Christopher Longuet-Higgins

In last year's Gifford Lectures my colleagues and I discussed The Nature of Mind. How much light we cast on the matter we must leave you to judge. None of us, as Waddington remarked in the last lecture but one, attempted to define ‘mind’; but I think we all dropped pretty heavy hints as to what we were talking about. John Lucas laid great stress on his rationality and moral autonomy. Tony Kenny set much store by his ability to speak and understand a natural language. Wad aligned himself with all those creatures which select and modify their own goals. And I claimed that, even if computers can't think, at least you and I can compute.

This year our subject is The Development of Mind. In case, by some oversight, we should fail to define ‘development’, let me say at once how I shall be using the word. I want to use it in its biological sense—or, rather, in its two biological senses. The first sense occurs in the name developmental psychology, which is the study of mental development in the individual from conception to senility. I deliberately say ‘mental development’ rather than ‘behavioural development’, because like Chomsky I value the distinction between subject-matter and evidence. A baby's behaviour may be a very good clue to its mental processes, but this does not imply that the object of our enquiry is the behaviour itself. In any case other kinds of evidence are also available to us; such as the baby's heart rate which goes up suddenly, so Tom Bower credibly asserts, when the baby is surprised. And if surprise is a form of behaviour rather than a mental state, then I am not a native speaker of English.

The other biological sense of the word ‘development’ is the evolutionary sense. Two billion years ago, we believe, there was a primaeval soup in which the first living things assembled themselves. Whatever one's views about the nature of mind, one can hardly deny that there is a good deal more mental activity going on now than there was then. What has caused this remarkable development?

The two questions—why do we develop minds as we grow up, and why has evolution produced minds at all—will be my subject. I shall suggest that the former question is a perfectly sensible one, but that we just don't happen to have much of an answer to it. We know, thanks to Piaget and his successors, quite a lot about how a child's mind develops, but we have little or no idea why it develops further than that of a chimpanzee, say. The other question, why did evolution produce beings with minds, I believe to be unanswerable, for reasons which I shall explain. But I see no reason to doubt that we shall discover how it happened; and furthermore, that when we know the evolutionary history of our species, we shall be in a position to understand why our minds develop along the lines that the psychologists are beginning to map out.

I must explain what I mean by ‘how’ and ‘why’. I remember being told by my elders and betters that science could answer ‘how’ questions but not ‘why’ questions, and I believed this for many years. The suggestion was that science offers us general laws about the way in which things happen, but gives no reasons why these laws should hold rather than others which we might think up. In a sense this is true, but it misses one very important point: that the laws of science are not logically independent assertions about how the world behaves, but have to fit together as far as possible. The chief difference, between natural science on the one hand and unbridled speculation on the other, is that scientific assertions must harmonise, not only with experience, but with one another. If an otherwise plausible hypothesis in biology clashes with one in physics, then one or other must go by the board. Dissension can be tolerated for a while, but not indefinitely.

The fact that science is a coherent body of ideas, and not simply a disconnected set of observations, makes it possible for science to answer not only ‘how’ questions, but also certain sorts of ‘why’ questions. The answer to a ‘why’ question will require an appeal to laws more general than those requiring explanation, and there will come a point at which the laws being appealed to cannot be explained scientifically. But a very few primary laws may suffice to account for a very wide range of events and phenomena, and the most remarkable thing about the physical world is the fewness of the principles on which it seems to operate. I say ‘the physical world’ advisedly, because it is not at all obvious that the same must be true of the living world and the world of human affairs. So we had better consider the relation between these worlds, or rather, between the ideas which we use for thinking about them.

Last year I talked at some length about ‘reductionism’, and criticised the suggestion that psychology was really physiology, which was really chemistry, which was really physics, so that psychology was really physics. I now realise that this simple message was open to misunderstanding. In his book, Beyond Reductionism, Koestler seems to suggest that we can expect little help from physiology in our attempts to understand how we think or perceive; and this I would wish to deny. But of course the relation between a pair of sciences, such as psychology and physiology, or chemistry and physics, which obviously border one another, needs very careful definition. What I was trying to say last year was that the concepts of the ‘higher’ science—adopting the convention that psychology is ‘higher’ than physiology—cannot be arrived at by analysing those of the lower science. What I would now add is that although psychological questions must be posed in psychological terms, questions about the mechanism of thought or perception can only be answered in physiological terms. In other words, if the questions which arise in any science cannot be answered within that science, they can only be answered by a lower science, not by a higher one. It would, for example, seem distinctly strange to offer a psychological explanation for the propagation of impulses along nerve fibres!

With these warnings in mind, let us see what modern psychology can tell us about the mental development of the human child. Here I tread with the utmost delicacy, because Edinburgh is a centre of excellence in the study of babies and small children. The first problem is to find a way of describing a child's mental faculties at any stage in its development: what concepts it has at its disposal, what perceptual tasks it is able to perform, and how it sets about solving practical problems such as arranging wooden sticks in order of length; The greatest challenge to scientific description is the astonishing feat of first language acquisition; and once the child, at three or four years old, is able to talk to the psychologist the range of possible studies is enormously increased, and so is the quality of the available evidence. I will not attempt to review our present knowledge of human mental development, particularly as Waddington will be talking later about the work of Jean Piaget. I will only say that in Piaget's view all children go through very much the same stages, in the same order, just as we all go through the same stages of physical development, in the same order. The difficulty with this assertion is not in believing it, which is very easy, but in substantiating it, which is very difficult: how is one to define a ‘stage’, so that one can recognise without doubt when two children are at the same stage? Piaget's special achievement, as I understand it, has been so characterise these stages, or ‘structures’ as they are called by the Geneva school, in terms which are relatively unmistakable to the trained observer.

In passing, I must admit to a slight unease about the use of the word ‘structure’ in this connection. The word suggests something entirely passive, like a bit of scaffolding. Surely psychological development consists of the progressive mastery of different skills, and of the concepts required for them. For this reason I am attracted by Seymour Papert's interpretation of Piaget's ‘structures’ as programs1 or routines which arise in the child's mind and are to some extent, but not fully, open to introspection and conscious modification. Papert, on the basis of his own observations, believes that one can teach children to ‘debug’ their own thinking programs by teaching them, at a very early age, to write and run simple programs on computers. This might turn out to be the best use to which computers can be put.

For those who are not familiar with the jargon of computing I should explain that a ‘bug’ is simply a programming error. The most insidious bugs are those undetected errors which cause a program to fail intermittently. One has to be continually debugging one's mental programs. When we went over to decimal currency, most people managed to re-program their monetary thinking without too much trouble, except that a lot of people had difficulty in persuading themselves that the florin was worth only ten new pence, not twenty. In my case I am pretty sure that this was because of the visual resemblance between the symbol for two shillings and the symbol for twenty; as soon as I spotted this bug I stopped making that particular mistake. But the sixty-four-dollar question—to think in American currency for the moment—is this: what causes Piaget's structures or Papert's programs to arise in the child's mind in the first place?

The answer to this question might be superficial, or it might be beyond the reach of scientific investigation; I believe it is neither. It might be superficial if all our mental skills were directly imparted to us by our parents and teachers, as some of our skills undoubtedly are: for example, the kind that an apprentice learns from a master craftsman. But there is overwhelming evidence that our most primitive skills, such as our ability to interpret visual impressions in spatial terms, and our ability to learn the grammar of our native language, are inborn. Psycholinguists have made much of the apparent uniqueness of the human being in his ability to discover the grammar of his parents' language. Whether we are indeed unique in this respect doesn't seem to me to matter very much; I would be more than happy to think that a chimpanzee such as Sarah or Washoe could learn a language of the kind which human beings use. It does seem extremely unlikely, though, that we could ever teach cats or rabbits to read, or even to communicate with us in sign language; one's pessimism is based on the feeling that these animals lack some absolutely basic mental capacity without which they could never even get to Square One. It is this mental capacity, that Chomsky describes as our knowledge of universal grammar, which we are fortunate enough to bring into the world with us.

The trend in developmental psychology at the moment seems to be away from the idea that the human being is little more than the product of a lifelong schedule of operant conditioning, and towards the view that his mind, just like his body, develops largely from the inside, as it were. It would be most surprising if this were not so, because the vehicle of our thoughts, namely the central nervous system, continues to grow and ramify, though more and more slowly, during the first few years of our lives. The distinguished psycholinguist Jacques Mehler, alive to the acute difficulties facing any ‘instructive’ theory of mental development, has gone so far as to suggest that the new-born child knows all the answers, as it were, and merely forgets those which turn out to be irrelevant to his circumstances. Perhaps it is a sign of the times that such an idea should be advanced in the name of science; it was actually expressed by William Wordsworth nearly two centuries ago.

The conclusion to be drawn from these remarks is that the development of the child's mind seems to be entirely in line with what we know about the development of other biological functions in a growing organism. All animals, as they grow from birth to maturity, develop not only in stature but in wisdom: the particular kind of wisdom they need for the particular styles of life they lead. Though a philosopher might run into dualistic difficulties about the relation between the animal's physical growth and its mental development, the zoologist suffers from no such hang-ups. He takes it as a matter of course that every biological function—and thinking is no exception—must be performed by some organ, or organs, and that until these are fully developed the function must remain immature or unrealised. But there is something else which the zoologist knows, and which is not at all obvious to a mere pet-fancier. He knows that the growth of an animal—and indeed of any organism—is controlled and directed by a program, immensely longer and more informative than any program ever fed into a computer. The longest man-made program I know of is about as long, on paper, as an ordinary detective story. The program for constructing a human being in the womb runs to about the length of the Encyelopeadia Britannica. The program for making me is written in every cell of my body; if you looked you could just see, with a powerful microscope, the reels of tape on which it is written, though the individual letters are far too small. So if we are ever to understand in full detail why our brains, and hence our minds, develop as they do, we shall have to understand the instructions which are written on these reels of tape: the chromosomal DNA.

Put in this way, the problem of understanding how our bodies and minds are formed looks totally insoluble: even if we could read the billion letters of a human blueprint, how could we ever make sense of them? And so it would be, were it not for the other things that the biologist has to tell us, not only about the way in which organisms grow, but also about their evolution.

Everyone knows the main outlines of the theory of evolution. The problem, very roughly speaking, is to understand why there are men and dogs about, but not dodos or dinosaurs. The reason is simple: the dodos died out, and the men haven't yet, presumably because intelligence has so far proved a more valuable asset to human beings than elegance to the dodos. The problem reduces to that of explaining how a species can, as it were, conduct experiments in survival. Nowadays we know that the range of experiments which a species can try out is essentially determined by its genetic resources, enshrined in the chromosomal DNA of all individuals of the species. The DNA is pretty faithfully transmitted from one generation to the next, but now and again there are errors of transcription, called mutations, which may result in the appearance of a new breed. So the history of any existing species, such as our own, is to be thought of as a random walk through a space of possible forms, with the special property that every form along the route had to be fully viable.

Now this constraint, that all one's ancestors must have been fit enough to survive to maturity, has an analogue in developmental biology; namely that at any stage before an individual reaches maturity he must be in good enough working order, biologically speaking, to reach the next stage. And this kind of constraint is not unique to biology; it is thoroughly familiar to the civil engineer who has to make sure not only that his bridges will not blow down when built, but that they will not collapse during construction. What I am leading up to is the thought that it may be an extremely tricky problem—fortunately Nature's problem, not ours—to put together a highly complex organism; and perhaps the only way of doing so will be to follow pretty closely the ancestral walk through the space of possible forms, starting with the most primitive and ending with the most mature. This is the thought which enables us to understand dimly why ontogeny, the development of the individual, roughly recapitulates phylogeny, the evolution of the species.

The structure of my main argument may now be apparent to you. If it is really true that ontogeny must largely recapitulate phylogeny, then we do not have to wait until the molecular biologist has read our chromosomes before we can begin to understand why our minds develop as they do. We may, instead, be able to gain some enlightenment from studying our evolutionary ancestors and cousins. The missing links recently found by Robert Leakey in East Africa, and the partial success of the Premacks in teaching Sarah to read and write, hold up the mirror to our nature much better than any molecular biological study possibly could; though without molecular biology we would be unsure of the validity of such comparative studies, because we would not properly understand the evolutionary process.

Let me expand briefly on the way in which an evolutionary view might affect our thinking about mental development in human beings. One immediate consequence might be that we took a distinctly less academic view of such concepts as intelligence, rationality, morality and so forth. Take, for example, intelligence. If the human mind is the product of evolution by natural selection, then the quality for which Nature has selected us is probably not the capacity for abstract thought but the capacity to use our heads in tricky situations. On this view, the poet, the philosopher and the scientist enjoy faculties which have been earned for them by the hunter, the craftsman and the man of action. A romantic and a salutary thought. Or, if we are thinking about morality, we may surmise that the survival of our species has in the past depended upon its members helping and protecting one another, rather than merely telling one another not to do this or that. But here I am quite out of my depth, knowing only that the moral sense is a very delicate plant which is all too easily blighted by harsh conditions in early childhood.

The only trouble with the evolutionary approach is that homo sapiens now seems to be evolving in a different way from any other species. Claims for the uniqueness of man might have been open to question two million years ago; but whether it is to our credit or not, we seem to have set a genuine precedent in inventing the idea of a civilisation, or cultural tradition. The existence of civilisation rests crucially upon our ability to record our thoughts in writing and, more recently, in other media as well. We seem to be no longer dependent upon the vicissitudes of genetic variation, but to have replaced them by the vagaries of human politics. It seems that we are on our own now; having taken over from Nature we must accept the responsibility for moulding our biological future.

In introducing my main thesis I remarked that we already know, through the work of the developmental psychologists, quite a lot about how the child's, mind develops but that we should not succeed in understanding why it develops as it does until we had acquired a better evolutionary perspective. I hope my reasons for this assertion are now rather clearer. The development of our brains, like that of our bodies, is very largely controlled by the programs embodied in our chromosomal DNA; and these programs record the biological secrets which enabled our ancestors to escape extinction. So if we are concerned to understand the most striking features of our own mental development, such as our ability to master our first language, we should look not only at the way our brains develop but also at our recent evolutionary history. At present we can only observe that our brains continue to develop both in size and in interconnectedness as we advance from infancy to maturity. But how well such a pattern of physical development, leaving adequate scope for the impressions of early childhood, would suit the needs of a species whose biological prosperity depended upon the handing on of a cultural tradition!

The question why evolution has produced creatures capable of talking English, building cities and visiting the Moon, is an altogether more forbidding one. All the evidence points to the view that nothing can create new chromosomal DNA except completely random events; this is the Central Dogma of molecular biology. On the other hand, an initially random variation will tend to be perpetuated if it suits the established life cycle of an animal, which is why species tend to get better and better at what they are already good at. So once a species has found a particular niche, it will tend to exploit that niche to the utmost. Unfortunately—or perhaps fortunately—this principle seems to be of little or no predictive power. Organisms have a way of discovering niches and exploiting them in ways which no-one would have thought of, just as, in our own era, history has a way of taking totally unexpected turns which only afterwards do we recognise—if we are good Marxists—for their historical inevitability. All one could have confidently said, in surveying the primaeval soup, was that evolution would almost certainly produce something interesting. We like to think that human beings are specially interesting.

Whether this is an acceptable proposition in natural theology or not, I must leave my colleagues to discuss. At least it is very nice to know that no scientific obstacle to our existence would be posed by the lack of a Creator with our interests specially in mind. In other words, there is much to be said for having a theory of the origin of mind which does not involve us in an infinite regress. In this respect the origin of mind is a similar problem to the origin of life; it would be commonly regarded as a confession of intellectual defeat to conclude that life on earth must have come from somewhere else.

If I stopped at precisely this point you might suspect me of exceptional insensitivity to the mysteries of our existence. So I will say one more thing before I close. There are different ways of talking about Man and Nature; some of them emphasise Man and some Nature. Traditionally, the scientist treats of Nature and of Nature's indifference to Man; and in much older tradition, the poet sings of Man and of his dependence upon Nature. Need we shut our ears to either?

Discussion

LUCAS

Let me start with points on which I agree with Christopher. The point I found most interesting was his bridge-building analogy which explains the recapitulation, in the development of the individual, of the development of the species. I think that the account he gave was a valuable and illuminating one, and one that has a certain moral for the philosopher; in particular, although this is not an entirely new point, that we should not think of the mind simply, as John Locke thought, as a blank sheet of paper waiting for the impress of outside experience, but rather as an internally programmed—although, as I would say, not entirely programmed—system which generates a great deal of our intellectual activity. Particularly to those studying philosophy, who are brought up on the English Empiricists, I stress the point how wrong and how unempirical the Empiricists were. We don't sit and wait for things to be impressed on our consciousness: we try out things and see what happens. That is to say, we are always putting Nature to the question; we set the questions, but Nature gives the answers. In this respect, I am disagreeing with John Locke, and I think it is an important point to bring out.

There is another thing Christopher said, which makes a point Locke often made, namely, the relative unimportance of academic activities. Locke, putting it in a theological mode which Christopher would not adopt, said that God gave us sufficient understanding for the practical affairs of life, but not so as to fit us to speculate on matters with which we had no proper concern. Perhaps that goes a bit too far. What is, however, of great importance is that when we come to try and understand the mind, we too easily take an excessively academic attitude to it; and then with Kant we find soon that it is necessary to abolish knowledge in order to make room for faith. The proper answer to this is to remember that Kant offers us a critique of only the pure reason, and that Christopher is asking us to do what a study of evolution is forcing him to do; to consider the mind first and foremost as a practical matter. It is a matter of making choices, when you are hunting, shooting, or carrying on some line of business; it is a matter of choosing aright. And this is a second point which I feel we could very well learn from the biologists.

Now for some disagreements. When Christopher was giving his account of the evolution of mind (p. 8), he gave us the Central Dogma: nothing can create new chromosomal DNA, except completely random events. A little bit earlier, he was explaining evolution in terms of a random walk. I want to pick on this word ‘random’, which is a very slippery concept. It is negative and it is equivocal. It is negative in the sense that for a thing to be random it must be inexplicable. But the concept of explanation is itself equivocal and ambiguous; there are many different sorts of explanation, and you want to know the sort of explanation that someone has in mind before you are able to say what is random or not. Take the example of a pin-table (see diagram). There are lots of pins (B), a supply of bagatelle balls (A), and then at the bottom, a number of different slots into which they can fall (C). We start dropping bagatelle balls in. This is, in one sense, a model of a random process; but in the long run, if you start seeing how the balls go into the slots, you find that the distribution is a Gaussian distribution, e to the minus x squared. Now there are two different questions we could ask about this curve. One suggestion is that it has got this shape because a small boy, like a Maxwell demon, followed each ball all the way through and put it into one of these slots. If you are asking with that sort of explanation in mind ‘Is it random?’, the answer is ‘Yes, it is random’. But you could ask ‘Is this a random distribution?’ in a second sense. You are asking this with respect to the global features of the whole structure; you are asking it with various other possible explanations that you might have in mind, to see whether it is something special about this e to the minus x squared, or whether it could have been an e to the minus x, cosh x, or e to the minus x to the power of 4, or various other things. Could it have been one of those? And then the probability merchant will say: ‘No, it couldn't have been one of those; it had to be e to the minus x squared’. That is, it is not random with regard to that sort of question, and that sort of explanation. And this is the point that I want to raise on the apparent randomness of the random walk which has led us here. Yes, of course it is random in one way, that is it was not carefully contrived. Later, I shall give a philosophical, even a theological, justification of why it should not have been contrived (pp. 123–30). But in another sense, it's not random; as Christopher himself said after he had introduced the Central Dogma, evolution was pretty well bound to produce something interesting. And on that point I shall stop; this, I think, is where our dispute will lie and this is why it is going to be possible to accept the randomness of the random walk in one sense without being committed to complete randomness in another. To borrow from the title of a well-known book published in the last year, it is random in the sense that it is not one sense a necessity that everything that has happened had to happen just as it has; but it is not random in the sense that it is quite unintelligible why it should have happened—that it is entirely in every sense a pure matter of chance.

LONGUET-HIGGINS

Well, I find awfully little to disagree with in what John Lucas has said. I think that he made only one mistake, namely, when he spoke of a particular distribution of the pin-balls being random. The concept of randomness cannot describe any particular situation; it describes our expectations about the average results of infinitely many trials. In fact the concept of randomness should really be interpreted as applying to processes, not to particular detailed situations. What I was trying to say was that the processes of modification which occur in the chromosomal DNA, and which give one the variation upon which evolution can then get a hold, are random. The question ‘Why did that cosmic ray arrive at that particular moment to produce that particular mutation?’ seems to be a silly one, and I think that you would agree that it was. In saying this I am not denying that in surveying the primaeval soup, one might have said with reasonable confidence that once some self-replicating system had got going, something interesting was virtually bound to happen; and it's for that kind of reason that people look quite optimistically these days for life in other regions of the Universe, because they think that it's more than likely that there are certain general principles which ensure that things get more and more interesting provided you don't have absolutely impossible conditions for them to contend with. So I don't really want to disagree very much with you.

WADDINGTON

I don't want to take up most of the points that Christopher has made, because I shall be speaking later about the ontogenetic development of mind, and also about its evolution. But the question of randomness in evolution has been raised and I should like to say some things about it. The word ‘randomness’ has been applied to several different facets of evolution, and as John Lucas said, in each context it partly applies and partly does not apply.

For instance, it is often said that changes in the chromosomal DNA are random changes. But there is an elaborate science of mutagenesis; if you read the recent numbers of any of the main journals in genetics, they are largely devoted to analysing the causes and character of mutational changes. We have in this University a very famous research group, under Dr Charlotte Auerbach, engaged precisely in this study. The changes that can happen to a piece of DNA are restricted in kind; there are only a certain number of things that can happen to it, and this is definitely a limitation on the randomness of mutational events. When people say that the changes in the DNA are random, what they usually mean is not that they are random from the point of view of biochemistry, but that they have no causal connection with the environmental factors which the adult organism is going to meet, and which will exert natural selection on it. There is a real disconnection between the changes occurring in the DNA and the tests that are going to be applied to that DNA during natural selection; to speak of ‘randomness of mutation’ is an old-fashioned, and by now very unfortunate, way of referring to this disconnection.

Now let's consider another context. Christopher pointed out that the evolution of higher organisms is a matter of changes occurring in populations which contain very large numbers of genes, and a great amount of variation in those genes. The populations have highly heterogeneous gene pools. When, in a situation of this kind, natural selection exerts some new pressure, so that the population has to evolve to meet a new test—perhaps a new ice-age has started or something of that kind—does the population have to wait for the occurrence of a new mutation, random in the sense we have just discussed? I think the answer is that it does have to do so if what it needs is something which involves only one or a few genes. If it needs to change the sequence of amino acids in some particular proteins such as haemoglobin, it has got to wait until a piece of DNA turns up that will put just the right amino acid into the right place. But if the situation requires that the animals get longer legs, or can run faster, or fly better, or fulfil some other complex requirement of that kind, there will probably be dozens of different combinations of genes that can produce an adequate result. The selection pressure is operating on some activity of the organism which involves at least several tens, and possibly several hundreds, of genes. In this situation the appearance of new DNA mutations in the gene pool may be totally unnecessary. The population can probably produce the desired result from the resources already available in its highly heterogeneous gene pool.

In an earlier lecture I used the analogy that if you need to build a very simple structure, such as a prehistoric dolmen or Stonehenge, you would have to wait till chance brought your way stones of suitable size and shape. But if you wish to erect a modern concrete building, you need not worry very much about the precise sizes and shapes of the stones in the aggregate, and the fact that these also have been shaped by chance is of very minor import. Much the same argument applies to the evolution of complex functions in higher organisms; it is true to point out that the basic elements are genes which have been produced by mutations which are random in the sense of the first paragraph, but this is trivial, and gives us almost no help in understanding the interesting questions.

Finally there is a third context in which ‘random’ has been used. Christopher spoke of ‘a random walk’ through the space of possible types of organism. From a large gene pool, in which mutation is continually occurring, a very large number of different genotypes could be constructed, and these could give rise to a great variety of organisms, larger or smaller, with longer legs or shorter legs, hairy or bald, and so on. Some of this vast array of possibilities will actually have been realised in the ancestors of any present-day species, so that one can say that the species has evolved along a path through this space of possibilities. Christopher used the phrase ‘a random walk’ to describe this path through the possibility space. Here I think I really disagree with him. I do not think this is a random walk at all. I think the walk is very much confined by the tests that natural selection is putting on the system. For instance, once an ancestral horse has evolved to the stage of running away instead of standing its ground and fighting off an attacking wolf, it will evolve structures and behaviours suitable for this strategy. This makes it very difficult to reverse the strategy, and start evolving instead the kind of fore-legs that would make effective weapons if it tried to stand and fight. It is much more likely that the horse will evolve further along a path consistent with its previous strategy. Its progression through the space of possible forms is therefore not at all random, but on the contrary very much conditioned by what it was doing before.

KENNY

May I join my voice to the chorus which is suggesting that there are more senses of random than were distinguished by Christopher? I would like to take up something which he said when he was partially recanting his rejection of reductionism last year (p. 2). He said that one would not expect there to be an explanation in terms of a higher science of events at a lower level of scientific explanation, and that we would not expect for instance the transfer of energy along neural pathways to have a psychological explanation.

In one sense this seems true: one would not expect there to be general laws of a psychological type governing such events. I think that is what he meant, but surely in another sense what he said must be quite false. His own reading of his paper involved a considerable amount of activity of the kind he had in mind and the explanation of that must be psychological; I mean there could be no explanation of it which did not include the counterpart of the thoughts he had and his desire to communicate them to us.

It seems to me that the same sort of point can be made about the role which he wanted to attribute to random mutations in the history of the development of the human species. I think he was consciously putting this forward as an alternative to the idea that there was some superhuman designer who caused or supervised or designed the process which ends up with human beings like us. It seems to me that it isn't necessarily in any way an alternative. Those who believe in a divine design would think that the divine design was related to the particular mutations which are in various respects random, in the same way as the psychological factors in Christopher's lecture were related to the physiologically random events involved in his opening and shutting his mouth and moving his tongue.

LONGUET-HIGGINS

Let me explain a bit more fully what I meant when I said that one wasn't going to find an explanation for the phenomena of a lower science within a higher science. I was talking about the general rather than the contingent. I meant that it would be quite misguided to look for a psychological explanation of how, in general, impulses travel along nerve fibres. For a particular occurrence of this kind one might very well find a psychological why explanation—when I decide to lift this pen, the pen rises into the air, and I have explained a contingent lower-level fact in terms of a higher level science—but it would be quite another matter to explain a low-level generalisation in terms of a higher level science; and I think that is the way I would want to wriggle out of Kenny's rope.

On Wad's point I stand here under correction, because Wad is a very distinguished student of these matters and I am a rank amateur. But I am not sure if what Wad said actually strikes at the root of any of my main arguments. Of course, if one is going to lecture on the theory of evolution, one needs a whole course of lectures, and one would have to go into the question in what sense can a new combination of existing genes be regarded as a new biological possibility, as opposed to the appearance of a new gene produced by a mutation at a particular position in the DNA. Obviously a new combination of existing genes can provide enormous power, and it is the role of sex in biology to enable new combinations to be produced and then to be selected for. Wad also pointed out, quite correctly, that the ancestral walk which I spoke of—the random walk through the space of possible forms—is only random in a sense, and I should obviously have attempted to be more precise about the meaning of the word ‘random’. But I don't think I left the matter there; I remember saying that the process was subject to a very important constraint. Perhaps the one that I mentioned wasn't strong enough, that every one of one's ancestors should have been a good biological specimen, good enough to have at least lived to the age of reproduction. But all the mutations which actually occur are probably lethal, as von Neumann established in his discussion of self-replicating machines. The mutations which we actually notice are a miserably small fraction of the total number of mutations that actually occur, and in that sense there is a very heavy selection going on right from the word go, so that these mutations are not random from that point of view. I think one of the most interesting things about the structure of science as a whole, is the role which random events seem to play in the development of processes which are very far from random. For example, if we heat one end of a copper rod and cool the other, we can be absolutely certain that heat is going to flow through the rod. The reason we can be absolutely certain is that we can treat the thing mathematically as a completely random state of affairs subject only to extremely general constraints, namely, that the centre of kinetic energy is displaced from the middle of the rod towards one end rather than the other. I don't think there is any need to get schizophrenic about the fact that the world has, from one point of view, got a lot of randomness to it, and from another point of view, seems to be extraordinarily orderly, well controlled, and intelligible. My main thesis is simply that there is all the difference in the world between saying these things after the event and saying them beforehand; between reviewing human history or human evolution afterwards to see how it went, and seeing it from the other side of time as something which inevitably had to happen.