Truth in the silence of mushrooms?

Noam Chomsky is fond of describing linguistics as a branch of psychology, ultimately biology. This places the study of language in the area of the natural sciences, grouping it with genetics and chemistry, rather than poetics or literary criticism, and inviting evaluation of its theories by the standard scientific criteria of simplicity, elegance and truth. Lyle Jenkins's book is a reasoned defence of this position, and an attack on those who have misrepresented aspects of the generative enterprise. The core problems of (bio)linguistics are: to define what constitutes knowledge of language; to determine how that knowledge is acquired and put to use, and what the brain mechanisms underlying it are; and to investigate how it evolved in the species. Jenkins devotes considerable space to each of these, but it is the question of evolution that has recently elicited most (mostly uncomprehending) attention, which attracts the largest part of Jenkins's fervour.

The background is Chomsky's "minimalist-internalist" view of language. It is minimalist because it eschews any constructs that are not conceptually inevitable or forced by empirical necessity (so no more deep structure, for instance); it is internalist because the notion of "language" that supports any coherent theorising is as an internal construct of the individual mind. This construct is largely innately determined: we are born knowing the principles of grammar, and the huge variety of the world's languages results from the setting of a number of "switches" or parameters, on the basis of triggering data from the environment. The claim of innateness is motivated inter alia by universal properties of language and by the fact that all, and only, human infants learn the language they are surrounded by, and do so on the basis of a rather impoverished input.

If knowledge of language is largely innate and common to the species, then it is presumably genetically determined. If it is genetically determined, but our relatives - from chimpanzees to mushrooms - have no comparable language faculty, then it must have evolved. But now there is a problem, as Chomsky is widely quoted as denying that natural selection could have produced human language. Furthermore, as there are supposed to be only two possible explanations for evolution - "God or natural selection" - Chomsky must either be appealing to divine providence or be a mystic. Jenkins documents the debate at length, and establishes beyond reasonable doubt that this crass characterisation of Chomsky's views is radically, perhaps maliciously, misguided.

Chomsky's position is not that natural selection has played no role in the evolution of language, but that natural selection is only one factor. In all evolution, whether of language, the eye, or an antibody, an essential contribution is also made by physical and developmental factors working on the variation provided by random mutation. Consider whales and sunflowers. Whales are so big that they would not be viable on land. As has been known since Galileo, there are different growth constraints on skeletal form for aquatic animals, because buoyancy diminishes the effect of gravity. It is uncontroversial that natural selection played a role in the emergence of 150-ton mammals, and it is equally clear that this emergence took place in an environment that was physically constrained. Sunflowers show a different facet of the complexity of evolution. Sunflowers have two interlaced families of spirals in the head, one going clockwise, one anti-clockwise. The numbers of florets in each spiral are typically adjacent terms in a Fibonacci series (the series obtained by starting with 0, 1 and adding the previous two terms to get the next: 0, 1, 1, 2, 3, 5, 8, 13...). It is implausible that the genes determine that the flower produces 21 seeds in one direction and 34 in another: that is not the sort of thing genes do. In fact, it is now known that this mathematically fascinating property of sunflowers and other organisms arises as a result of the timing of the emergence of the primordia during development.

How is this relevant to the evolution of language? The standard criticisms of Chomsky with regard to evolution are based on a simplistic analysis of the range of possibilities available, and of Chomsky's position on them. His realistic evaluation that evolution is extremely complicated and that we know next to nothing about it is represented as a denial of any role for natural selection, despite frequent assertions to the contrary. This travesty often comes with the suggestion that refusing natural selection an exclusive role is equivalent to mysticism.

Jenkins demonstrates in meticulous detail the real state of affairs and what Chomsky's admittedly speculative contribution is. Properties of language have evolved. Natural selection must have played a role in this evolution, but so too have elementary physical constraints, such as the size of the human head. Many other factors have also, presumably, been involved, such as the adoption of a trait in one domain for use in another: an example might be the exploitation of discrete infinity by both the number sense and the language faculty. Our knowledge is vanishingly small in this domain, but the questions are coherent, the issues are empirical and the problem of unifying the various bits of our knowledge with the rest of the natural sciences is standard. Jenkins has done a signal service in setting this out explicitly. He writes with passion and a controlled fury at the frequent misrepresentation of Chomsky's views. Moreover, he shows mastery of the linguistics literature; his comparisons and analogies with data drawn from a huge range in mathematics and the hard sciences are illuminating and reliable. Sadly, however, his tone is unnecessarily strident, his style is repetitious, and his proof-reading is sloppy. This is a pity, as the book is much needed and systematically excellent.

Neil Smith is professor of linguistics, University College London.