The comparison with mammals raises the biological metaphor that inevitably comes into discussion of the death of languages. We talk about ‘extinct’ languages, ‘threatened’ languages and ‘endangered’ languages, precisely the terms used for the animal and plant world. Some talk about the revival or even the resurrection of languages in ways that are reminiscent of talk about finding some ‘viable’ DNA and the ‘keys’ to bring back the Tasmanian tiger. Hagège is acutely aware of the power and ubiquity of such metaphors, but early in the book he insists on their limitations. He draws on Saussure’s distinction between langue and parole to point out that while words, meanings and even the speaking of a language (its expression as parole) may cease, it does not necessarily follow that the langue (its abstract structure or, if we insist on the metaphor at least at this point, its DNA) is dead: ‘Thus the distinction between langue as a system and parole as an activity leads us to this essential conclusion: a language termed dead is nothing other than a language that has lost, if we may use the expression, the use of speech. But we have no right to equate its death with that of a dead animal or plant. Here, the metaphors reach their limit. Because a dead language continues to exist.’

Seemingly dead languages can continue to exist. Latin is not ‘alive’ in the sense that it gives voice to a culture, but it is not completely ‘dead’, since it is still taught in the school curriculum and is the language of the Roman Catholic Church. Coptic and Sanskrit are no longer spoken, but they are studied and their structures can be known. At an extreme, a language may be known only from inscriptions on ancient tablets, but it is not ‘dead’ in the same way that most people believe the Tasmanian tiger (or, at least, the woolly mammoth) to be dead. Even so, some languages have disappeared without trace, and more will do so.

The central part of the book gives a clear and compelling account of why languages die. There can be relatively straightforward reasons. For example, Hagège attributes the loss of our knowledge of Tasmanian languages to the annihilation of the people who spoke them. Similarly, in parts of Australia Aborigines were shifted from their traditional lands and mixed with other tribal groups, and in that situation their languages disintegrated. In the latter case, the reason for the disintegration is partly the uprooting of the people and their culture, but also the consequent shift in power relationships and perceptions of prestige, and these are the major causes of the destruction of languages.

Hagège is passionate about the importance of bilingualism and multilingualism, but he also stresses how a ‘bilingualism of inequality’ can lead to the extinction of a language. In a bilingual situation, if one of the languages carries greater social status, greater opportunities for social advancement and greater attractions for the younger generation because of the social values this language represents, the stronger language will gradually swamp and dismantle the weaker one, leading to its extinction. At the moment, English, because of its power and prestige, is the most common destroyer, but it is not the only one – in parts of Africa, for example, the destroyer is Swahili.

The final third of the book is about the possibility of resurrection in languages. Most of this section is taken up with an account of the process that led to Hebrew becoming a spoken language again. The unusual feature here is that Hebrew was not a spoken language in decline, but a language that had ceased to be spoken and existed only in a written tradition. This serves as a remarkable paean of hope, but it remains very much an isolated and special case.

Why should we be worried about the death of languages? Hagège has two central reasons, which can be usefully explained by drawing on the distinction between langue and parole. He refers to ‘the linguistic genome of the human species’, and clearly believes that each individual language is a different expression or manifestation of what the human capacity for language is capable of – to lose one of these languages is therefore to lose forever one of the pieces of the linguistic genome. This is langue at the level of a system that can generate all the instances of particular languages that have ever existed and that will ever exist. The particular languages can be seen as instances of parole. The language of a community is an expression of its culture and the carrier of its history, and the death of a language therefore means the death of a linguistic community and its culture. The death of a language limits our knowledge of what human beings are capable of linguistically; it also destroys a particular culture forever.

Towards the end of the book, Hagège raises the issue of paternalism, especially the role that linguists play in recording threatened and endangered languages. This issue is especially acute in Australia, where the question has been raised about whether a policy of preserving Aboriginal languages and their cultures at any cost is necessarily in the best interests of these Aboriginal peoples. It is easy to assume that the two aims of mapping the linguistic genome and preserving particular languages and cultures are in accord, when it is possible that in particular social situations they may come into conflict. Is there a danger that we seek to preserve particular languages and their cultures in our own interests rather than in theirs? This is a question that those who wish to preserve languages and cultures at any cost will have to confront more directly in the coming years.

On the Death and Life of Languages ranges effortlessly over centuries, over linguistic commentators and theorists, and over numerous languages. This English translation is of a slightly updated version of the original 2000 French edition. The translation by Jody Gladding is elegant and highly readable, although there is no doubt some irony in the fact that the French book must be translated into the ‘conquering’ language to reach a wide audience.