Do you believe in 'zombies'?

How could conscious experiences affect brains? The simple answer is that they could not and do not. Yet we feel as though they do. This clash between the way things seem and the way they rationally must be has fuelled one of the age-old conundrums of philosophy. The question forms the title of this slim volume from the Journal of Consciousness Studies , consisting of a target article, a series of commentaries and a final response.

The main article is by Max Velmans, professor of psychology at Goldsmiths College, University of London, and well known in the field of consciousness studies for his books and his proposed "programme for a nonreductionist science of consciousness". The commentaries are by philosophers, psychologists and cognitive scientists of widely differing opinions. Velmans also provides an appendix and a final reply to his critics.

Briefly, Velmans' argument goes like this. Our conscious minds seem to control at least some of our actions. There is evidence from psychosomatic medicine that imagery, hypnosis, biofeedback and other mental factors can affect medical conditions, but there is no accepted theory of how.

Reductionist theories cannot explain how conscious experiences could have bodily effects, so Velmans wants a nonreductionist theory, but this raises three problems. He solves these problems by proposing a theory of biological determinism that is compatible with experienced free will.

His theory is a version of dual-aspect theory and can be summed up by the dictum "ontological monism combined with epistemological dualism"; that is, one kind of stuff in the world but two ways of knowing about it. As Velmans describes it, there is only one underlying reality but it can be observed from two perspectives. From the third-person perspective we see brains, neurons firing, and all sorts of external events in the world; from the first-person perspective we see the same things but as conscious experiences. This, he claims, solves the famous "hard problem" of consciousness and allows it to be true that conscious experiences have bodily effects.

But Velmans gets himself into some fearful tangles along the way. Take the example of mind-body effects in medicine; just calling them that already carves a distinction between mind and body that may not be warranted. Then Velmans discusses biofeedback, the placebo effect and the effects of imagery and hypnosis on health as evidence that the body can be "influenced by conscious states". But where is the evidence that the "conscious states" (if they exist) are relevant at all? Yes, we see that people engaging in imagery practices can control their bodily states; yes, we find that dummy pills and impressive machines can alleviate illness, but this is a long way from evidence that it is the conscious states themselves that are causal.

Velmans also falls into some classic traps. For example, in describing the brain in terms of information processing, he writes: "In principle, the same function, operating to the same specification, could be performed by a non-conscious machine." But perhaps it could not. Perhaps any machine with that specification would necessarily be conscious. This is the old "zombie" problem that has bedevilled consciousness studies for years. The philosophers' zombie is an imaginary creature that would look, act and speak like a normal conscious human and yet would be totally unconscious.

Many arguments have shown that believing in the possibility of zombies leads only to confusion, and yet Velmans makes his claim without giving even a passing mention to such classic difficulties.

I have deliberately picked out problems that the commentators do not mention. For their part, they mostly tackle Velmans' proposed three problems with non-reductionist theories. These are: that the physical world appears causally closed so there is no room for interventions by consciousness; that one is not conscious of brain processing so how could one consciously control it; and that conscious experiences come too late to affect causally the processes they relate to.

The commentators variously accuse him of circularity, non sequiturs, premature closure, tentative speculation, category mistakes and, perhaps most important, of not replying to the standard objections to positions such as his own.

Some of the most powerful arguments are made by the philosopher Robert Van Gulick, who admits that he finds Velmans' arguments puzzling. He suggests that Velmans comes close to nonreductive physicalism but then rejects that view without making his reasons clear. As for Velmans' three challenges, Van Gulick argues that two are misconceived and are not really problems at all, while the argument about causal closure is not only important but devastating to Velmans' theory.

Coming from the perspective of artificial intelligence, Ron Chrisley and Aaron Sloman point out that few of Velmans' arguments are new and his proposal would benefit from being located within existing discussions. By carefully unpicking his argument, they show that Velmans ends up with the converse of what he was aiming at, that is "ontological dualism, but epistemological monism".

In the last section of the book, Velmans puts up a spirited defence of his position, but he cannot really evade the serious objections of some of his commentators.

The volume is full of interesting arguments and provides a manageable and at times amusing introduction to a classic argument. I do not think, however, that Velmans will convince many people that he has solved one of the great consciousness problems.

Susan Blackmore is a psychologist and writer.