Saying the Unsayable about Saving the Unsavable

by Peter Wells

The last speaker of Ayapaneco: The Guardian

Saving endangered languages is an activity the virtue of which seems to be beyond question among Western liberals. Endangered languages, as conventionally identified, are usually spoken by poor and vulnerable people in remote areas of developing countries, so our sympathies are engaged. The least we could do, it seems, to alleviate their misery and show solidarity with them, would be to preserve their language.

At the same time, it is also urged that the endangered languages are valuable in themselves. Followers of the Sapir-Whorf Hypothesis (Linguistic Relativity) assert that each individual language can express concepts which no other language can express. Therefore, they insist that it is important not only to record these languages before they die, but also to keep the language communities alive so that they continue to use them.

I wish to suggest that this project, while well-meaning, is ill-conceived. It seems to employ an unscientific definition of the term ‘language’, and to rely upon a disputed theory about the nature of language. It is unrealistic in terms of what carrying it out would involve, and finally, and most importantly from the point of view of liberalism, it would tend to reduce, rather than increase, freedom.


In the first place, it is not always clear what definition of ‘language’ is being used when the world’s endangered languages are being counted. There are said to be about 7,000 languages in the world, of which about half are said to be in danger, e.g. UNESCO. However, we are not usually given information about the distinctiveness of these ‘languages’ – how different they are from each other. Is the difference like the difference between English and Chinese? English and French? American and British English? Scouse and Yorkshire (two British dialects)? It seems overwhelmingly likely that many of the clusters of ‘endangered languages’ found around the world are much more like what we would call ‘dialects’ than ‘languages.’ Indeed, the names of some of these vanishing ‘languages’ support this speculation, for example, ‘Western Cham,’ ‘Scottish Gaelic’ or ‘Mariupolitan Greek’ (from the UNESCO table in the Guardian). That being the case, the number of important ‘endangered languages’ would have to be drastically reduced.

Or perhaps there are many more ‘endangered languages’ than even the most enthusiastic preservers realise. The vague popular term ‘language’ could profitably be replaced by the more scientific expression ‘variety’ – the term for a distinguishable system of grammar, vocabulary, and pronunciation. A ‘variety’ could be a standard language, like French or Chinese, a dialect of a superordinate variety, like Scouse or Yorkshire, a ‘mode’ of language (written or spoken), a diachronically determined variety, like the English of the 1950s, or a variety based on class, gender, or trade. It could even refer to the variety unique to one person – the ‘idiolect.’ To be differentiated from another variety, a variety needs to differ in only one item of vocabulary, grammar or pronunciation. Giving a variety a name creates the illusion that it has some special status, even when there is no evidence that it does. If a language with half a dozen speakers in a remote area of Australia has distinctive features that are worth preserving, so does every other variety of every language that has ever existed, right down to idiolects, which become new languages every day, as their speakers adopt, modify or discard items, for all ‘languages’ are, by definition, in a constant state of flux.

Which version of a vanishing tribal ‘language’ are we supposed to preserve? The one that anthropologists described ten years ago, the one in use now, or the one that will exist tomorrow? The one used by the women or the men, by the old or the young, by the northerners or the southerners? The fact that a named variety is attached to a small ethnic or tribal group does not entitle that variety to special attention in education or language study.

Linguistic Relativity

Language preservers adopt a strong version of the Sapir-Whorf Hypothesis, according to which “Language shapes human experience—our very cognition—as it goes about classifying the world to make sense of the circumstances at hand” (National Geographic). However, this ‘strong’ form of Linguistic Relativity is no longer academically respectable.

The proponents of Linguistic Relativity claim that some Native American languages (for example) have single words for designs or patterns that take many words to describe in English. This, they urge, shows that members of this culture are able to think thoughts that European minds cannot grasp. But in fact Westerners can describe these figures, given enough words. We just don’t need to do so very often, whereas apparently the Native Americans do. Thirty odd years ago a rare disease emerged which had a ten-syllable name using Greek and Latin scientific vocabulary. When we began to need to talk about it frequently, we just called it AIDS. Language automatically changes to fit the actual needs of its users.

Chinese verbs are not marked for tense or number. This does not mean that Chinese people can’t count, or have no idea about time. Japanese has a gender-neutral suffix (-san) covering Mr, Mrs and Miss (or Ms). This, however, has not succeeded in making the Japanese non-sexist. Welsh has one word covering blue and green, instead of two different words, but Welsh people are perfectly able to see the difference between a blue dress and a green one. Enforcing politically correct language does not seem to have had much effect, if any, on people’s underlying attitudes towards minorities – except possibly to aggravate prejudice. Wherever you look, the argument that language determines thinking is unconvincing.

We are not prisoners of our languages. Where indigenous languages, in various parts of the world, have been abandoned in favour of a world language, because of colonialism or economic pressures, the speakers have generally succeeded in transferring those parts of their culture that they wished to retain into the local form of the new language. It is of the essence of language that it will provide its speakers with the means to express whatever they want to express, whether pleasant or unpleasant.

The illusion that a language, in its entirety, is somehow imbued with the spirit of its finest poets and thinkers (the nation’s ‘culture’) is easily explained. We use our native language to bond with our family, to be at ease with our neighbours, and to communicate with our co-workers; we feel comfortable in it, and feel it is part of our ‘identity.’ But language is amoral. It is just a tool. It will provide you with words for comparing your beloved to a flower, or a member of a despised racial group to dung. The disappearing languages will be very strange if they do not contain terms of abuse as well as terms of endearment, and I’m not at all sure that the preservers should want to preserve the former.

The Task

It may be of interest to record aspects of dead or dying languages. They may have something to teach linguisticians about how languages change. However, this is not what the language preservers are proposing. They want to ensure that the declining language continues to be used. When the Endangered Languages Fund website speaks blithely of the “establishment of orthographies and literacy materials to be used by endangered language teaching programs all over the world” this almost certainly means that children should be obliged to learn the language in school (not that middle-class adults might learn it as a hobby!), and that pressure should be put on them not to speak the language they normally use. The aim of language preservers is: “to maintain or revitalise their mother tongues and pass them on to younger generations” (UNESCO).

I do not know exactly how children in remote Indonesian villages would respond to being forced to learn a language that almost no one in their area speaks, but I do know how English pupils react to being forced to learn Latin, and I have heard what English-speaking children in Wales think about being forced to learn Welsh. In apartheid South Africa, children in the ‘Bantustans’ were originally forced to learn their supposed native languages, such as Xhosa or Zulu (as part of the policy of ‘separate development’). As soon as they acquired some degree of independence, the governments in these so-called ‘nations’ replaced this policy with the teaching of English, a language that would enable the children to participate in world learning. Rightly so.

The Liberal Agenda

However, the task of preserving a language is not merely difficult, time consuming and expensive (and repressive); taken seriously, it is impossible. In order to preserve an endangered language, you would have to place a reasonably sized language community – say, a small village – into a sort of quarantine.

If we imagine applying this cumbersome procedure to a remote area such as Northern Queensland, in order to preserve one of its aboriginal languages, we will see how language preservation is a contradiction in terms. To oblige a group of local children to learn the language identified as ‘theirs,’ they and their families will have to be cut off from all contact with speakers of other languages, and with every aspect of the modern world. For languages change all the time, the principal source of change being contact with other languages, and the challenges of new situations and inventions. There are also endogenous changes that emerge within language communities, rather like genetic mutations or fashions in clothing, and are taken up by everybody. Are we to allow these changes, or do we insist that only the code of a certain generation must be used? If we allow changes, the original language will be lost. However, if we do not allow them, it will be dead.

Sadly, language preservers are not the philanthropic saviours of the marginalised that they may feel they are. Rather, they are to be compared to the illiberal prescriptive grammarians and reactionary language teachers who fulminate against changes that have irrevocably taken place, and who want to preserve the English (for example) of the 1960s in opposition to the English of the 2020s. Language change is not merely inevitable, it is to be welcomed as a sign of life. The only languages that do not change are languages that are already dead, such as Latin, or that were never alive, such as Esperanto.

I’m afraid that the apparently liberal project of saving endangered languages is not as liberal as it first appears. Achieving the preservation of living languages, insofar as it is possible at all, would involve bringing at least some pressure upon children of minority racial groups in developing countries to sacrifice a certain amount, perhaps a large amount, of their opportunity to engage with the wider world represented by the major world languages and their associated arts and sciences. The main aim seems to be the preservation, for the benefit of the world at large, of some of the ecological and cultural information known by the speakers of those languages. Thus, the life chances of some of the poorest people in the world are to be diminished for the benefit of the richer ones. Yet the benefits are extremely dubious, for either the information can be translated (in which case there is no need to preserve the language) or it cannot (in which case there is still no point in preserving the language, for we shall never understand it).

While we are right to be outraged when alien oppressors prevent a community from using its mother tongue, the appropriate liberal response to language decline, in my view, is not to oblige people to use what we consider to be their ‘correct’ language, but to allow them to use the language they want to use.