Abstract
The movement to protect threatened languages from extinction is based on misleading metaphors and false sentiment. Of course, linguists should do their utmost to document languages that are in danger of being entirely abandoned. But whether speakers of such languages should continue to use them or adopt more current languages is a decision no outsider can make for them. Threatened languages are not like threatened species: they do not die out but are abandoned by those who used to speak them. Small language communities do not necessarily add to cultural diversity, they may isolate their speakers from all cultural alternatives, they may be oppressive and restrictive. The adoption of more current languages may increase the chances of education and employment. The many social dilemmas facing small language communities can only be understood within the encompassing anthropological context.

Endangered languages, sociolinguistics, and linguistic sentimentalism

Linguistics has always had a tendency to impose its norms. Linguists were never entirely content simply to study how people spoke or wrote, or what they had to say; they also wanted to tell them how to do it. But the rule makers have changed their tune, and these days prescriptive linguistics – in the Netherlands at any rate – impinges on the consciousness of the general public only during periodic spelling reforms or campaigns for non-sexist usage. To real linguists it is something of an embarrassment. Recently, however, a whole new generation of pedants has turned up in linguistics, or in one of its offshoots, sociolinguistics. This is another area where it would be sensible to start by taking a good look at how people actually use language when they talk to one another , when trying to distinguish themselves from others, or to adapt to them, or when they meet speakers of other languages, or when they try to learn another language. These are questions studied by sociolinguists, and in that work they meet kindred spirits from adjoining social sciences - mainly sociologists, but also political scientists, geographers, anthropologists and economists. It is this that makes sociolinguistics such a lively, inspiring, varied and interesting discipline.

But the present generation of sociolinguists has a bad conscience. They themselves speak and write flawlessly one or more major languages. This maximizes their potential readership and their prospects of a university career. At the same time, many of them prefer to study languages that do not have such advantages to offer their users. They may decide to research the linguistic behavior of small, isolated peoples who have not yet plowed and asphalted their surroundings, and who still have to survive with the resources and hardships of the rainforest, steppe, tundra, or desert.
Other sociolinguists study the language of immigrants who have come from less prosperous and generally less democratic countries of the developing world [to the richer and generally more democratic countries of the West. The language of their country of origin may be a millennia-old language of culture (such as Mandarin or Hindi, Turkish or Arabic), or it might be a languishing language, unwritten and unregarded. Whatever the case may be, having come a long way with their native language, they discover it will not get them any further.

Other researchers apply sociolinguistics to speakers of a regional dialect that, however indigenous, however venerable, is being supplanted by the official national language, the use of which has been extended and imposed by the state. Even so, no sociolinguist would dream of writing about it in Breton or Welsh. These languages remain the exclusive province of the people from that region. Researchers choose languages that will expand their readership and their reputation. So when they write about the linguicide being perpetrated upon Welsh they do so in English, and when they describe the glottophagia being inflicted upon Breton they do so in French. Linguistics has found a new mission: not to prescribe correct language, but to protect endangered speeches.

What does it actually mean to say that a language is endangered, that it is dying out, or dead? Simply that its speakers use it less and less, start to neglect its finer points , resort more and more to another, rival language, and eventually stop teaching the original language to their children and largely forget it themselves. A language only exists through the use that people make of it with one another. That is the basis of sociolinguistics. A language cannot die, let alone be murdered. The people who speak it simply stop doing so. This is best described not as language death but as a process of language abandonment.

People who abandon their native tongue do so because they move elsewhere or take up something else and in this new existence they have higher expectations of a different language. Or they neglect it because another language is preferred at school, by public authorities, or in courts of law, and their own language is treated with disdain. Or they have to stop using it because they are ruled by another nation that imposes its language on them, and, having lost heart, they no longer take care to preserve their own language. Sometimes an entire people is simply exterminated. Then the language dies along with its speakers.

The abandonment of a language may be attributable to any number of causes, some of them bitterly tragic, and others rather gratifying. One thing is certain: there is always more involved than a simple shift away from the original language. It is therefore inappropriate to discuss the abandonment of languages invariably in mournful terms. Sometimes it brings relief , release and enrichment: a disembarrassment of speeches. Nor is it right to discuss such social developments exclusively in terms of the loss and extinction of a language. Each case should be judged on its own merits, and in its own particular social context.

Every language is a product of the collective creativity of people expressed over scores hundreds, or thousands of years. Whether it is written or unwritten, its disappearance is an irreversible loss of culture. A language is a piece of cultural heritage comparable to the Egyptian pyramids or to medieval cathedrals, to African polyrhythm or European polyphony. It is obvious that any language that is in danger of falling into disuse must be recorded and described for posterity in every possible way. That is certainly a job for linguists. When Uhlenbeck et al. launched a program for endangered languages in 1992, their main stated aim was to record these languages, although right from the outset there were intimations of an effort to preserve them (Dorian, 1981; Fishman, 1991; Robins and Uhlenbeck, 1991; Language, 1992; Brenzinger, 1992).

Even in this start-up phase, Peter Ladefoged (1992) published a caveat. He pointed out that people might have very good reasons for abandoning one language for another, citing the example of the Toda in India:
‘They also realize that with less than 1,000 speakers they are unlikely to remain a distinct entity. Many of the younger people want to honor their ancestors, but also to be part of a modern India. They have accepted that, in their view, the cost of doing this is giving up the use of their language in their daily life. Surely, this is a view to which they are entitled.’ (p. 810)

Whether the Toda are acting wisely is not a matter on which linguists, with their specific fund of expertise, have a right to pronounce. It would be presumptuous to advise any particular course.
Ladefoged dissented on another, equally fundamental, point. He did not believe that cultural diversity diminished if fewer languages remained: ‘different cultures are always dying while new ones arise.’ He continued:
‘In the popular view the world is becoming more homogeneous, but that may be because we are not seeing the new differences that are arising. Consider two groups of Bushmen, the Zhu|oãsi and the !Xóõ, who speak mutually unintelligible languages belonging to different subgroups of the Khoisan family, but otherwise behave in very similar ways. Are these groups more culturally diverse than Appalachian coalminers, Iowa farmers, and Beverly Hills lawyers?’ (p. 810).

And here he is referring only to a handful of American subcultures within the English-speaking world, which also includes countless other subcultures, for instance in Ireland, India, Britain, Australia, New Zealand, Canada, Nigeria, Hong Kong, Ghana and South Africa. Linguistic diversity, then, is only one dimension of cultural diversity. In fact it is perfectly tenable to maintain that this multifariousness comes into its own better within a single language area, where the different subcultures confront each other far more directly, than in communities that scarcely interact because of the language difference. Monolinguism and multilinguism correlate only up to a point with homogeneity and diversity.

Jan Blommaert (2001) adds an important argument in an incisive recent article on the Asmara Declaration on the linguistic rights of African languages: the imbalance of power, income gap, and differences in social status overlap only marginally with differences between language communities, in his view. According equal rights to languages can achieve little. Huge differences exist between the speakers of every language; recognizing linguistic rights cannot improve the position of a disadvantaged minority group: ‘Granting a member of a minority group the right to speak his or her mother tongue in the public arena does not in itself empower him or her.’ What is at issue is the availability, accessibility, and acquisition of specific language skills, such as a command of the standard written variant, the power version of a language. This will not be news to any sociolinguistics graduate. Yet the pioneers in the struggle for the rights of disadvantaged languages have evidently lost sight of this fact.

Linguists must not allow an obsession with language to lead them into linguistic solipsism, writes Blommaert. Sociolinguistics is based on a few basic premises: that languages exist through the use that people make of them together, that languages function within a given milieu, and that people are continuously determining the form and content of that milieu through their speech.
This means that the movement for language rights, against language imperialism, against ‘linguicism and linguicide,’ has become unstuck from sociolinguistics. It favors imagery instead of theory: languishing languages are compared to plant and animal species facing extinction. This is a misleading metaphor, a fallacy [1]. Yet it has spawned an entire movement [2]. Blommaert exposes the fallacious reasoning in Terralingua’s website (2001, 11): ‘We know that the diversity of species lends stability and resilience to the world’s ecosystems.
Terralingua thinks that a diversity of languages does the same for the world’s cultures.’ But as Ladefoged has reminded us, a decline in linguistic diversity does not necessarily lead to a decline in cultural diversity. There is another more fundamental difference between species and speeches: to preserve a species, it is sufficient to leave its habitat undisturbed. But preserving a language requires a more active response: people must continue to speak it, even under radically changed circumstances, against their will if needs be. Languages are made by people.
In the absence of actual observation or theory, what remains is an appeal to emotion: ; the panda, that sweet, cuddly creature, is all but extinct; Breton, that lovely language, is seriously endangered; the ancient language of Sorbian seems likely to perish, and in spite of the high patronage of the Fryske Akademy Frisian has not been well for quite a while. All this comes down to sentimentalism, linguistic sentimentalism: an exaggerated appeal to familiar feelings with the aim of eliciting the traditional response of sympathy.

Take the Asmara Declaration on African Languages of January 17, 2000. Article 1 states: ‘African languages must take on the duty, the responsibility, and the challenge of speaking for the continent.’ I quote it, but I don’t understand it. Languages with duties and responsibilities? As if languages were sentient creatures endowed with superhuman powers enabling them to speak for an entire continent. Those familiar with the genre will have recognized the originators of this school of thought: Tove Skuttnab-Tangas and Robbert Phillipson (e.g. 1999), the village enthusiasts of linguistics, whose views, though seldom taken seriously, are rarely openly refuted either.

The ink had scarcely dried in Africa before it began to smudge in Europe, in the small Dutch town of Oegstgeest. The Declaration of Oegstgeest of January 30, 2000 (printed in an appendix to Extra and Gorter, 2001) opens by alluding to a misconception that Ladefoged had already proven false: ‘the intrinsic relation between multiculturalism and multilingualism in Europe, as expressed in the vitality of regional, minority and immigrant languages.’ But this relationship is not intrinsic, nor indeed extrinsic. Expressed in this simple, direct fashion, it simply does not exist at all. Another curious thing about this quotation is that it presents Europe’s minority and regional languages as throbbing with vitality, whereas elsewhere [ they have been declared comatose. Must get to the bottom of this.

The Oegstgeest Declaration is a plea to strengthen the position of disadvantaged languages, especially in education and the media, again without any empirical references or theoretical foundations except for the highly questionable assertion quoted above. Aside from this, the Declaration invokes a whole string of directives, conventions and manifestoes, none of which add to its sociolinguistic weight.

The range of languages offered in education, article 7 stipulates, must be based on the local demographic composition and the stated preferences of parents and students. This at least represents progress in relation to the linguistic paternalism that was so offensive to Ladefoged. Children from migrant families and other ethnic minorities must be pandered to by rewarding their knowledge of their home language with a grade. That is an easy addition to the curriculum. One examination topic less. Still, I think that parents and children would prefer to learn the national language as well as possible than to forever float about in the wash water of their native language.

Again and again, it appears that parents (and children) prefer a language that they assume will maximize their opportunities on the labor market, rather than the minority language that linguists and educationalists so well-meaningly prescribe [3]. It is obvious that children who do not speak the national language at home should be given every opportunity to learn it, and that those teaching them must take their pupils’ native language into account. Poor or authoritarian countries are all too often deficient in this respect, which provokes opposition (however ill-advised) to the principle of education in the national language.

A recent policy proposal in the Netherlands (Manifest, 2002) suggested that all Dutch primary school children should be required to learn a third language of their choice in addition to Dutch and English; ethnic minority children would therefore be able to choose their native language as a school subject. This would make the program more demanding for children who speak Dutch at home and relatively less so for ethnic minority children, thus narrowing the opportunity gap between them. Leo Prick attacked the suggestion in two of his columns in the daily newspaper NRC Handelsblad (2002). I can think of several valid educational arguments in favor of the idea, and several equally valid (if not more so) against it. The decisive factor, however, is its unfairness : the proposal makes life harder for the children of Dutch parents and easier for the children of immigrants. That would breed resentment.

It can get even sillier. The language educationalist Ingrid Gogolin proposed at the conference Which languages for Europe? (1998) that all German schoolchildren be required to learn Turkish. Not that they would ever need this language, but as a form of punishment for having been raised by German parents in Germany, speaking German. It would be hard to conceive of a better way of hardening the hearts of German children against their Turkish classmates. I’m sure that was not the intention. Overwhelmed by linguistic sentimentalism Gogolin lost sight of the real dynamics of the classroom.
All these manifestoes, declarations, and resolutions present multilingualism as an enrichment, an asset (that may be ‘ redeemed’), and as the prerequisite for a multicultural society that in turn embodies intrinsic value. But multilingualism and multiculturalism are two different things. When Dutch society was compartmentalized along denominational and ideological lines, Catholic, Protestant, conservative, and socialist forces fought out their differences in a single language. The Netherlands was far more multicultural in those days than it is today. That was because of the strength with which people held their different beliefs.

Even so, one of these manifestoes (the Taalkundig Manifest of June 23, 2000) includes a sentence that I thoroughly endorse: children whose home language is different from the language of school should be made to realize in the classroom that they have a particular skill, something extra that their classmates do not possess. This difference could be taken as the point of departure for practical language teaching. The sociological and educational problem remains, of how classroom techniques can convert society’s implicit disdain for the use of a different language into respect and interest.

Multilingualism need not necessarily lead to multiculturalism, and multiculturalism is in itself neither desirable nor detestable. It is a social fact of life. Human rights and freedoms, which safeguard the diversity of thoughts, feelings, and convictions, do indeed embody intrinsic ethical values.
Linguistic sentimentalism identifies each language with a particular group, and seeks to preserve the group’s cohesiveness by preserving its language. But this is to overlook the differences of power within each group, the gaps in status between the registers of each individual language. It is quite possible to preserve group cohesiveness if the language is no longer spoken. You can perfectly well be Breton without Breton, Irish without Gaelic, Jewish without Yiddish, Frisian without Frisian, and Catholic without Latin [4].

Why should someone from Friesland have to speak Frisian, or act Frisian? She might prefer to be a punk, or a Buddhist. Language communities may be highly restrictive and stifling. Lois Kuter (1989, 79) recalls Breton peasant women’s rejection of Breton: ‘Women, who were affected strongly by the drudgery of farmwork, were the first to seek escape from a Breton identity which they felt chained them to this lifestyle.’ There are other documented cases of rural women abandoning their language. And soldiers returning to Brittany after the First World War were also eager to shed Breton idiom, because it cut them off from opportunities in the outside world. The preservation of a language community very often means the continued oppression of women, children, young people, the dispossessed, deviants, and dissidents.

Language communities are very often faith communities. Perhaps it is precisely in the profession of that faith that their language is most irreplaceable. This applies all the more to the religions of small, fairly isolated peoples, whose faith is professed and practiced only in a particular language, and cannot be separated from it. Yet those who present themselves as defenders of endangered languages care nothing for the destruction of these religions. On the contrary, one of the fiercest defenders of minor ethnic languages, the Summer Institute of Linguistics (Dallas, Texas), was founded to promote the evangelical mission, a form of Christianity that steamrollers every indigenous religious practice it comes across in its uncompromising zeal.

Exactly the same litany as is recited about minor languages – the relentless threat to their survival, deliberate attempts at suppression, and looming demise – could be recited about minor religions. Exactly the same metaphors could be pressed into service about the extinction of species, the decline of diversity, and the irreversible loss of unique human knowledge. But no one knows how many minor religions initially existed, how many still have adherents today, what truths they proclaim, on what points they differ, or how many are nearing extinction. This evidently has less power to arouse righteous indignation; it does not have the same sentimental appeal to linguists. No one has been adopting resolutions or manifestoes about it, and no one is offering any subsidies for research on the subject. That would run counter to the triumphal march of Islam and Christianity.

Identities are not fixed; they are multiple and changeable, they are fabricated and presented, conveyed, feigned, or concealed, in constant interaction with others who are equally concerned with their presentation. Most people identify with one or another aspect of their life according to their circumstances. Languages are only one part – and not even the most important one – of this identity.

On the other hand, languages do mark their speakers’ identity. Language use serves as a proof of identity – a badge. But that is not its only function. A language is also a means of communication. This word now enters the present discussion for the first time. That is because linguistic sentimentalism ignores the communicative functions of language: the more languages the better. All these languages must be fully equipped and accorded absolutely equal rights. How any sort of public debate could be carried on in such a society , is a question not given any thought. Many linguistic sentimentalists even oppose the introduction of a single language as a medium for public exchanges of opinion. To escape a Bable-like confusion they breezily suggest that everyone learns many languages – preferably three, four, or five. But not everyone is endowed with the gift in which linguists take such pride.

‘The more languages the better’ remains the battle-cry. But this multiplicity of languages actually subverts diversity: the more languages that compete, the more English will take hold. In the general confusion of tongue, in which no indigenous language can predominate, English automatically imposes itself as the sole, obvious, solution. That is what has happened in India and South Africa, in Nigeria, and in the European Union. The hegemony of English is being hastened and consolidated by the promotion of a multiplicity of languages - by the European Commission, for instance.

Some would define English as the great glottophage, the insatiable devourer of languages. This would make the spread of English the purest form of ‘linguistic imperialism’ (Phillipson, 1992). In fact it is improbable that the confrontation with English, or with any other world language, is the main factor in the disappearance of minor languages. Most languages, and the most vulnerable among them, are spoken by small nations, often hunter-gatherers who are dominated by neighboring sedentary peoples. This inequality leads them to adopt the language of their powerful farming neighbors and to give up their own. They do not even get as far as learning English, Arabic, or French (Dimmendaal, 1989; Batibo, 1992; Rottland and Okombo, 1992).

It might appear that this entire debate about languages is just a mental exercise for academics. Far from it. South Africa recently introduced legislation inspired by the language rights movement, egged on by foreign linguistic sentimentalists. The new Constitution formally accords equal status to eleven languages. Ndebele and Pedi, Tsonga and Afrikaans, English and Sotho are all equal before the law. If a member of parliament chooses to speak Tshivenda in the House instead of English or Afrikaans like everyone else, the Speaker has to find a messenger who happens to come from the northeast of the country to translate the speech. So no one speaks Swazi or Tsonga or one of the minor indigenous languages on an official public occasion, except to make some point about folklore. Nor is anything being done to equip all these languages with textbooks and grammars and everything else that goes with the modern use of a language. There are simply too many of them, and there are no funds for it.

But could no one have predicted this? Not the linguistic sentimentalists. But South African linguists pointed out the problem long ago. Back in the 1940s, Jacob Nhlapo put forward a plan for the unification of related indigenous South African languages into a single language. But since missionaries had been the first to make simple vocabulary lists and elementary grammars in their areas, each with his own European language at the back of his mind, they had unintentionally emphasized the discrepancies between one region and the next. The local inhabitants, moved by a ‘narcissism of minor differences,’ were happy to contrast their own idiom to that of neighboring peoples. It was in the interests of traditional leaders to accentuate such differences, and the Apartheid regime found the resulting fragmentation very convenient. Neville Alexander (1992), who chaired the Advisory Panel on Language Policy and Planning in the post-Apartheid regime, on the other hand, followed Nhlapo in urging that Bantu languages be formed into two groups: Nguni and Sutu. Kwasi Prah (1998) supported him and even suggested integrating them all into a single language. Why did the ANC choose so blatantly to disregard its own experts and advisors? I suspect I know the answer.

The more languages , the more English. The African National Congress consists primarily of intellectuals and veterans of the liberation struggle, many of them exiles, who learned to speak English at school or as activists. English is also the most obvious lingua franca for the country as a whole, and the best means of curbing its only rival, Afrikaans. And English is what ‘non-white’ South Africans want to learn, because they think it will maximize their chances of getting on in life. They are quite right.
Thus works the ruse of history: the opponents of English linguistic imperialism, at the height of their influence, have accomplished precisely the opposite of what they hoped to bring about: by granting equal rights to so many languages, they only strengthened the hegemony of English even further. Anyone could have predicted it.
Anyone?
The European Union pronounced 2001 the Year of Languages. As the reader will perhaps recall, this linguistic disaster area has eleven official languages and dozens of minor ones. Soon it will have twenty official languages and far more disadvantaged ones. In the Year of Languages, the European Commission ‘celebrated’ linguistic diversity (arguing that it promoted cultural diversity, as indispensable as biodiversity). Minor languages too deserve to be supported, promoted and strengthened. The Commission urged Europeans to learn as many languages as possible. Is it churlish to wonder how, if Europeans were all to learn different languages – with the Irish studying Latvian, Cypriots learning Hungarian, and Czechs mastering Finnish – it would help them to communicate?

I have not come across this objection in any of the Commission’s propaganda. Full-page advertisements urge school students to learn Italian (‘so useful for asking the way when you are on vacation in Tuscany’), Spanish, Swedish, or any other language. But this is pure deception of Europe’s youth. It is perfectly obvious that the language that will be most useful to young people wanting to communicate with others is English, with French in southern Europe and German in eastern Europe as good alternatives. So the Commission deliberately gave its youngsters the wrong advice.
Students and their parents knew better. Almost ninety percent of students in Europe learn English as their second language. They are quite right. Within any particular constellation, people opt for the language that will enable them to communicate with the largest number of people (prevalence) and that has the highest percentage of multilingual speakers (centrality). This language has the highest ‘Q value,’ as I have called it, as a means of communication. People do not actually calculate it in this way, of course, but they are quite adept at assessing it nonetheless. I have discussed this subject at greater length in Words of the World: the Global Language System (De Swaan, 2001). When people decide to learn a particular language, they are motivated more by communication functions than by functions of identity – although it must be said that English is also attractive to adolescents as an identity marker, as an indicator of cosmopolitanism and youth.

Why did the European Commission spend an entire year, so ostentatiously and so emphatically, giving young Europeans the wrong advice? Because it cannot discuss the EU’s language problem. All the EU’s official languages are sacrosanct, at any rate on all public and ceremonial occasions, and in any regulations that affect citizens directly. The languages used in the Commission’s inner chambers are French (less and less) and English (more and more). English is the primary means of communication in Europe, where the language problem is gradually resolving itself, just as in South Africa. But to say so openly is just as much taboo in the European Union as it is in South Africa. The more languages , the more English . The effect of the campaign conducted in the European Year of Languages was therefore precisely the opposite of what was intended.

Like all forms of sentimentalism, linguistic sentimentalism is disingenuous at the core. The European Commission is not in a position to launch a debate on the language issue in the European Union. If it did so, France would insist on maintaining its privileged position, and Germany would demand equal rights. Spain would refuse to be left behind, and so would Italy. Even the Netherlands would have to defend its linguistic interests. So while supposedly conducting a major campaign to recommend all languages equally, the Commission is in fact doing nothing, since the direct impact of its campaign is predictably zero. And by doing nothing, the Commission is actually helping to consolidate the position of English as the only way out of the confusion of tongues.

This European language policy, or non-policy, favors sociolinguists and the protectors of minor, regional and migrant languages . Just about the only tangible result of the EU’s language policy is a permanent circuit of conferences on the endangered and disadvantaged languages of Europe, which keeps the experts and the activists fully occupied. The experts have another interest in the EU’s policy of linguistic diversity: it creates jobs for language teachers and experts. This does not refute the assumptions on which that policy is based. There is no need to refute them, since there aren’t any. Like the language rights movement, the EU’s language policy is based on metaphors and sentiment. Aren’t minor languages just like endangered animal species? How pathetic! And don’t we simply choose the easy way out, by keeping to the national language, and by using English to boot. It’s really a shame!

Kurt Baschwitz, in his Du und die Masse. Studien zu einer exakten Massenpsychologie (1938), wrote of ‘the power of the paralyzing idea,’ by which he meant an idea that has a superficial air of plausibility, but that on further inspection proves groundless, and that nonetheless remains uncontested because of vague feelings of fear and guilt. The determined efforts to propagate linguistic diversity when a shared language of communication is indispensable fall into this category. It is hard to swim against the tide. After all, who would want to see something as precious as a human language being abandoned and gradually disappearing? Who would want the awesome diversity of languages and cultures to wither or fade? Who would want to force his fellow human beings to learn a single language, even if it does oil the machinery of mutual comprehension? Who does not feel vaguely uneasy about the hegemony of the West? And isn’t English the carrier of globalization, imperialism, capitalism, consumerism, commercialization or whatever? Exactly.

If the speakers of a language are abandoning it, there is every reason to describe it and record it as well as possible for posterity. We should do everything possible in this respect. But how people living in small linguistic communities should make their way in the wider world cannot be prescribed beforehand, or in general. Choice of language is only one aspect of their predicament. They themselves often want to learn the language that will maximize their opportunities on the labor market. There is every reason to support them in this endeavor. They will often neglect the language they learned at home. The first priority should be to understand their reasons for doing so. Maybe they can be convinced that their language is worth preserving.

Fishman (2001, 452) maintains that his program for what he calls ‘reversing language shift’ is geared toward ‘the recovery, recreation and retention of a complete way of life, [his italics yes] including linguistic as well as non-linguistic features.’ And David Crystal (2000, 154) states emphatically, ‘only a community can save an endangered language.’ Both present a well-considered and elaborate defense of the drive for language preservation. But in most of such communities in transition, where language and culture are in danger of being abandoned, this goal is in itself controversial. Very often these communities contain advocates of assimilation and integration into the surrounding society, people who are quite prepared to adopt the dominant language and even to give up their native language altogether. When such divisions exist within a specific community, there is no good reason for seeking to pre-empt the outcome of the internal debate by supporting one side or the other.

Where a language is at risk, the linguistic community is itself endangered. Those who belong to such a group constantly face with the dilemma of whether to opt for assimilation into the surrounding society or preservation of identity in isolation . Assimilation often holds out better individual prospects in the immediate future, but it presents the risk of a collective loss in the longer term: once a language is no longer understood, its communal culture will no longer be accessible. On one side of the scales is short-term individual gain, on the other side the loss of a collective cultural capital, a community’s common heritage. This is a classical dilemma of collective action. It also suggests an appropriate theoretical framework for analyzing endangered linguistic communities [5]. This perspective also has normative implications. Those concerned face a dilemma that neither linguists nor sociologists or anyone else can resolve for them. It is not up to others to preach or admonish; the only helpful response is to clarify the dynamics of the dilemma.

Abram de Swaan, European Review, 12. 4, October 2004, pp. 567-580.

The Author acknowledges a grant from the Royal Netherlands Academy of Sciences for the translation from Dutch by Beverley Jackson