Too much eclectic hand-waving

Cognitive psychologists investigate the processes underlying thinking and reasoning, the properties of memory, and the heuristics of problem-solving. Neurophysiologists study the anatomy of the brain and the functioning of neurons and neurotransmitters. Few would dispute that scientists from both disciplines are working on opposite sides of the same coin, but there are few good theories that provide a convincing story of the relationship between cognitive processes and their neuronal implementation.

The lack of such theories is, of course, due to the enormous difficulties facing any attempt to bridge the chasm between mind and brain. This chasm has not deterred William Calvin, however, from having a shot at the problem, for in The Cerebral Code: Thinking a Thought in the Mosaics of the Mind, he lays out a wide-ranging and innovative theory linking the neural structure of the cortex to thought, language and consciousness.

Calvin's source of inspiration is somewhat surprising. One expects a book such as this to draw heavily on neurophysiology and psychology, but his principal inspiration is from Darwin's theory of evolution. He argues that the dynamics of mental processes reflect a Darwinian process of evolution. The idea is not simply that our brains are the result of a process of Darwinian evolution. That is not the issue. The idea is that, as Calvin puts it, the brain is a Darwin machine, that thinking is the evolution of thought - that evolutionary processes of competition, reproduction, mutation, and selection shape the development of our thoughts - and our train of thought is no more nor less than the continuous combination of "species" of existing thought that thrive, compete, reproduce and die out in the constantly changing environment of the brain. The Darwin machine is not just a metaphor. It is, Calvin postulates, what links mind and brain.

In the first half, neurophysiological evidence is employed in arguing that the necessary mechanisms for evolution are present in the brain, and the second half shows how such a machine can account for a variety of everyday psychological phenomena. To me, the presentation is both fascinating and worrying. It is fascinating because from a computational perspective the Darwin machine is a radical concept. Who knows what computational abilities such a device might possess? It shares little with the current computational approaches to psychological functioning, or the current approaches to computation in general, and is surely a concept worthy of investigation. But my praise is shrouded in caution. The argumentation worries me.

Behind the central idea, stunningly thought provoking as it is, there is a great deal of handwaving. Virtually every argument is presented through a musical analogy, but I was constantly searching for substance, unsure if I really understood what was being said, and wondering if it had any real explanatory force. The book is eclectic because the topic is eclectic. Few readers will have the breadth necessary to feel at home with all of the strands woven together (though a useful glossary alleviates many potential difficulties). Calvin's treatment of those areas in which I do have more than a passing knowledge only serves to compound my worries.

The discussions of language, and universal grammar in particular, are a case in point. The Darwin machine provides only the grossest account of pronominal binding and embedded structures (in that a Darwin machine can, allegedly, support the computational mechanisms implicated in such phenomena), but says nothing about where the numerous constraints on these processes, and the many other aspects of universal grammar, come from. Handwaving is hardly a substantive objection to any work designed for a general audience. However, Calvin explicitly states that this volume is primarily aimed at "fellow scientists". Once again, I am not convinced.

These ideas deserve to have an impact on cognitive science, but cognitive scientists will have to look elsewhere for detailed argumentation supporting Calvin's position. Nonspecialists will probably find The Cerebral Code perfectly adequate as a fascinating and readable presentation of a novel and radical approach to bridging the gap between mind and brain.

Richard Cooper is lecturer in psychology, Birkbeck College, London.