Tales of mystery and imagination

There is a great deal of fuss these days about consciousness. Yet, 10 years ago or so, the fussometer reading would have hovered around the zero mark, where it had been stuck for most of the past century. Just a few philosophers worried about a problem that was totally ignored by the scientists.

So what has caused this change? Some great scientific breakthrough? Alas, no. The problems pondered by the philosophers are still with us - the contemporary excitement about consciousness tells us more about the Zeitgeist than the Geist . I put much of it down to those pretty pictures of bits of the brain being active, as shown by neuro-imaging. The power of this method has convinced scientists that now they can watch the brain in action, they can forget the hoary old issues raised by philosophers and get on with the real job by the simple expedient of describing what the brain does.

But the philosophers were way ahead of the scientists on this one. They long ago considered what it would mean for our understanding of consciousness if we knew in detail every correlation between every event in the brain and every event in consciousness. They concluded that this alone would tell us nothing fundamental about the nature of consciousness - about how it relates mechanistically to activity in the brain or, indeed, how it fits at all with the rest of natural science. Correlation does not establish cause. Contemporary science has no idea of how neuronal firing can cause conscious percepts. Worse still, science has no idea why the brain needs to cause conscious percepts anyway.

I believe in the "hard problem" of consciousness (see earlier review on page 33), something that will take more than a collection of data, no matter how sophisticated, to solve. What is needed is a conceptual revolution, a stance that only Max Velmans among these authors shares (although the other two, in varying degrees, do pay it lip service). That is perhaps because he is the one most at home with philosophy. Indeed, his Understanding Consciousness presents a lucid, indeed masterly, account of the philosophical issues involved. It is strong on historical scholarship and conceptual analysis, can be easily digested by the beginner, and is an excellent review for the jaded. New readers should definitely start here.

But the book is much more than a philosophical treatise. Velmans is a scientist as much as a philosopher. He brings the philosophy to bear on the interpretation of scientific fact and theory and vice versa, to the enrichment of both. This is reflected in the clarity of his argument that the real world, in all its visibility and tangibility, forms part of the problem of consciousness, since the brain of each one of us constructs this world and projects it into a 3D frame that it equally constructs.

Given that no one yet has a solution to the "hard problem", it follows that if you believe in it, then one way or another, you have to cop out. Velmans' cop-out is to posit complementarity between "first and third-person accounts of consciousness and its neural correlates". According to this view, both accounts are correct; they simply exist side by side so you can choose the one most useful for your immediate purpose.

This, of course, is exactly the way neuroscientists function, taking one perspective in the lab and the other, like everyone else, out of it. But this is a statement of the problem, not a solution. You might try seeking respectability for brain-consciousness complementarity by analogy with wave-particle complementarity in physics. But this does not help. The mathematical apparatus and predictive theory of the latter just does not exist for the former. Until we have such a theory, brain-consciousness complementarity is an empty promissory note.

Still, full marks for Velmans up to the cop-out. But I wish he had not written his last three pages. After a perfectly rational 7 pages, we suddenly trip over Carl Jung, not to mention "SHU the dweller within the one million beings". Those of us willing to admit to the hard problem constantly court the danger of being lumped in with, to use Daniel Dennett's caustic phrase, "the New Mysterians". These characters leap on science's inability to solve the problem of consciousness as evidence that "there is something else". This makes it harder for the rest of us to tread the fine line between accepting that science has not yet solved the hard problem, while not doubting that it will do so one day.

By reading Velmans first, you may benefit from a healthy scepticism as the other authors merrily sweep the hard problem under the carpet. But let your scepticism be tempered by the following consideration. The most likely scenario for the eventual emergence of a solution to the hard problem is that the accumulation of experimental observations will force specific issues into new and unsuspected moulds, altering the questions until they become tractable to science.

Who in the early 19th century could have foreseen the development of post-Newtonian physics - a development that could not have taken place had the physicists of the time not collected the data that the old physics could no longer handle? It is of little moment whether, in the 21st century, scientists who collect and organise data from neuroscience, psychology, computer science and so on, do so believing that their activities will be sufficient in themselves to crack the problem of consciousness, or whether they see themselves as preparing the ground for a radically new theory to come along. When that new theory is written, which, if any, of the books reviewed here might be seen as its harbinger? My money is on Consciousness: How Matter Becomes Imagination by Gerald Edelman and Giulio Tononi.

This is a remarkable book, sparkling with ideas, steeped in scholarship, offering rich food for thought to specialists, yet written in a manner and with a brevity that allows the general reader to start at many different squares-one and still be in there at the finish, grappling with the whole daunting complexity of the scientific problem posed by consciousness.

The basic strategy pursued by Edelman and Tononi is to map a range of high-level properties of consciousness, for example the capacity to integrate a diversity of elements into a single unified experience - its serial nature, its limited capacity and so on, onto equivalent properties of the activities of the brain. The latter are first subjected to scrutiny from many perspectives, including those of neuroanatomy, neurophysiology, mathematical modelling and Edelman's previously formulated "neuronal Darwinism": the theory that many connections in the brain are formed by virtue of their consequences rather than by genetic ordinance. To make each of these perspectives clear, along with the sometimes idiosyncratic but always absorbing inferences the authors draw from them, is a real tour de force .

It is possible that some of the ideas developed by Edelman and Tononi will figure in a genuinely explanatory theory- when we have one. But this book does not take us as far as its authors would have us believe. To demonstrate correspondence between properties of consciousness and properties of the brain is one thing. To treat them as an explanatory theory is quite another. There is a figure on page 108 in which this critical boundary is flagrantly crossed. Amid a number of boxes marked with the names of regions of the brain there is one, linked by inoffensive arrows to the rest, named "primary consciousness". I am reminded of the cartoon in which an Einstein stands in front of a blackboard covered in equations, one of which reads: "And then there is a miracle!" Critically, however, the authors' model of how the brain creates consciousness is spelled out in sufficient detail to be open to experimental test. I was particularly struck by the assertion, based on well-formulated theoretical arguments, that the anatomy of neuronal interconnections of one kind (found in the cerebral cortex and thalamus) is able to sustain conscious activity, whereas the anatomy found in other regions of the brain is not. I shall be surprised if this assertion does not provoke useful experiment. That is the real news. For the first time, ideas about consciousness are moving from philosophical argument into the experimental laboratory. It is there that progress will be made.

Igor Aleksander's How to Build a Mind offers a narrower, and hence less demanding, overview of consciousness. His book is a quirky mix of autobiography and imaginary playlets featuring great figures of the past (from Anaximenes to Wittgenstein) and present (Roger Penrose to Melvyn Bragg) as a means of setting out rival views and descriptions of research into computer simulations of complex behaviour, which, he tells us, will lead to the construction of machines with consciousness.

One of the playlets has "Susan Greenfield" summarising the difference between her approach and the author's: "The answer must lie in the neurons. Not Igor's artificial neural nets - messing about with computers gets you nowhere." This is neatly dramatised, but in fact Aleksander accepts the value of neuroscience and Greenfield that of computer models. It would be nice to tell you which of these tracks will get you to the heart of the forest faster. But having accepted that there is a hard problem to be cracked, I am unable to tell you who will do it.

The solution, most likely, will require a combination of small advances in a range of empirical studies. But the majority of scientists with an interest in consciousness are looking for key breakthroughs to come from either neuroscience or computer science, so reading Aleksander's book will bring you up to at least half-speed. It is entertainingly written and readily digestible for the non-specialist. Indeed, it makes it all seem rather easy. This is what worries me.

If there is a hard problem, we should not let the scientists get away with telling us that just by elaborating their computer models or physiological experiments they will shortly deliver an understanding of the most central experiences of our lives. Occasional disclaimers aside, that is exactly the impression this book gives. This passage will give you a flavour: "I assume that for a system to be conscious ... it must ... possess neural machinery capable of creating ego-centred representations in some of its firing patternsI It seems to me evident that whenever neurons fire to contribute to states that reflect the cohesiveness of the world, one says that the organism 'feels', 'perceives', or 'imagines' the world." Far from being "evident", this remark (typical of the "functionalist" tradition in which Aleksander's work squarely lies) ignores reams of philosophical argument (well summarised by Velmans) showing that the mere identification of neuronal firings or information processing with conscious sensations just does not work.

Indeed, the very notion of "representation" that Aleksander relies on is, philosophically speaking, problematic. "So what," many scientists will snort, "who cares what the philosophers say?" But conceptual problems do not disappear just by being dumped into someone else's discipline. And even within this pair of science-oriented books there is an antidote to Aleksander's reliance on representations as the equivalent of conscious experiences. On well-argued empirical grounds to do with the way the brain actually works, Edelman and Tononi launch a powerful attack on the notion that it contains "representations" of any kind. I suspect Velmans would side with Edelman and Tononi here. Indeed, if you take seriously (as I do) his claim that the whole experienced world is constructed by the brain, there is nothing else out there - apart from the world as described by physics - for our experiences to "represent".

The debate in this crucial field of inquiry is as lively as it has ever been, turning on both conceptual and empirical issues. It is from the interplay between these approaches that some day the truth will emerge, driving out old mysteries and New Mysterians alike.

Jeffrey A. Gray is emeritus professor of psychology, Institute of Psychiatry, King's College, London.