Wolfram Hinzen's volume is a sweeping but often detailed and contentious account of the foundations of the so-called Minimalist Program. The principal audience is philosophers and linguists, although there is much here for theorists in related fields. For the philosophers, Hinzen spells out a philosophically unorthodox ‘naturalistic’ picture of language and mind, and by way of a high-level discussion of Minimalist theory exemplifies how enquiry might fruitfully proceed. For the linguist, the same material admirably serves as an explication of the philosophical motives behind Chomsky's (and others') theorising and even offers some original thoughts on syntax. The volume succeeds splendidly in meeting the author's desiderata.

The book is divided into three parts. The first part, entitled ‘Naturally human’, is primarily philosophical and includes a valuable discussion of evolutionary theory. The second part, ‘Deducing variation’, covers foundational ground in linguistic theory; while the third part, ‘Rational mind’, is speculative, integrating aspects from the foregoing.

Chomsky has long battled against philosophers' objections to the generative enterprise. Rather than merely offer a survey of these ongoing debates, Hinzen articulates a Chomsky-informed ‘methodologically naturalistic’ conception of enquiry into mind/language that is very much at odds with the prevailing philosophical mood (for Chomsky and Hinzen, a ‘metaphysical naturalism’ holds sway in contemporary philosophy, where the standard for ‘reality’ is set beyond and independent of ongoing enquiry). Hinzen's principal charge is that many philosophers are guilty of ‘methodological dualism’, that is, the thought, often tacit, that empirical enquiry into mind/language should cleave to constraints that do not obtain in any other field of enquiry. Most often, such dualism shows its face when philosophers demand that mental states must be conscious or that the evidence for mentalist theories must be observable behaviour or that claims about ‘psychological reality’ require data beyond informants' judgements. In a careful and thorough discussion, Hinzen rightly rejects all such claims in favour of a ‘methodological monism’ that sees all enquiry as operating by the same standards; in particular, linguistics is no more obliged to clear metaphysical hurdles than is physics or chemistry. Three aspects of Hinzen's rich discussion bear special mention.

Firstly, Hinzen rejects ‘functionalism’ in the philosophy of mind (51–54). Briefly, functionalists contend that mental states should be understood in abstraction from the biological character of the brain: mental states are state types (defined by their mediation of inputs into outputs) that can be indifferently realised in a biological or non-biological substrate. Hinzen points out that functionalism most often uncritically accepts a commonsense taxonomy of mental states and so amounts to a faux naturalism. More crucially, functionalism detaches the mind from the brain, as it were, for the sake of including a rag bag of entities (humans, aliens, dogs, etc.) into an equivalence class of those with mental states. The far more sensible model Hinzen offers is that we should view enquiry into the mind as functional in the sense that it is abstracted from the details of neurophysiology, but that this is a methodological prerequisite at the current stage of enquiry. Our ultimate goal is to understand how the brains of organisms of all types beget their special functions, and this goal is not furthered by the creation of equivalence classes that merely encode commonsense categories. So understood, functionalism becomes a formalist approach to biological design.

Secondly, Hinzen articulates and defends a ‘rationalist’ conception of science stemming from the foundational work of Galileo and Descartes (66–79). Central to this discussion is the notion of ‘perfect design’. The basic thought here is that science does not offer a compendium of the facts, nor even constrains itself to deduce all relevant facts. Rather, science operates via an ongoing conception of how the universe must be; deviation from such perfection calls for explanation (perfection is akin to a form of necessity, where a ‘final’ theory would be one that uniquely explains the facts in an optimally simple and economical manner, reflecting the thought that the universe itself is so elegant). In this sense, it might be that successful theories will dispense with explanation of seemingly relevant facts in favour of principles whose explanatory reach is deep rather than broad. On this model, science is ‘minimalist’ in that it sets perfection as a standard for explanation, rather than letting coverage of the heterogeneous array of potentially relevant data be the standard. Deserving particular praise is Hinzen's acute discussion of Descartes' conception of the relation between theory and evidence.

Thirdly and correlatively, Hinzen takes issue with the common philosophical understanding of the great scientific revolutions of the seventeenth and eighteenth centuries (32–55). Especially welcome is Hinzen's discussion of Rorty's influential but absurd understanding of early modern philosophy. Rorty is not alone in thinking of philosophers from Descartes to Kant as positioning themselves outside of science in order to evaluate its deliverances. In effect, such gross historical misunderstanding is a species of methodological dualism read onto a period of enquiry that saw all phenomena as open to the same naturalistic understanding. (It bears emphasis that while Descartes famously hypothesised a dualism of substances, the hypothesis was straightforwardly empirical and did not appeal to methodological principles unique to putative non-material phenomena (61f.).)

In a long chapter that concludes part 1 (‘Biological internalism’), Hinzen deals with the topic of evolution, with particular reference to language and cognitive function. In some quarters, Chomsky has been assailed for being anti-Darwinian or even theological. To my mind, Hinzen provides a definitive answer to this ideological canard. For Hinzen, we should be ‘pluralists’ about evolution, which amounts to no more than taking into account not just environmental contingencies and mutation but all of the factors that make for organic design: physical constraints, developmental processes, and adaptive pressures. As Hinzen emphasises, such a pluralism is not optional, for no course of natural selection operates in a vacuum; it must proceed along a physico-chemical channel and must factor in developmental constraints that are only just being revealed by ‘evo-devo’ research (the dominance of population genetics has occluded these crucial constraints). Of course, Chomsky has suggested that, in the case of language, physico-chemical factors of design are perhaps more pertinent than adaptive function. But this is not an ideological nostrum; since a plurality of factors are involved in the evolution of cognitive function, it is a theoretical matter to discover in any given domain which factors are explanatorily most fecund. It is mere a priori dogma to insist on adaptive function as being the explanation. It is one of the more regrettable features of current cognitive science that a chapter such as Hinzen's is required.

The second part of the volume takes up issues more familiar to the working linguist. An excellent opening chapter (‘Prior to function’) offers a defence of Chomsky's cognitive conception of linguistics. The arguments here are familiar, although expertly marshalled. The remainder of the chapters discuss recent Minimalist theory, with special reference to the syntax-semantics interface and the Principles-and-Parameters model of language acquisition. While this material is very well presented, I have two slight worries. Firstly, it would have been useful to many readers, I should imagine, to have had a brief history of how we ended up with the Minimalist Program. Hinzen does say something about past models in relation to phrase structure, but only in passing. A historical overview of the changing face of, say, constraints on syntactic movement or Case would have served this purpose well. Secondly, for one sceptical of Minimalist theorising but potentially sympathetic, perhaps not enough is done to assuage one's concerns. For example, Hinzen might well have taken up the recent criticisms about computational explosion (valid or not), which might have led to an interesting discussion of the relation of Minimalist-style theory to parsing and linguistic processing more generally. As they are, Hinzen's chapters constitute an interesting take on what is crucial to linguistic theory as informed by the philosophy of the previous chapters. Especially welcome is Hinzen's accurate statement that generative theory has always been concerned with semantics (the so-called ‘autonomy thesis’ notwithstanding) (151). The issue for everyone is how ‘structure’ relates to ‘meaning’.

The final part of the book deals with a number of outstanding issues. Two of them deserve some comment.

Firstly, Hinzen takes up the recent ‘level-free’ derivational models of the language faculty offered by Chomsky, Uriagereka, and others. What we might call first-wave Minimalism dispensed with the levels of representation of D-Structure and S-Structure. This radical move was motivated on empirical and theoretical grounds. More recent theorising has sought to extend such reasoning to the interface level with the conceptual-intentional system, Logical Form (LF); in other words, the general idea is to see if the conditions hitherto defined over LF (e.g. binding, scope taking, etc.) can be met by a derivation that is open to the interface at various points or even at each stage of Merge, rather than just at the ‘completed’ level of LF. Reciprocally, a theoretical motivation is that LF too often plays the role of a set of global conditions to which the derivation somehow ‘looks ahead’. The search demands on such a model are far too high, if not explosive. Thus, dispensing with LF admits the thought that there are no conditions on the derivational output other than those that the derivation satisfies locally (the derivation becomes ‘crash-proof’). These are intriguing ideas and, I dare say, require a great deal of further enquiry. Hinzen is clearly very sympathetic and the thrust of the proposals is flush with his general anti-functionalism. He writes:

A number of issues arise here. Firstly, intuitively, a derivation clearly does require representations, that is, structures that might be built and transformed. Doing without levels of representation is not to do without representations (I suspect that this might be a mere terminological disagreement). Secondly, while it is clearly desirable not to have LF as a global condition that serves as a focus for ‘look-ahead’, a crash-proof syntax is not inconsistent with a level of LF or any other levels. Indeed, LF would not necessarily be redundant, if it were the case that the usability conditions that the external systems impose call for certain structural conditions to be met, even if these apply to less than full clausal structures. Thirdly, what seems to be behind Hinzen's thinking is that the syntax serves no inherent function. Yet a functionalist perspective on the syntax might still be of explanatory value, as the faculty does support whatever functions emerge through its extant embedding in an ensemble of systems. At the limit, under a non-functionalist proposal, there would be a constant on-line access to the derivation, but this seems both empirically and theoretically unsound, at least on the present understanding of uninterpretable features and the probe–goal/remerge matching process, for it would leave a structure accessible to the semantic interface even though it retained uninterpretable features, yet to be matched by corresponding interpretable features of a higher head. If access is only open at certain points, such as the CP and vP phases, as Chomsky proposes, then, in a sense, this condition is functional vis-à-vis the external systems, with the un/interpretable distinction reflecting the divergent demands of the interfaces. Besides, without a far better understanding of the ‘interface strategies’ that lead to usability, it is not clear to me why Hinzen's anti-functional model should even be a regulative ideal. Still, these are intriguing speculative ideas and Hinzen's hunches might well prove to be correct.

A second interesting stance of Hinzen's concerns lexical content (262–271). Hinzen defends the familiar atomism of Fodor, under which lexical content is broadly innate and non-decomposable. While this position is certainly a serious one, I remain unconvinced. Firstly, it is true that lexical concepts do tend to be definitionally non-decomposable, but the patterns of polysemy, aspect, and inferential connection remain to be explained; kicking them elsewhere does not relieve the explanatory deficit. Secondly, lexicalisation tends to distort conceptual adicity; thus, inferences running either way between concepts and lexical items require a far richer understanding of the role of the lexicon than we presently have. Thirdly, we should be unmoved by a priori arguments premised on a definitional model of concept constitution that tell us that all concepts (more or less) are innate. Again, without a clear distinction between concepts we might well share with at least other primates and lexicalised items, it seems to me obscure just what we are talking about. However, these are complex issues for future research.

Overall, Hinzen has done a tremendous job of blending many of the leading ideas of a range of disciplines into an always thoughtful and, to my mind, largely convincing picture of where our best science of language and mind is and should be going.