealand, there are numerous complicating factors
surrounding the treaty and its language which make its interpretation and legal
status fraught with difficulties. The terms of the Maori version of the treaty
guaranteed to the Maori te tino rangatiratanga o ratou wenua o ratou kainga me o
ratou taonga katoa, which could be translated as “the full authority of chiefs over
their lands, villages, and all their treasures”. Maori activists interpret this as a
guarantee rather than cession of Maori sovereignty and have pressed land claims as
well as support for the Maori language. The Crown acknowledged Maori claims
that the treaty obliged it not only to recognise the Maori language as a part of the
country’s national heritage and a treasured resource on a par with lands but to
actively protect it. We have here another instance in which legislation and practices
at one level (the English Laws Act) are in conflict with those at another (the Treaty
of Waitangi). Recognition that the Crown had broken its promise required
affirmative action rather than passive tolerance. As Benton (1985) said in

testimony before the tribunal, “rights which cannot be enforced are illusory, and
protection which cannot sustain life is no protection”. This points to the need for
strong linkages between language policy and economic planning, which have
generally been lacking. Benton (1999, 7), for example, recognises that the
revitalisation of the Maori language is primarily a matter of sustainable cultural and
economic development.
Unfortunately, potential sources of support for Maori language activities have felt
that tribal resources should not be used to subsidise what is regarded as state
responsibility under the terms of the treaty. Benton and Benton (2001, 439) write
that if it could be shown that supporting Maori increased tribal monetary wealth,
Maori Trust Boards and land corporations might feel more inclined to give it
priority.
3. Timing: Too Little Too Late?

Much probably depends on the timing of policies and legislation. Planning in many
domains, linguistic or otherwise, faces inevitable charges of “too little too late”.
Few communities are concerned about language transmission when all is
proceeding normally, and even when it is not, various factors impede recognition
of the impending loss and its consequences. In Quebec, however, Bourhis (2001,
105, 111) says that language planners were well placed to intervene in the 1970s in
favour of French with strong intergenerational transmission on their side, even
though a sociolinguistic analysis would have led to the conclusion that such
planning was unnecessary. More than 80 per cent of the population had French as a
mother tongue and more than three-quarters were monolingual French speakers.
Moreover, francophones controlled most of the provincial administration, even
though they lacked control of the major business and financial institutions. It was
the threat to French survival in the long term in the face of declining birth rate and
increased immigration of anglophones and others likely to assimilate to the
anglophone population that provided the ideological impetus to mobilise. In many
other cases, however, where erosion is painfully evident, communities may not
recognise or wish to confront the impending loss, or feel that other concerns are
more pressing.
One can contrast the case of Quebec with that of Irish, with its far weaker
demographic base for reproduction of the language, where similarly aggressive
legislative policies in favour of Irish have not significantly reversed language shift.
Only 18 per cent (actually an overestimate) of the population was reported to be
Irish-speaking in the 1926 census, just after the foundation of the Irish state. The
newly independent government in 1922 promoted policies directed at altering the
linguistic market, to enhance the social and legal status of Irish by declaring it the
national language, to maintain it where it was spoken, and to extend its use
elsewhere. Irish was required in public administration, law and media, domains in
which it had not been used for centuries. Ó Riagáin (1997, viii), however, says that
207 Suzanne Romaine
the problem was not the small demographic base, but rather the social distribution
of Irish, confined as it was to peripheral rural communities. The state had hoped,
not unreasonably, that by supporting the agricultural sector, it would support Irish,
whose speakers were primarily engaged in farming.
Meanwhile, schools were supposed to replace English with Irish as a medium of

education. The policy of Gaelicising the schools was increasingly effective from
the 1920s up to the 1950s, at which point just over half the state primary schools
were offering an immersion programme of a full or partial type (i.e. teaching all or
part of the curriculum through Irish to children whose mother tongue was English).
Subsequently, the amount of bilingual or all-Irish education declined. Public
opinion polls conducted in the 1960s indicated that compulsory Irish instruction
was not popular. Even in the 1930s many teachers were opposed to teaching
English-speaking children through the medium of Irish (Ó Riagáin 1997,19, 31).
By the early 1960s, when it was clear that supporting agriculture was not working
to stem out-migration and the viability of farming, economic policy shifted to
encourage small industry and the export market. By the 1970s, however, when
more young people began to look towards education for upward mobility, state
language policies had shifted so that Irish ceased to be a compulsory subject in
public examinations at the end of secondary schooling. Hence, incentives for
achieving Irish competence were weakened at a time when they were needed. In
the earlier period, relatively few young people were affected by the incentives for
Irish built into the education and civil service sectors. Ó Riagáin’s analysis
underlines a disjunction between economic policy and language policy before the
1960s and after the 1970s.
Over the past few decades the thrust of policy, in so far as there is any explicit
statement of it, has been towards maintenance rather than restoration. Official
rhetoric has shifted meanwhile to talking of survival rather than revival. Today the
largest proportion of Irish speakers is to be found among those between 10 and 20
years old. Ó Riagáin’s (1997, 283) sobering assessment, based on his examination
of a century of language policy in Ireland, reveals how timing enters into the
equation in another sense too:
“Language patterns are but aspects of highly complex social systems. They are
the outcome of slow, long-term processes. If language policies are to have any
significant impact, they will require resources on a scale which has not been
hitherto realised. Effective language policies will and must affect all aspects of
national life and will have to be sustained for decades, if not forever.”
4. Factors other than Legal Status

The deficiencies in the formulation and implementation of policy examined by Ó
Riagáin are by no means unique to Ireland, but are typical of language-planning
experiences more generally. They point back again to the autonomy fallacy.
Skutnabb-Kangas (2000, 303) concludes that there is an urgent need for more
research before we can start understanding the importance of various factors in

supporting or not supporting the world’s languages.
Factors other than legal status are often more important. Again, Ó Riagáin (1997,
170–1):
“ … the power of state language policies to produce intended outcomes is

severely constrained by a variety of social, political and economic structures
which sociolinguists have typically not addressed, even though their
consequences are profound and of far more importance than language policies
themselves.”
Carrington (1997, 88) comments that “real status is achieved when official action
confirms an already existing situation in which significant objectives of official
recognition are already operationally in place”. As an example, he cites the
granting of official status to Papiamento in 1985 in the Netherlands Antilles, which
came long after the language was used in newspapers, signs, etc. Likewise,
Gardner (1999, 86) comments that laws regulating language matters are often
limited “to sanctioning what has already become reality or enabling what
sociological dynamics could potentially make reality. What it cannot do in a
reasonably democratic society is fulfil a coercitive [sic SR] function in any major
way”. Any policy for language, especially in the system of education, has to take
account of the attitude of those likely to be affected.
Ó Riagáin (1997, 174, 279) notes that while the compulsory element in pre-1973
policies enhanced the practical or economic value of Irish, many people opposed
them. Although support for Irish was ostensibly high, the public was not prepared
to back policies that would discriminate strongly in favour of Irish and could
potentially alter the linguistic landscape. In his view, the major constraint on policy
development was the absence of sustained public support and not state action per
se (Ó Riagáin 1997, 23).
Strubell (2001) suggests a similar lack of support in Catalonia, while as far as

Basque is concerned, Gardner (1999, 85–7) argues that the need for a monolingual
heartland is paramount, but legally unobtainable. At the same time, he stresses that
“granting monolingual official status to a minority language ultimately affects
prestigious but relatively marginal uses of the language. Declaration by decree of a
monolingual enclave cannot ensure its existence in practice”. He concludes that the
problem is not the limits imposed by present laws, but lack of proper awareness of
priorities by language planners and Basques more generally (Gardner 1999, 88). As
recent legislation in Navarre illustrates, language rights are not timeless
declarations, but are time-limited, subject to shifting political regimes.
Evidence from various quarters indicates that grass-roots initiatives are often more
effective than top-down directives. A case in point is the PROPELCA (Projet de
Recherche Opérationnelle pour l’Enseignement des Langues Camerounaises)
project in Cameroon, one of the best-documented and most complete examples of a
209 Suzanne Romaine
literacy programme which includes materials development, a teacher-training

programme, and evaluation (Gerbault 1997). The project’s working principle was
to use local languages as instruments for scientific and technical training, rather
than to maintain languages per se. Teacher training began in 1981 as an
experimental programme in two Roman Catholic schools; by 1986 there were
eleven experimental schools in four different provinces teaching in four of
Cameroon’s 236 languages. Although the pedagogical approach and its
development by local specialists has been exemplary in sub-Saharan Africa,
Gerbault (1997, 182) says that it has met with lack of involvement of official
institutions typical of this part of the world. Only private institutions and local
communities have supported the programme, even though the project workers have
recommended that existing practices of using Cameroonian languages for the first
three years of primary school should be made official.
Meanwhile, there had been a remarkable recent rise in entirely voluntary Irish-
medium schooling in Ireland with over 150 all-Irish primary and secondary
schools, and more than twice that number of all-Irish pre-schools. In Northern
Ireland, a deliberately created community in Shaw’s Road in urban Belfast, where
parents who were not native speakers of Irish, has succeeded in raising children
who are (Maguire 1991).
5. Conclusion: The Proof is in the Pudding

My assessment of the efficacy of policy in assisting endangered languages has
perhaps been unduly pessimistic in an effort not to minimise the complexity and
enormity of the task. In his reappraisal of the scene ten years after his 1991 book,
Fishman (2001a, 478–9) observes that none of the dozen individual cases studied
in the late 1980s and early 1990s has experienced “dramatic successes”. Naturally,
there are many reasons why that is the case, as the individual chapters show, so it
would be hard to pinpoint policy as the unique cause or source of either success or
failure. Indeed, Fishman’s (2001a, 480) overall conclusion is an ambiguous one:
although the general climate of opinion on threatened languages has improved “in
an amorphous and largely still ineffectual sense”, the prospects for reversing
language shift have not improved much and have even deteriorated.
This does not mean that advocates of linguistic diversity should abandon the
struggle to obtain legal measures at all levels supporting languages. On the
contrary, we must redouble our efforts. However, we must do so in the knowledge
that without well-focused action on a variety of other fronts, these will not
guarantee maintenance. It is political, geographical and economic factors that
support the maintenance of linguistic and cultural diversity. Holistic ecological
planning of the kind advocated by Nettle and Romaine (2000) works towards
international, regional and national policies that empower indigenous peoples and
promote sustainable development. This is the key to preserving local ecosystems
essential to language maintenance. Because the preservation of a language in its
fullest sense ultimately entails the maintenance of the group that speaks it, the
arguments in favour of doing something to reverse language death are ultimately

about preserving cultures and habitats.
Finally, however, the proof is always in the pudding. In the interests of justice, it is
incumbent on liberal democracies to accommodate cultural and linguistic diversity
to the fullest extent possible. Kymlicka (1995) argues that respecting minority
rights is essential for enlarging the freedom of individuals, a cornerstone of liberal
democracy. The issue of language rights has begun to receive serious international
discussion within the last decade (see, for example, Skutnabb-Kangas et al. 1994;
Benson et al. 1998; Skutnabb-Kangas 2000; International Journal on Multicultural
Societies 2001a, 2001b). Although survival cannot depend on legislation as its
main support, legal provisions may allow speakers of endangered languages to
claim some public space for their languages and cultures from which we can all
benefit.
References
ADEGBIJA, E. 2001. “Saving threatened African continent”. In: Kazuto Matsumura,
languages in Africa: A case study of Oko”. ed., Studies in Endangered Languages.
In: J. A. Fishman, ed., op. cit., 284–308. Papers from the International Symposium on
Endangered Languages Tokyo November 18-
AMERY, R. 2000. Warrabarna Kaurna! 29, 1995, 85-100. Tokyo: Hituzi Syobo.
Reclaiming an Australian Language. Lisse:
Swets and Zeitlinger. CARRINGTON, L. D. 1997. “Creoles and other
tongues in Caribbean development”. Journal
BAMGBOSE, A. 1991. Language and the Nation: of Pidgin and Creole Languages 8: 125–35.
The Language Question in Sub-Saharan
Africa. Edinburgh: Edinburgh University CLYNE, M. 2001. “Can the shift from immigrant
Press for the International African Institute. languages be reversed in Australia?” In: J. A.
Fishman, ed., op. cit., 364–91.
BENSON, P., P. GRUNDY. and SKUTNABB--
KANGAS, T. eds. 1998 Special issue on COBARRUBIAS, J. 1983. “Ethical issues in status
Language Rights. Language Sciences 20: 1. planning”. In: Juan Cobarrubias and Joshua
A. Fishman, eds., Progress in Language
BENTON, R. 1985. Testimony. Waitangi Planning: International Perspectives, 41–85.
Tribunal, Department of Justice, Wellington, Berlin: Mouton.
New Zealand.
CONVERSI, D. 1997. The Basques, the Catalans
BENTON, R. 1999. Maori Language Revitali- and Spain. London: Hurst.
zation. Final Report, Wellington, New
Zealand. FISHMAM, J. A. 1991. Reversing Language
Shift. Theoretical and Empirical Foundations
BENTON, R. and N. BENTON. 2001. “RLS in of Assistance to Threatened Languages.
Aotearoa/New Zealand 1989–1999”. In: J. A. Clevedon, UK: Multilingual Matters.
Fishman, ed. op. cit., 423–51.
FISHMAN, J. A. 1997. “Maintaining languages.
BOURHIS, R. Y. 2001. “Reversing language shift What works and what doesn’t”. In: Gina
in Quebec”. In: J. A. Fishman, ed., op. cit., Cantoni, ed., Stabilizing Indigenous
101–42. Languages, 186–98. Flagstaff, Ariz.:
BRENZINGER, M. 1998. “Various ways of dying Northern Arizona University.
and different kinds of death: Scholarly FISHMAN, J. A. 2001a. “From theory to practice
approaches to language endangerment on the (and vice versa): Review, Reconsideration
211 Suzanne Romaine
and Reiteration”. In: J. A. Fishman, ed., op. MAGGA, O. H. and T. SKUTNABB-KANGAS.

cit., 451–83. 2001. “The Saami languages. The present and
the future”. Cultural Survival Quarterly 25
FISHMAN, J. A. ed. 2001b. Can Threatened (2): 26–31, 51.
Languages Be Saved? Reversing Language
Shift, Revisited: A 21st Century Perspective. MAGUIRE, G. 1991. Our Own Language: An
Clevedon, UK: Multilingual Matters. Irish Initiative. Clevedon, UK: Multilingual
Matters.
GARDNER, N. 1999. Basque in Education in the
Basque Autonomous Community. Vitoria- MCCARTY, T. and L. WATAHOMIGIE. 1998.
Gasteiz: Eusko Jaurlaritzaren Argitalpen “Indigenous community-based language
Zerbitzu Nagusia. education in the USA”. Language, Culture
and Curriculum 11 (3): 309–25. Special issue
GARDNER-CHLOROS, P. 1997. “Vernacular on Indigenous Community-Based Education,
literacy in new minority settings in Europe”. edited by Stephen May.
In: Tabouret-Keller et al., eds., op. cit., 189–
221. Native American Languages Act. 1990. P.L.
101–407. Washington, D.C., United States
GERBAULT, J. 1997. “Pedagogical aspects of Congress.
vernacular literacy”. In: Tabouret-Keller et
al., eds., op. cit., 142–85. NETTLE, D. and S. ROMAINE. 2000. Vanishing
Voices. The Extinction of the World’s
HENRARD, K. 2001. “Language rights and Languages. New York: Oxford University
minorities in South Africa”. International Press.
Journal on Multicultural Societies 3 (2).
Ó RIAGÁIN, D. ed. 1998. Vade-Mecum. A guide
HICKEY, T. 2001. “Mixing beginners and native to legal, political and other official
speakers in minority language immersion; international documents pertaining to the
Who is immersing whom? Canadian Modern lesser used languages of Europe. Brussels:
Language Review 57 (3): 443–74. European Bureau for Lesser Used Languages.
International Journal on Multicultural Ó RIAGÁIN, D. 2001. “The European Union and
Societies. 2001a. Lesser Used Languages and lesser used languages”. International Journal
the Law in Europe. 3 (1). on Multicultural Societies 3 (1).
International Journal on Multicultural Soci- Ó RIAGÁIN, P. 1997. Language Policy and
eties. 2001b. The Human Rights of Linguistic Social Reproduction. Ireland 1893–1993.
Minorities and Language Policies. 3 (2). Oxford: Oxford University Press.
KAMWANGAMALU, N. M. 1998. “Preface: PEÑA, J. L. M. 2001. “Bilingualism under threat
Multilingualism in South Africa”. in Navarre. Navarre government language
Multilingua 17 (2/3): 119–25. Special issue policy”. Contact Bulletin 17 (2): 8–9.
on Aspects of Multilingualism in Post- Brussels, European Bureau for Lesser Used
Apartheid South Africa. Languages.
KYMLICKA, W. 1995. Multicultural Citizenship. REYHNER, J. 2001. “Cultural survival vs. forced
Oxford: Oxford University Press. assimilation: the renewed war on diversity”.
KYMLICKA, W. 2000. Politics in the Cultural Survival Quarterly 25 (2): 22–5.
Vernacular: Nationalism, Multiculturalism ROMAINE, S. 1991. “Introduction”. Language in
and Citizenship. Oxford: Oxford University Australia, 1–24. Cambridge: Cambridge
Press. University Press.
LO BIANCO, J. and M. RHYDWEN. 2001. “Is the ROMAINE, S. 1994. “From the fish’s point of
extinction of Australia’s indigenous view”. International Journal of the Sociology
languages inevitable?” In: J. A. Fishman, ed., of Language 110: 177–85.
op. cit., 391–423.
ROMAINE, S. 1998. “Language contact in
LUCAS, P. F. N. 2000. “E ola mau kākou i ka relation to multilingualism, norms, status and
‘ōlelo makuahine: Hawaiian language policy standardization”. In: Unn Royneland, ed.,
and the courts”. Hawaiian Journal of History Language Contact and Language Conflict,
32: 1–28.
9-25. Volda, Norway: Volda College. STRUBELL, M. 2001. “Catalan a decade later”.
In: J. A. Fishman, ed., op. cit., 260–84.
SCHIFFMAN, H. 1996. Linguistic Culture and
Language Policy. London: Routledge. TABOURET-KELLER, A., R. B. LE PAGE., P.
GARDNER-CHLOROS and G. VARRO eds. 1997.
SKUTNABB-KANGAS, T. 2000. Linguistic Vernacular Literacy. A Re-Evaluation.
Genocide in Education – or Worldwide Oxford: Oxford University Press.
Diversity and Human Rights? Mahwah, N.J.:
Lawrence Erlbaum Associates. UNESCO. 1953. The Use of Vernacular
Languages in Education. Monograph on
SKUTNABB-KANGAS, T and R. PHILLIPSON in Fundamental Education VIII. Paris, United
collaboration with MART RANNUT. eds. 1994. Nations Educational, Scientific and Cultural
Linguistic Human Rights. Overcoming Organization.
Linguistic Discrimination. Berlin/New York:
Mouton de Gruyter. WEBB, V. 1998. “Multilingualism as a
developmental resource: Framework for a
STRUBBELL, M. 1999. Language, democracy & research program”. Multilingua 17-2/3:125-
devolution in Catalonia. In: Sue Wright, ed., 154. Special issue on Aspects of
Language, Democracy and Devolution in Multilingualism in Post-Apartheid South
Catalonia, 4–39. Clevedon, UK: Multilingual Africa.
Matters.
About the Author

Suzanne Romaine has been Merton Professor of English Language at the
University of Oxford since 1984. Her research interests lie primarily in historical
linguistics and sociolinguistics, especially in problems of societal multilingualism,
linguistic diversity, language change, language acquisition, and language contact in
the broadest sense. She has conducted extensive fieldwork in Europe as well as in
Papua New Guinea and Hawaii. She is the author of numerous articles and books,
the most recent of which is Vanishing Voices. The Extinction of the World’s
Languages (jointly with Daniel Nettle, Oxford University Press 2000). Further
information at http://users.ox.ac.uk/~romaine.
Address: Merton College, University of Oxford, Oxford OX1 4JD, United

Kingdom; email [email protected]
Minority Matters: Issues in Minority
Languages in India
RAJESHWARI V. PANDHARIPANDE
University of Illinois
T his paper discusses the following major issues relating to minority languages
in India: (a) the definition of minority languages; (b) their status; (c) the
factors contributing to their retention or attrition; and (d) the role of speakers’
attitude towards their language.
The paper demonstrates that the definitions of minority languages proposed in the
current literature are inadequate to define minority languages in India. It further
argues that minority languages can be defined on the basis of two major features:
(a) their functional load; 1 and (b) their functional transparency in the various
domains of society. Minority languages are typically those which carry relatively
less or marginal functional load and functional transparency. The concept of
“functional load” in this context refers to the ability of languages to successfully
function in one or more social domain. The load is considered to be higher or lower
on the basis of the number of domains it covers. The higher the number of
domains, the higher the load. For example, in India the English language covers
almost all the major public domains such as business, education, national and
international communication, and technology. In contrast, the tribal languages
control only one (rapidly diminishing) domain, that of home. The regional
languages cover private domains such as home, as well as public domains such as
intra-state communication, education, government and law.
The “functional transparency” feature is important in determining the degree of

functional load. Functional transparency refers to the autonomy and control that the
1
The term “functional load” is taken from phonology where it is used to determine the degree of
contrast between phonemes. “For example, in English, the contrast between /p/ and /b/ would be
said to have higher functional load than between /j/and /z/. The former contrast distinguishes many
minimal pairs whereas the latter contrast distinguishes only a few. Several criteria are used in
making such quantitative judgements, such as the position within a word at which the contrast is
found, and the frequency of the occurrence of the words in the language” (Crystal 1985, 130). The
term “functional load” in this paper is used to provide a quantitative base to evaluate the notion of
“power” of the languages in a society in order to distinguish between major and minor languages.
The language that successfully functions in relatively more domains is considered to have a higher
functional load. Moreover, functional transparency is another concept used here as a parameter to
measure “power”.
214 Rajeshwari V. Pandharipande
language has in a particular domain. Thus the functional load is higher if the
language does not share the function with other languages, i.e. there is an
invariable correlation between the language and the function. In other words, if it is
perceived as the most appropriate language to carry out that particular function, the
language is considered to be “transparent” to the function. For example, Sanskrit is
most transparent to its function of expressing Hinduism. Regional languages are
most transparent to their function in state government. Similarly, English is
transparent to the function of “modernity”. If the function is shared by other
languages, the transparency is lowered and the functional load is also lowered. For
example, the function of regional languages in the domain of education is shared
by English in many states, which lowers the transparency of their function and
consequently lowers their functional load.
I argue that there is a hierarchy of functional load in India, where multilingualism

is part of the ecology. This hierarchy coincides with the power hierarchy of
languages. The higher the functional load, the more powerful the language is
perceived to be. Thus, minority languages are those that carry a lower functional
load and thereby hold a lower position in the power hierarchy. The hierarchy of
power (political, economic and cultural) of languages in multilingual India needs to
be taken into account in order to fully define and explain the status of minority
languages. It is further demonstrated that decrease versus enhancement of the
functional load can be seen as the major factor in the status of minority languages.
The above definition of minority languages allows us to evaluate the role of factors
such as language planning and policies, and the attitudes of speakers in India
towards either protecting, maintaining and promoting minority languages or
causing their decay and attrition. Those factors contributing towards increasing the
functional load are identified as those promoting sustenance and promotion of the
languages, while those reducing the functional load are identified as those causing
decay or attrition.
1. Definitions of Minority Languages

The Constitution of India recognises eighteen languages as “scheduled languages”2
(listed in Schedule VIII, Articles 343–51) while those languages not included in the
scheduled eighteen are listed as “minority languages”. A close examination clearly
shows that the criteria used to divide languages into “scheduled” and “non-
scheduled” 3 (minority) languages fail to account for the status of languages in
India. The Constitution does not provide a clear criterion for defining minority
languages. The Supreme Court of India, in 1958, presented a parameter for
defining a minority language as “the language of the minority community” (which
is defined as a community numerically less than 50 per cent). However, this
2
See Annex Table 1.
3
See Annex Table 2.
Minority Matters: Issues in Minority Languages in India 215
parameter is not applicable at the national level because “there is no linguistic

group in India which can claim the majority status” (Chaklader 1981, 14). Hindi,
the official language of the Union, is the language of only one-third of the total
Indian population. Thus, as Chaklader (1981, 14) correctly points out, “the
majority-minority question is considered in reference to the state only”. In this
context, Chaklader (1981, 14) argues for adopting a definition of minority
languages at the state level. For example, a minority language can be viewed as the
language of the population which is less than 50 per cent of the total population of
a state and which is different from the language of the majority community and the
language of the state. This parameter turns the numerical majority languages into
minority languages (Bhatt and Mahboob 2002). Kashmiri, which is spoken by
53 per cent of the total population in the state of Jammu and Kashmir, is not the
state language (which is Urdu). Moreover, Urdu, the official language of Jammu
and Kashmir, is spoken by less than 1 per cent of the total population of the state.
Similarly, English, the official language of Meghalaya, is spoken by 0.01 per cent
of the total population. Thus the parameter of defining minority languages on the
basis of their numerical strength is not appropriate in the context of India.
Other parameters have been proposed based on the dominance or “power”

(political, economic and/or cultural) of languages (Bhatt and Mahboob 2002;
Chaklader 1981; Williams 1964, among others). Languages lacking political,
economic or cultural power tend to be included in the list of minority languages. A
good example is that of the tribal languages,4 speakers of which constitute 7.08 per
cent of the total population of India. These languages lack political, economic and
cultural power at the state or national levels, therefore they belong to the category
of minority languages. In contrast, Sanskrit, which is perceived as a language of the
cultural heritage of India (but not spoken natively in any state) is not labelled as a
minority language. Similarly, English, though numerically a minority language, is
not viewed as such owing to its high economic value at the national as well as the
international level.
A very broad definition of minority provided by the United Nations captures the
salient features of minority languages: “The term minority includes only those non-
dominant groups in a population which possess and wish to preserve stable, ethnic,
religious or linguistic traditions or characteristics markedly different from those of
the rest of the population.”5 The two features, “non-dominant” and “different from
the rest of the population”, are generally shared by the minority languages of India.
Moreover, this definition points out that a language receives its minority status due
to the minority status of the speech community to which it belongs. It allows a
language to be labelled as a minority language if the community using it is
numerically large but non-dominant.
4
See Annex Table 3.
5
UN Yearbook for Human Rights 1950, 490; quoted in Chaklader 1981, 16.
Another phenomenon which complicated the definition of minority languages in

India was the large-scale reorganisation of the states according to the concentration
of languages in different parts of India (based on the Report of the States
Reorganisation Commission of 1955). Although the policy behind the
reorganisation was to minimise the number of linguistic minorities (and to some
extent it did so), it created new minorities as no state was completely unilingual.
Speech communities were distributed across state boundaries, therefore an
official/majority language in one state could become a minority language in
another state. For example, Telugu is an official/majority language in Andhra
Pradesh while it is a minority language in Tamil Nadu.
Srivastava (1984) provided a new approach towards defining minority-majority

languages based on two principles, “quantum” and “power”, as shown in the
diagram.
Power
+ −
+ (a) majority (b) Janta
Quantum
− (c) elite (d) minority
According to this view, a language can be of four types: (a) powerful as well as
majority (e.g. Marathi in Maharashtra State); (b) powerless but majority (e.g.
Kashmiri in Jammu and Kashmir); (c) minority but powerful (English in all states);
(d) minority and powerless (tribal languages in all states).
The above discussion shows that definitions of minority languages are based on
either numerical or functional criteria. While the numerical criterion marks a
language as minority if the number of speakers of the language (i.e. the speech
community) is relatively low, the functional criterion marks a language with
relatively low power of dominance in the economic, political and social domains.
The numerical criterion (based on the size of the speech community) is inadequate
to describe the status of minority languages in India. The criterion of dominance
fails to take into account the fact that, in a multilingual country such as India,
different languages are dominant in different domains. For example, Sanskrit is
dominant in religion but not in economics, politics and business. The regional
languages are dominant at home, but in higher education and business at the
national level they are not. English is dominant in higher education, business and
politics but not in religion. The criterion of dominance will indicate the same
language as dominant and non-dominant in different domains.
In the light of the above, I propose that a different framework needs to be

formulated which will take into account the multilingual profile of India, the
functional distribution of languages across domains, the size of the speech
community and the notion of dominance. Moreover, the framework should be able
to explain various types of minority in the country, and why the same language can
have the status of minority as well as dominant language simultaneously (in

different states). For example, minority languages can be divided into three groups:
(a) those which have “minority (non-dominant)” status in their native state; (b)
those which are reduced to “minority status” in their non-native states; and (c)
languages which do not have a native state but are distributed across states (e.g.
Sindhi and Konkani). This framework clearly shows that a language acquires
minority status when its functional load is reduced (in a non-native state where the
dominant language of that state is different, and used in many public domains),
while it continues to enjoy the status of a dominant (non-minority language) in its
native state.
The concept of functional load of a language provides a framework within which a

comprehensive definition of “minority languages” can be presented. In this context,
I argue that all the above definitions of minority languages have one feature in
common – minority languages (regardless of whether they are numerically a
minority or not) carry a marginal functional load, or none at all, in the public
domains of society. Thus, English, though numerically a minority language, cannot
be called a minority language as it carries a heavy functional load in the public
domain (education, business, international and intranational communication,
religion, etc.). In contrast, Kashmiri, a majority language in Jammu and Kashmir, is
viewed as a minority language because it does not carry a heavy functional load in
the public domain of the society within which it is located. The tribal languages are
numerically minority languages, and carry a marginal functional load in the
domains of education, business and inter-group communication. This definition of
minority languages further allows us to identify the factors (sociopolitical) that are
instrumental in creating minority languages. Moreover, it has a predictive value, in
that a language which is in the process of being eliminated from the public domain
(its functional load is decreasing) will be reduced to the status of a minority
language. Also, this definition implicitly assumes that a stable or increasing
functional load is conducive to language retention, while a decreasing functional
load leads to language attrition. It also predicts that a minority language can
acquire the status of a dominant language if its functional load increases in the
public domain.
2. Factors Influencing the Status of Minority Languages: Language

Planning and Language Policies
The following discussion indicates the factors that have contributed towards
reducing the functional load of minority languages in the public domain, and
thereby led to the shift of these languages to the dominant languages. In a number
of cases minority languages (especially tribal languages) are facing rapid attrition.
These factors are: (a) language policies; (b) modernisation; (c) speakers’ attitudes
towards their languages; (d) separation of the link between language and identity or
a change in the speech community’s perception of its identity. I point out below
how these factors can be seen as mechanisms through which the marginalisation of
minority languages is taking place.
First, the impact of language policies on minority languages is discussed. The

policy of reorganisation of states on a linguistic basis was seen as a strategy to
homogenise a state where the language spoken by the majority (over 50 per cent of
the total population of the state) would become the official language. It was
assumed that this policy would bring the administration and the people together, in
contrast to the British policy that had imposed English as the language of
administration in India and thereby severely inhibited the growth and development
of the indigenous vernaculars. In order to implement this policy, each state
developed a programme to ensure the use of the majority vernacular in major
domains such as legislation, education, administration and other state-controlled
operations such as public transport, banking, etc. In the context of legislation at
state level, the official language of the state was used for (a) introducing Bills by
the Governor under Article 213 of the Constitution; (b) introducing by-laws passed
by the state government or by Parliament or the state legislature, all official
notifications issued by Parliament or the state government; and (c) for other official
correspondence within the state (see Chaklader 1981, 45 for further discussion).
In the domain of education, the Education Commission set up in 1966 6

recommended the use of the state language at university level. However, for high-
school education, a “three-language formula” was proposed and approved by the
Central Advisory Board of Education in 1957 and was fully endorsed by the chief
ministers of the states in 1961. The implementation of the formula was complex.
The three languages were introduced at different phases of high-school education:
(a) at lower-primary level (grades I–IV), either the mother tongue or the official
language; (b) at higher-primary level (grades V–VII) two languages – mother
tongue or regional language and Hindi (national language) or English; (c) at lower-
secondary level (grades VIII–X) three languages – mother tongue/regional
language, Hindi and English; (d) at higher-secondary level (grades XI–XII) any
two languages including a classical language.
In the third and major domain of administration, regional vernacular languages

were promoted for intra-state communication in all contexts, such as the official
Gazette of the state government, the judiciary, employment procedures, and all
official documents had to be in the official language of the state. For inter-state
communication, the use of the associate language English was permitted. This
situation continues today with varying degrees of implementation. Another major
domain where the reorganisation of the states influenced the status of languages
was the conditions of employment. Under Article 309 and item 41 of list I in
6
Report of the Education Commission 1964–66, 7, New Delhi, National Council of Educational
Research and Training, 1971.
Schedule VII, the states organised Public Commissions 7 to determine the

conditions of employment within the states. The members of the Commissions
were appointed by the state governors. Though it was agreed by chief ministers that
language should not be an obstacle to recruitment, a number of states (with the
Commissions’ approval) sanctioned the legislation that knowledge of the state
official language should be made mandatory for all state employees. Maharashtra
(Marathi), Orissa (Oriya), West Bengal (Bangla), Gujarat (Hindi and English),
Haryana (Hindi) and Punjab (Hindi and Panjabi) were the early advocates of this
requirement for knowledge of the respective state languages.
Regardless of the degree of success in implementing this measure in different

states, the three-language formula, and the recommendations of the various
Commissions towards language use, had an enormous impact on minority
languages. The “functional load” of the numerically minority languages was
drastically reduced in the public domain and as a result their status as non-
dominant/powerless was further confirmed. They were almost completely
eliminated from the dominant public spheres. Four types of response to these
policies were observed: (a) language movements against the policies; (b) segre-
gation from the “mainstream” communities; (c) assimilation with the larger,
majority language communities; (d) adoption of multiple strategies. Thus the
Bengali speech community in Assam (with Asamiya as the state language)
demanded autonomy and rights to education in Bangla, while many communities
of numerically minority languages such as Konkani (in Maharashtra and
Karnataka), and many tribal languages in the north-east, have adopted a separatist
attitude and maintained their languages. However, most of the minority-language
speech communities have adopted the third choice, of assimilation with the
majority or dominant languages within their respective geographic regions or
states. Kundu (1994) explains why several tribal language communities are losing
their languages in a process of assimilation with the dominant language in the
north-eastern parts of India. Lack of educational facilities such as textbooks,
teachers, schools with the tribal language as the medium of instruction, lack of a
standard language (and script), and most importantly, marginalisation or exclusion
from the major domains of social behaviour, have severely curtailed the sustenance
of tribal languages. A similar situation exists with Yerva in Kerala, or Bhumj and
Rajbamshiin in West Bengal. The adoption of multiple strategies (using their
language at home and the dominant language at school and other public domains)
to maintain their languages is seen among the minority languages in diaspora.
These languages have a stable cultural and linguistic base elsewhere that provides a
constant motivation for their retention.
The languages spoken by a numerical minority have clearly become non-dominant

and powerless minority languages under the above language policies. Again, it
7
16th Report, Commissioner for Linguistic Minorities, 11–20, New Delhi; quoted in Chaklader 1981,
51.
should be noted that their non-functionality in the major domains of society may be
seen as the reason for their low status.
3. Constitutional Safeguards
The Indian Constitution adopted several safeguards to protect linguistic minorities
in the country. Articles 350(A) and 350(B) were adopted in addition to the earlier
Articles 29(1), 30, 347 and 350 in order to safeguard the interests of minorities.
Article 29(1) notes: “Any section of the citizens residing in the territory of India or
any part thereof having a distinct language, script or culture of its own shall have
the right to conserve the same.” This clearly guarantees the right of minorities to
conserve their cultural as well as linguistic traditions. The first clause of Article 30
of the Constitution guarantees all minorities based on religion or language to
establish and administer educational institutions of their own in order to preserve
their linguistic and/or cultural heritage. The second clause of Article 30 prohibits
the state from discriminating against minority educational institutions in giving
financial aid on the grounds that they are under the management of minorities.
Thus minorities are allowed to secure state funds for their educational institutions.
Article 347 allows the use of minority languages for official purposes.
Accordingly, a state should be recognised as unilingual only if one language group
within the state constitutes 70 per cent or more of the total population. Moreover,
where there is a minority of over 30 per cent or more of the total population, the
state should be recognised as bilingual for administrative purposes. A similar
principle applies at the district level.
Minority languages can be majority languages at the local level. Clear cases of this
are Karbi and Dimasa in the autonomous districts of south Assam; Tibetan in the
Ladakh region, and Baltistan in the north, of Jammu and Kashmir; Nepali in
Sikkim; Hindi in the north-eastern region of Maharashtra, etc.
Article 350(A) proclaims, “ [I]t shall be the endeavour of every State and of every
local authority within the State to provide adequate facilities for instruction in the
mother tongue at the primary stage of education to children belonging to linguistic
minority groups.” Moreover, Article 350(B) gives power to the President to
appoint appropriate officers and use proper methods to investigate and safeguard
the rights of linguistic minorities. Wadhwa (1975) points out that the 12th Report
of the Commissioner for Linguistic Minorities shows that education in the minority
languages is provided at the primary level in the following states and union
territories: Andhra Pradesh, Bihar, Gujarat, Kerala, Maharashtra, Karnataka,
Nagaland, Rajashtan, Tamil Nadu, Uttar Pradesh, West Bengal, Andaman and
Nicobar Islands, Dadra and Nagar Haveli, Delhi, Goa, Daman and Diu,
Pondicherry.
The above discussion shows that the Constitution of India attempts to guarantee
linguistic minorities the right to use their languages in administration and edu-
cation. According to our hypothesis, this government strategy would result in

increasing the functional load of the languages. The above safeguards proposed by
the Constitution guarantee the use of minority languages in the domain of
education, thereby identifying and guaranteeing a certain functional load to those
languages. However, the implementation of these policies is not always successful
for various reasons, both external and internal to linguistic minorities, which are
discussed in the following sections.
4. Bilingualism and the Status of Minority Languages

13 per cent of the Indian population is bilingual and over 42 per cent of the
minority population is bilingual (Singh 2001). Singh and Manoharan (1993) point
out that among the 623 tribal communities with which they worked, only 123 were
monolingual while 500 were bilingual. They further note, “[T]he second or third
language may be either a minor language, a scheduled language or even a regional
language of the area in which they reside. … Apart from the official language of
the State, regional languages like Chattisgarhi, Halbi, and Tulu are also spoken for
inter-group communication by Tribal communities.” Each state in India is multi-
lingual but the rate of minority languages varies from 10 per cent (Gujarat) to
44 per cent (Panjab) (Bhatt and Mahboob 2002, 22). Moreover, the three-language
formula has further contributed to the high rate of bilingualism among minority
linguistic communities. Traditionally, tribal communities lived isolated from the
cities and villages and their occupations included cutting firewood, hunting, fishing
and farming (for further discussion see Parvathamma 1984). The languages of
those communities have been maintained due to their isolation from the main-
stream population which did not interact with them. In the fifty years since India’s
independence in 1947, it has become necessary for tribal communities to interact
with the mainstream population owing to the following changes caused by
modernisation (Pandharipande 1992, 258): (a) mechanisation of the professions of
farming, fishing, tanning of leather, etc., (b) deforestation and urbanisation of
villages, and (c) the policy of state governments to promote education in these
communities (through the three-language formula), which has accelerated the
speed of learning the dominant regional language among these communities. As a
result, a majority of tribal languages are shifting to the “dominant language” in
almost every functional domain. The functional domain of these languages is
restricted to home and intra-group communication. Several studies – Biligiri
(1969), Karunakaran (1983), Khubchandani (1983), Roy Burman (1969), Raju
(1977), Abbi (1995) – show that due to the lack of script, the paucity of teaching
materials and the small number of speakers, a large number of tribal languages are
facing attrition.
The discussion here shows that the reduction in their functional load in the public
domain is leading minority languages towards attrition. It is important to note that
there is a hierarchy in the shift of the minority to the dominant languages. While
Kui in Andhra Pradesh and Bhili in the Nagpur area (Maharashtra) show a very
high degree of shift, Santali in Bihar and West Bengal shows a relatively lower
degree of shift. In contrast, some of the tribal languages in Kerala show negligible
shift or none at all.
Like tribal languages, the minority languages of diaspora in different states also
face pressures from state or regional languages in their respective state of
immigration. Pandharipande (1992) points out that the maintenance versus shift of
these languages is determined by their prestige or importance at the national level
or in their native states. An example is Hindi in its non-native state of Gujarat.
Although the number of Tamil and Hindi speakers in Gujarat is similar (about
1.6 per cent), the degree of maintenance of Hindi is much higher than that of
Tamil, because Hindi is a national language while Tamil is only a regional and
state language. Similarly, English is a minority language in every state. However,
its maintenance is very high. The two cases of Hindi and English support the
hypothesis of the correlation between a higher functional load and the maintenance
of languages. Another important factor to note is that the implementation of the
three-language formula is almost impossible when the mother tongue of the
speakers is tribal and does not have a script, a standard code or literature. In the
absence of these, it is not possible for the education department to produce teaching
materials to ensure teaching of the mother tongue, even at the elementary/primary
level. Young children who are speakers of tribal languages tend to begin to learn
the state language at the primary level of education, and soon become bilingual.
The use of the state language in school further causes the reduction of the domain
of use of their first (tribal) language because bilingual children tend to use the state
language (as opposed to their mother tongue) in most public domains. After a
couple of generations, the language of home (of the tribal communities) is
gradually replaced by the dominant state language, thus causing severe attrition of
the tribal language. In contrast, those children who do not go to school tend to
preserve their languages (tribal languages) as their use at home is maintained. This
phenomenon supports the hypothesis that a guaranteed functional load (i.e.
sustained use in a domain) guarantees maintenance of a language while the
reduction and/or elimination of functional load leads to language attrition.
5. Language Attitudes, Functional Load and Minority Languages

This section covers some of the internal reasons for the reduction of the functional
load of minority languages resulting in their rapid shift and attrition. One of the
major factors affecting the maintenance or shift of minority languages is the
speakers’ perception of their own languages. Modernisation of Indian society has
resulted (in addition to the mechanisation mentioned above) in the need to acquire
a certain type of linguistic capital for sustained upward mobility in society.
English, Hindi and other regional state languages (in that order) present a hierarchy
of the power of linguistic capital. Technology for communication at the state,
national and international levels has promoted unprecedented vigour in the use of
English (although regional languages are catching up). In the domains of
production, sustenance, promotion of any product (both material and ideological),

linguistic capital plays an important role. In India, the labour market in all domains
is dominated by English and regional languages.
Power hierarchy
English High
Regional/state languages
In other words, the functional load of English and the regional languages is
extremely high compared with that of minority languages, therefore it is not
Minority
surprisinglanguages
that speakers of minority languages
Low perceive their languages as “power-
less” in terms of their functionality in society. Several studies show that speakers of
minority languages do not think that it is useful or important to learn their first
language. Singh (2001) points out that out of the total 7.8 per cent tribal population
in India, only 4 per cent speaks tribal languages. Breton (1997, 30–31) also
illustrates the phase of transition of a large number of tribal languages towards the
respective dominant languages. Razz and Ahmed (1990) claim that half of India’s
tribal population have already lost their languages, and that people have assimilated
with the dominant linguistic group, adopting the dominant language as their mother
tongue. Abbi (1995, 177) supports the above claim: “It is sad that the Kurux and
Kharia languages are quickly disappearing from most of the urbanised area of
Ranch district. This trend indicates that the urban tribals seldom consider it their
privilege to speak their mother tongues. On the contrary, ignorance of the tribal
languages is regarded as an enhancement of status and prestige. In speaking Hindi
they feel superior in comparison to other fellow-tribals who cannot speak it.” This
negative attitude towards their languages has resulted in their shift to the dominant
languages and a drastic reduction in their use.
The study in Pandharipande (1992) shows that the dialects of Marathi spoken
around the Nagpur area corroborate the above claims about the attitudes of
minority language speakers. As part of a survey, educated farmers in the 30–35 age
group were interviewed. They controlled both standard Marathi and their dialect
(Varhadi) of Marathi. These subjects, unlike their parents, had replaced the use of
their dialect by standard Marathi, even at home. They readily admitted that the
retention of their own dialect would hamper their socioeconomic success in the
rapidly urbanising society of Maharashtra. However, they did not think that the loss
of their code would result in the loss of their (sub)cultural identity. In fact, they
thought that they could retain their identity through their rituals, foods and their
“unique values” towards life. The minority speakers feel that they must control the
dominant code in order to compete and succeed in the dominant culture. A similar
case is that of the Hindi dialects in the northern parts of India. These dialects, Braj,
Bhaka, Bangru, Bundelkhandi, and other closely related languages such as Maghai,
Maithili and Bhojpuri, are rapidly being replaced by Khadi Boli (Standard Hindi)
which is the dominant language in the area. Most speakers of the dialects can also
speak Hindi.
An important point to note here is that there is not an invariable correlation

between maintenance of language and maintenance of culture. Diachronic evidence
supports this claim. The Persian community that migrated to India in the seventh
century has lost its language but has meticulously maintained its ethnic identity
through preservation of a religious and cultural identity separate from the dominant
culture and society.
In contrast, some minority linguistic communities seem to have strong language

loyalties which they use for retention of their ethnic identity as well as to secure
sociopolitical rights. The Santali language movement is a case in point (Mahapatra
1979). Santals demanded the establishment of a separate province for the tribes of
Chota Nagpur and introduced Santali as the language of schools. Similarly, Sindhi
and Konkani are preserved due to the extreme loyalty of the speakers towards their
languages.
6. A Changing Equation of Language and Culture

Another dimension of speakers’ attitudes towards their languages is a changing
perception of their own cultural identity. The modernisation and technological
development of the country has created a new vision of homogeneous culture with
modern amenities available to all, where individuals are judged by their ability to
succeed in the (apparently) fair competition. The road to success, in this view, is
carved out through science and technology. As a result, languages such as English
and Hindi are perceived as mechanisms to achieve the “dream of success”. This
overarching vision of culture is commonly shared by all, majority as well as
minority communities. Their choice of language is therefore determined by their
view of their “imagined or aspired” identity. The Bhils and Gonds in Maharashtra
are keener to move up the economic ladder than to retain their tribal identity. When
I asked a Tulu (minority language) speaker (a maid) in Mumbai why she did not
speak Tulu to her children, she said, “I want her to go to law school. I do not want
her to be a maid when she grows up. She should know English and Marathi.” With
great pride, she asked her daughter to recite an English poem to me, as if
proclaiming her victory over the linguistic barrier!
7. Functional Load, Functional Transparency and Language

Maintenance
I argue above that minority languages are prone to attrition as they are being
replaced by other dominant languages in almost all public domains. In other words,
attrition of minority languages is directly related to their reduced functional load.
In the following discussion I provide evidence to support the assumption of a
correlation between functional load and language maintenance. A language with a
higher functional load shows a higher degree of maintenance than a language with
a lower degree of functional load. For example, the regional languages in India are
used in many more domains than the tribal languages. While regional languages
are maintained, tribal languages face attrition, leading to death. American-Indian
languages in North America and tribal languages in Australia are rapidly being
replaced by the dominant languages in every domain (see Fishman 1991).
Haugen’s classic work on the Norwegian language in the United States (Haugen
1953) also shows that, over a period of time, the Norwegian language spoken by
Norwegian immigrants was gradually replaced by English in almost all domains,
leading to shift of the Norwegian language. Similar cases are also noted by Dorian
in her 1982 work on loss and maintenance in contact situations, which points out
that English and Russian are displacing many indigenous languages in Australia
and the Soviet Union, respectively, while English is not endangering the native
languages of India (Fishman 1977). The reason is that the indigenous languages in
India have retained their functional domains (i.e. official context, local business,
schools, etc.). Dorian (1982) also refers to Hebrew as an example of revival of a
language by the national/political policy of making it functional in virtually all
domains of use (Dorian 1982, 44). Derhemi (2002 and forthcoming), in her case
study of Arbresh in Italy, points out how the language is in a dangerous phase of
attrition in Italy due to its displacement by Italian in many public domains such as
school, media, business, etc. Crystal (2000, 83), discussing why languages die,
claims that in South America the indigenous languages are left alone as they are
not viewed as a serious threat to national unity. However, as he points out, these
languages are not used in any major public domains of prestige. “People find they
have fewer opportunities to use their language, because it has been marginalised. It
is not found in official domains such as local offices of civil service, and the local
banks. It is not found in the media. It is not found as the language of higher
education” (Crystal 2000, 83). The presence of these languages in unimportant
domains creates what Fishman (1987) calls, “the ‘folklorisation’ of a language –
the use of indigenous languages only in irrelevant or unimportant domains”
(Crystal 2000, 83). Crystal further claims, “ And with each loss of a domain, it
should be noted, there is a loss of vocabulary, discourse patterns, and stylistic
range. It is easy to see how languages would eventually die, simply because,
having been denuded of most of its domains, there is hardly any subject matter left
for people to talk about, and hardly any vocabulary to do it with.” In his
monumental work on Reversing Language Shift, Fishman (1991) describes the case
of the Irish language, which is being consciously revived by making it functional in
the public domains of musical recitals, drama, school education, workplace, etc.
These examples clearly demonstrate two points: (a) languages are endangered or
die when their functional load is reduced in the public domain; and (b) they are
maintained when their functional load is retained or increased.
In the above discussion, it is claimed that “functional load” provides a parameter

for defining minority languages. It is assumed that the degree of functional load
can be measured by the number of functional domains of the languages, i.e. the
higher the number, the higher the functional load. I would argue here that the
number of domains is not the only parameter for measuring the degree of
functional load, but that “functional transparency” is another important parameter.
Functional transparency can be explained as follows: if a language A is the only
language used to perform a particular function in a particular domain, then
language A can be said to have “functional transparency” vis-à-vis that function. In
contrast, if the same function is performed by more than one language, the
languages involved are said to be not transparent (but opaque) to that function. A
language with higher functional transparency can be said to have a higher
functional load compared with a language that does not have functional
transparency. For example, the only language used for science and technology in
India is English. Therefore, English can be said to be transparent to this function.
Similarly, regional languages (in their native states) are almost exclusively used at
home, thereby command functional transparency in that domain. In Mumbai, the
pidgin Hindi (Bazaar Hindi) is almost exclusively used as the “market language”,
thereby claiming transparency to the function of a link language (in the
multilingual community in Mumbai). I argue that the invariable correlation
between the language and its function makes the language transparent to that
function.
In contrast, two languages are generally used as alternatives by immigrants in their

non-native context. That is, they begin to use the dominant language (of the
country/place of immigration) along with their native language in various domains
(home, social gatherings, etc.) where they earlier used their native language
exclusively. In this case, their native language does not remain transparent to the
function. Though the number of domains in which their native language is used is
higher than the domain of Bazaar Hindi, its functional load will be said to be lower
than Bazaar Hindi. This situation is fairly common within minority languages in
India. Many minority languages spoken exclusively at home at one point in time,
gradually begin to be accompanied by the dominant language when children begin
schooling in that language. This use of two languages (minority and dominant)
reduces the functional transparency of minority languages.
Some other cases fall between the two extremes, where a language may not be
exclusively used for a function but there is a high correlation between the language
and its function. A good example of this is the Sanskrit language, which in India is
closely connected to the context of Hinduism (although other languages also

perform the same or a similar function).
The hierarchy of functional load can be presented as follows:
High functional load

1. + functional transparency + number of domains
2. + functional transparency − number of domains
3. − functional transparency + number of domains
4. − functional transparency − number of domains
Low functional load
The above diagram shows relatively high/low degrees of functional load.

Languages such as English and regional languages in India fall into category (1) as
they all carry a high degree of transparency as well as a high number of domains.
Sanskrit and Bazaar Hindi belong to category (2), where the functional
transparency is high but the number of domains is low. Categories (3) and (4) show
the phases of attrition of minority languages. In the first phase (3), minority
languages are used along with the dominant language (thus losing functional
transparency); and in the second phase (4), the dominant language displaces
minority languages, leading to their disappearance.
The question of maintenance and shift of languages is related to the above. Can we
assume that a high degree of functional load is a necessary as well as an adequate
condition for the maintenance of a language? The answer is as follows: a language
with a higher functional load has a better chance of survival than a language with a
lower functional load. For example, the regional languages, with their higher
functional load, are more likely to be maintained in India than the tribal languages
with a very low functional load. However, a language with a higher degree of
transparency (and low number of domains, see category (2)) has a better chance of
survival than a language with a high number of domains but low transparency.
Evidence to support this hypothesis comes from the fact that languages involved in
a diglossic situation generally show a high degree of maintenance compared with
languages used to perform identical functions. In a multilingual country such as
India, each (multilingual) community maintains stable bi/multilingualism as long
as functional transparency is maintained across languages or, in other words, the
situation is di/multiglossic.
8. Conclusion
The above discussion shows that minority languages can be defined on the basis of
their low prestige, which is the result of their low functional load in the public
domain. “Functional load” can be used as a diagnostic tool to predict maintenance
or attrition of languages. It is further shown that external factors (language policies,

modernisation) as well as internal factors (attitudes of speakers) contribute to the
enhancement or retardation of minority languages. Two main points emerge: (a)
culture can be maintained without the language; and (b) perception of the (desired)
identity changes over time and therefore the choice of language to express that
identity also changes. The paper brings out the complexity of the issues related to
definitions and the desirability of language maintenance. The hypothesis proposed
makes a strong case for the need to raise the functional load of minority languages
to prevent their shift and/or attrition.
Annex
Table 1: Scheduled Languages
Language Number of Speakers

1 Assamese 13,079,696
2 Bengali 69,595,738
3 Gujarati 40,673,814
4 Hindi 337,272,114
5 Kannada 32,753,676
6 Kashmiri 56,693
7 Konkani 1,760,607
8 Malayalam 30,337,176
9 Manipuri 1,270,216
10 Marathi 62,481,681
11 Nepali 2,076,645
12 Oriya 28,061,313
13 Panjabi 32,753,676
14 Sanskrit 49,736
15 Sindhi 2,122,848
16 Tamil 53,006,368
17 Telugu 66,017,615
18 Urdu 43,406,932
Source: Census of India, 1991: excludes figures for Jammu and Kashmir.
Table 2: Non-scheduled Languages
Language Number of Language Number of

speakers speakers
1 Adi 158,409 49 Kuki 58,263
2 Anal 12,156 50 Kurukh/Oraon 1,426,618
3 Angami 97,631 51 Lahauli 22,027
4 Ao 172,449 52 Lahanda 27,386
5 Arabic/Arbi 21,975 53 Lakher 22,947
6 Bhili/Bhilodi 5,572,308 54 Lalung 33,746
7 Bhotia 55, 483 55 Lepcha 39,342
8 Bhumij 45,302 56 Liangmei 27,478
9 Bishnupuria 59,233 57 Limbu 28,174
10 Bodo/Boro 1,221,881 58 Lotha 85,802
11 Chakesang 30,985 59 Lushai/Mizo 538,842
12 Chakru/Chokri 48,207 60 Malto 108,148
13 Chang 32,478 61 Mao 77,810
14 Coorgi/Kadagu 97,011 62 Maram 10,144
15 Deori 17,901 63 Maring 15,268
16 Dimasa 88,543 64 Miri/Mishing 390,583
17 Dogri 89,681 65 Mishmi 29,000
18 English 178,598 66 Mogh 28,135
19 Gadaba 28,158 67 Monpa 43,226
20 Gangte 13,695 68 Munda 413,894
21 Garo 675,642 69 Mundari 816,378
22 Gondi 2,124,854 70 Nicobarese 26,261
23 Halabi 534,313 71 Nissi/Dafla 173,791
24 Halam 29,322 72 Nocte 30,441
25 Hmar 65,204 73 Paite 49,237
26 Ho 949,216 74 Parji 44,001
27 Jatapu 25,730 75 Pawi 15,346
28 Juang 16,858 76 Phom 65,350
29 Kabui 68,925 77 Pochury 11,231
30 Karbi/Mikri 366,229 78 Rabha 139,365
31 Khandeshi 973,709 79 Rengma 37,521
32 Kharia 225,556 80 Sangtam 47,461
33 Khasa 912,283 81 Santali 5,216,325
34 Kheza 13,004 82 Savara 273,168
35 Khiememnugan 23,544 83 Sema 166,157

36 Khond/Kondh 220,783 84 Sherpa 16,105
37 Kinnauri 61,794 85 Tangkhul 101,841
38 Kisan 162,088 86 Tangsa 28,121
39 Koch 26,179 87 Thado 107,992
40 Koda/Kora 28,200 88 Tibetan 69,416
41 Kolmi 98,281 89 Tripuri 694,040
42 Kom 13,548 90 Tulu 1,552,259
43 Konda 17,864 91 Vaiphei 26,185
44 Konyak 137,722 92 Wancho 39,600
45 Korku 466,073 93 Yimchungre 47,227
46 Korwa 27,485 94 Zeliang 35,079
47 Koya 270,994 95 Zemi 22,634
48 Kui 641,662 96 Zou 15,966
Source: Census of India, 1991.
Table 3: Numerically Significant Minority Languages in each State and Union

Territory of India
State/ Number of Per- State/ Number of Per-

Territory speakers centage Territory speakers cent
age
India Andhra Pradesh
Hindi 337,272,11 40.2 Telugu 56,375,755 84.8
4
Bengali 69,595,738 8.3 Urdu 5,560,154 8.4
Telugu 66,017,615 7.9 Hindi 1,841,290 2.8
Arunachal Assam
Pradesh
Nissi/Dafla 172,149 19.9 Assamese 12,958,088 57.8
Nepali 81,176 9.4 Bengali 2,523,040 11.3
Bengali 70,771 8.2 Bodo/Boro 1,184,569 5.3
Bihar Goa
Hindi 69,845,979 80.9 Konkani 602,626 51.5
Urdu 8,542,463 9.9 Marathi 390,270 33.4
Santhali 2,546,655 2.9 Kannada 54,323 4.6
Gujarat Haryana
Gujarati 37,792,933 91.5 Hindi 14,982,409 91.0
Hindi 1,215,825 2.9 Punjabi 1,170,225 7.1
Sindhi 704,088 1.7 Urdu 261,820 1.6

Himachal Karnataka
Pradesh
Hinndi 4,595,615 88.9 Kannada 29,785,004 66.2
Punjabi 324,479 6.3 Urdu 4,480,038 10.0
Kinnarui 61,794 1.2 Telugu 3,325,062
Kerala Madhya Pradesh
Malayalam 28,096,376 96.6 Hindi 56,619,090 85.6
Tamil 616,010 2.1 Bhili/Bhilodi 2,215,399 3.3
Kannada 75,571 0.3 Gondi 1,481,265 2.3
Maharash Manipur
tra
Marathi 57,894,839 73.3 Manipuri 1,110,130 60.0
Hindi 6,168,941 7.8 Thado 103,667 5.6
Urdu 5,734,468 7.3 Tangkhul 100,088 5.4
Meghalay Mizoram
a
Khasi 879,192 49.5 Lushai 518,099 75.1
Garo 547,690 30.9 Bengali 59,092 8.6
Bengali 144,262 8.1 Lakher 22,938 3.3
Nagaland Orissa
Ao 169,837 14.0 Oriya 26,199,346 82.8
Sema 152,123 12.6 Hindi 759,016 2.4
Konyak 137,539 11.4 Telugu 502,102 1.6
Punjab Rajasthan
Punjabi 18,704,461 92.2 Hindi 39,410,968 89.6
Hindi 1,478,993 7.3 Bhili/Bhilodi 2,215,399 5.0
Urdu 13,416 0.1 Urdu 953,497 2.2
Sikkim Tamil Nadu
Nepali 256,418 63.1 Tamil 48,434,744 86.7
Bhotia 32,593 8.0 Telugu 3,975,561 7.1
Lepcha 29,854 7.3 Kannada 1,208,296 2.2
Tripura Uttar Pradesh
Bengali 1,899,162 68.9 Hindi 125,348,49 90.1
2
Tripuri 647,847 23.5 Urdu 12,492,927 9.0
Hindi 45,803 1.7 Punjabi 661,215 0.5
West Andaman and
Bengal Nicobar Islands
Bengali 58,541,519 86.0 Bengali 64,706 23.1
Hindi 4,479,170 6.6 Tamil 53,536 19.1
Urdu 1,455,649 2.1 Hindi 53,536 17.6
Chandiga Dadra and
rh Nagar Haveli
Hindi 392,054 61.1 Bhili/Bhilodi 76,207 55.0
Punjabi 222,890 34.7 Gujarati 30,346 21.9
Tamil 5,318 0.8 Konkani 17,062 12.3
Daman Delhi
and Diu
Gujarati 92,579 91.1 Hindi 7,690,631 81.6
Hindi 3,645 3.6 Punjabi 748,145 7.9
Marathi 1,256 1.2 Urdu 512,990 5.4
Lakshadw Pondicherry
eep
Malayalam 43,678 84.5 Tamil 720,473 89.2
Tamil 282 0.5 Malayalam 38,392 4.8
Hindi 217 0.4 Telugu 34,799 4.3
Source: Census of India 1991: excludes figures for Jammu and Kashmir.
References
ABBI, A. 1995. “Small languages and small CRYSTAL, D. 2000. Language Death.
language communities”. International Cambridge: Cambridge University Press.
Journal of the Sociology of Language 116:
175–85. DERHEMI, E. 2002. Albanian immigrants of the
Arbresh diaspora: a sociolinguistic analysis.
BHATT, R. M. and A. MAHBOOB. 2002. In: B. International Conference on Albanian
Kachru and S. N. Sridhar, eds., The Minority Migration and New Transnationalisms,
Languages and their Status. Languages in Sussex Centre for Migration Research,
South Asia. London: Cambridge University University of Sussex, UK, 6–7 September.
Press.
DERHEMI, E. Forthcoming. Communicative
BILIGIRI, H. S. 1969. “Problems of tribal patterns of the endangered Arbresh dialect of
languages in India”. In: N. Ray, ed., Piana degli Albanesi. Ph.D. thesis, University
Language and Society in India, 245–50. of Illinois at Urbana-Champaign,.
Simla: Institute of Advanced Study.
DORIAN, N. C. 1982. “Language loss and
BRETON, R. J. L. 1997. Atlas of the Languages maintenance in language contact situations”.
and Ethnic Communities in South Asia. New In: R. D. Lambert and B. F. Freed, eds., The
Delhi: Sage. Loss of Language Skills, 44–59. Rowley,
Mass.: Newbury House.
CHAKLADER, S. 1981. Linguistic Minority as a
Cohesive Force in Indian Federal Process. FISHMAN, J. A. 1977. “The spread of English as
New Delhi: Associate Publishing House. a new perspective for the study ‘language
maintenance and language shift’”. In: J. A.
CRYSTAL, D. 1985. A Dictionary of Linguistics Fishman, R. A. Cooper and A. W. Conrad,
and Phonetics. Oxford: Basil Blackwell.
The Spread of English, 108–33. Rowley, Amsterdam: John Benjamins Publishing

Mass.: Newbury House. Company.
FISHMAN, J. A. 1987. “Language spread and PARVATHAMMA, C. 1984. Schedule Caste and
language policy for endangered languages”. Tribes. New Delhi: Ashish Publishing House.
In: Proceedings of the Georgetown
University Roundtable on Language and RAJU, D. R. 1977.Study of Tribal languages in
Linguistics, 1–15. Washington: Georgetown Andhra Pradesh, 159–68. Seminar on
University Press. Sociolinguistics. Hyderabad: Telugu
Academy.
FISHMAN, J. A. 1991. Reversing Language Shift.
Clevedon, UK: Multilingual Matters. RAZZ, M. and A. AHMED. 1990. An Atlas of
Tribal India. New Delhi; Concept
HAUGEN, E. 1953. The Norwegian Language in Publications.
America: A Study in Bilingual Behavior. 2
vol. Philadelphia: University of Pennsylvania ROY BURMAN, B. K. 1969. “Languages of tribal
Press. communities in India and their use in primary
education”. In: N. Roy, ed., Language and
KARUNAKARAN, K. 1983. Sociolinguistic Society in India, 251–69. Simla: Indian
Patterns of Language Use. Chidambarm: Institute of Advanced Study.
Velan Press. (All India Tamil Linguistic
Association Publication 20.) SINGH, K. S. and S. MANOHARAN. 1993.
Languages and Scripts. Delhi: Oxford
KHUBCHANDANI, L. 1983. Plural Languages, University Press.
Plural Cultures: Communication, Identity,
and Sociological Change in Contemporary SINGH, U. N. 2001. “Multiscriptality in South
India. Honolulu, East West Center: Asia and language development”.
University of Hawaii Press. International Journal of the Sociology of
Language 150: 61–74.
KUNDU, M. 1994. Tribal Education: New
Perspectives. New Dheli: Gyan Publishing SRIVASTAVA, R. 1984. “Linguistic minorities
House. and national language”. In: Florian Coulmas,
ed., Linguistic Minorities and Literacy:
MAHAPATRA, B. P. 1979. “Santali language Language Policy Issues in Developing
movement in the context of many dominant countries, 99–114. Berlin: Mouton de
languages”. International Journal of the Gruyter.
Sociology of Language 75: 73–8.
WADHWA, K. K. 1975. Minority Safeguards in
PANDHARIPANDE, R. 1992. “Language shift in India. New Delhi: Thompson Press.
India: Issues and implications”. In: W. Fase,
K. Jaspaet and S. Kroon, eds., Maintenance WILLIAMS, R. M. JR. 1964. Strangers Next
and Loss of Minority Languages. Door. New Jersey, Prentice Hall.
About the Author

Rajeshwari V. Pandharipande is Director of the Program for the Study of Religion
and Professor of Religious Studies, Linguistics, and Comparative Literature at the
University of Illinois at Urbana-Champaign, United States. She received her first
Ph.D. in Sanskrit Literature from Nagpur University, India, and her second Ph.D.
in Linguistics from the University of Illinois at Urbana-Champaign. She has
published four books and over fifty research articles on sociolinguistics, the syntax
and semantics of South Asian languages, and the language of religion. Her two
major works are The Eternal Self and the Cycle of Samsara: Introduction to Asian
Mythology and Religion (Ginn Press 1990), and A Grammar of Marathi (Routledge
1997). Professor Pandharipande received the title “University Scholar” (1992–93)

for her outstanding research, the Harriet and Charles Luckman All Campus
Distinguished Undergraduate Teaching Award, and the William Prokasy Award for
excellence in undergraduate teaching at the University of Illinois (1996).
Address: University of Illinois, Department of Linguistics, Urbana-Champaign,

Illinois, USA; email: [email protected]
Uchumataqu: Research in Progress on the
Bolivian Altiplano
PIETER MUYSKEN
Katholieke Universiteit Nijmegen
In this paper the current linguistic situation of the Uru, who live near
Lake Titicaca (Bolivia) is discussed. An overview is given of earlier
studies of the language, in the late nineteenth and twentieth centuries.
Then I focus on possible causes of the decline of the language of the
Uru, Uchumataqu: persistent droughts in the 1930s, intermarriage
with the surrounding Aymara and ethnic reorientation. In addition to
losing ground to Aymara and Spanish, Uchumataqu has undergone
considerable Aymara structural influence. Subsequently, I summarise
my own research on the language, and the possibilities of linguistic
studies serving community goals. Finally, the chances for survival are
discussed, which depend in part on large-scale developments, or the
absence thereof, in the Bolivian economy and society.
I n this paper I describe and comment on the linguistic situation of the Uru of
Iru-Itu (hispanicised as Irohito), a small ethnic group on the borders of Lake
Titicaca, in the highlands of Bolivia. The colonial denomination for this group is
Juchusuma or Ochosuma, which may be the basis for Uchuma, the first part of the
compound Uchuma-taqu (taqo or taqu means “language”), the name of the group
for their language, often also called Uru. Local sources suggest that Juchusuma is
the traditional name for the Río Desaguadero, and thus Uchumataqu would mean
“language of the Desaguadero river (people)”.
I have carried out linguistic research with the Uru on three successive visits in 2001
and 2002, and have been exploring ways, together with the community leaders, to
preserve the language, which has almost been lost. The history and prospects of
Uchumataqu cannot be seen separately from the development of the indigenous
peoples of Bolivia and surrounding countries.
I mentioned that the Uru live on the borders of Lake Titicaca, but this is slightly
inaccurate. Properly speaking, they live on the banks of the Río Desaguadero, the
river through which the excess water from Lake Titicaca flows towards Lake
Poopo and then onwards to the salty marshes of the southern Altiplano. The Uru
are surrounded by Aymara-speaking campesinos (peasant farmers). There were
Uru communities on Lake Titicaca proper as well, but these communities have now
236 Pieter Muysken
become Aymara-speaking.1
On Lake Poopo there are several Murato communities, ethnically related to the
Uru. However, the Murato no longer speak a separate language, but have adopted
Aymara, preserving a number of original words from an Uru-like language. In the
Murato oral testimonies published in Miranda Mamani et al. (1992) a number of
these words appear.
Finally, south-west of Lake Poopo on the salty marshes near Lake Coipasa there is
another group related to the Uru, the Chipaya. They live in one community, Santa
Ana de Chipaya, and number about 1,500. Their language has been preserved. It
has been documented by Olson (e.g. 1967) and Porterie-Gutierrez (1990), and is
currently being studied by Rodolfo Cerrón Palomino (Pontífica Universidad
Católica Peruana, Lima). Apaza Apaza (2000), in a Aymara dialect study of the
region between the salares (salt lake basins), of Uyuni and Coipasa, suggests that
there may be lexical traces of Uru there as well, but the lexical evidence he adduces
does not yet match the Uchumataqu data I have collected. Most of the non-Aymara
words he found are actually Quechua, rather than Uru-like.
There has been considerable confusion about the genetic affiliation and identity of
the three original Uru languages: Uchuma-taqu, Murato or Chholo (Miranda
Mamani et al. 1992, 171), and Chipaya or Chipaj tago (Porterie-Gutierrez 1990,
160). In some colonial sources mention is made of the Puquina living along the
shores of Lake Titicaca. The Puquina language is now extinct, but it was once
important enough to receive the status of lengua general (general language), along
with Quechua and Aymara, in the early years of the Spanish occupation.2 Puquina
has been tentatively classified as Arawakan. Since they were spoken in roughly the
same area, Uru and Puquina have been subsequently confused as being the same
language. This mistaken assumption was reinforced by Créqui-Montford and Rivet
(1925–27), and since then many publications and museum displays link Uru to
Puquina and the Arawakan language family. However, linguistically, this link is
unmotivated. The grammar of the Uru languages does not resemble that of
Arawakan. There is no trace of this in the Uru languages. Furthermore, what we
know of the Puquina lexicon is completely unlike that of the Uru languages.3
Until we know more of all the languages of Amazonian northern Boliva, with
which there are some lexical resemblances (Fabre 1995), it is best to treat the Uru
1 Furthermore, preliminary results from an Aymara lexicographic dialect-survey of the people living
on the banks of the Desaguadero between Lake Titicaca and Lake Poopo (Filomena Miranda,
Universidad Mayor de San Andres, La Paz, personal communication) have uncovered a number of
non-Aymara words in this area. These suggest that there once was an Uru-speaking population all
along the Desaguadero.
2 This meant that the language could and should be used as a missionary language, even with
speakers of smaller languages.
3 Torero (1987) has shown that some Puquina lexicon survives in Callahuaya, a ritual healing
language from the Charazani region north of La Paz (cf. e.g. Muysken 1996; Adelaar and
Muysken in prep.).
Uchumataqu: Research in Progress on the Bolivian Altiplano 237
languages as a separate group. Earlier attempts by Olson (1964, 1965) to link the
Uru languages to Mayan have been shown to be without foundation.
The well-known French ethnohistorian Nathan Wachtel has written a detailed

“regressive” history (i.e. going back in time) of the Uru peoples (1990). He shows
that at the time of the conquest by the Spaniards of the area now called Bolivia, the
Uru peoples occupied a large territory from the Pacific coast in Chile through the
Bolivian altiplano to the borders of Lake Titicaca. Progressively, they lost ground
and many of their communities underwent ethnic restructuring. Ticona and Albó
(1997) have published an absorbing account of the recent history of the Jesús de
Machaqa region, the larger area that the Uru community of Irohito forms part of.
They document the resurgence of the Uru of Irohito from a marginalised and
dispirited small band in the 1940s to a highly self-confident and progressive group
at present (see also below).
1. Earlier Research
There are several early-twentieth-century sources for the Uru languages. These
include Polo (1901), a general description of the Uru people with a vocabulary (full
of inaccuracies, unfortunately); Bacarreza (1910), a general description of the
Chipaya; Posnansky (1915), a preliminary description of the language of the
Chipaya; Créqui-Montford and Rivet (1925–27), who visited Irohito in the early
1920s; Métraux, with both linguistic and ethnographic observations (1935). How-
ever, the richest early material, largely unpublished, is probably that gathered by
the German ethnographer and archaeologist Max Uhle on two successive visits in
1894 and 1896 (Uhle 1894–96). Together with the Uchumataqu material gathered
by his student Walter Lehmann in 1928, it is deposited in the Ibero-American
Institute in Berlin (Lehmann 1928). The material contains word lists, a sketch for a
grammar, comparative studies and ethnographic notes.
However, probably the most important source on Uchumataqu is the intensive and
detailed work of the French doctor Jehan Vellard, who visited the Uru on numerous
occasions in 1938, the 1940s and early 1950s and left a very rich set of source
materials, including detailed vocabulary lists, short phrases, stories (1949, 1950,
1951, 1967), and a French monograph, dramatically titled Gods and Pariahs of the
Andes. The Uru, those who do not want to be men (1954).
His central thesis is that a great drought occurring between 1939 and 1948
destroyed the fluvial ecosystem on which the Uru depended, and spelt the end of
them. “But the people of the lake have been struck dead. The last group of Uru will
not reform itself” (1954, 12).4 The loss of the language has been interpreted in
magical terms, according to Vellard (1954, 103): “They always consider their
forgetting the mother tongue as a punishment accompanied by the loss of the gifts
of magic and of prophesy. By having allied themselves with men, the last Uru have
4 “Mais le peuple du lac a été frappé à mort. Le dernier groupe ourou ne se réformera plus.”
238 Pieter Muysken
lost their language and are no longer respected.”5
In more recent times, a team organised and financed by UNICEF visited the
community in 1995, involving the French linguist Colette Grinevald (Grinevald et
al. 1995). This team collected vocabulary, worked with the community on an
orthography, and generally rekindled enthusiasm for reviving the language among
the Urus.
In 1985 Lorenzo Inda, the most interested community leader, published a history of
the Urus, including much detail about cultural practices and Uchumataqu
vocabulary.
Crucially, only a small fraction of the research on Uchumataqu is available to the

Bolivian people, let alone to the speakers themselves. It is either unpublished,
published in obscure sources, or transcribed in notations based on German and
French pronunciations, and the translations given are often in German or French.
2. Causes of the Decay of the Language

The dramatic and gloomy predictions of Vellard did not come true. The Uru are
there and continue to form a distinct group, as noted in the previous section, but
they certainly do not differentiate themselves as much as they once did from the
surrounding Aymara. One of the manifestations of this is that the Uru no longer
speak their own language, but most of the time Aymara, in addition to some
Spanish.
Vellard was right in that the shift from Uchumataqu to Aymara as the daily
language of the people took place before 1950, but the language did not disappear
altogether. One clear cause of language loss is that after the drought the reduced
size of the group remaining in the community forced marriages with Aymara from
neighbouring villages. In 1942 only six men and a few women, all elderly, were
left in the community. As Vellard puts it (1954, 93): “Fifteen years ago, more than
fifty persons spoke Uru fluently. With the dispersion, mixed marriages have
accelerated the decay of the language. The Aymara women married to Uru refuse
to speak the language of their husbands: the children, Aymaras through their
mothers, do not want to be taken for Uru by speaking a despised language.”6
In addition, as population size increased again after the 1950s, many Uru were
forced to seek work outside the community, and lost contact with potential
5 “Elles considèrent toujours l’oubli de la langue maternelle comme un châtiment accompagné de la

privation du secours magique et du don de la prophétie. Pour s’être alliés aux hommes, les derniers
Ourous ont perdu leur langue et ne sont plus respectés.”
6 “Il y a quinze ans, plus de cinquante personnes parlaient couramment l’ourou. La dispersion venue,
les mariages mixtes ont acceleré la ruine du langage. Les femmes aymaras mariées aux Ourous
refusent de parler la langue de leurs maris; les enfants, étant aymaras du fait de leurs mères, ne
veulent pas être pris pour des Ourous en parlant une langue méprisée.”
speakers of Uchumataqu. They functioned in Spanish, Aymara, and for those who
migrated to the Cochabamba area, Quechua.
3. Present Project
The primary aim of my own project is to document the language as well as
possible, with the community itself as the primary beneficiary in mind, and the
wider public, including the linguistic research community, as the secondary bene-
ficiary. However, even this modest initial aim turned out to imply much more,
because documenting a language in this stage of decay requires wide community
support and interest in the language.
The 1992 census, analysed by Albó (1995), revealed that over half of the
community in residence in Irohito, eighty-seven persons, claimed to speak
Uchumataqu. This finding contrasts sharply with what I found in April 2001, when
in fact no single person spoke the language well enough to do fairly simple
vocabulary work with me.7
However, it does point to a self-perception on the part of the community as linked

to Uchumataqu. When I contacted the community to do fieldwork, their enthusiasm
for the language and its preservation gained the upper hand over their distrust of
foreigners, particularly foreigners without lots of aid funds and connections with
aid agencies. After lengthy meetings a contract was drawn up with the following
main clauses:
(1) I am not to get involved with Uru women.8
(2) The authorship of the resulting publication lies with the community; my
name appears as asesor lingüista (linguist advisor).
(3) I am to leave US$500 as a guarantee that I will return with a draft

vocabulary on 1 August 2001.
(4) The community provides a team of consultants every night at seven to

work on the language with me. They are compensated for their time with
US$1 per person per hour.
On the basis of this contract we worked for a few weeks, using all the words in the
Vellard material as an initial stimulus. I left with about 800 recognised words,
organised in twenty-five themes (semantic fields), ranging from “the family” to
7 At that point women were not asked to participate in the fieldwork, unfortunately, by the village
leaders, I suspect because they know more than the men.
8 Some Uru say that Jehan Vellard left the community with an Uru woman. This story of a romance
finds some support in his field notes, e.g. when he exemplifies: ampt’e wira k’ucha chuni pini
pek’uchay “toi, j’aime bien beaucoup, blanc” (1951, 21) [you, I love quite a bit, white man]
(spelling adapted to modern Uchumataqu orthography as introduced in 1985).
240 Pieter Muysken
“existence and possession”. I also asked children in the local school (twenty-seven
pupils) to make drawings with black felt-tipped pens.
On 1 August I returned with several copies of the draft version of the vocabulary,
which also includes the phrases and expressions I had gathered, illustrated with the
children’s drawings. Reception was good, although the community leaders
probably thought the drawings awkward and childish. A new series of evening
sessions started, this time with the oldest and most knowledgeable of the April
consultants, Teodora Vila, and his elder sister who had just arrived from La Paz,
Julia Vila. Indeed, she may be described as the only reasonably fluent speaker of
the language, having been brought up in it by her grandmother. The group of
speakers that Colette Grinevald had worked with had all died. Julia Vila was able
to correct the pronunciation of many words I had elicited earlier, provide the
Uchumatau word for many items for which we had only the Aymara equivalent so
far, and give full clauses.
Currently, I am reworking all this material and attempting to write an accessible

introduction to the language which takes the concrete themes as its point of
departure (Muysken 2000, 2001; Distrito Nacionalidad Indígena Urus de Irohito
2002). I returned to Irohito in early 2002 to present the new version to the
community. This version also includes songs written in Uchumataqu by Lorenzo
Inda, and a transcribed story told by Julia Vila.
The question remains, however, whether this effort will contribute to the
revitalisation of the language, and how this may be achieved. On the negative side,
I must mention the fact that it is extremely hard for the Bolivian highland
communities to survive at all, let alone preserve their language. Bolivia as a whole
remains an economic black hole, far away from any growth poles. It borders on
two poor countries, Paraguay and Peru, and on three less poor countries, Argentina,
Brazil and Chile. However, the regions of these countries bordering on Bolivia are
all underdeveloped. While Bolivia as a whole has zero or negative economic
growth at present, in fact all new economic activity (cattle, oil, gas, tropical
agriculture) is concentrated in the lowlands, and the highlands economically slowly
starve to death. This leads to tremendous labour migration, and much political
unrest. There is a complex system of cargos (annually rotating ritual obligations)
operating in the highland communities, forcing a number of adult males to stay in
the community and occupy a political function (president, vice-president, secretary,
head of school committee) by rotation for one calendar year.9 Apart from a small
group of committed adult males, the only people permanently present are women,
children and older people. However, many women, like Julia Vila, have also
migrated and only come back occasionally. All adolescents are elsewhere as well,
in school or working.
On the positive side, a number of factors may be mentioned. First, even after fifty
9 Of course, they may refuse, but then they lose their status as community members.
years of disuse the language has not yet disappeared. The eighty-seven members of
the community who claim to speak Uchumataqu all know a number of set phrases
and most of the vocabulary related to boating, fishing and hunting of waterfowl. In
fact, the Uru (in Uchumataqu qot suñi – “people of the lake”) use their original
language most when they are on the water.
Second, there is a clear interest in the language. Some years ago, the community
went to the expense of paying for the time and travel costs of a Chipaya to come
and teach them his language. The travel time is about a day and a half. It did not
work out well, because there are a number of differences between Chipaya and
Uchumataqu (roughly as between Italian and French), and perhaps also because the
Chipaya involved had no experience in language teaching. The episode does
illustrate the seriousness of the desire to recuperate the original language.
Third, it should be mentioned that Irohito is exceptional within its region. Far from
being the destitute and down-trodden group evoked by Vellard (1954), they are
now the most advanced community in the region, looked upon with some jealousy
and respect by their Aymara neighbours, and not without political influence. This
was the first community in the region (and the only one so far) to hoist the white
flag of 100 per cent literacy (in Aymara, ironically), due to the enthusiasm of
young community members who attended a secondary school nearby and returned
home to carry out an alphabetisation campaign. Also, it is the only community with
a number of solar panels. They applied for, and received, a computer and printer,
and recently a generator was installed. A few younger members of the community
have developed basic computer skills.
A fourth factor is the political constellation of the indigenous groups in Bolivia,

which itself is a reflection of continent-wide, or even global, developments. While
the 1952 revolution had emphasised class status (miner, peasant, etc.), in the mid-
1990s a strong political movement stressing ethnic pluralism came to the fore,
which in the area of education was a champion of bilingual schools. A number of
special programmes were made available particularly earmarked for small groups
with a separate status. This makes it advantageous for ethnic groups to strengthen
their distinct character. The Banzer-Quiroga government (1998–2002) only paid
lip-service to this policy, without actually discontinuing it. From 2000 onwards,
however, the nationalist mobilisation of particularly the Aymara leading to much
political unrest is highlighting ethnicity in politics again, and it is very possible that
the outcome of the 2002 elections will be that pluralism is once again high on the
political agenda. I return to this below.
A fifth factor is the growth of small-scale ethnotourism in Bolivia. Several ethnic

groups are experimenting with tours visiting their communities as a way of
generating extra income. The Uru are currently debating this as an option, and
already have a small museum with a reed boat, fishing nets, etc.10 Even though
10The Uru are traditionally associated with the technique of reed boats made famous by Thor
242 Pieter Muysken
most tourists would hardly hear the difference between rural Spanish and Aymara,
let alone between Aymara and Uchumataqu, it is clear that the Uchumataqu heri-
tage will be one of the assets of Irohito. This holds a fortiori if the ethnotourism
also involves secondary school and institutional outings from nearby La Paz, since
most Paceños do know some Aymara and would be curious to learn about Uchu-
mataqu. Note that I am not claiming that the presence of occasional tourists,
“ethno” or otherwise, would itself induce the Uru to speak Uchuamataqu in their
daily lives, but rather that tourism would turn the language into an asset and could
constitute a base, also financial, for teaching facilities and materials in the
language.
These five factors conjointly could play a role in the revival of Uchumataqu. It
would not be a purely automatic and unconscious reversal of a process of language
shift, of course. That shift took place much too long ago for that, and the language
is too far gone. It would be a conscious effort to give the language its place along-
side, not in place of, Aymara and Spanish. It would involve the activities of a small
group of cultural brokers, community leaders, and be linked to processes such as
folklorisation and musealisation of Uru culture. It would also need to be a modern
development, relying on literacy and possibly even on modern media.
For some, this makes the possible revitalisation of Uru unreal, artificial or suspect.
However, it may be the way in which many such revitalisation processes take place
in different parts of the world. Situations such as that of the Uru and the
Uchumataqu language cast doubt on traditional notions of authenticity and
spontaneity, and show that even rural communities are capable of “language
planning”.
4. Changes in the Language: Aymarisation and Simplification

One aspect not discussed so far concerns the linguistic features of the Uchumataqu
that have survived. It is fairly clear, when comparing the pronunciation recorded in
Vellard’s materials with that of younger speakers, that Uchumataqu has undergone
quite a few changes in the course of time, in part perhaps under the influence of
Aymara. Such changes come as no surprise, given the recent history of the
language and the extensive bilingualism with Aymara (and, for many, Spanish).
The changes are clearest in the pronunciation of the language. Modern speakers
have a tendency to reduce the five-vowel system of the language to a three-vowel
system. Uchumataqu e merges with i, and o merges with u. However, the exact
phonological environments favouring this merger have not yet been determined.
Given that five-vowel systems of the Uchumataqu type a, e, i, o, u are not marked
and the presence of the Aymara three-vowel system a, i, u, it is likely that the
mergers are triggered by Aymara. Notice that Spanish also has a five-vowel system
Heyerdahl, even if now most of these boats on Lake Titicaca are made by the Aymara, and most
boats used by the Uru themselves are crafted from wood.
of the original Uchumataqu type; it could not have triggered the merger.
Another change involves the nature of the consonants but is harder to define
exactly. Vellard (1994, 100–1) writes: “The language strikes one at first sight by its
sweetness, in contrast with the harder Quechua, richer in gutturals (there are four
different forms of k in Quechua) and even more with Aymara. It is a whispering
language, with countless sibilants and hissing sounds, tch, ch, sh, ts, etc.”11 The
lexical data I gathered in April 2001, however, did not contrast significantly with
Aymara words in their pronunciation. The full range of gutturals (presumably
velars and uvulars) was present, and the number of sibilants is only slightly larger
than in Aymara. A more detailed analysis of the precise phonetic form of the words
in Vellard’s transcription in contrast with the present form of these words can
resolve this issue.
Another complicated phonological problem concerns glottalisation. Uchumataqu

shares with Aymara the series of simple and aspirated stops:
p t ch k q
h h h h
p t ch k qh
However, the glottalised equivalents do not match entirely:
Aymara: p' t' ch' k' q'

Uchumataqu: ? ? ch' k' q'
The question mark indicates the one or two words with this pronunciation in
Uchumataqu that have a marginal status. This gives rise to the hypothesis, no more
than that at present, that glottalisation in Uchumataqu was borrowed from Aymara.
Further research, involving a detailed comparison with Chipaya (which is also in
close contact with Aymara, however), analysis of the earlier sources for Uru, and a
reconstruction of the proto-phonology of the Uru language family, will need to
clarify this issue.
A fourth change in the pronunciation of Uchumataqu concerns the structure of the

syllable. A few examples of the contrast between the words gathered in April 2001
and the pronunciation of these same words by Julia Vila in August 2001 will
suffice:
11“La langue frappe au premier abord par sa douceur, en contraste avec le quichoua plus dur et riche
en gutturales (il y a quatre formes différentes de k en quichoua) et plus encore avec l’aymara. C’est
une langue chuchotante, avec d’innombrables sifflantes et chuintantes, des tch, ch, sh, ts, etc.”
244 Pieter Muysken
April August gloss

sikuru-chay sqknu-chay “tie up”
tars-chay chatsna-chay “shake”
tik-chay t’oxsna-chay “cover”
It is not so much that the sound inventory has been reduced as that complex
syllable clusters and word forms have been simplified, and some substitutions
made.12
For morphology and syntax, things are less clear. In the realm of morphology, it is
difficult, even in the recorded speech of Julia Vila, to discover all the suffixes
mentioned in Vellard’s work. As for word order, Vellard writes (1954, 102): “In
the construction of phrases, the determining complement is placed before the noun.
Without this being an absolute rule, the verb is ordinarily relegated to the end of
the clauses. These are very short and all discourses are composed of little, very
simple clauses.”13 The pattern described by Vellard coincides typologically with
that of Aymara, i.e. head final. Consistent with this is the presence of postpositions
and possessor-possessed constructions. If there has been Aymara influence in the
realm of word order, it must surely predate the 1950s. Other aspects of
Uchumataqu syntax have not yet been sufficiently studied .
In any case, it may well be that if Uru is revitalised, it will be a simplified form of
the language that survives as a second language, as it has survived these last fifty
years.
5. Can we Look Ahead?: Grass-Roots and Government Support

Now that the first steps towards documenting what survives of Uchumataqu have
been taken, it is time to plan ahead. In addition to the thematically organised word
list mentioned above, it will be worthwhile publishing whatever traditional texts
have survived in an accessible orthography, so that reading material becomes
available. Vellard (1949) contains half a dozen recorded texts, mostly descriptions
of fishing and hunting practices, and a few accounts of recorded recent history. A
third step would be the establishment of a programme of regular meetings about
the language and courses for younger people.
Whether these efforts will eventually lead to the revival of the language depends on
the economic survival of Irohito as a viable community and on the politics of
12Notice also that there is a possible case of re-etymologisation in the case of sikuru-chay. It looks
suspiciously like Spanish asegurar (secure).
13“Dans la construction des phrases, le complément déterminatif se place avant le substantif. Sans
être une règle absolue, le verbe est d’ordinaire rejeté à la fin des phrases. Celles-ci sont très courtes
et tout le discours est composé de petites phrases fort simples.”
ethnicity in the region and in Bolivia as a whole. These questions are shrouded in
uncertainty.
In 2001 the Nación Originaria Uru (NOU) was formed in Oruro, a city in the centre
of the Altiplano south of Irohito, representing six groups. In the spelling of the
foundation document these are:
- Chipaya
- Murato
- Iruhitu
- Koro (= San Juan de Coripata, Carangas Prov., Bolivia)
- Isluka (= Isluga, near Chipaya, but in northern Chile)
- Uroz (= Urus of the “floating islands” on Lake Titicaca near Puno, Peru).
Not much is known about the possible survival of Uru languages with the last three
of these groups. Most probably the situation is as with the Murato: general use of
Aymara, Quechua or Spanish, and knowledge of individual words with Uru
etymologies. None the less, politically the formation of a larger ethnic unit in the
form of NOU is important, for two reasons. First of all, it can give small groups
more self-confidence and channel the exchange of expertise in revitalisation efforts
(high on the NOU agenda).
Second, it may stimulate the Bolivian Government to start educational programmes

specifically aimed at the Uru cultures and languages. In the mid-1990s the Reforma
Educativa (Educational Reform) was launched to introduce the three major
indigenous languages, Quechua, Aymara and Guaraní, into the curriculum. While
preliminary research on the indigenous Amazonian and Chaco languages was
carried out, establishing alphabets and studying educational needs, no programmes
were set up in this area. The Uru languages, the only small highland languages that
survive, were completely left out of consideration. One of the arguments was that
these languages are spoken by too few people to warrant serious teaching efforts.
The NOU initiative may help to draw the attention of the planners at the Section of
Curriculum Development at the Ministry of Education to the Uru languages and
communities.
Currently, the community itself is engaged in organizing one hour of Uru teaching
a week for the children in the school, using the materials we have prepared over the
last two years. In a teacher-training institution 14 km away, some students are
246 Pieter Muysken
interested in spending a semester in Irohito as a part of their final practice period,

developing teaching materials.
Bolivia is culturally and linguistically an extremely rich country. Many of the

Amazonian languages in the country are linguistic isolates or belong to extremely
small local language families. It is likely that the ethnolinguistic situation in
Bolivia and the adjacent Brazilian state of Rondônia directly reflects a very old
stage in the language development of the continent. Thus understanding the
languages and cultures of Bolivia offers a key to the early history of the continent
as a whole.
References
ADELAAR, W. F. H. and P. C. MUYSKEN. In MÉTRAUX, A. 1935. “Contribution à
preparation. The Languages of the Andes. l’etnographie et à la linguistique des indiens
Cambridge: Cambridge University Press. Uru d’Ancoaqui (Bolivie)”. Journal de la
Société des Américanistes de Paris 28: 75–
ALBÓ, X. 1995. Bolivia plurilingüe. Guia para 110.
planificadores y educadores. Cuadernos de
investigación 44. 2 vol., maps, appendices. MIRANDA MAMANI, L., D. MORICIO CHOQUE.
La Paz: Unicef–Cipca. and ALVAREZ DE MORICIO, S. 1992.
Memorias de un olvido. Testimonios de vida
APAZA APAZA, I. 2000. Estudio dialectal del Uru-Muratos. Recopilado por Rossana
aymara. Caracterización lingüística de la Barragán (ASUR). La Paz: ASUR/HISBOL.
región intersalar de Uyuni y Coipasa. La
Paz: Instituto de Estudios Bolivianos. MUYSKEN, P. 2000. “Drawn into the Aymara
mold: the Uru in Jehan de Vellard’s texts”.
BACARREZA, Z. 1910. Fragmento del informe In: Simon van de Kerke and Hein van de
de la provincia de Carangas del Voort, eds., Papers ... Studies in Indian
Departamento de Oruro. Boletín Oficina Languages of Latin America. Leiden:
Nacional de Estadística, Vol. 6. La Paz. CNWS/Leiden University.
CREQUI-MONTFORD, G. DE and P. RIVET. 1925– MUYSKEN, P. 2001. El Uchumataqu (Uru) de
27. “La langue Uru ou Pukina”. Journal de la Irohito. Observaciones preliminares. Lengua
Société des Américanistes de Paris 17: 211– Nación Originaria Uru (2001) Kiriwill
44; 18, 111–39; 19, 57–116. Qamanak_tan nij_ cheqanchištanpacha tiy
Distrito Nacionalidad Indígena Urus de Irohito. wajtha Qalltiniki Uru_ / Estatutos orgánicos
2002. El idioma Uchumataqu. Manuscript, y reglamentos de la Nación Originario Uru.
Irohito. Año 0-2000. Oruro: Centro Diocesano de
Pastoral Social (CEDIPAS).
FABRE, A. 1995. “Lexical similarities between
Uru-Chipaya and Pano-Takanan languages: OLSON, R. D. 1964. “Mayan affinities with
Genetic relationship or a real diffusion?” Chipaya of Bolivia I”. International Journal
Opción 11: 18, 45–73. of American Linguistics 30: 313–24.
GRINEVALD, C. et al. 1995. Curso de OLSON, R. D. 1965. “Mayan affinities with
uchumataqu. Manuscript, Irohito/La Paz. Chipaya of Bolivia II”. International Journal
of American Linguistics 31: 29–38.
INDA, L. 1985. Historia de los urus. Comunidad
Irohito de Yanapata. La Paz: HISBOL. OLSON, R. 1967. “The syllable in Chipaya”.
International Journal of American
LEHMANN, W. 1928. Vocabulario uru de Linguistics 33: 300–4.
Irohito. Manuscript. Berlin: Ibero-American
Institute. POLO, J.-T. 1901. “Los indios Urus del Perú y
Bolivia”. Boletín de la Sociedad Geográfica l’Institut Français d’Études Andines, Tome I

de Lima 10: 445–82. (Paris-Lima), 145–209.
PORTERIE-GUTIERREZ, L. 1990. “Documentos VELLARD, J. 1950. “Contribution à l’étude des
para el estudio de la lengua chipaya. Indiens Uru ou Kotsuñs”. Travaux de
Presentación de Rosaleen Howard- l’Institut Français d’Études Andines, Tome II
Malverde”. Amerindia 15: 157–91. (Paris-Lima), 51–89.
POSNANSKY, A. 1915. La lengua Chipaya. La VELLARD, J. 1951. “Contribution à l’étude des
Paz. Indiens Uru ou Kotsuñs”. Travaux de
l’Institut Français d’Études Andines, Tome
TICONA ALEJO, E. and X. ALBÓ CORRONS. III (Paris-Lima), 3–39.
1997. La lucha por el poder comunal. Jesús
de Machaqa: La marka rebelde. La Paz: VELLARD, J. 1954. Dieux et parias des Andes.
Cedoin and CIPCA. Les ourous, ceux qui ne veulent pas êtres des
hommes. Paris: Éditions Émile-Paul.
TORERO, A. 1987. “Lenguas y pueblos
altiplánicos en torno al siglo XVI”. Revista VELLARD, J. 1967. Contribución al estudio de
Andina 5: 329–406. la lengua uru. Buenos Aires: Centro de
Estudios Lingüísticos, Universidad de
UHLE, M. 1894–6. Manuscripts on Uru Buenos Aires.
Vocabulary and Grammar. Berlin: Ibero-
American Institute. WACHTEL, N. 1990. Le retour des ancêtres. Les
indiens Urus de Bolivie XX–XVIe siècle.
VELLARD, J. 1949. “Contribution à l’étude des Essai d’histoire regressive. Paris: Gallimard.
Indiens Uru ou Kotsuñs”. Travaux de
About the Author

Pieter Muysken is currently professor of linguistics at Nijmegen University (the
Netherlands), after having taught at the University of Amsterdam for many years
and a brief stint at Leiden University. His main specialism is language contact,
which he has studied in the context of the Caribbean Creoles, the relation between
Quechua and Spanish, and code-switching among immigrant groups in Europe. His
publications include Mixed Categories. Nominalizations in Quechua (with Claire
Lefebvre, Kluwer 1987), Language Contact and Bilingualism (with René Appel,
Edward Arnold 1987), Pidgins and Creoles, an Introduction (ed. with Jacques
Arends and Norval Smith, Benjamins 1996), One Speaker, Two Languages (ed.
with Lesley Milroy, Cambridge 1996), and Bilingual Speech. A Typology of Code-
mixing (Cambridge 2000).
Address: Linguistics, Katholieke Universiteit Nijmegen, Postbus 9103, NL-6500

HD Nijmegen, The Netherlands; email: [email protected]
The Endangered Arbresh Language and the
Importance of Standardised Writing for its
Survival:
The Case of Piana degli Albanesi, Sicily
EDA DERHEMI
University of Illinois at Urbana-Champaign
This study covers two main topics: a presentation of the condition of

the endangered Arbresh language and analysis of those features of
attrition that demonstrate the importance of writing and codification
of a language; and an examination of the implementation of recent
linguistic policies in Piana degli Albanesi, Sicily, in interaction with
the complex environment of a bilingual and diglossic minority whose
language is undergoing attrition. The research concludes that the
revitalisation of Arbresh in its current condition requires standard-
sation of the written language and the inclusion of this language in
Piana schools.
T his paper discusses the condition of the Arbresh language, an Albanian

dialect spoken for some 500 years in Sicily (Italy) in the community of Piana
degli Albanesi. The decaying features of the language are analysed in relation to
possible steps that the community and the cultural elite can undertake with the
support of recent language policy.
Although Arbresh is the language used in everyday informal communication and is

not stigmatised either by its speakers or by non-Arbresh, it shows clear features of
language attrition. A growing number of gaps in the Arbresh linguistic system
reduce its utility and frequency of use for communication, which leads in turn to
further decay. If linguistic decay is not stopped, the use of the language will
continue to diminish, and the effects will spread to reinforce patterns of further
linguistic loss, weakening the position of Arbresh in relation to the dominant
language, Italian.
Grenoble and Whaley (1998, 32–5) present an overview of the main scholarly
positions on the relationship between endangered languages and literacy. The
dominant view argues that literacy is essential to nationalism and to language
survival in the modern world, but there are also opposing opinions that literacy
International Journal on Multicultural Societies (IJMS), Vol. 4, No. 2, 2002: 248 -269
ISSN 1817-4574, www.unesco.org/shs/ijms © UNESCO
249 Eda Derhemi
facilitates language loss. The authors of this overview maintain that literacy has a
strong effect at the macrolevel, the larger and external context of linguistic en-
dangerment, but that its effect on language vitality is primarily a result of
microvariables, which are specific characteristics of each community with an
endangered language (27). Both authors claim that communities with a written
tradition are certainly in a stronger position to revitalise a language (34), which
may need reconstruction of lost or degraded material.
Piana Arbresh does not have the advantage of a written tradition, although there
have been sporadic efforts, especially in the translation of religious texts. Here I
argue that the way to stop the vicious circle of structural degeneration and loss is
the stabilisation of a normative form of the language and the immediate teaching of
a written codified form of the language in Piana schools. For this purpose, Piana
community leaders and other officials must make the best use of the opportunities
afforded by the 1999 Italian law for the protection of Arbresh communities, which
prescribes some specific actions and provides funding for the necessary changes.
At the end of the paper some suggestions are made on certain aspects of the
language policy as it functions in Piana today.
More generally, data on Arbresh show that the process of attrition can advance
towards eventual death, even in cases when the language still has an important
function in the community and the speakers feel positive about its use and
existence. Analysis of Arbresh’s endangerment indirectly suggests that commu-
nities trying to save their languages and traditions must raise awareness that a
minority discourse can vanish, and not only when there are political groups that
actively and deliberately fight the existence of a certain language and culture.
Minority languages can also die when societies that use these languages are in-
different and lack effective institutional intervention to protect them. In such
situations the commitment of the community and its efforts towards revitalisation
are decisive for the future of the language.
The next section of this paper presents the history of the community of Piana,
focusing on the original language and ethnicity of this community and on the
specific socio-economic consequences of these characteristics. Section 2 analyses
the features of language use in Piana, in the complexity of bilingual, minority and
endangered linguistic environment. In the third section I present evidence on
linguistic instability and loss at the structural levels, which suggest the need for a
codified written language in Piana. Section 4 analyses the situation in this
community since legislation for the protection of the Arbresh language and ethni-
city was passed in 1999, and the steps undertaken by Piana institutions and
community leaders with respect to the changed circumstances. I conclude with
some observations on the future of efforts to revitalise the endangered Arbresh
language.
The Endangered Arbresh Language 250
1. Short History of the Linguistic Community

Piana degli Albanesi is a small community of 7,000 people whose ancestors settled
in Sicily more than 500 years ago, with Albanian roots, language and culture, and
an authentic Greek-Byzantine religious rite. The people of Piana call themselves
and their language “Arbresh” and about 90 per cent of the community speak it,
although the speakers show very different levels of linguistic competence. The
word “Arbresh” is historically related to the word “Albanian” and the other
versions of the same word used in Greece, “Arvanitika”, and in Turkey, “Arnaut”.
The communicative patterns of Arbresh have shown rapid erosion in the last
decade, and the language may be at risk of disappearing. The same process of
gradual erosion leading to language death has already occurred recently in towns
and villages of Albanian origin scattered in other Italian regions, which lost first
their language and later their Orthodox Christian religion and their original
customs, costumes and folklore. From about 100 small Arbresh communities
recorded in eight Italian regions in 1837, only 50 were recorded in 1963, and 40
after another three years, in 1966 (Gjinari and Shkurtaj 1997, 255). From my
observations during fieldwork in Italy between June 2001 and January 2002, more
than half of these communities have lost their language. In most cases, the
particularities of their dialects and cultures died uncollected and unstudied. Those
that still maintain their language have lost a great number of speakers and use
Arbresh in restricted situations and settings. In some dialects the language is near
complete death. A concomitant of Arbresh language loss is a gradual shift to the
dominant language, Italian. This situation is characterised by a proficiency
continuum where one finds speakers with very different linguistic abilities, in
which the relation between the speaker and the languages in the repertoire of a
given community is determined by age, linguistic attitude, contact with the
language(s), etc. The decrease of linguistic competence occurs first in more formal
language and slowly spreads into the informal structures and vocabulary; thus it is
the opposite of a “bottom-to-top death” process specified in the typology of
language attrition by Campbell and Muntzel (1989, 185). The main feature of
Arbresh today is the widespread existence of “semi-speakers” or non-proficient
speakers, a common characteristic of languages undergoing attrition (Dorian 1981,
115).
However, Piana is one of the last strongholds of Arbresh, still showing its original
characteristics in spite of linguistic decay. There are other Arbresh communities
around Piana, such as Palazzo Adriano, Santa Cristina and Contessa Entellina, all
in a more advanced stage of language attrition. Palazzo lost the language a few
decades ago, but the oldest and most important church in the town is still of Greek-
Byzantine rite, and memories of being of Albanian origin are fresh. The other two
towns are in a more advanced stage of language attrition, but parts of the
population still speak the language, usually the older generation. Piana is the centre
of the Greek-Byzantine dioceses and also the cultural centre of Arbresh
communities in Sicily, sometimes envied by the other communities for its active
251 Eda Derhemi
social and cultural life, its enthusiastic young generation and its love for and pride
in the Arbresh tradition.
In general, the Arbresh communities in Sicily have lived in a peaceful relationship

with the surrounding Sicilian communities, but have often created mild negative
stereotypes of Sicilians and have very rarely accepted intermarriage with them.
This last characteristic has lost some of its force recently, but is still active.
The Arbresh communities in Italy are in general far from the stage of “primordial
ethnicity” understood as a state of little language consciousness (Fishman 1972,
179). The community of Piana in particular has traditionally been a symbol of
linguistic and cultural awareness, an example of pride in Arbresh culture and of the
fight for official recognition as a minority. Historically, Piana has given to the
Arbresh world well-known poets, writers, educators and researchers who have
enriched both Arbresh and Italian culture with their work. There have been periods
of greater or lesser awareness, but the feeling of being different – and, as
interpreted by Piana people, “therefore better” – has never died in Piana. Unlike
some Arbresh communities in Italy with severe economic problems and patterns of
linguistic and cultural self-depreciation, Piana has enjoyed relatively high
economic prosperity and has constantly regarded its different language and culture
as a source of prestige and self-appreciation. But the Pianioti (as the residents of
Piana are called) have never claimed or desired separation, as has been the case
with other minorities in northern Italy. The Arbresh of Piana have always
considered themselves to be Italians who in addition have Albanian origins,
although they still feel a little discomfort at being considered Sicilians. Some
decades ago not all the community spoke Italian, but now even very old speakers
master Italian. Today the community demonstrates stable and widespread
bilingualism and diglossia (Fishman et al. 1985, 42–3). But the weight of each
language in the linguistic repertoire is changing as a consequence of recent
linguistic attrition. While Arbresh speakers formerly acquired the language and
continued to employ it at home throughout their life, now they often replace it with
Italian, especially very young speakers.
In the last ten years which have seen a large Albanian immigration to Italy, the
Arbresh of Piana have tried to distinguish themselves from Albanians from
Albania, who today are stigmatised as “criminals”, “related to prostitution”, and so
on almost everywhere in Italy. The new situation has caused changes in the
linguistic attitude of the Piana elite and Arbresh speakers towards the use of
Standard Albanian.
Piana, only 25 km from the city of Palermo, is situated high in the mountains,
overlooking a nearby lake. The characteristic oriental religious celebrations of the
Greek-Byzantine rite, with its colourful original costumes that cannot be found in
non-Arbresh communities, bring many tourists from all over Italy during feasts and
other events, especially in summer, which generates money and jobs for the
Pianioti. The local people have also taken advantage of their proximity to Palermo,
which provides work and schooling opportunities for a large number of them. This
town has also had the luck and merit to possess a strong cultural elite to guide the
community in its determination to remain different, from the distant past to the
present. Today there are community activists of various specialisations, teachers,
writers, doctors, linguists, priests and others who are leading efforts to prevent the
loss of the Arbresh language, rites and traditions. The most important organiser of
these efforts is now the Piana Public Library, which has taken on the role of centre
of research and other activities on behalf of Piana and other Arbresh communities.
The business class of Piana, mainly small shopkeepers, artisans and restaurateurs,
is very interested in maintaining the distinct language, customs and traditions,
perceived by this group as further potential for investment and promotion of
tourism. Hopes for the inclusion of Arbresh instruction in schools have brought
Arbresh-speaking teachers who are jobless or do not have a permanent position in
Piana schools, into the effort to maintain Arbresh. They have a greater chance of
being appointed in Arbresh schools because this labour market excludes
monolingual Italian speakers, who until now have had the same opportunities to
teach in Piana schools. Other professionals who have part-time jobs in the libraries,
museums, archives or other institutions that make use of Arbresh, are also
interested in language maintenance and revival. The cultural elite of Piana believes
that future changes in the status of the language can create even more
opportunities, in many sectors, for the people of Piana.
2. Use of Arbresh as a Minority Language in Piana Degli Albanesi

The relationship between language and power reflects, in fact, the relationship that
speakers of different languages used in a community have with each other and with
the institutions that govern the community. What usually happens in these cases is
that the speakers of the subordinate language develop a negative attitude towards
their language, are ashamed to speak it or even to know it. A classic case is that of
Albanian speakers of the Arvanitika dialect in Greece described by Trudgill and
Tzavaras (1977) and Tsitsipis (1998). Such linguistic minorities are usually treated
less favourably by society than are the speakers of the dominant language and have
less access to institutional power. But Piana Arbresh, even though it is an
endangered and for this reason subordinate language, does not fit the socio-
linguistic framework of subordinate languages described above: the people of
Piana are proud of their language and culture, and they still use it, although their
linguistic competence and the functional domains of use of Arbresh are decreasing
significantly. In a survey I conducted with 100 participants, Arbresh from Piana
between 15 and 65 years old, ninety claim that they feel proud of being Arbresh
and not just Italian, six do not answer the question and only four say they would
have preferred to be just Italians instead of Italo-Albanians. Seventy-one partici-
pants in the same survey believe that their being Arbresh makes Italians more
interested in them, and only three participants think the contrary. These last three
are part of the group of four participants who would not have wanted to be
Arbresh. This high level of language loyalty is also demonstrated in the results of a
survey at the beginning of school year 2001 organised by the school district, in
253 Eda Derhemi
which 98 per cent of first-grade children and their parents declared themselves in
favour of Arbresh instruction in school.
It is very unusual for the speakers of a dying language to demonstrate this attitude
towards the language and their ethnicity. An example of a situation in complete
contrast to that in Piana is described by Trudgill and Tzavaras (1977, 177–8), who
claim that Arvanitis, the Arvanitika speakers of Greece, try to hide the fact that
they speak Albanian and to deliberately lose the language. Of 200 children between
the ages of 5 and 14 who were asked whether it was an advantage or a disad-
vantage to speak Arvanitika, only thirteen said that it was an advantage.
Yet, in spite of the positive attitude towards Arbresh in Piana, the dominant
language in the community remains Italian. As Arbresh is not written and has never
been systematically taught at school, Italian has replaced it in a wide range of areas
that require a more formal language. Important political or cultural activities that
are conducted orally use Italian exclusively. The lack of wide use of Arbresh has
gradually diminished its expressive force, so that it can now only partially fulfil the
needs of everyday conversation, which has strengthened the language shift towards
Italian that is taking place in Piana. In a translation task of three Italian texts, each
one paragraph long (first a simple conversation between two persons, the other a
simple description taken from a daily newspaper, and the last, a formal analysis of
a writer taken from a high-school anthology text), given to twenty Arbresh
speakers who had completed high school, only the first text was partially translated
with many grammatical deviances and inconsistencies. Most of the speakers failed
to translate the second text and none was able to translate the third text. As this
example clearly shows, the relationship between linguistic competence eroded
from attrition and the functional range of a language is very strong. The Arbresh
data reinforce one of the five major findings in the study of language death, as
analysed by Lukas Tsitsipis, that “significant structural restrictions in grammar
have been correlated with reduction in speech genres” (Tsitsipis 1989, 117). The
school plays a very important role in the development of speech genres, which
increases the functional abilities of a language. As mentioned above, the restricted
use of language itself becomes a reason for further structural decay, in this way
reinforcing the process of language attrition.
Arbresh has never been an official language of instruction, but there are still
children in the early grades of elementary school who feel more comfortable when
class communication is in Arbresh rather than Italian. A decade ago the number of
children who had a stronger competence in Arbresh than in Italian when they
started elementary school was even higher, and few decades ago the children
would start school without any significant knowledge of Italian. The lack of the use
of a written form of the language that can allow people to write and read it,
therefore enlarging its functions and uses, and the lack of instruction of Arbresh in
school, is the cause of a radical drop in the linguistic abilities of Arbresh speakers.
In the survey of 100 Piana residents mentioned above, seventy-four said that they
used Arbresh more easily than Italian when they started elementary school. But
only twenty reported using Arbresh more fluently than Italian at the end of middle
school. There were only three participants in elementary school who felt equally at
ease speaking both languages, and only four at the end of middle school. The
difference is that in elementary school about 80 per cent feel more comfortable
with Arbresh, while in high school about 80 per cent feel more comfortable with
Italian.
This dramatic change occurs at school, during the years in which Arbresh speakers,
while enriching and developing their Italian skills, do not add anything to their
knowledge of the language. On the contrary, in a competition for domains, Italian
wins and Arbresh loses. From interviews and observations in different settings of
the community, I have noticed that although almost everybody in simple informal
communication is naturally inclined to use Arbresh, it is often difficult to maintain
its use in a long conversation. The language rapidly becomes overloaded with
elements from Italian vocabulary and expressions, until a complete switch to Italian
occurs. Often young Arbresh, because of gaps in their linguistic knowledge, feel so
uncomfortable speaking Arbresh that they switch to Italian even in relaxed and
informal conversations. 1 But even when the switch to Italian does not occur,
Arbresh shows clear signs of grammatical and lexical attrition. Despite the
functional role of Arbresh as the informal means of communication in Piana, a
large number of speakers, especially younger ones, speak it less, and less fluently,
than older generations. While still being used in informal settings, the language is
slowly decaying and losing its expressiveness.
At present, although there are no sanctions against Arbresh in the Piana job market,
neither is there any reward for knowing the language. An Arbresh speaker who
wants to teach in the schools of the community or work in its offices does not have
priority over an Italian candidate who does not speak Arbresh, even though a large
majority of the people of Piana speak Arbresh in preference to Italian. Something
written in Arbresh, as rare as that may be, is obligatorily translated into Italian even
when it is intended for Piana use only, such as an advertisement for a local show;
but if it is written in Italian no translation is considered necessary. The monolingual
Italian media, especially television, and the lack of any Arbresh media, facilitate
the shift to Italian, especially for young people and children who spend a lot of
time watching television. The media pressure is forming a new negative pheno-
menon: the modern world is indexed in Italian, the old one in Arbresh. An old man
with beautiful Arbresh but not very cultivated Italian, who needed to solve a
problem in the post office – the post offices are important institutions in Italian
towns and the main mediators between individuals and state or private companies –
asked his niece, a university student, to go there and solve his problem. “He could
not speak in Arbresh there, although the employee spoke Arbresh. And it is not
1
This is not the kind of code-switching to which Milroy and Muysken (1995) refer as one that “does
not usually indicate lack of competence on the part of the speaker”, nor one that tends to
accommodate the interlocutor (Giles and Coupland 1992) or to fulfil complex interactive strategies
(Gumperz 1982). This code-switching occurs because speakers cannot express themselves in
Arbresh and therefore have to switch to Italian.
255 Eda Derhemi
even useful complaining in Arbresh: nobody would solve your problem. While a
person who spoke a refined Italian could be more convincing and show that he was
right.” Literacy is often translated as ability to develop networks of relations with
the state system. The relation between the written Italian and the unwritten Arbresh
produces the same negative ideological discourse described by Tsitsipis (1998:19)
about Arvanitika dialect in Greek villages: "The belief that a written language has a
superior status has come to be accepted in the local level through the influence of
schooling and the media." A contradictory discourse between Arbresh and Italian
begins to replace a non-antagonistic one, and in the relations between the two
languages, Italian appears as the authoritative language.
The inequalities between Italian and Arbresh arise in the practical usage of the
languages, in which Italian is fast replacing Arbresh, while Arbresh slowly
degenerates structurally and the competence of its speakers declines. The
inequalities are not based on the stigmatisation of Arbresh or on differences in the
socio-economic status of speakers. The case of Piana shows that linguistic attitude
plays an important role in the process of maintenance of a language, but it is not
decisive in this process. Nor, in the case of Arbresh, is the relation of the
endangered language to one particular functional discourse decisive.2 Arbresh is
still the main means of communication used in everyday conversations among
members of the Arbresh community, although at very different degrees of
competence. But the language is undergoing attrition because the level at which it
is functioning does not have the prestige of an informal level and because its
function as an informal means of communication is becoming increasingly
restricted. In their preface to Endangered Languages, Grenoble and Whaley (1998)
claim that a pervasive predictor of the use or the loss of a language is the prestige
attached to it. They also list the reasons that give prestige to a language, such as
“government support”, “large number of speakers”, “association with rich literary
tradition”, “use in local or national media of communication”, “use in
economically advanced commercial exchanges” and “use in a widely practised
religion” (11). The informal use of Arbresh does not fit any of these characteristics
believed to derive prestige for a language. The development of a written
standardised form and the use of Arbresh in school, together with the return of past
and lost literary traditions, are key factors in the process of raising the prestige and
the usage of Arbresh, therefore to the revitalisation of the language.
As I mentioned earlier, the people of Piana generally do not write their language,
although they gain some knowledge of Arbresh spelling in elementary school from
teachers who themselves are not trained to teach Arbresh but who do it as a labour
of love. However, people have never completely stopped writing the language.
Over the years, used now and then by one intellectual or another, the written form
of Arbresh has gained a symbolic power, precisely because very few were able to
write it. But it has never reached the level of a codified language. Every writer has
2
The clergy sometimes use Arbresh to serve the separate Greek-Byzantine rite of Piana for special
occasions and as a symbolic gesture, but other languages such as Greek, Latin and Italian are also
used.
followed personal views about the choice of alphabet and correct forms. The only
exception to this variability of forms and systems is the writings of some poets over
the last three decades who have used Standard Albanian instead of Arbresh, and
therefore show a normative consistency and lack the grammatical deviances that
have now spread to different degrees throughout the community. In the struggle
between written and oral, the prestige that comes through the mechanism of
codification of a language goes to the languages written in the community, to
Italian and, when used, to Albanian. The very sparse knowledge of written Arbresh
and Arbresh grammar was once transmitted to the community through the Church.
Now, as the Church no longer has much control over young members of the
community, who are also the most vulnerable from a linguistic perspective, this
duty falls exclusively to the schools.
3. Some Instances of Variability and Loss

Below I analyse some expressions of decay, first at the morphological level and
then at the phonetic level, that demonstrate the indispensability of Arbresh
instruction at school, and suggest how the inclusion of Arbresh should be
structured during the implementation of language policy. The importance of the
written model and a linguistic norm is also emphasised as ways to create models
for speakers in the linguistic confusion created by attrition and to protect linguistic
knowledge from further loss. The lack of a codified norm and written form is not
the only factor that causes and accelerates the instability and loss of forms in the
decaying language, but their existence would certainly have been decisive for the
process of language survival and restoration, in the climate of the very positive
attitude of its speakers.
The lack of a written norm for Arbresh grammatical categories and forms that is
distributed throughout the community has led to the existence of multiple versions
of a great number of words and grammatical structures, and a high degree of
variability in the actual use of the language. The children hear words pronounced in
different ways by different people and sometimes in different ways by the same
person. They reflect this confusion when they communicate, particularly when they
try to write in Arbresh: the phonetic and morphological image of the expression in
their mind is weak and blurred. But aberrant forms at all linguistic levels are
common, independently of age and education.
Some examples from the variability in the verbal system shows the advanced
degree of loosening of the system and, on the other hand, the importance that a
codified language and the written form has today in protecting the language from
decay. In speakers between 19 and 45 years old I found five participle forms of the
verb “to eat” in complete free variation: hëndur, hëndër, hëndrur, hëngër and
ngrënë, while among older speakers who in general demonstrate a higher linguistic
competence, I found only the form ngrënë. This is also the form testified in older
written texts. The third person imperfect indicative shows variability as well: most
speakers use the form ending in -jë, while some use the one ending in –ëj (e.g.
257 Eda Derhemi
prisjë and prisëj– “he waited”). The most recent grammar of Arbresh, Udhëtimi
(2001), considers the latter only as the form of the imperfect tense. Even the third
person singular of the verb “to be” appears to have three forms. The lack of a
written language has caused confusion about the use of the progressive aspect
among young speakers. The present and imperfect forms of the progressive aspect
in Arbresh are shown below.
1. na jemi e biem
we are and fall (first person plural, present tense)
we are falling
2. ai ish e pasjar
he was and walk (third person singular, imperfect tense)
he was walking
But the process of relaxation during informal oral communication has blurred the
morphemic boundaries between the aspectual forms into the fused forms:
1’. namebiem (we are falling)
2’. shepasjar (he was walking)
Young speakers are often unaware of the relation between forms 1’ and 1, and 2’
and 2, and they have never seen the complete written forms of these expressions.
Therefore they often perceive these forms as separate lexical items and cannot
recognise them as parts of the same grammatical paradigm.
The lack of use of some verbal forms has led to their loss in Piana: the optative
mode now only appears in a very few texts still used in occasional religious
ceremonies, and it is remembered by Piana speakers only in those two or three
expressions. Outside these contexts it is not active today. The forms of the
conditional mode, although they are considered to be present in the Arbresh of
Piana by Udhëtimi (2001), no longer exist in the specific conditional forms but
have been replaced by other modes such as indicative and subjunctive. Other parts
of speech forgotten through lack of use are the forms of the gerund still alive
among the old. Even the forms of the imperative show erosion and the mode seems
to be active only in the case of very frequently used verbs.
Another problem reflected in the spelling of many young people is the lack of any
awareness of the peculiar features of the sound system of Arbresh that do not occur
in Italian. This might seem an easy task for a bilingual, but it is not so easy,
particularly in the young, especially in a decaying language. Young speakers often
lose these sounds from their phonetic inventory, if they had ever acquired the
sounds, replacing them with similar sounds that occur in Italian. Sometimes the
occurrence of these authentic Arbresh sounds in their speech is sporadic and
inconsistent. 3 Such sounds are th and dh, voiceless and voiced interdental
fricatives, that are reduced to s and z,; and q and gj, voiceless and voiced palatal
stops that are reduced to kj or k and g+j or g. If the speakers were exposed to an
Arbresh writing system, it would be easier for them to realise that different
graphemes must represent different sounds of the language. This is not a guarantee
of using these sounds, but at least Arbresh speakers would be more aware of their
existence and it would exercise a corrective pressure on them.
In one of the best journalistic expressions of the local press in Arbresh, the
newspaper Mondo Albanese (Albanian World) published from 1981 to 1984, more
than one satirical poem and story was published on the loss of these sounds by
young speakers, which tended to raise awareness of loss and tempt people into
making a deliberate effort to maintain sounds.
4. Implementing the Legislation

Now that the language has reached a critical level of endangerment, a very well-
researched, flexible and co-ordinated language policy is necessary. An intensive
effort began in the community of Piana degli Albanesi immediately after some
significant institutional steps were taken towards the recognition of the Arbresh
linguistic communities in Sicily. First, in 1998 a regional law (No. 26) for the
protection of linguistic minorities was passed, but it was amended and weakened
by the State Commissar who thought that certain rights should remain within the
competence of the state rather than the region. This legislation was further
modified by a regional bill but the result, although an improvement, was still
considered to be a very tardy response to a long-standing demand. What finally had
a powerful impact on community efforts to revive the language was a national law
(No. 482) for the protection and promotion of Albanian linguistic and cultural
historic minorities that was passed in December 1999, as it was more complete and
radical than the regional law. The work of implementing this legislation and
breaking it up into practical projects eligible for state funding is still under way.
As community leaders have emphasised for decades, while trying to gain

recognition as a minority and hence be entitled to institutional protection, the only
way to change the situation of Arbresh is intervention by the Italian state. One of
the main promoters of Arbresh language and culture, Pietro Manali (Damiani 1999,
5), referring to the complexity of the rights of linguistic minorities, says that the
only solution to this problem is, in euphemistic terms, time: “often, as in the Italian
(state intervention) case, there has been no other solution … but time, which,
3
Some young speakers still maintain the full inventory of Arbresh phonemes, although in current use
of the language there are inconsistencies in the occurrence of such sounds. But among middle-aged
and young speakers the sound system is very well maintained. The decay in the phonetic system is a
sign of an advanced degree of attrition of the language: Hamp (1989, 200) observes that “in a
healthy language conservative phonology is the expected thing”.
259 Eda Derhemi
always generous, fixes things and resolves the problem”. But time alone is a
passive factor, and it will not change the situation of Arbresh. Obviously, Manali
gives no credit to the state for its very delayed action when he speaks of the
intervention of time. In fact the efforts for recognition mounted in the regional and
national parliaments by community leaders from Piana and other Arbresh
communities in Sicily, particularly Contessa Entellina, have been unsuccessful for
decades.
There is no doubt that the root cause of the recent degeneration of the Arbresh
linguistic system is the long-term lack of institutional intervention. As the use of
Italian intensifies in all spheres, its position strengthens in relation to Arbresh. In
this competition of vocabulary and grammar, lack of use has caused Arbresh to
lose many lexical units and grammatical details, especially among younger
generations. The only way this loss can be recovered is through the introduction of
Arbresh in the school system as an obligatory course. But such a course would
require at least one good grammar textbook that sets out a normative system and a
teaching model, as well as some trained instructors.
One achievement of the Piana community is that legislation now allows for the
instruction of Albanian and its use in the teaching process in elementary and
middle schools, and the training of the necessary instructors. Efforts by community
leaders towards the implementation of the law began immediately after it was
passed, awakening cultural life in Piana and generating funds that in turn
accelerated the economic life of the community. In a period when Italy is in
continuous economic and political turmoil, this type of prosperity is quite unusual,
particularly for a small southern community like Piana. The process is now
beginning to involve other Arbresh communities in Sicily, five of which are
included in official statistics of Arbresh minorities in Sicily, Piana being the largest
among them. Recently their action has been echoed in other parts of Sicily and
Italy, even in communities that long ago lost the language, customs and religion
but still have some vague remembrance of it. The municipalities of these
communities are actively trying to find ways to be part of the movement centred on
Piana.
Piana community leaders, especially public library activists, municipal officials,

leading teachers in Piana schools, local writers, priests, and the faculty of the
Department of Albanian at the University of Palermo, took immediate advantage of
the favourable national and international situation after the legislation was passed.
With funding from the European Commission and the authorities of Palermo
Province, Palermo Commune and Piana Commune, they have constructed a three-
stage project. This project is designed to create the foundation for the future
protection of the language and culture of this community, especially for the
addition of Arbresh as an obligatory language course in Piana schools. Below I
analyse how much the legislation has achieved towards the creation of a codified
language and its use in Piana schools, both goals that would contribute a great deal
to the Arbresh revival, as argued above.
After the 1999 law, the Piana project was organised in three stages: “Skanderbeg
3000”, “Kastriota 2001” and “Brinjat”. All these names are symbolic: the first two
refer to George Kastrioti Skanderbeg, the Albanian national hero who fought
against the Turks just before the ancestors of the Piana Arbresh left Albania for
Italy, and the third to an Arbresh place name. The main goal of the first stage was
the production of a textbook for Arbresh instruction in elementary and middle
schools. The goal of the second was two other volumes that are now in press: one
is a guide for schoolteachers and other Arbresh instructors; the other is a full
grammar at a higher level than the grammar produced at the first stage. The third
stage has diverse goals, including encounters among students and teachers of
different Arbresh communities in Sicily, and the publication of Piana writers who
represent the strength of the past cultural and written tradition. Some of these
events have already occurred. Others, such as the publication of a CD-ROM and
the inauguration of various exhibitions, are ongoing.
The first stage concluded with the production of a basic text to help with Arbresh
instruction in elementary and middle schools. Community leaders are aware that,
as argued in the linguistic analysis earlier in this paper, the main problem today is
the creation of linguistic norms for Arbresh, and the insertion of these norms and
other grammatical and lexical restorative devices in the schools at all levels. This
would end the haphazard and inconsistent use of forms and words and would be a
first step towards neutralising and then defeating the process of obsolescence. After
the law for the protection of Arbresh was passed, the commission responsible for
its implementation began a complex effort to produce a manual to teach elementary
schoolchildren the elements of correctly writing and reading Arbresh. Development
of the contents was assigned to three authors: Giuseppina Cuccia, a prominent
schoolteacher and community leader, and two of the main poets (and teachers) of
Piana mentioned earlier: Giuseppe Schirò Di Modica and Giuseppe Schirò Di
Maggio. Besides these three, a scientific and a technical committee of nine was
appointed to oversee the work. An international seminar was held before the work
began, to ensure that the book would be based on the experience and good practice
of those who had worked on similar issues before.
The resulting book, Udhëtimi, was published in 2000, in an edition of 2,000 copies.
Its 240 pages include an ABC, a grammar and an anthology of Arbresh pieces, with
illustrations.
This book, although the principal result of the first stage of the project, has not yet
been regularly used in Piana schools. It has been criticised by many community
members who claim that its parts do not combine to make a coherent whole. It is
unclear what norm the book represents: clearly not Arbresh grammar or Arbresh
vocabulary, but not Albanian either. Even members of the committee involved in
the compilation are dissatisfied with the result. Some instructors continue to use
materials that they have collected themselves, which creates even more incon-
sistencies than the book itself.
261 Eda Derhemi
The criticisms of the book have some merit, as for example whether it reflects
Standard Albanian or Arbresh grammar, or a third standard based on both
languages. Standard Albanian grammatical categories have been maintained, but
Piana Arbresh has lost many forms that would fit into these categories. In the
grammatical tables, and in various chapters, notes such as “rare” appear over
certain forms, implying that the grammar is based on Arbresh. However, there are
no such notes on other forms that today do not exist at all in Piana (for example,
the verbal forms mentioned in Section 3), but that are erroneously considered
extant in the paradigmatic tables. On the other hand, the readings in the book are
neither in Standard Albanian nor Arbresh. In a group of seven boys and girls from
Piana middle and high schools, none could understand the meaning of two non-
conversational pieces chosen from the book. It also lacks a final overall editing to
avoid inconsistencies in the use of words, forms and constructions. The direction of
language attrition in Piana Arbresh, and the multiple aberrant forms described in
Section 3, fully support the Dauenhauer and Dauenhauer claim that in language
endangerment situations “if literacy is taught, it should be standard and consistent”
(1998, 90).
The result is that, although the book is the main result of efforts to produce a
codified model for the first stage of the project, schools are not yet using it. It is
nevertheless the only textbook available. Any critical comments will be helpful in
producing future works on the language, and should be taken into account when
using it as a textbook. Community members who do not agree with the choices
made in the book need to concentrate on its positive aspects, in that it is better
researched and more complete than any of the groups of materials that teachers
have assembled over the years. Besides, the book is relatively attractive to children,
with its colourful pages and the nicely organised rubrics covering exercises, drills,
grammatical rules, lexicon items and idiomatic knowledge. Those parts that cannot
be understood by different users could be treated as a challenge for those who want
to learn more. Hence the book can still be used, albeit critically, until a better
version comes along.
On the other hand, some of these problems could be partially offset by the two
books that are the goal of the second stage. Teachers who will be using the
textbook can consult these other books to seek clarification and answers to any
questions they may have. This assumes of course that the forthcoming guide for
Arbresh instructors and the comprehensive grammar reflect in a realistic way the
linguistic knowledge of the community and have taken careful, studied steps to
replace the components of Piana Arbresh that have been lost in recent decades.
Although different people were responsible for developing the guide and the
grammar, if these books share the same principles and position on grammatical
choices concerning the lost forms of Arbresh and the way they should be replaced,
together with the textbook from the first stage they will constitute a good starting
point for the addition of Arbresh to the school curriculum. But even if there are
differences between these books, as they are the only ones available at the moment
they need to be used and they need to be used soon. The following admonition is
appropriate to the current situation of Arbresh: “to have to argue such points now is
to take time out to ‘rediscover the wheel’ when the real issue is to use all kinds and
sizes of wheels more effectively and more interactively” (Fishman 1985, 54).
As argued above, immediate intervention is needed to target the young generations

in particular. This can only be achieved through schools. Based on the texts dis-
cussed above, the introduction of Arbresh instruction to all Piana schools would
significantly reduce the inequalities of use between Arbresh and Italian. As many
community leaders have noted, Arbresh should not be an optional course (as it was
before the law for the protection of Arbresh was passed), but obligatory like Italian.
Later, when teachers and students have reached a more advanced stage in the
mastery of the language, the teaching of other courses in Arbresh as well as in
Italian can be taken into consideration. The Director of Piana Schools, Pasquale
Ferrantelli, points out that this process will be very slow and it will take years
before instruction in Arbresh begins. The funding received for this purpose is only
40 per cent of the amount requested by the schools to fulfil this mission. However,
Mr Ferrantelli says: “We are happy. This is better than nothing.” The schools
received nothing from state institutions until two years ago, and the efforts of
teachers to form groups of students to study Arbresh were voluntary. The
compilation of different types of grammars and other Arbresh textbooks and the
use of the language in schools are now possible financially and legally thanks to
the 1999 legislation. As for the means of achieving these ends, more than one type
of bilingual education has been suggested for minority bilingual children
(Skutnabb-Kangas 1981; Skutnabb-Kangas 1999), but the “Fishman type 4”
characterised as “complete bilingualism” (Skutnabb-Kangas 1981, 124), seems to
be the method that this community is seeking to apply. It prescribes the use of both
languages at school in all linguistic functions (understanding, speaking, reading,
writing, thinking) in all domains and for all subjects. The specific goal of this
method is to maintain and develop the minority language.
At present, since the passing of the minorities legislation and the financial support
received, there is a wave of intense activity in Piana. But, as some community
leaders have pointed out when interviewed, this wave mainly involves the upper
level of the community, the intellectual elite that gathers at national and
international congresses, but little is passed on to the wider community. These
community leaders nostalgically recall the late 1970s and early 1980s, when
Arbresh radio and the Mondo Albanese newspaper involved the local people. Such
comments point to a Weberian social closure, described as the action of social
groups that “restrict entry and exclude benefits to those outside the group in order
to maximise their own advantage” (Bilton 1996, 669). In sociological literature the
desire for “closure” and the need for “disclosure” is seen as occurring not only
among intellectual groups, but all kinds of groups that consider themselves to be
privileged in a certain direction (Lamont 2001). Although the cultural elite of Piana
has reason to be proud of its work and leadership, there is always need for
awareness of possible closure, which can be fatal in conditions of language
263 Eda Derhemi
endangerment. If the movement for language reaffirmation and revitalisation is

restricted to the elite, there is little hope of changing the present state of affairs in
Piana. Reaching the grass roots of a community should be the main goal of a policy
that targets changes in the language of that community.
But the work has begun; state institutions should satisfy the requirements
prescribed by the law for the protection of Arbresh language and culture; and funds
to support the change are being made available. The climate of “unprecedented
European support for multilingualism and an overspill of protective enthusiasm for
smaller languages” has made it possible that “even minority languages within the
EC countries have gained a certain increased recognition and at least a few
economic benefits” (Dorian 1998, 19). The new political and economic changes at
European level 4 have favoured the realisation of the long-standing hopes and
efforts of the Piana community. Future plans of the community include the
reinforcement of an Arbresh “linguistic market” (Bourdieu 1991, 49), that consists
first of all of new jobs for teachers of the Arbresh language, teachers of other
courses who are Arbresh speakers, workers in other cultural spheres of Piana life
related to the language and ethnicity. As the law allows for the use of Arbresh in
institutional offices and its use in schools not only in language courses but in other
courses, Arbresh-speaking teachers will no longer have to compete with mono-
lingual Italian teachers from other towns and regions, and will no longer have to
leave Piana in search of work. Community leaders also foresee economic growth
related to the new conditions, which will not only promote the further use of
Arbresh but will create better living conditions in Piana and end the dispersal of the
Pianioti around Italy. Along with these improvements, Piana will be able to attract
more tourism, drawn by its unique language, customs and religion. The small
merchants in Piana’s shops, bars and restaurants depend heavily for their existence
on the tourists who often visit Piana, especially at weekends. Eventual loss of the
language would probably be followed by the loss of other characteristic features, as
has happened recently in tens of other small Arbresh towns, and Piana would lose
its attraction for tourists, a major wealth-generating factor in this small and non-
industrial town.
5. Some Notes on the Future …

The work of many researchers, particularly Fishman (1991), clearly shows that the
reversing of language shift requires reconstruction of the decaying language and
language planning. Although new opportunities raise problems that are not always
solved in the best possible way, there is a great deal of activity in the Piana
community. To fulfil the wishes of the community requires the careful, flexible and
democratic organisation of all the positive elements that now are coming together
for the first time.
4
For more on changes in the EC, see Niamh Nic Shuibhne, 2001.
The inclusion of Arbresh teaching in Piana schools, accompanied by strengthening

of the written form of the language – or an enriched version of it – will gradually
raise speakers’ competence. It will contribute to enlarging their vocabulary and the
sphere of Arbresh use. The grammars and texts, aiming at the systematic
presentation of Arbresh and filling gaps in lost knowledge, clearly cannot afford to
add their inconsistencies to those of individual Arbresh speakers. The grammar
should represent a single normative position chosen by the community. Examples
of uncontrolled fluctuation of forms, some of which are discussed in Section 3 of
this paper, show the need for a code with corrective normative pressure. Un-
fortunately, a norm for Piana Arbresh has still to be discussed and selected. The
spread throughout the community of the written form of the language and the
restoration of lost components should be one of the main objectives for the future.
There is still a chance to save Arbresh and to preserve and maintain the original
characteristics of this community that are so important for its cultural and
economic survival. My survey with 100 Piana Arbresh speakers introduced in
section 2 shows that about half of Arbresh speakers think that their language has no
problems. Often, speakers who in a survey claim that they can express themselves
in Arbresh in any given situation, are unable to do so when interviewed and asked
certain questions, or observed in natural conversation. The awareness of language
problems is not so high among community members as is the awareness of
belonging to a minority group. Perhaps this is the point where the work should
begin in Piana: allow people to see the linguistic problem and realise the real
danger of losing their language.
The community elite, who until the early 1990s supported the use of Standard
Albanian in Piana, now supports the use of Arbresh as the language of Piana
schools. This idea has travelled with incredible speed throughout the community
and has been embraced by the mass of Arbresh speakers. There is a good reason for
this: the Pianioti wish neither to change their language nor to learn another. But is
there a way to change the status of Arbresh, from a language that does not satisfy
the natural needs of its speakers to a healthy language, without a huge investment
of resources in all types and forms of linguistic communication? Albanian, a
language that has been written for centuries and functions today in all domains,
levels and registers, has all the necessary resources that Arbresh needs. Kosovo, an
Albanian-speaking community in Yugoslavia, makes full use of Standard Albanian
and considers it to be the standard language of the community.5 Although of a very
different character, the Kosovar dialects are not much closer to Standard Albanian
than are the Arbresh dialects. If the Piana community wishes to have its own
standard language, the chances of successfully creating such a language based on
Arbresh, and maintaining it with all that a language needs to function normally, for
a population of 7,000 people, are very low.
5
The difference is that Kosovo has undergone a long, forced, severe pressure for assimilation, while
the Arbresh communities have not. For Kosovars the need to grasp Standard Albanian was a vital
patriotic and political act. Now that Kosovo feels freer and Serbian pressure is felt less, a movement
for its own standard language has begun, although it is still limited and weak.
265 Eda Derhemi
Thoma Rrushi, one of the members of the commission for the linguistic
implementation of the 1999 law, feels that this is not the best approach for Arbresh.
Instead he suggests the Kosovar way, with interaction between Albanian and
Arbresh based on Standard Albanian. He considers the teaching of Standard
Albanian in Piana schools to be indispensable for the successful maintenance of
Arbresh (Rrushi 2000). Giuseppina Cuccia, another member of the commission,
thinks that the goal for the future written language of Piana should be Standard
Albanian, but a gradual passage from one language to the other should be planned
and studied. On the other hand, other members of this commission, such as Schirò
Di Maggio and Schirò Di Modica, think that Standard Albanian could be used as
an additional resource, but the codification of Arbresh should be based on the local
dialect. Another active implementer of the linguistic part of the new law, a
dedicated teacher and diligent promoter of Arbresh among the young, Giuseppe
Scalia, follows the same line of focusing on the local dialect. There is a basis for
their opinion: the Arbresh people find it very difficult to understand Standard
Albanian. But the language they propose in their grammar is in fact not understood
by the community either, because of the natural tendency of the authors to fill the
gaps in Arbresh with Standard Albanian, a language they know well and are able to
use, and even to be creative. The language that has served as a model for Udhëtimi
is not a codified language with a normative grammar, orthography and
pronunciation. It is a simple mixture of features from Arbresh and Albanian. This,
clearly, is no solution for the language of Piana.
Although implementation of the new language policies has been in progress for
two years, there is no agreement yet on the selected code. A better approach would
be to combine the two main views of the commission for the implementation of the
legislation. One way to do this would be to adapt Standard Albanian as a written
language, maintaining the oral version of the Piana dialect, thus preparing the
ground for the natural combination of both. At present, Standard Albanian sounds
like a foreign language to Pianioti, but after some contact with it the great
similarities between the two languages will slowly become obvious. This is the
mirror image of the process that has faced all the Albanians who have had contact
with the Arbresh of Piana: they cannot understand a word the first day, but in a
week or so they can see many similarities and in few weeks they speak Arbresh. I
am confident that the written code, after being taught at school for some years, will
contribute to the oral language. The way I see the future oral Arbresh of Piana is
similar to the language used by Gerbino (in Biblos, 2001) in his translation of
Dante. Unlike other poets of the community who use Albanian beautifully as their
language of poetic expression, Gerbino translates 136 lines from the first canto of
Dante’s Divine Comedy in a very carefully and cleverly enriched dialect of Piana.
This could be a way of avoiding forcing Standard Albanian on Arbresh speakers,
and at the same time strengthening Arbresh with the help of a codified language.
There are two other theoretical possibilities concerning what code to select. They
both assume the creation of a non-existent language. The first is to create a written
version of Arbresh based on a mixture of old Arbresh, current Arbresh and
Standard Albanian. Gerbino’s translation serves as an example. The other possi-

bility is to elaborate the current “degenerating” mixture of Arbresh and Italian in a
way that will create a healthy language from a decaying one. This task would be
even more difficult, as differences between Italian and Arbresh exist at all
linguistic levels. Both methods would be difficult to realise, with little chance of
succeeding as natural languages, and would be extremely costly in the present
climate of communication standards.
With no agreement on a selected code, efforts to stabilise the language will lead
nowhere. The authorities in Piana responsible for implementing the law need to be
aware of the importance of the selection of the language to be taught in schools,
and of the pros and cons of their decisions, otherwise funds will be lost together
with Arbresh’s chances of survival. One should keep in mind that “an ethnic
language once lost is far less easily recovered than other identity markers, and the
cultural content that language carried is never fully recoverable” (Dorian 1999, 34).
Decisions must be made not only to solve the current problems in the simplest
possible way, but also in a way that will resist the passage of time and have
meaning in the future. This would make the language policy of Piana a sustainable
process that will satisfy the community not just today, but in the long term.6
In this paper I have analysed the extent of Arbresh’s endangerment, focusing on the
need for a written and codified form of the language. I see the process of
standardisation as the basis of language reconstruction and therefore as the first
step in language shift. “Standardisation is the single most technical issue in
language reinforcement. Unless it is accomplished, literary production and the
expansion of literacy will always be problematic, because people need both, good
models and a certain amount of technical reference materials to be comfortable
with literacy” (England 1998, 113). This assertion was made concerning the Mayan
language, but it fits the Arbresh situation perfectly.
I support the use of Standard Albanian as a basis for only the written form of
Arbresh to be used in the schools of the Piana community. The reasons for this are
related to the current conditions of Arbresh and Albanian:
(1) Arbresh is significantly damaged by attrition and needs a normative form to

help the community to create a correct model at almost every linguistic level.
(2) Arbresh is very limited in its literary functions and other oral domains, and has
a very restricted amount of publications. Albanian is a cultivated language in
6
This section presents the situation in Piana in February 2002. From my contacts in the community I
have learned that the two grammars from Schirò Di Modica and Schirò Di Maggio have recently
been published, one under the name Udhëtimi paralel (Parallel Travel), the other Udha e mbarë
(Have a Good Trip), but I have not yet been able to consult either. The training of Arbresh teachers
in Piana has also begun this summer (although it consisted of a few hours only), and surprisingly it
has been conducted mainly in Standard Albanian. Instruction in Arbresh as an obligatory language
at school has not yet begun, but the book Udhëtimi has been used in a few courses taught this
summer on a non-curricular basis, as in the years before the legislation was passed.
267 Eda Derhemi
all forms and has publications on a large scale both within and outside
Albania. It possesses the necessary “reference materials” so important to the
survival of Arbresh. I do not see any possible functional expansion of Arbresh
as it is today. Albanian would connect the 7,000 members of Piana with a
much larger community of speakers, readers and writers – in spite of the
tension between the Arbresh community and recent Albanian immigrants.
(3) The cost of using a language that is alive and ready for use is lower than the
cost of reconstructing a language and then trying to make it available to a
community of speakers – even if such a reconstruction would work.
(4) I do not propose the replacement of Arbresh with Albanian, but the existence
of both in parallel, with Arbresh stronger in oral discourse and Albanian in
written discourse. The contribution of Albanian, as I see it, will consist mainly
of reconstructing the grammatical structure of the language, which has a very
similar base. The lexical interaction, where the differences between the two
languages are greater, is secondary and can proceed very gradually. The goal
is not a merger of the two, but rather the use of Albanian elements to support
the reconstruction of Arbresh.
The main problems raised by my proposals are:
(1) The differences between the two languages must be dealt with, although there
are fewer than in other languages with similar links, such as Jewish languages
around the world and Hebrew (King 2001, 214). The main issues to be
overcome here are the differences in pronunciation and vocabulary, but as I
propose the use of Standard Albanian only in written discourse, ways can be
found to create a natural interaction between the two languages.
(2) A relatively unfavourable sociolinguistic situation has arisen in Piana during

the last decade with regard to anything that relates to Albania, and the
continuous flux of Albanians coming to Italy reinforces that tendency. I
consider the matter of the social meanings of any use of Albanian literacy,
described as attitudes, beliefs and values of a community (Grenoble and
Whaley 1998, 33), as a very delicate and complex issue that needs more
attention than the first problem I have raised.
I do not consider that my proposed strategy would work for every minority
language or even for every minority endangered language. In fact the opposing
opinion, that the school-selected language does not need to be a
normative/standard/codified language, is not new among linguists (Spolsky 1986,
184–5). But I think that my approach takes into consideration the increasingly
endangered situation of Arbresh and its specific features, including the low
proficiency of its speakers, particularly the young. The decay of the language, its
grammatical inconsistencies, and its variability from speaker to speaker are the
main factors supporting the need for a codified written form.
In spite of my optimism regarding the survival of Arbresh, I prefer to end this

paper with a very important warning from a researcher who has contributed so
much to the field of endangered languages, Nancy Dorian: “The existence of a
writing system and even the existence of a notable literature do not necessarily
ensure that a language will survive as a living speech form, much less thrive”
(1998, 11). The communities and researchers who work on issues of endangered
languages should bear in mind how vulnerable this domain is and the importance
of every interventional step to the future of the languages.
References
Biblos. 2001. Anno VIII, nn. 17–18–19–20. ENGLAND, N. 1998. “Mayan efforts towards
Palermo: Comune di Piana degli Albanesi. language preservation”. In: Grenoble and
Whaley, eds., op. cit.
BILTON, T., BONNETT, K. and P. JONES. 1996.
Introductory Sociology, 3rd ed.. London: FISHMAN, J. A. 1972. “Varieties of ethnicity and
Macmillan. varieties of language consciousness”. In: Dil,
ed., op. cit.
BOURDIEU, P. 1991. Language and Symbolic
Power. Cambridge: Harvard University FISHMAN, J. A. 1991. Reversing Language
Press. Shift: Theoretical and Empirical Foundations
of Assistance to Threatened Languages.
CAMPBELL, L. and M. C. MUNTZEL. 1989. “The Bristol, Pa.: Multilingual Matters.
structural consequences of language death”.
In: Dorian, ed., op. cit. FISHMAN, J. A., ed. 1999. Language and Ethnic
Identity. Oxford: Oxford University Press.
DAMIANI, G. 1999. Il diritto delle Minoranze
tra individuo e comunità. Quaderni di Biblos: FISHMAN, J. A. et al. 1985. The Rise and the
Società e Istituzioni 9/3. Palermo: Comune di Fall of the Ethnic Revival. Berlin/New York:
Piana degli Albanesi. Mouton de Gruyter.
DAUENHAUER, N. and R. DAUENHAUER. 1998. GILES, H., ed. 1977. Language, Ethnicity and
In: Grenoble and Whaley, eds., op. cit. Intergroup Relations. New York: Academic
Press.
DIL, A. S., ed. 1972. Language in Socio-
cultural Change. Essays by J. A. Fishman. GILES, H. and N. COUPLAND. 1992. Language:
Stanford: Stanford University Press Contexts and Consequences. Milton Keynes:
Open University Press.
DORIAN, N. 1981. Language Death: The Life
and Cycle of a Scottish Gaelic Dialect. GJINARI, J. and G. SHKURTAJ. 1997.
Philadelphia: University of Pennsylvania Dialektologjia. Tirana: ShBLU.
Press.
GRENOBLE, L. and L. WHALEY, eds. 1998.
DORIAN, N., ed. 1989. Investigating Endangered Languages: Current Issues and
Obsolescence. Studies in Language Con- Future Prospects. Cambridge: Cambridge
traction and Death. Cambridge: Cambridge University Press.
University Press.
GUMPERZ, J. J. 1982. Discourse Strategies.
DORIAN, N. 1998. “Western language Cambridge: Cambridge University Press.
ideologies and small-language prospects”. In:
Grenoble and Whaley, eds., op. cit. HAMP, E. 1989. “On signs of health and death”.
In: Dorian, ed., op. cit.
DORIAN, N. 1999. “Linguistic and ethnographic
fieldwork”. In: Fishman, ed., op. cit. KACHRU, B. and C. NELSON, eds. 2001.
Diaspora, Identity, and Language Commu-
nities. Special Issue of Studies in the
269 Eda Derhemi
Linguistic Sciences, 31:1. Department of SKUTNABB-KANGAS, T. 1981. Bilingualism or

Linguistics, University of Illinois at Urbana- Not: The Education of Minorities. Clevedon,
Champaign, USA. UK: Multilingual Matters.
KING, R. 2001. “The paradox of creativity in SKUTNABB-KANGAS, T. 1999. “Education of
diaspora”. In: Kachru and Nelson, eds., op. minorities”. In: Fishman, ed., op. cit.
cit.
SPOLSKY, B., ed. 1986. Language and
LAMONT, M. 2001. “Symbolic Boundaries”. Education in Multilingual Settings. San
International Encyclopedia of the Social and Diego, Calif.: College-Hill Press.
Behavioral Sciences, edited by Neil J.
Smelser and Paul B. Baltes. London: SPOLSKY, B. 1986. “Overcoming language
Pergamon Press. barriers”. In: Spolsky, ed., op. cit.
MILROY, L. and P. MUYSKEN, eds. 1995. One TRUDGILL, P. and G. TZAVARAS 1977. “Why
Speaker, Two Languages: Cross-disciplinary Albanian Greeks are not Albanians:
Perspectives on Code-switching. Cambridge: Language shift in Attica and Biotia”. In:
Cambridge University Press. Giles, ed., op. cit.
NIC SHUIBHNE, N. 2001. “The European Union TSITSIPIS, L. 1989. “Skewed performance and
and Minority Language Rights”. (UNESCO) full performance in language obsolescence:
International Journal on Multicultural The case of an Albanian variety”. In: Dorian,
Societies, Vol. 3, No. 2. ed., op. cit.
RRUSHI, TH. 2000. “La parlata di Piana e TSITSIPIS, L. 1998. A Linguistic Anthropology of
l’albanese commune”. In: Scanderbeg 3000, Praxis and Language Shift; Arvanitika
op. cit. (Albanian) and Greek in Contact. Oxford:
Clarendon Press.
SKANDERBEG 3000. 2000. Quaderni di
Biblos: manuali e guide 12/2. Palermo: Udhëtimi. 2001. Palermo: Comune di Piana
Comune di Piana degli Albanesi. degli Albanesi.
About the Author

Eda Derhemi is a doctoral student in the Institute of Communications Research at
the University of Illinois at Urbana-Champaign, United States. She did her
undergraduate work in language and literature at the University of Tirana, Albania,
and was a member of the faculty at the same university from 1985 to 1990. From
1990 to 1995 she lived in Sicily, and worked as a journalist for Deutsche Welle. In
1995 she entered the graduate programme in linguistics at the University of
Illinois, receiving her Master’s degree in 1997. In 2000 she worked in Paris as an
intern on UNESCO’s Management of Social Transformations (MOST) Pro-
gramme. In 2001 she did extensive fieldwork in Sicily on the endangered Arbresh
language. Her interests are in language contact, language change, and endangered
languages and cultures. Recently she has also been working on issues of Albanian
migration after the fall of communism and the cultural integration of Albanian
immigrants.
Address: Institute of Communications Research, University of Illinois at Urbana-

Champaign, United States; email: [email protected]
Endangered Minority Languages:
Prospects for Sociolinguistic Research
JOSHUA A. FISHMAN
Yeshiva University; New York and Stanford University, California
I have worked on endangered minority languages all my life (being a native

speaker of one such language and having been raised – and continue to live – in
a “language activist” family and social environment), and my professional
involvement in this topic goes back forty years. While I do not claim any
relationship whatsoever between age or years of involvement and the fruitfulness
of commentary on work in this field, reading the five papers of this issue has
certainly prompted a few reflections which may be of general interest. Obviously,
in the limited space available it is not possible to comment directly on all the
papers. Less obvious, however, may be the fact that I do not intend to excuse
myself from any of my criticisms. It is precisely because the problems that I
mention are so endemic in our field of inquiry that I make bold to mention them
here.
1. Parsimony and Complexity

Many of the authors comment that providing assistance to threatened languages is
more complicated than most of us had originally thought it would be, and indeed it
is. But the origin of this complexity seems to reside not only in the phenomena
under consideration but also in the theories or conceptual frameworks through
which we approach these phenomena. I remember being told (over and over again)
by a psychoanalytic critic of the “non-psycho dynamic” social sciences that the
problems we were studying were “far more complicated” than we imagined.
Finally, I screwed up the courage to interject “but perhaps the greater complexity
that you note is in your own theoretical baggage rather than in the problem per se”.
To some extent, of course, this is related to the well-known issue of finding the
proper “level of analysis” for the topic or results being explained. There is no way
of being entirely sure in advance that the explanatory variables are at the same
level of analysis as are the consequent variables that we are trying to account for.
Finally there is the problem of adopting a research design that permits us to tell, “at
the end of the day”, how much of the variation (or “variance” as it is referred to
technically) in any consequent variable that happens to be the focus of inquiry has
actually been explained by the antecedent variables that we have employed and
271 Joshua A. Fishman
how much remains unaccounted for. Only if the latter proviso obtains (and it can
obtain usefully only when the first two also obtain) can we really answer the
question of whether it “pays” to either add additional antecedent variables or,
indeed, even switch to a different level of analysis entirely. Needless to say, this
type of research design has rarely (hardly ever) been utilised in connection with
research on endangered and minority languages. Such being the case, we have each
gone our own way, methodologically and conceptually, and little cumulative
meeting of minds has been arrived at.
There is a long-standing preference for theoretical parsimony in scientific work.

Occam’s razor teaches us that the theory that explains most by taking into
consideration the least number of predictors or “causes” is the best. This principle,
derived from the early days of astronomy and chemistry, holds even more today
than it did then, because quantitative and computerised approaches to data-analysis
have allowed us to literally take an endless number of variables into simultaneous
consideration. As a result, the considerations of both parsimony and complexity
have increasingly come to the fore. And this is so not only in quantitative research.
My psychoanalytic discussant was not urging more quantification upon me, but,
rather, urging the introduction of many more qualitative variables. Indeed, the issue
of parsimony versus complexity far transcends the issue of quantitative versus
qualitative research. In either case, there is a need for the simultaneous appreciation
of the “greater complexity” underlying the accelerated rate of decline and demise
of endangered minority languages, while we also search for theoretical parsimony
in our conclusions about what is going on. Obviously, both of these directions of
inquiry usually cannot be going on simultaneously within the efforts of any given
researcher. Are we pursuing incommensurables? My review of the papers of this
issue has convinced me that such is not the case, although we are still far from
grasping either the full complexity or the underlying parsimony of the factors that
are in operation.
2. “From Whence Cometh my Help?”

It seems to me that in empirical science parsimony must be arrived at inductively.
Only the close examination of a large number of presumably different cases (via
“secondary analyses” or “meta-analyses”, if you will) can lead to the ultimate
recognition of the basic underlying similarities and differences between them. Thus
complexity and parsimony are mutually interdependent, the first leading to the
second inductively and the second suggesting further refinement deductively. At
the present juncture in the study of attrition in the world’s endangered minority
languages, we still lack the basic constituents from which parsimony can be
derived: a plethora of detailed case studies. It would be too good to be true to hope
that these studies might ever be conceived and executed in accord with a common
design. That is simply not the way social research is done in the modern
democratic capitalist world, or anywhere else for that matter. Nevertheless, from a
wealth of case studies (conducted at the same time or at different times) a
parsimonious theory should be inductively derived from such studies. In a sense,
Prospects for Sociolinguistic Research 272
that is what I attempted in my 1991 and 2001 books, the first based on thirteen
cases and the second on eighteen. These volumes contain all the case study
information, the theoretical integration derived from these cases and even several
independent critiques of that integration. These cases can now be subject to review
and interpretation by others, but, even more urgently, their number must be added
to, so that the empirical baseline for further theory can be augmented. Another
desideratum, even if pie in the sky, would be to have a common research site which
many different investigators could get to know well. This would overcome the
regrettable redundancy between unique site and unique investigator from which our
research currently suffers because the two cannot be disambiguated.
Of course, most of the foregoing comments apply to almost all social research, on
any site-related topic whatsoever, and I begin with them so that it will be crystal
clear why it is premature to be very happy about such research on our topic and,
accordingly, much too early to take seriously any claims as to their applied value or
prospects.
3. Metaphors
Where theory is weak, metaphors flourish. We already have a surfeit of metaphors
in the field of minority/majority intergroup relations. Such a plethora of visual
imagery is a sign of (inter-)disciplinary conceptual limitation. Where once we
spoke of the “melting pot” and of the “fruit salad”, we now tend to favour
“globalisation” and “killer languages”. While I am pleased to see both of the latter
receive their deserved come-uppance in the papers under discussion, there is also
the opposite danger of overly discrediting them instead of refining them so that (as
has already occurred with post-Second World War inquiry into “race”) no
acceptable role remains for them. Thus, while it is beneficial that we realise that the
price of our over-reliance on “globalisation”, a virtual deus ex machina of late, is a
debilitating one, it is also desirable that we continue to study the relationships
between variance in age, education, location within and between countries, social
class, rural/urban residence, religious and ideological variation, on the one hand,
and variance in acceptance of consumerism and its attendant values and lifestyles
(including language shift or repertoire enlargement and contraction), on the other.
The explicit recognition of globalisation as a continuous variable, rather than
merely as a dichotomy (“yes” versus “no”) will add precision to our research and
provide both the possibility of recognizing that some indicators are stronger than
others and of disconfirming hypotheses or at least realizing the degree to which
they are supported.
On the other hand, globalisation is by far “not the only process transpiring on the
language front” (Fishman 1999), as some of our contributors have pointed out. It is
precisely because “globalisation” and “localisation” are so commonly co-present
that the designation “glocalisation” has been coined. However, it too needs to be
calibrated and I am sure that it increasingly will be as time goes by. What, other
than “localisation concurrently with globalisation”, describes almost all our efforts
and sentiments on behalf of endangered lesser-used languages? Indeed, the

ongoing tensions between independence and interdependence, between withdrawal
and interaction, are at the very heart of all language planning per se (Fishman
2000), about which a little more is said below.
To some extent, our tendency to mystify and metaphorise our endeavours derives
from our embarrassment that “language” is not yet a fully understood variable, not
even in the so-called “language sciences”. Misery loves company, it is said, and in
connection with not fully grasping the significance of their central concern, we are,
for once, in good company. Anthropology is still struggling to define “culture”,
psychology to define the “mind”, sociology to define “society” and linguistics to
define “language”. In accordance with such lack of definite central definition, there
is also a tendency for metaphors to replace one another. The suggested characteri-
sation of language as a “parasite” is a case in point. It is doubly difficult to
precisely investigate and conceptualise the relationship between language and
culture, for example when both variables are substantially metaphorised and
thereby simplified, if not even more basically misunderstood.
A language is simultaneously indexical of both the material and the non-material

realia of its traditionally associated culture, symbolic of that culture (and of
membership in that culture) and, therefore, like all symbols, easily politicised, and,
finally, language is also part and parcel of the bulk of any culture (note the
complete interdependence of language and laws, religion, education, jokes, riddles,
songs, blessings, curses, greetings and the thousand pleasantries of everyday life).
Culture and language are in large part identical rather than merely the co-
occurrences or “fellow-travellers” that they are all too often taken to be.
Our usual problem in giving language its due is that we are trying too hard to right
the centuries-old neglect of language by the social sciences as a whole. We, in the
modern social sciences of language behaviour, necessarily focus on language. As a
result, we easily slip into implying, without necessarily intending to do so, that
language functions as a truly discrete and separable variable. We come to conceive
of it ourselves and to foster the view of it among our students and readers as a
“dependent variable” or as an “independent variable”, rather than grasping it in its
ubiquitous embeddedness, in its part-whole functioning within both society and
culture. As there is also much to gain by adopting the “independent cause and
consequence” conceptualisation of the sciences, we must always remember to
correct ourselves and to realise that this is not really the whole story in any
language in a social behaviour setting. Certainly we do not help matters by positing
yet other distinctions between social-cultural-historical settings (for example,
between former colonies versus non-colonies. or between “types of colonies”, etc.),
when such distinctions make it even more difficult to come to grips with our basic
dilemma of differentiating between etic distinctions and emic differences with
Prospects for Sociolinguistic Research 274
respect to a particular manifestation, such as reversing language shift efforts, of

language embeddedness in social behaviour1.
4. Language Planning
I hesitate to add yet another consideration in connection with the papers under
review, particularly so prominent a consideration as language planning about
which so much has already been written. Certainly language planning must now be
seen as part of a potentially sequential path between original planning, de-planning
and re-planning, such as that which Michael Clyne has illustrated so tellingly
(1997). Each of the stages along this sequential path may derive from the
authoritative allocation of resources to language, no matter how different and
ideologically opposed to one another the authorities involved may be. Let us take
care to remember that oppositional language planning vis-à-vis RLS (Reversing
Language Shift) is still language planning. But even this stagewise path, testifying
eloquently to the changeability and diversity of human goals and values as it does,
does not move us closer to fathoming (let alone fostering) unplanned and
unplannable language use in functions of lesser and greater formality and power.
However, it is precisely unplanned, informal, spontaneous and unritualised
language use that constitutes the bulk of normal language use. Accordingly, it is
exactly such unplanned language use that must become the crux of our investi-
gations. This too will require a correction in perspective vis-à-vis our previous
over-concentration on the language behaviours of governments and institutions.
Once again, the stress must be on the speech network and the speech community.
We need to learn to keep our eyes on the ball and to more often study minority and
other threatened languages in situ, where language behaviour actually and
unselfconsciously lives. Of course, we need to study authority structures, reward
systems and organisations too, as most of us have long been doing, but the balance
is now too far in that direction and some redirection of emphasis would seem to be
very much in order.
The five papers under consideration here deserve to be read and their authors to be
congratulated. They can certainly lead us to reflect on several directions in which
we have to move if the study of minority and threatened languages is to become a
source both of deeper understanding and of more efficacious assistance.
1
As far as I know, no one has ever followed up my 1989 finding that former Spanish/Portuguese
colonies had by then become significantly less multilingual than had former British or French
colonies (Fishman 1989, 59). For the continued high rate of multilingualism in most former
American (as well as in most former British) colonies, see Fishman, Conrad and Rubal-Lopez
(1996).
References
CLYNE, M. (ed.). 1997. Undoing and Redoing 435–42. Bloomington: Indiana University
Corpus Planning. Berlin: Mouton de Gruyter. Press.
FISHMAN, J. A. 1989. Utilizing societal ––––. 2000. “The status agenda in corpus
variables to predict whether countries are planning”. In: A. Ronald Lambert and Elana
linguistically homogeneous or Shohamy, eds., Language Policy and
heterogeneous. Language and Ethnicity in Pedagogy, 43–51. Amsterdam: Benjamins.
Minority Sociolinguistic Perspective, 580–
601. Philadelphia: Multilingual Matters. FISHMAN, J. A., ed. 2001. Can Threatened
Languages Be Saved? Philadelphia:
––––. 1991. Reversing Language Shift. Multilingual Matters.
Philadelphia: Multilingual Matters.
FISHMAN, J. A., A. CONRAD and A. RUBAL-
––––. 1999. “The new linguistic order”. In: LOPEZ, eds. 1996. Post-Imperial English:
Patrick O’Meara, Howard D. Mehlinger and Status Change in Former British and
Matthew Krain, eds., Globalization and the American Colonies, 1940–1990. Berlin:
Challenges of a New Century: A Reader, Mouton de Gruyter.
About the Author

Joshua A. Fishman is Distinguished University Research Professor of Social
Sciences at Yeshiva University, and Visiting Professor of Linguistics and
Education at Stanford University. His research has made him an international
leader in issues of language and ethnicity, language planning, and bilingual
education, and he is also well known for his work on societal aspects of Yiddish.
He is the founder and General Editor of the International Journal of the Sociology
of Language. He has published many books and articles, including his recent books
The Handbook of Language and Ethnic Identity (2000) and Can Threatened
Language be Saved (2001).
Email: [email protected]

International Journal On Multicultural Societies (IJMS)

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

International Journal On Multicultural Societies (IJMS)

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

International Journal On Multicultural Societies (IJMS)

Uploaded by

Copyright:

Available Formats

International Journal

Vol. 4, No. 2, 2002

Vol. 4, No. 2, 2002

The Impact of Government Policies on Territorially Based

MATTHIAS KOENIG, Editorial 148

EDA DERHEMI, “Thematic Introduction: Protecting Endangered Minority Languages: 150

SALIKOKO MUFWENE, “Colonisation, Globalisation, and the Future of Languages in 162

SUZANNE ROMAINE, “The Impact of Language Policy on Endangered 194

RAJESHWARI PANDHARIPANDE, “Minority Matters: Issues in Minority Languages 213

PIETER MUYSKEN, “Uchumataqu: Research in Progress on the Bolivian Altiplano” 235

he present issue of UNESCO’s International Journal on Multicultural

As noted in the Thematic Introduction, the debate over the preservation of

The thematic thread of linguistic diversity will be pursued in a subsequent issue

1. Main Foci of this Issue

2. Relevant Sociolinguistic Research on First-Language Attrition

At a time when it is estimated that “80% of the world’s 6,000 or so living

1. Grinevald Craig (1997) presents a general discussion of language endangerment.

loss. Dorian (1989) assembles a large number of sociolinguistic contributions,

Besides communicative functions and other sociolinguistic parameters of EMLs,

Thus research on linguistic endangerment is characterised by a holistic approach

Analysis of language policy is another important trend of sociolinguistic research

Concerning language policies and threatened languages, a “role dilemma” is

Skutnabb-Kangas (2000, 249), being herself a sociolinguist who supports

and even researchers as “a primordialist dream, creating employment for the

3. Theoretical Studies Presented in this Issue

Mufwene attempts to answer the question: Is it possible and necessary to interfere

Romaine’s paper is a response to some of the issues raised by Mufwene. Unlike

Romaine brings a large number of examples of problematic understanding and

In her paper, Pandharipande analyses the main sociolinguistic factors that

Pandharipande emphasises that these sociolinguistic factors must be included in the

As Pandharipande’s paper deals with the complex situation of EMLs in India,

Muysken discusses Uchumataqu, the language of the Uchuma in Bolivia, from an

5. Fishman’s Criticism and Some Conclusions for This Issue

Below I briefly summarise the main conclusions of sociolinguistic research on

• Sociolinguists are divided on whether they should press for intervention

• Language policies will have a stronger impact on EMLs if they are

• In order to grasp the nature of the phenomenon of language attrition and to

• In the process of research on linguistic endangerment and language policy,

civil responsibilities, sentimental positions, even the ability to follow a

FISHMAN, J. A. 1991. Reversing Language SKUTNABB-KANGAS, T. 2000. Linguistic

About the Author

Programme. In 2001 she did extensive fieldwork in Sicily on the endangered

Address: Institute of Communications Research, University of Illinois at Urbana-

The typical academic discourse on language endangerment has

Exceptions to the above observations include volumes such as Brenzinger (1998),

non-European languages in some extra-European colonies during the seventeenth

Like colonisation, the terms globalisation in English and mondialisation in French

Globalisation and mondialisation have typically been assumed to be cross-

Globalisation need not be universal, as in global warming, or regional, as in global

To my knowledge, Caribbean territories reflect some of the earliest experiences of

Things are not necessarily so different in economically more affluent former

2. Importance of Distinguishing Different Colonisation Styles

In the context of this article, in which socio-economic ecology is invoked to

Mufwene (2001) distinguishes between trade, settlement and exploitation colonies.

It is also worth determining the extent to which settlement is advanced in a

Instituting economic systems that generally reaped raw materials to be processed in

Once demarginalised and now relatively absorbed minorities, the indigenous

The closest approximation of European values is evident in the development of