Why we talk: The evolutionary origins of language (review)

If we wish to know what is included in language—a biological trait—it is necessary to investigate its evolution, development, and functions, along with the mechanisms that are responsible for its acquisition and use. Such an ethological strategy, among other virtues, sidesteps a problem posed by purely structural, inorganic approaches to language: a linguistic product that can, and to date has, eluded orthodox evolutionary explanation.

Unlike their early predecessors, many of the new theories on linguistic evolution are susceptible to constraint by empirical observation, and so may be expected to assume an important place in the growing literature on human behavior evolution. In Why we talk, Jean-Louis Dessalles makes an important contribution to linguistic evolution by setting a plausible ecological-informational scene for its investigation, and by offering a functional hypothesis on which it may be possible to acquire relevant evidence from contemporary human speakers.

The book was originally published in 2000 and only recently translated from French (by James Grieve). As a consequence, there is, for a new book, an oddly 'unaware' quality in some places, and the supporting material is dated. Still, there is much in D's interesting thoughts to recommend the English edition. It begins with what he doubts. D is skeptical of any early and formative role of culture in linguistic evolution, and is also alarmed by those (e.g. Noam Chomsky and Derek Bickerton) who, at the time of the original edition, had taken the 'absurd' position that language evolved in an abrupt or catastrophic way. D argues that language is far too complex to be treated as a spandrel. After Pinker & Bloom 1990, he holds that language evolved gradually in response to recognized neo-Darwinian principles. D also abhors the fact that language has been treated as an isolated phenomenon with few if any links to the physical and social environment from which it emerged.

D is also to be applauded for his intentions. His preeminent goal is to develop a cognitive model of linguistic behavior that would make it possible to infer the biological functions of language in our evolutionary ancestors. Toward that end, D offers a communication-driven theory of linguistic evolution. Language would not exist, he suggests, if it did not enable speakers to create in their listeners' minds thoughts that are related to their own.

For D, much of the functionality of language is tied up in conversation and, more specifically, narration. These applications of language, he thinks, were responsible for its evolution. It is here that he notices a paradox that seems to have inspired his major claim—people spend a great deal of time giving away information, much of it hard won and potentially useful to others.

But it is not just any information. A devotee of Claude Shannon, D knows that old information is no information, and that mundane facts are frequently worthless. There are 'stories we can tell and stories we cannot tell', D says. 'I cannot just recount that I got up this morning, had my breakfast, listened to the wireless, got washed and dressed and then left the house. Anyone listening to such a story knows something essential is missing' (283).

The missing element is salience or, following on from Sperber & Wilson 1995, RELEVANCE. Facts that are relevant are those that, among other things, are true, significant, informative, and germane to the topic—features that combine to make them interesting. But, he asks, why should people give others relevant information, especially when doing so could level the playing field, possibly even give their rivals the competitive edge?

The answer lies in what donations do for the donor. Individuals who have interesting facts did something to get them, and succeeded because they had unusual fact-acquisition abilities. Recognition of this by others enhanced their prestige and status. This seems reasonable enough—any conspicuous trait that increased the status of our evolutionary ancestors is likely to have been selected, and may now be considered...