9780195136968

Chapter One

What this book is about

The mind is its own place, and in itself Can make a Heav'n of Hell, a Hell of Heav'n. Satan, in Milton's Paradise Lost

Out of sight, out of mind; times out of mind; put it out of your mind; you must be out of your mind; mind out! It's all in the mind; it comes to mind; it's on my mind; at the back of my mind; in my mind's eye. Of sound mind; of like mind; all of a mind. Absence of mind; presence of mind; strength of mind. Make up your mind; know your own mind; speak your mind; change your mind; never mind! Do you mind?

From time to time, most of us use most of these phrases, or phrases like them. We use them spontaneously and without anxiety, and they appear to convey information without ambiguity or confusion. So we seem to be able to understand what we mean by the word mind . Of course, what we mean depends on context, for both as a noun and as a verb mind has a family of related meanings. The mind that I change when I change my mind is not quite the same as the mind that is absent when I am suffering from absence of mind -- what is changed is (usually) my intention , what is absent is my attention . What is expected of the window cleaner with the ladder, when he is told to mind the baby, is different from what is expected of the baby-minder. Minding a dog is much like minding a baby, but minding dogs is quite different. If you want a friend to mind your dog for an afternoon, you had better choose a friend who doesn't mind dogs.

Our competence in using the word mind in its various meanings in everyday conversation might lead us to suppose that we have a thorough understanding of all of these meanings. Such a supposition would not be justified, and for two reasons. In the first place, competence in word usage does not imply any deep understanding of the concepts referred to by the words used. We need to know very little of harmony to be able to say, appropriately, that something is harmonious. The second reason is less obvious but not less important. Intelligible conversation requires that conventions be shared between those conversing; it does not require that those conventions be based on fact or on correct interpretations of fact. Provided that we are confused in the same way, our conversation is unlikely to draw attention to our confusion.

In this book, I want to discuss a number of interconnected questions that have to do with our minds. What kind of thing is a mind? What is the relation between our minds and our bodies and, more specifically, what is the relation between what goes on in our minds and what goes on in our brains? How did brains and minds originate? Can our brains be regarded as nothing more than exceedingly complicated machines? Can minds exist without brains? Can machines have minds? Do animals have minds? None of these questions is new, and some of them are extremely old. None of them can yet be answered in a way that is wholly satisfactory, though in the last few centuries, and more particularly in the last few decades, there have been some worthwhile partial answers. Such progress as there has been has come from work in a wide range of disciplines, and this itself creates problems, for experts in one discipline may be ignorant of or unsympathetic to the conclusions (and even the problems) of experts in another. Relevant disciplines include neurophysiology (the study of normal nervous systems), neurology (the study of diseases of the nervous system), psychology, psychiatry, philosophy, and artificial intelligence (the use of computers to do things that, when done by us, seem to need intelligent behaviour). But the questions are not only -- or even mainly -- the concern of neurophysiologists or neurologists or psychologists or psychiatrists or philosophers or computer experts. Because the answers to them inevitably affect the way in which we see ourselves, the way in which we treat others, our attitude to animals, our interpretation of the past and our expectations for the future, they are of interest to all of us.

In the chapters that follow, I shall try to show how, and to what extent, advances in various areas of knowledge have increased our understanding of the nature of minds in general and of our own minds in particular. I shall try to ensure that the arguments are accessible to the non-scientist, either by giving the necessary background or, where that is impossible, by making clear what the reader is being asked to take on trust. But before we start to consider the contributions made by individual areas of knowledge, it will be necessary to clear the ground, for we each approach these problems supported (or burdened) with a host of assumptions and preconceptions, which, whether they are right or wrong, have an influence that is all the more insidious because we take them for granted.

The first bit of ground-clearing will be to look at what might be called the common-sense response to questions about the mind. Though common sense can be wrong -- it would, after all, suggest that the earth is flat, or that white paper seen by moonlight reflects more light than black paper seen by sunlight -- to reject it without good cause is to embrace gullibility and court the seductions of every philosophical, religious and scientific mountebank. And there is no doubt that the common-sense view of the mind serves us very well in ordinary life -- it is the view even philosophers adopt when they are not actually engaged in philosophy. Yet we shall see that to accept it forces us to choose between two unattractive alternatives: either that the laws of physics that seem so universally effective in other spheres are incapable of explaining the apparent effects of our mental activities, or that our thoughts are without physical effects.

Anyone writing a book about the mind and rejecting the common-sense view has an obligation to establish what might be called a starting position. My own starting position can be summed up in three statements: first, that the only minds whose existence we can be confident of are associated with the complex brains of humans and some other animals; second, that we (and other animals with minds) are the products of evolution by natural selection; and, third, that neither in the origin of life nor in its subsequent evolution has there been any supernatural interference -- that is, anything happening contrary to the laws of physics. (This last is, if you like, a confession of biological uniformitarianism -- a belief that, so far as possible, it is sensible to try to explain the past in terms of the kinds of processes that occur in the present.) The first of these three statements seems to me self-evident. But if minds exist only in association with complex brains, then the origin of those brains is crucial; in which case the truth of the second and third statements is crucial. And no one can claim that those statements are self-evident. I therefore want to spend the greater part of this section in justifying them.

The notion that we are the products of evolution by natural selection is widely believed; but it is believed in the curious way that was foreseen by Darwin's friend Joseph Hooker in 1860. Writing at the height of the Darwinian controversy, he said, `There will be before long a great revulsion in favour of Darwin to match the senseless howl that is now raised, and ... as many converts of no principle will fall in, as there are now antagonists on no principle. A belief in evolution by natural selection is widely held, but even among scientists it is often held more as a matter of faith than of reason. At a time when the British Treasury feels that Charles Darwin is too controversial a figure to appear on a banknote, a potential American Presidential candidate advocates the compulsory teaching of creationism in American schools, and more than a quarter of first-year medical students at a major Australian university reject the theory of evolution by natural selection, it seems worth reviewing the evidence and showing that recent work, so far from casting doubt on Darwin's theory, provides impressive further support for it.

To the majority of educated people in the western world, my starting position about the origin of life is probably less acceptable than the idea of evolution by natural selection. Yet as a basis for trying to understand the nature of minds it is just as important. For if there is a supernatural element in the origin of life, a clue to the understanding of the mind might lie in that earlier mystery. If the origin of life can be explained without invoking any supernatural processes, it seems more profitable to look elsewhere for clues to an understanding of the mind. For this reason, I want to discuss the evidence that bears on the way in which life originated, a subject which, without much attention from the public, has been quietly transformed by work over the last three decades from an area of flighty speculation to an area in which plausible scenarios are carefully assessed.

Often in this book I have taken a historical approach. This is sometimes because the best way to grasp an intellectual position is to understand how it was reached, but there is another reason. Our understanding of the origin and machinery of brains and minds, in so far as we do understand them, is the product of a long history of investigation by people of different backgrounds working in many fields and in many different parts of the world. And the picture that has emerged, however incomplete, represents a major triumph in the history of thought. I hope, by talking not only about what we now believe but also about how we came to believe it, to convey something of the excitement and grandeur of that achievement, and also something of the problems and pitfalls. It is fashionable at the moment to decry the anthropocentric attitudes of our ancestors, to emphasize that we are one kind of animal among many, and to remind ourselves that our world is not the centre of the universe and that it was not created for our dominion. Such comments may be salutary, but the fact remains that the human brain is the most complex object we are aware of in the universe, and not the least of its many remarkable achievements is the partial elucidation of its own working.

Chapter Two

The Failure of the Common-sense view

Common sense is the most widely shared commodity in the world, for every man is convinced that he is well supplied with it. --Descartes

Consider the relation between our minds and our bodies. The common-sense view would be something like this: we all have bodies and we all have minds. It is true that introspection, which seems to give each of us such confidence in the existence of our own mind, is not available for examining the minds of others but, given the similarities in our bodies and in our overt behaviour, it would be perverse -- as well as arrogant -- to maintain that only one mind -- our own -- exists. It also seems obvious that our minds and our bodies can interact, and in both directions. We blush with embarrassment, we are sick with fear, and we shake with rage; Shakespeare, speaking through Shylock, tells us that there are men who, 'when the bagpipe sings i' the nose, Cannot contain their urine'. And in the reverse direction, when we bark our shins or our teeth rot we feel pain; when we eat strawberries or look at dappled sunshine in a wood we feel pleasure. Smells are particularly evocative. The smell of a particular furniture polish transports me vividly into the hall of my grandfather's house in London in the late 1940s; and we have all had experiences of that kind.

But when we look more closely at this mutual interaction between our minds and our bodies we run into a fundamental difficulty. By the 1870s, the successes of classical physics had made it increasingly awkward to suppose that interaction could be mutual. It seemed axiomatic that the explanation of each physical event lay in an unbroken series of antecedent physical events, yet the commonsense interactionist view made it necessary to believe that in humans, and probably in some of the higher animals, some physical events were explicable only in terms of conscious mental events. And that is a notion that is wholly at variance with the development of the physical sciences during the last three centuries. With the development of physics, mutual interaction between the world of physical states and events and the private world of mental states and events became, as the Oxford philosopher Geoffrey Warnock paradoxically puts it, `unbelievable and also undeniable'.

Attempts to escape from this uncomfortable position were suggested by a number of people around 1870, perhaps most clearly by an Englishman, Shadworth Hodgson; but before describing that attempt I want to make two other points.

The first is that the dilemma created by the successes of classical physics is not helped in any obvious way by the indeterminacy of quantum physics, which introduces a random element in the coupling between very small-scale physical events.

The second point is that, though the dilemma was made acute by the successes of classical physics, there were earlier intimations of it. What we now regard as the commonsense interactionist dualist view was first codified in the seventeenth century by Descartes, who thought that animals were automata, but that in humans the automatic machinery was influenced by the `rational soul'. In 1643 Descartes received a letter from Princess Elizabeth of Bohemia saying that she could not understand how the soul, being only a thinking substance, could determine the actions of the body. Descartes' reply was to ask her to consider three things: `as regards body', he said, `we have only the notion of extension [that is, the property of occupying space] ... as regards the soul ... we have only the notion of thought ... as regards the soul and the body together, we have only the notion of their union'. And, he argued, `we go wrong if we try to explain one of these notions by another, for since they are primitive notions, each of them can be understood only through itself'. Not surprisingly, this unholy and divided trinity failed to satisfy the intelligent princess, who knew a fudge when she saw one, and who in subsequent correspondence with Descartes had to be content with flattery and the discussion of less difficult problems.

It is time to return to Shadworth Hodgson. In a book with the delphic title The Theory of Practice , he proposed that mental events -- by which he meant conscious mental events (and I shall use the term in the same sense) -- were caused by physical changes in the nervous system, but could not themselves cause physical changes. Like the whistle of a railway engine (which does not affect the engine), or the chime of a clock (which does not affect the clock), they were caused by (and accompanied) physical events, but they did not themselves act as causal agents. In a slightly later terminology, they were epiphenomena -- a word taken over from the pathologists who had used it to refer to secondary symptoms of a disease.

The idea that a sensation might be regarded as a by-product of physical events in the brain -- events which could perhaps achieve their effects on the body (and its behaviour) without the mental accompaniment -- had been suggested by J. M. Schiff, in Berne, in a textbook of human physiology published in 1858; but Hodgson went further. Not only did he suppose that mental events could not cause physical events; he also supposed that they could not cause other mental events. Where one mental event appeared to cause another, what was really happening, he said, was that the physical events in the nervous system, associated with the first mental event, were causing the physical events in the nervous system, associated with the second mental event. A succession of nervous events, he argued, can give rise to a succession of mental events just as arranging coloured stones can create a mosaic picture. The stones, he compared to states of the brain; the colours, to `feelings' or `states of consciousness'. His analogy is peculiarly unsatisfactory, since apart from their colour the individual stones in a mosaic are often not distinguishable, but better analogies were supplied by William James (who did not accept Hodgson's theory) writing twenty years later. If Hodgson is right, says James, the `mind-history' of each man runs alongside his `body-history', each point in the first corresponding to but not reacting on a point in the second; `so the melody floats from the harp-string, but neither checks nor quickens its vibrations; so the shadow runs alongside the pedestrian, but in no way influences his steps'. (It was not without good reason that William and Henry James were described, respectively, as a philosopher who wrote like a novelist, and a novelist who wrote like a philosopher.)

How helpful is this theory that mind is an epiphenomenon? It is of course crucial to the theory, that mental events never occur in the absence of nervous events. Proving a universal negative is notoriously difficult, but the evidence on this point is strongly in Hodgson's favour. Claims for the existence of disembodied minds are made from time to time, not always fraudulently, but I am not aware of any that has been competently investigated and found to be convincing. So far as we can tell, mental activity is always associated with nervous activity, and later on we shall see that particular kinds of mental activity are associated with nervous activity in particular parts of the brain.

But though the epiphenomenon theory passes this first test, it is not satisfactory.

The most obvious, but least serious difficulty, is that the notion that mental events are epiphenomena -- that they are caused by, but incapable of causing, physical changes in the brain -- does not readily square with the results of our own introspection. Thomas Henry Huxley, who was a strong supporter of the epiphenomenon theory, gave a famous talk to the British Association meeting in Belfast, in 1874, in which he coined the phrase `conscious automata' to describe what we would have to be if the theory were right. But our immediate reaction to this phrase is to say that we do not feel ourselves to be conscious automata.

Consider this situation. I am driving home from the Laboratory when I remember that I have not posted the letter that my wife particularly asked me to post when I left home in the morning. I know that I will shortly pass a public post-box but I realize that there will be no further collections from it today. It then occurs to me that one of the Trinity College porters takes the last collection from the College post-boxes to the main Cambridge sorting office a little before seven o'clock. I reckon that I probably have time enough to get to Trinity and catch the porter, and as the car approaches the back entrance to the College I indicate that I intend to turn into it. The chain of private mental events has culminated in a public physical event. If at that moment I was stopped by a policeman and asked to explain in detail why I had indicated an intention to turn into the entrance, I should have no hesitation in recounting the links of the mental chain; indeed I should have no alternative explanation to offer.

Or consider Thurber's short story The Secret Life of Walter Mitty . In the course of five pages, the middle-aged, henpecked hero drives too fast, almost backs into a Buick, startles a woman in the street by saying `Puppy biscuit', mutters `Things are closing in!' when his wife finds him curled in a large armchair in the hotel lobby, and stands to attention in the sleet with his back to the wall of a drugstore. Most of these public physical events turn out to be the culmination of a sequence of dramatic but private mental events, the whole story being a succession of sequences in which the fantasist Mitty escapes from his wife's badgering by imagining himself, in turn, the commander of an eight-engined navy hydroplane, a great surgeon, a crack shot accused of murder, an intrepid bomber pilot leaving his Second World War dugout for a suicidal mission, and a man who, facing a firing squad, scorns both blindfold and cigarette. In the course of a few pages we have five examples of a chain of mental events leading to a physical event.

Even in those situations in which `conscious automaton' seems only too apt a description for us -- when, having thoughtlessly made a gaffe, we go beetroot -- we still feel that it is our embarrassment that causes our colour; that our mental state is responsible for our physical state.

But however discouraging the results of introspection may be for Hodgson's theory, they are not fatal. Introspection -- the examination of one's own thought and feelings -- is the wrong tool for investigating the theory, because the physical events accompanying mental events are not open to introspection; they are not the kind of thing that we introspect about. If it is these hidden events that supply the causal connections, then an understanding of those connections cannot be expected from introspection. That the results of introspection do not support the theory is just what the theory predicts.

A much more serious difficulty, first pointed out by William James in an essay entitled `Are we automata?' (a deliberate reference to Huxley's `conscious automata'), is that, if mental events are epiphenomena, they cannot have any survival value. Darwin's struggle for existence is a struggle in the physical world, and if mental events cannot cause physical effects they cannot affect the outcome in that struggle. But if they cannot affect the outcome -- if they have no survival value -- why should we have evolved brains that make them possible? You can, of course, make the ad hoc assumption that the physical events of which the mental events are the epiphenomena are themselves advantageous (or linked to events that are advantageous), but you then have to explain what it is about physical events in this class that makes them advantageous (or links them to events that are advantageous). That they make conscious thought possible is not relevant, for thought that merely accompanies behaviour without influencing it will be ignored by natural selection. It is astonishing that Huxley, who must have thought more about the implications of natural selection than almost any of his contemporaries, should not have commented on this difficulty in his talk to the British Association; but he did not.

And, as James pointed out, the position is even more awkward than appears at first sight. It is not just that we have evolved brains that allow us to experience conscious sensations; we must also explain the correlation between the nature of those sensations and the survival value of the activities associated with them. By and large, pleasant sensations are associated with behaviour that promotes survival (eating, drinking, making love); unpleasant sensations are associated with behaviour that is harmful. There are, of course, exceptions: a surfeit of lampreys did not promote the survival of King John, and three great institutionary dinners in a week killed off poor Dr Grant in the last chapter of Jane Austen's Mansfield Park . `Oh, show us the way to the next whisky-bar!', sings Jenny, in Brecht and Weill's The Rise and Fall of the City of Mahagonny , and though Jenny links her desire explicitly to survival -- `For if we don't find the next whisky-bar, I tell you we must die!' -- the message of the opera is that the pursuit of pleasure is a poor policy for living. But surfeits of lampreys, institutionary dinners, whisky-bars, and even whisky itself, arrived late in evolution, and cannot be expected to have had a significant influence. Despite such exceptions, a correlation between pleasant sensations and survival-promoting events, and unpleasant sensations and harmful events, remains to be explained; and if sensations are incapable of causing either physical or mental effects, the obvious explanation is ruled out. `The burnt child fears the fire' says the proverb, and we normally suppose that he fears the fire because the burn was painful, and that fearing the fire increased his ancestors' chances of survival. But if Hodgson is right, the feeling of pain and the feeling of fear are irrelevant. `Pain,' says Hodgson, `must be held to be no warning to abstain from the thing which has caused pain; pleasure no motive to seek the thing which has caused pleasure; pain no check, pleasure no spur to action.' Sensations could make us aware of events, he believed, but only nervous (as distinct from mental) activity was effective in causing both physical and mental events.

And there is yet another difficulty. Even if the notion that mental events are epiphenomena is true, it leaves unexplained what most needs explaining. Why should particular physical changes in our nervous systems cause feelings or thoughts? Even epiphenomena need to be accounted for. The smoke from the engine may not move the train but its presence is not a mystery. There's no smoke without fire, we are told, and we are confident of locating the fire in the engine's firebox. The shadow running alongside James's pedestrian may not influence his steps but there is no mystery about the shadow either. Feelings, though we are all familiar enough with them, seem altogether more elusive. Even if we were in a position to say `There's no feeling of embarrassment without such and such events in such and such bits of the brain', we would still think that to demonstrate the specified events in the specified bits was an inadequate explanation of embarrassment.

Smoke and shadows and feelings of embarrassment are in different categories. Smoke is a mixture of substances arranged in a particular way, and chemists can tell us why fires smoke; shadows, though not made up of substances, form part of the public physical world, and schoolchildren who have had their first lesson in optics can tell us why they occur; feelings of embarrassment are not made up of substances, they exist only in the private mental world, and no one can tell us -- or, at any rate, no one has yet been able to tell us -- why they should be associated with a particular pattern of nervous activity.

So, despite its promising start, the notion that mental events are epiphenomena has not got us out of the difficulties that a combination of commonsense and a proper respect for physics got us into. We shall return to this problem later.

I want to consider the question: can our brains be regarded as nothing more than exceedingly complicated machines? If we put this question to someone of robust common sense but with no special knowledge of physiology or psychology or philosophy, we can be fairly sure that the answer will be `no'; and this denial will probably be justified by a string of assertions such as: `a machine can never be intelligent', or `machines aren't conscious', or `machines don't feel pain', or `machines can't love or hate or be sad or jealous or over-the-moon', or `machines don't have free will'. And the tone of voice in which this string of assertions is delivered will betray just a hint of irritation, as if we had said something not actually improper but not quite right. This irritation may reflect nothing more than a disinclination to spend time answering Tom-fool questions, but it may also reflect a sense, however slight, of being threatened, a frisson of vulnerability.

The cause is probably twofold. First, all the machines with which we are familiar are manufactured; there is therefore a latent suggestion that, however remote the possibility, our brains (and presumably the rest of us) might be manufactured. And to be something that could conceivably be manufactured is to be well downmarket of something created by God in his own image. Secondly, because the machines with which we are familiar are indeed unintelligent, unconscious, unimaginative, unfeeling, incapable of passion and without free will, we suspect that the suggestion that our brains are machines might imply that the same list of unflattering terms could be used to describe us. And this seems tp threaten not only our self esteem but the validity of our deepest feelings. To suggest that there is something mechanical about the sound of lamentation in Rama -- that `Rachel weeping for her children' and refusing to `be comforted, because they are not' represents something less than the intolerable anguish of a bereft mother -- is to attack us in a way that we bitterly and rightly resent. But, although, and understandably, this makes us anxious, our anxiety can be allayed. The disturbing suggestion should not be made. The implication that we are unintelligent, unconscious, unimaginative and the rest is not justified. For a machine capable of doing what the brain does must, by definition, be different from familiar humdrum machines and must be as competent as the brain in enabling us to display all the qualities that we do display -- the love and the hate, the kindness and the cruelty, the serenity and the passion. To be told by a physiologist that one's brain is a machine that is different from familiar machines, and has all the properties we associate with brains, ought not to be threatening.

Except perhaps to one's patience. For the meaning of the statement is not easy to extract, as the following dialogue shows:

SCEPTIC: When you say the brain is a machine, you seem to be using machine in an abnormal sense. All machines, as the word is normally used, are contrived by someone for some purpose. Indeed the word `machine' is derived from the Greek word mekhos meaning contrivance. But no one supposes that brains have been contrived by anyone. So is machine merely a metaphor, as you might say `Little Jemima is a ray of sunshine' without implying that you could get a row of coloured Jemimas by passing her through a prism? Or do you mean merely that the brain is like a machine, or (what no one would dispute) is in some ways like a machine?

PHYSIOLOGIST: I am not speaking metaphorically, and, I am, of course, assuming that being contrived by someone is not part of the definition of a machine.

SCEPTIC: Despite the derivation of the word, and its normal usage?

PHYSIOLOGIST: Derivations don't define meanings. The word `mechanism' has the same derivation as the word `machine', yet we habitually talk about the mechanisms of biological processes -- the pollination of orchids or the clotting of blood -- mechanisms which no one has contrived. Usage is a better guide to meaning, but it does not follow that because all the machines that one usually talks about have been contrived by someone, being contrived by someone is a universal feature. Everyday cats have tails, but a Manx cat is still a cat; a zebra without stripes -- an albino zebra -- would still be a zebra.

SCEPTIC: So even if we accept your assertion that the brain is a machine, it lacks at least one feature -- having been contrived by someone -- that is common to everything that would be called a machine in everyday life. And in order to behave in the way we know it behaves, the brain must have various features that no other machine has. In insisting that the brain is a machine, you must be thinking of features that the brain has in common with other machines and that you consider justify the label `machine'. It is these features (or some of them) that you must wish to draw attention to when you say that the brain is a machine. What are these features?

PHYSIOLOGIST: There are two. The first is suggested by the dictionary synonyms for machine: apparatus, appliance, instrument . It is the existence of a function, or purpose, or role. For machines that are contrived, the role will usually be the same as that for which the machine was contrived, though occasionally it will not be. A fire extinguisher may be used as a doorstop, and a house (which Le Corbusier tells us is a machine for living in) may be used as a defence against inflation. The role of the brain is to control the behaviour of the organism of which the brain forms part, in such a way as to promote the survival, well-being and reproduction of that organism. Even if there is no contriver and therefore no conscious purpose, there can still be a role.

SCEPTIC: What is the second feature that you feel justifies the label `machine'?

PHYSIOLOGIST: The second feature is suggested by an early special use of the word machine: an apparatus for applying mechanical power, consisting of a number of parts, each having a definite function . We no longer think of machines as necessarily mechanical, but we do still think of them as functioning in accordance with the laws of physics. With all ordinary machines, i.e. machines that have been contrived by someone, the working of the machine -- its mechanism -- will necessarily be understood, though not perhaps by all of the machine's users, and not always in as much detail as we should like. With machines that form part of biological organisms, the extent of our understanding is very variable. We have no difficulty in understanding how the elbow joint acts as a hinge, or the hip joint acts as a ball-and-socket joint, though our knowledge of the characteristics of the lubricating fluid, and of the membranes that cover the bearing surfaces, is not complete enough for us always to be able to prevent inflammation and arthritis. We understand how the contraction of the walls of the heart's chambers, and the opening and dosing of the various valves enables the heart to act as a pump; and we know how the rhythm is maintained, and a good deal about the way in which the heart's muscle converts chemical energy into mechanical work. Because it is natural to think of hinges, ball-and-socket joints and pumps as machines, no one is disturbed if elbow joints, hip joints and hearts are referred to as machines; and when Mr Charnley introduced a very successful operation for replacing the arthritic hip with a contraption of steel and plastic, no one felt threatened, and those with arthritis were particularly pleased.

SCEPTIC: And brains?

PHYSIOLOGIST: With brains, of course, the position is rather different. What brains do is so complicated, and their structure is so complicated, that, although a great deal has been discovered about the way in which brains work, we are scarcely more than scratching at the surface.

SCEPTIC: If you have not plumbed the depths, how do you know that there are not lurking leviathans whose behaviour is not in accordance with the laws of physics, as we know them. Might not intelligence, consciousness and free-will be such leviathans?

PHYSIOLOGIST: We can't be sure. I don't think intelligence or free-will are strong candidates, though consciousness is more worrying. But even if we can't be sure, the best working hypothesis is to assume that the known laws of physics will be sufficient. A great deal has been discovered about what goes on in the nerve cells and synapses of different parts of the brain, and there is no suggestion of anything incompatible with physical laws. And some of the brain's overall behaviour is remarkably machine-like. For example, Professor Merton has shown that, in controlling voluntary movements, the nervous system acts as a servo-mechanism with variable gain -- to use engineering jargon.

SCEPTIC: Are you suggesting that, in the remote future, it is possible that a machine could be designed capable of doing what the human brain does?

PHYSIOLOGIST: In principle, yes.

SCEPTIC: What do you mean by `in principle'?

PHYSIOLOGIST: I mean that to design such a machine is not impossible, either in the sense that it is impossible to find a four-sided triangle -- that task is impossible by definition -- or in the sense that it is impossible to straighten the leaning tower of Pisa by stroking it with a teaspoon -- that task, though not one that has been attempted, is incompatible with the behaviour of towers and teaspoons predicted by the laws of physics, as we know them.

SCEPTIC: So `Yes, in principle' means `No, in practice'.

PHYSIOLOGIST: It depends on what you want to do. If you want to design a machine that does everything a human brain does, then I agree that, for all practical purposes, the task is impossible, and likely to remain so. But we already have machines that do some of the things that the brain does, including -- and this is the crucial point -- things that until recently would have been thought quite impossible without the intervention of a mind. In The Physical Basis of Mind , published in 1877, George Henry Lewes -- now remembered mostly as George Eliot's lover -- discusses the performance of Kempelen's chess-playing automaton. The public were surprised, he tells us, but `every instructed physiologist present knew that in some way or other its movements were directed by a human mind; simply because no machine could possibly have responded to the unforeseen fluctuations of the human mind opposed to it [italics added]'. Yet now we have chess-playing computers that can win against all but the very best players, and sometimes even against them. It is true that the methods used by these computers differ from the methods used by the brain, but the task performed is the task that common sense in 1877 held to be impossible without the intervention of a human mind.

SCEPTIC: You are saying that common sense is a perishable commodity.

PHYSIOLOGIST: Sometimes. And it doesn't come with a `use by' date, so you don't know in advance. A moment ago you conjured up three `leviathans' -- intelligence, consciousness and free-will -- and you implied that they could not be possessed by any machine. Yet already we have computers that can properly be described as intelligent, even if their intelligence is limited to a restricted field and not comparable with our own. Is it not possible that, in time, the other two will also lose their terrors? And wasn't the biblical leviathan a mythical beast, anyway?

SCEPTIC: Its status is uncertain. But what we do know is that it was not possible to

`draw out leviathan with an hook' -- not even a hook baited by a physiologist.