Patricia Churchland is less convinced. When I am asleep, I am unaware of the smell of cinnamon in the air. Distracted by a video game, I am unaware of the movement of my tongue. If a stroke renders me paralysed, I may be unaware of my paralysis. Each of these cases presents an opportunity to the neuroscientist trying to understand the nature of conscious experience. So what are the differences in the brain when I am aware of a stimulus, and when I am unaware of it?

Discovering the relevant differences marks an important starting point in discovering the mechanisms of conscious experience. To advance beyond the starting point will require a detailed understanding of the brain's anatomy and physiology.

At this early stage in the project, what is meant by "consciousness" is best specified by example, using well-attested instances. More contentious examples, such as whether frogs are visually aware, can be sorted out once cognitive neuroscience is a little more solid. In other words, we start with common sense, and see where the science leads.

Is the above approach hogwash? Possibly. Dualists, who believe that there is a nonphysical soul in addition to the physical brain, will certainly say so. Colin McGinn and Jerry Fodor, for example, have declared that the brain is more complicated than it is smart and that consciousness will forever be a mystery to us. Given how much remains to be discovered about the brain, not to mention the unpredictability of technological innovation and the doors thereby opened, it is surprising that they rely on poverty of the imagination.

³ÉÈËVRÊÓƵ

Now suppose we do find some phenomenon really mysterious. This is a psychological fact about us - not a metaphysical fact about the nature of the world. It is a fact about what we do and do not know, about where science has and has not reached.

For Roger Penrose, the key to consciousness lies in quantum events in tiny protein structures - microtubules - within neurons. Why there, and why quantum mechanical properties? Because the nature of mathematical understanding, Penrose believes, transcends the kind of computation that could conceivably by done by neurons and networks. As a demonstration of neuronal inadequacy, Penrose cites the Godel incompleteness result, which concerns limitations of provability in axioms systems for arithmetic. What is needed, according to Penrose, are unique operations at the quantum level. Quantum gravity, were it to exist, would do the trick. Granting that no theory of quantum gravity exists, Penrose and his colleague Stuart Hameroff argue that microtubules are about the right size to support the envisioned quantum events, and have the right sort of sensitivity to anesthetics to suggest they do sustain consciousness.

³ÉÈËVRÊÓƵ

The details of the Penrose-Hameroff theory are highly technical. Before investing time in mastering the details, most people want a measure of the theory's "figures of merit", as an engineer might put it. Specifically, is there any hard evidence in support of the theory, is the theory testable, and if true, would the theory give a clear and cogent explanation of what it is supposed to explain?

The figures of merit are not encouraging. First, mathematicians generally disagree with Penrose on what the Godel result implies for brain function. Additionally, the link between conscious experiences such as smelling cinnamon and the Godel result is obscure at best.

Now, is there any significant evidential link between microtubules and awareness? Hameroff believes microtubules are affected by hydrophobic anesthetics in such a way as to cause loss of consciousness. But there is no evidence that loss of consciousness under anesthesia depends upon the envisaged changes in microtubules, and only indirect evidence that anesthetics do in fact (as opposed to "could conceivably") have any effect on microtubules. On the other hand, evidence points to proteins in the neuron membrane as the principal locus of action of hydrophobic anesthetics.

Is there any hard evidence that quantum coherence happens in microtubules? Only that it might. Surely the presence of cytoplasmic ions in the microtubule pore would disrupt these effects? They might not. Surely the effects of quantum coherence would be swamped by the millivolt signalling activity in the neuronal membrane? They might not be. Can the existence of quantum coherence in microtubules be tested experimentally? For technical reasons, experiments on microtubules are performed in a dish, rather than in the animal. If tests under these conditions failed to show quantum coherence, would that be significant? No, because microtubules might behave differently in the animal, where we cannot test for these effects. Does any of this, supposing it to be true, help us explain such things as recall of past events, filling in of the blindspot, hallucinations and attentional effects on sensory awareness? Somehow, it might.

³ÉÈËVRÊÓƵ

The want of directly relevant data is frustrating enough, but the explanatory vacuum is catastrophic. Pixie dust in the synapses is about as explanatorily powerful as quantum coherence in the microtubules. The theory needs work.

What I share with Penrose - and essentially all neuroscientists - is perplexity. So far, cognitive neuroscience really does not have its hands on a conceptual framework for explaining how experience is based in brain operations. Significantly, awareness is by no means the lone enigma. For virtually no higher function do we have a theoretical framework adequate to yield genuine, full-blooded explanations - of how, say, one recalls the punch line of a joke Uncle Bart told last month, or how a skill like typing becomes automatised or even how eye movements are controlled in reading. For none of these cases do we have a theory that genuinely explains how the effect is achieved.

To be sure, most new ideas are bound to go the way of the three-legged trout. But the climate should not be so harsh as instantly to snuff out any contender that looks outlandish. For this reason alone, I applaud the boldness of Penrose and Hameroff. Having looked closely at the details of their proposal, however, I am inclined to pin my explanatory hopes on cognitive neuroscience.

Patricia Churchland is director of the Experimental Philosophy Lab, University of California, San Diego.

³ÉÈËVRÊÓƵ