You are here

Third Lecture. The Development of Mind

C. H. Waddington

I am going to talk about the development of mind, using ‘development’ in only one of the two senses which John and Christopher have given it, during the life-time of an individual; what is technically called the ontogeny of mind. I shall go on to the other sense of development, namely evolution over many life-times, in my next lecture, so when I use the word development this evening I shall always be meaning development during one life-time.

The development of mental activities in a baby has been studied by a very large number of people, and there have been a few studies on other animals. We have several distinguished developmental psychologists in this University. Let me assure them that I am not going to try to teach them their own business. My task is more modest. If the four of us are going to discuss the nature of mind at all thoroughly, at some point we must look at what is known about its development. My job is to draw attention to some of the salient features that are emerging from these studies, to feed them in as grist to be processed by the powerful intellectual mills of my colleagues and to be thought about by the audience. I shall only have time to give a very sketchy outline, and I shall be doing this not with inside knowledge, but as someone in a different but related field, namely the development of the material structure of the body as contrasted with its mental activities. I think there may be a number of points in which phenomena occurring in the general development of man's body may be suggestive in relation to the development of his mind. I mean not only that, if we knew a great deal more about the establishment of connections between the nerve cells in his brain, we should undoubtedly come across things closely related to the development of mental abilities; I hope to show that the character of the developmental processes occurring in many other parts of the body besides the brain may be relevant in a more general way.

One further preliminary remark; many of the most important stages in the development of mentality go on before the child is able to talk, or certainly before it is able to think critically about its own mental processes and describe them to others. Much of the study of developing mental abilities has therefore to be based on observation of a child's activities in various situations, and deductions drawn from these.

When one looks at the literature on the development of mind, one soon comes to realise that there are very few, if any, generally accepted theories about it. There are different schools of thought, often arguing quite fiercely even about relatively specialised topics, let alone about the most general characteristics of mental development. However, I think that in some cases where rival schools have been seen as mutually exclusive alternatives, they can often be regarded as opposite ends of a spectrum of continuously varying interpretations, with intermediate views which are in some sense mixtures of the two extremes. I shall consider a number of the more obtrusive pairs of alternatives, or dimensions along which there is a spectrum of opinions, which one finds in reading the recent literature.

The first is concerned with the relative importance of internal and external factors in the development of mental ability. Is the mind of a new-born baby or other animal a complete blank, which would remain blank except for the influence of inputs from the external environment? The American behaviourist B. F. Skinner comes near to the end of the spectrum of possibilities in this direction. Right at the other end would be the possibility that all mental abilities are fully described in the genes which the child has inherited, and will duly make their appearance at some time in the life history, either before or some time after birth, without requiring any assistance from the environment other than that it should permit the individual to live and grow up. Some students of animal behaviour (ethologists) hold views near this end of the spectrum, at least in relation to the behaviour of some insects and some birds. In fact, there certainly are many examples in the animal world in which some simple and not very specific factor in the external environment will act as a ‘releaser’ to set going a complicated performance in the responding animal. This behaviour may be shown by animals which have been reared in isolation, or in abnormal situations, in which it was impossible for their responses to have been gradually built up by earlier environmental influences. They seem to be truly innate patterns of response.

However, I do not think any students of human development would claim that there are many such patterns in man, except at the level of simple reflex actions, which we would be rather unlikely to dignify with the name of mind. Reaction to the external environment, that is to say ‘learning’ in the broadest sense, certainly plays an enormously greater role in human mental development than in that of the bee, ant or some of the song birds. On the other hand, most people find it very difficult to go all the way with B. F. Skinner and accept his claim that everything is learning. He has undoubtedly shown that by appropriate teaching one can train animals to do the most unlikely things; but one must ask the question, is this the process by which they acquire mastery of doing normal things in the course of their normal lives? This seems much more doubtful. You can, for instance, train a pigeon to turn round three times and then stand on tiptoe before looking for the next pellet of food in its food tray, but one cannot apparently train a normal pigeon to carry out the sequence of behaviours which constitutes the courtship ritual of the ring dove. Behaviour of that normal kind seems to depend much more strongly on internal innate factors.

The development of the human mind therefore seems likely to come somewhere in the middle ranges of the spectrum. Undoubtedly learning from the environment is important, but most people think that internal factors are also very important. What is the nature of these internal factors?

One factor which I think everyone agrees must be of some importance is maturation. The number of cells in the human brain is still increasing at birth, and goes on doing so, at an ever diminishing rate, for some time. Moreover the insulation of nerve fibres by the deposition of myelin sheaths around them is still very incomplete in the young infant, and goes on for quite a long time after birth. Clearly one could not expect a fully functioning mind when the machinery in the central nervous system is not yet fully assembled.

There is, however, still quite a lot to argue about. In particular, does the maturation of a child have to go through a definite unalterable sequence of stages? Defining a stage is not easy and is partly a matter of convenience, but people can agree on fairly broad stages, such as being able to crawl, able to walk, able to count, realising that a ball of clay remains the same size, whether it is rounded into a fistful or squeezed out into a long thin worm, and so on. The argument is about whether these stages have to be gone through in a definite order. Piaget tends to emphasise that the stages follow one another in a definite sequence. Another very eminent developmental psychologist, Jerome Bruner, argues that the sequence is very flexible. In fact he is sometimes accused of asserting that you could teach a child anything, at any age, if you set about it in the right way. That would almost certainly be going too far; I am not certain whether Bruner has ever actually committed himself to such an extreme notion, but there must be some things which cannot be done unless some preliminary has already been accomplished. You have to learn to walk before you can run, though I am well aware that this is a doctrine which seems more convincing to elderly professors than to young students. The question is to discover which are the necessary sequences, and which are flexible. Piaget emphasises the importance of a logical sequence of developments; Bruner stresses that the child may be able to acquire a number of separate abilities not in logical sequence, and only fit them together into a logically coherent whole at a later stage.

This raises the whole problem of how things which are both complex and organised are brought into existence during development. This is one of the places where it may be useful to look at what we know about the development of organised complex material structures in the development of human anatomy and physiology.

The organised complex entities which appear during the development of an embryo will go on changing after their first appearance. They are not just organised structures, they are organised pathways of change, which I have called chreods. The point I want to make now is that, when we study the development of the material structures to the body, we find that there are two rather different and contrasting ways in which chreods may come into existence. I do not remember having seen this point made in the material I have read about the development of the human mind, although I can't claim to have done more than sample a few of the more general works by some of the best known authors, so I thought it worth explaining.

The distinction I want to make is between chreods which make their first appearance as definite coherent statements, which will then be elaborated into a whole series of details; and on the other hand situations in which a lot of isolated unrelated bits and pieces appear here and there, and are only later co-ordinated into a chreod which hangs together as a unitary organised system. The first kind, which I would call a ‘generative chreod’, may be compared to a piece of music which starts with a clear statement of a melodic theme which is then developed into a concerto or fugue. The second, which I should call an ‘assimilative chreod’, is perhaps more like what used to happen in the jam sessions of classical jazz, when half a dozen players might start doing their own thing until something gradually gelled out and the whole orchestra was working together.

In vertebrate animals the central nervous system, which gives rise to the brain, the spinal cord, and eventually all the other nerves, starts off as a clear statement, which appears on the surface of the embryo as an area with a definite shape. It appears at a definite time, and cannot be persuaded to do so much earlier or much later. It depends for its appearance on an inducing stimulus from some other part of the embryo, rather as some of the behaviour studied by ethologists depends on the ‘releaser’. But again, as in that case, the shape of the initial rudiment of the central nervous system does not depend at all closely on the inducer. It is inherent in the organ itself. One can take the cells of the inducer and mix them up into a totally disordered assemblage, and, if it still acts as an inducer, what is induced does not reflect the disorder in any way but comes out as a normal well-organised system.

Just to show what I am talking about, the diagram shows the appearance of the central nervous system and brain in the embryo of an amphibian, actually a newt, over a period of about 36 hours.

I find this sort of performance, that living things seem to be able to put on, really very astonishing. There is another example, which has always got under my skin as I just don't see how it's done. In a developing bird embryo, the limbs first appear as little swellings which then grow out into either wings or legs. Right from their very first appearance they have the character, they make the statement, if you like to put it so, that they are either wing or leg; and they do this long before any of the details are decided. If a little piece of material is taken out of the region of the leg swelling which will later turn into the thigh, and placed in the tip of the wing swelling in a region corresponding to the toe, what does it do? It elaborates its basic theme, leg, into the detail appropriate to the region of the limb in which it finds itself, namely the toes; and when the chick grows up it has a toe growing out of the tip of its wing. It is the basic fundamental theme which is stated first, leaving the details to come along later.

But, as I said, that is not the only way in which an organised chreod can put in an appearance. You may find a number of different things appearing in different parts of the embryo, disconnected from one another at first, but later being assimilated to form a coherent system. One example would be the various organs and glands concerned with sexual activities in vertebrates. The primordial germ cells appear in one region, and they have to migrate quite a long distance through the embryo before they get into the testes or ovaries. Then the external sex organs appear somewhere else, and so do various hormone-secreting glands, such as the pituitary. It is only after they have each appeared and started doing their own particular thing, that the separate elements are woven together into a coherent system which provides the material basis for sexual activity and reproduction.

As an example, consider the little animal, Hydra, which is rather like a miniature sea anemone though it lives in fresh water. If you chop up a number of Hydras and mix the pieces together, you have an artificially disorganised lump, but eventually the pieces will assimilate to each other and produce one or more well organised Hydras. Another example, of an experimentally disorganised system tending to organise itself gradually into a decent chreod, can be found in the developing eye of a normal newt. An eyecup, which will become the retina, and the lens of the eye appear in the embryo. Surgically, one can substitute for the normal lens a much bigger and faster-growing one from another species of newt. One finds that the large, fast-growing lens starts growing more slowly than it should, presumably because of some effect of the eyecup; while the eyecup starts growing faster than it should, presumably because of some influence of the lens. The final result is an eye with about the normal relationship between size of eyecup and size of lens, though the whole thing is a bit bigger than it should be in this species of newt.

Between generative and assimilative chreods there is another spectrum of varying opinions about how the notions apply to the development of the human mind. At the generative end we have the well-known and important figure of Noam Chomsky; in fact I have taken the word generative, to describe one type of chreod, from his expression ‘generative grammar’. He argues that the essential factor which enables the child to learn its first language at such an early age is an innate mastery of a universally applicable system of grammatical rules, which enable it to generate an infinite set of sentences, including ones which it has never heard or used before. Tony Kenny discussed his ideas in some detail in the second lecture he gave last year, and I shall not go further into them here. I will only remark that it is perhaps unfortunate, in the light of what we have been saying here, that Chomsky's ideas are discussed usually in terms of indicative or descriptive sentences. We have tended to conclude that language should be fundamentally regarded as imperative, conveying commands, programmes or instructions, rather than descriptions. Christopher has made this point, but I should like to see somebody spell out a reinterpretation of Chomsky's ideas in terms of a grammar of imperatives rather than a grammar of descriptions.

As an important figure towards, but by no means at, the other end of the spectrum one might take Jerome Bruner. One of his studies is concerned with the sucking behaviour of babies. Sucking actually involves two different kinds of action. One is an inward suck, giving rise to a negative pressure on the thing sucked; the other is an outward pushing with the lips, giving a positive pressure. When babies suck at the nipple, both these actions may produce some flow of milk with variable effectiveness. A baby normally develops a certain characteristic rhythmic alternation of positive and negative sucking. Bruner showed that if you provided it with a bottle with a teat so adjusted that it gave milk only on the negative or only on the positive sucking, then the baby very rapidly adjusted its behaviour to suit the circumstances. However, they do not remember this lesson between feeds, and at the next feed start again with their normal, presumably internally determined, rhythm. We thus have some interplay between an internal rhythm, which can be regarded as generative, and learnt behaviour, which assimilates the responses of the environment, in this case of the feeding bottle, into a modified adapted form of behaviour.

This type of adaptation is more fully expressed in another experiment carried out by Bruner, to find out how babies, one to three months old, bring together and organise into coherent behaviour two or more different types of activity. He let the babies look at a moving picture projected on to a screen, and fixed up an apparatus so that the picture only remained in focus when the baby sucked at more than a given rate. It was found that the babies soon learned to suck the picture into focus, and that they obviously like to have it in focus rather than out of focus. However, at this age they do not seem to be able to suck and look at the picture simultaneously, so they are faced with the problem how both to get the picture decently in focus, and to stop sucking long enough to concentrate on looking at it. The babies do actually succeed in assimilating these two originally diverse actions into a coordinated procedure which solves the problem. However, Bruner makes the important point that different babies adopt quite different strategies in achieving this. One nine-week-old baby, for instance, goes in for short bursts of sucking. Another keeps the picture in focus for much longer, but can presumably give less attention to it, while a third keeps the picture only roughly in focus by giving occasional well-spaced sucks. It is I think, because of the variableness in the way in which these two activities are brought together into one assimilation chreod that Bruner, as we saw before, is sceptical of the idea that mental development must involve regularities of the kind suggested by Piaget's unalterable succession of stages.

In the present context Piaget stands somewhere in the middle, between those like Chomsky who emphasise the generative character of chreods, and those like Bruner who are more impressed by their possible assimilative character. Piaget admits that the formation of a new organised mental activity or behaviour does depend to some extent on interaction within whatever the environment happens to offer. That is to say it has an assimilative aspect; though he also maintains that it is much affected by internal processes of maturation in the central nervous system. He claims that when the maturation has reached the right stage the assimilation of diverse bits of experience acquired from the environment into a coherent whole proceeds very rapidly. The main point he insists on is that the character of the coherent whole formed (or schema as he calls it) is very largely dependent on its own internal logic. The external elements are fitted together in one particular way because that is the only way in which they do fit together. It is as though there is quite a short period when all the instruments in an orchestra sound a few notes independently of one another, but very quickly settle on to some melodic theme which is held together by necessary internal harmonies. Once such an organised schema has arisen it can act like a generative chreod, capable of developing into all sorts of elaborations and details within the general framework set by its basic character.

Piaget develops this idea in a direction which seems to me of fundamental importance in connection with the appearance of mental ability. He argues for a process, the title of which has been translated into English as ‘Convergent Reconstructions with Overtaking’. By convergent reconstructions, he means the process by which a lot of originally divergent items of experience and behaviour become integrated into a coherent activity—the process which we have called the formation of an assimilative chreod. The point I want to draw attention to now is that referred to as ‘overtaking’; the French word is depassement. My French is not very good, but what I think Piaget means here is something much more like ‘surpassing’ than ‘overtaking’. He means that when a chreod is formed, it is often much more unified and coherent than is actually necessary.

This is the same sort of process as one which I have discussed in relation to evolution. It is easiest to see what is meant with a very simple example. Let us invent a model animal which is confined to two dimensions; say it floats on the surface of water, and it is assaulted from all directions by waves and wind. The best protection it has at its disposal are three stiff rods of some material, joined together to form a long structure which can lie along part of its circumference as a protection. If natural selection were to improve this protection, there would be two changes which could appear. The joints might be made stiffer and less prone to collapse; and the rods might be made longer. As soon as the sum of the lengths of the two shorter rods became more than the length of the longest rod, the structure could form a triangle. When it does, the strength of the whole structure suddenly increases many times over its former value, which depended only on the difficulty of bending the joints. The stability of the triangle may far surpass anything actually required as protection. When I discussed this notion in connection with evolution, the expression I used was that natural selection had brought into being a new archetype, a form which could provide the basis for a whole new range of evolutionary modifications.

As a biological example, consider the number of legs in organisms which have a hard outer skin. There are millipedes and centipedes with lots of legs, but there are not many of them, and one would not call them much of a success as groups of animals. With eight legs, there are spiders; they do a bit better, but again they are rather a minor group. When you come down to three pairs of legs, these are the insects, from many points of view by far the most successful type of animal ever produced; there are more different species of insects than of any other animal, and many of these species contain vastly more individuals than the human race. The insect form is some type of archetypal form, though nobody has any clear idea why.

Piaget argues that several human abilities depend on this type of ‘hitting the jackpot’, on the formation of a chroed which surpasses the needs of integration which brought it into being. In particular he uses this idea in connection with the development of logical and mathematical abilities. From a purely practical point of view, it is obviously useful to have some rough ideas about quantity, and logical argument, but man has developed mathematico-logical concepts, and rules of argument, which are incomparably more powerful than anything necessary for the day-to-day existence of a hunter or farmer, which our ancestors were.

It is not impossible, I think, that the universal generative grammar spoken of by Chomsky originated in a rather similar way; in some set of rules, which were at first favoured by natural selection because they made it possible to elaborate slightly on the degree of communication that could be carried out by ungrammatical grunts, shouts and so on, but which then turned out to be enormously more powerful than required by the immediate needs of the situation.

It seems, then, that the development of any particular mental ability involves factors of several different kinds. There are the external influences acting on a more or less recipient mentality, as emphasised by Skinner, and the internal factors, whose description is built into the genotype which the organism inherits, stressed by some ethologists; there is an internal process of maturation; and there is the final formation of an organised mental or behavioural system, a chreod, which may be attained by a generative process or by the assimilation of disorderly precursors. Probably factors of all these kinds are always involved, but with differing inportance in each particular case.

Contrast, for example, the developments of the human abilities of walking and swimming. Babies do not have to be taught much about how to walk. Well before they are strong enough to support their weight, they make movements of the limbs which foreshadow walking. When maturation, of their muscles, skeleton and nervous system, has gone far enough, they need only minimal help to start walking on their own. There is a short period of assimilating the earlier disorganised movements, then the normal walking pattern appears fully fledged, and begins to act in a generative manner; it can easily develop into running, hopping, skipping, long-jumping and so on, which use essentially the same arm-leg co-ordinations. It is difficult, but not impossible, to train people to use another system of co-ordination, such as the arm-and-leg-together style depicted in Greek vases of athletes, but on the whole the normal counter-swinging rhythm is fairly deeply embedded as an internal generative factor, and the influence of detailed inputs from the environment is small. In some other animals, such as sheep, the inbuilt factors are even more important; lambs can walk a very short time after birth.

Human swimming is in considerable contrast. We might remind ourselves that co-ordinated swimming movements are firmly incorporated into the genotypes of animals like fish or newts, which only have to mature to the appropriate stage of physical development to swim perfectly, even if they have been kept under anaesthetic from birth, and therefore have no external inputs. But it is very different with man, who normally does not start to swim until well after the period of maturation is over. He probably does not have much of an inbuilt system of appropriate co-ordinations; little more than some sort of ‘dog-paddle’. The young swimmer has to do a much larger job of assimilation before he achieves anything like an effective co-ordinated activity. In this process, external factors play a much larger role in determining the final outcome than they do in walking. Some of these arise from interactions with the non-living environment; what happens when you try to breath in while your head is under water is a major influence in bringing about a suitable relation between body movements and breathing. But the major external inputs are from other people, either examples to be imitated, or teachers using speech. And a swimmer can be taught, equally easily, any number of quite different styles: breast stroke, back stroke, crawl, and so on, depending on who teaches him.

One of the most important questions about the ‘higher’ and more interesting human mental abilities is, how far are they like walking, based on a chreod which is rather completely specified by internal factors and susceptible to only minor modifications by environmental inputs; or how much do they resemble swimming and need to be shaped by external factors, including cultural ones? I do not think we can yet answer this at all fully about most mental abilities, such as logical or mathematical thought, anticipation of outcomes, balancing several factors against one another, and so on. But I find it easier to think that most of them are more like swimming than walking.

Even in the case which is most in the limelight nowadays, the development of speech, dare one suggest that Chomsky has allowed the undoubted importance of a generative factor to persuade him into some degree of over-emphasis? Of course his Universal Grammar only applies to the ‘deep structure’ of sentences, and Chomsky fully realises that from any one deep structure, many surface structures may be developed, differing not only in the actual words used but, for instance, in word order, as between English and German. But a normally sceptical reader is likely to ask whether, and how, the Universal Grammar applies to such languages as Chinese in which it looks, at first sight at least, as though the distinction between nouns and verbs is made in a different way, if it can be said to be made at all.

Then there are the claims that the grammatical structure of some languages is more appropriate to express certain kinds of philosophical outlook than others. I think it was one exponent of such views, the American anthropologist Whorf, who claimed that the only language in which it is easy to talk about the world in terms of Einsteinian space-time is one of those belonging to the Indians of the American South West, Navajo as far as I remember. Apparently they do not analyse their experiences into the separate categories of space and time in the manner we do, and their language reflects this. But if so, its grammatical structure must be considerably influenced by their cultural outlook as some very basic level, so that the generative factor cannot be all-important.

There has been at least one attempt, by a leading theoretical physicist, David Bohm, to construct a language whose grammar enshrines a Whiteheadian process of view of the nature of the world, rather than the conventional common-sense view; and the mere possibility of doing this is evidence that a deep level of a generative grammar may be modifiable by cultural factors. Chomsky's Universal Grammar may not be so independent of culture as it appears to be in most discussions of it. Possibly we need to distinguish some intermediate level between the very deepest level which Chomsky tries to describe, and the surface. This would be concerned with the fundamental philosophical assumptions of the culture in which the language evolved. It would allow for the possibility that, in the development of language-ability, something more is involved than an innate generative universal grammar, and external factors which affect only such quite superficial matters as whether you use English, French or German words and word-orders.



Wad's fascinating paper was very dense, and I've only time to touch on one of the many interesting things which he had to say. Clearly much the most interesting notion which he introduced was the distinction between generative and assimilative chreods: the generative chreod in which, as in a fugue, the theme is stated first and then developed; and the assimilative chreod in which everything is much more disordered to begin with and gradually order is imposed on it. I've no doubt that this distinction is one which can be very fruitful when applied to the study of the development of mentality. However, it does seem to me that the particular application which he made of this distinction, namely, to a comparison between the approach of Chomsky and that of Bruner, was in some respects misleading.

I think that even the name is misleading if the notion of a generative chreod is meant to be generative in the same way as Chomsky's generative grammar. The notion of a generative chreod, as Wad introduced it, is a historical notion and a causal notion: in a generative chreod we have something which appears first and then directs a later stage of evolution. A generative grammar is not generative in this sense, if I've understood Chomsky rightly: the way in which the sentences of English are generated by a generative grammar is the logical sense in which all possible moves in a game of chess are generated by the rules of chess. It is a logical relation such as that between axioms and theorems, rather than a causal relation of the kind Wad illustrated for us so vividly.

I think also that Wad exaggerates the degree to which Chomsky attributes the learning of language to innate factors.

I'm particularly inclined to think this in view of his remarks at the end of his paper about the different cultural influences on the learning of language.

The universal grammar of Chomsky is not supposed to be, as it were, an elementary grammar of a rudimentary Esperanto which can evolve into English or French or German as the case may be. Rather, it is a set of constraints on the types of grammar which can be internalised. According to Chomsky the grammar, say of English, is a competence whose output is the performance of the English speaker. Universal grammar is, as it were, a black box, a feature of human nature which is postulated to explain how English grammar can be produced as an output from the input which is the linguistic data available to the child; what its parents and other people in the environment have to say to it. If I can use an analogy which for once is not drawn from computing, the universal grammar is related to particular grammars as the shape of a baking-tin is to the cake which comes out; whereas Wad thought that they were related as the ingredients, or one particularly important universal ingredient, is related to the cakes which come out.

On the contrary, it is the linguistic input to an English baby that is related to the linguistic output of an English baby as the ingredients of a gingerbread are related to a gingerbread. The universal grammar is the circular cake-tin which makes sure that the ingredients that go in come out a certain shape whether they are a gingerbread or an angel cake. For this reason, I think that Wad's fascinating distinction and the analogies which he drew from the development of the body to the development of mental activities, are misapplied if they are applied to Chomsky's theory of universal grammar.


I would like to say something about the comments Tony has made while we still have them in mind. He made two objections to my analogies, and they seem to me to be almost direct opposites of each other. At the end he talked about Chomsky's grammar as a cake-tin which simply shapes the cake that is coming out of it, but does not produce any of the cake's constituents. Now I gave, as an example of a generative chreod, a picture of the shape of the central nervous system when it first appears in the embryo. I think the analogy so far—cake-tin shaping the cake, and pear-shaped outline shaping the nervous system—is really not at all bad.

I was more impressed by Tony's earlier objection, in which he said that Chomsky's grammar is the basis for the generation of an infinite set of sentences, ‘in the logical sense in which all possible moves in a game of chess are generated by the rules of chess’. Here there is certainly an important difference from what happens in development. The beginning of the central nervous system ‘generates’ a very complicated nervous system with an elaborate brain, all the nerves running down the spinal cord, and the peripheral nerves reaching out to the fingertips and so on; but although this is complicated it is certainly limited and not infinite. The repertoire of a developmental chreod is always finite. If a generative grammar can generate an infinite number of sentences, then I admit the analogy breaks down at this point.


Could I just answer quickly? I don't think universal grammars generate sentences at all. It is particular grammars which generate sentences, and universal grammar is merely a set of constraints on particular grammars. A universal grammar doesn't generate sentences any more than a cake-tin bakes cakes.


I would like to take up Wad's contrast between walking and swimming, and start by pouring a bit of cold water on this. About ten years ago, when my first son was born, a colleague of mine, also an eminent biologist like Wad, but committed to the doctrine of the aquatic ancestry of man, held, as one of the pieces of evidence, that babies, like Wordsworth's infants, came into the world with a capacity to swim, which they then lost—they were made to forget it by being taught to walk—and so, could he, please, borrow the baby?

These reasons apart, I'm not sure whether swimming is the best contrast to walking; but the main analogy is apt. There is one way in which we could try and push the analogy further to the direct consideration of the development of mind; and this is that in some cases the jackpot phenomenon does seem to occur. I think it is particularly so in the humanities; sometimes with poetry or with novelists or with great playwrights, there's a strong sense of ‘yes, of course he's telling me what I knew before’; only, you really had to listen to someone first—Plato or Augustine, Dostoievski or Shakespeare—you didn't know it before, because nobody had actually put it in words for you. It is because the playwright is able to take you where you have already half gone, that we get the peculiar power of the literary imagination. I would like to air the possibility that this is something akin to Wad's walking. We have an ability which is already waiting to be realised; given some slight clues, you then are able to lead on much further. Not all intellectual disciplines involve this sort of ability; for example, some sorts of abstract algebras, which seem totally non-intuitive. (Not that this contrast is just that between arts and sciences. Some reasoning in the humanities is fairly opaque to the understanding and some parts of mathematics are highly intuitive.) This contrast between different sorts of intellectual abilities is like the contrast between walking and some other activities: in one case we seem to have a natural ability to develop on very slight training—we take to it very easily; in the other it comes to us only with great difficulty. I think there are applications of this contrast when we consider not only our physical capacity but also our intellectual abilities; and this is one point I'd like to see developed further.


I've just a couple of comments to make about Wad's very interesting exposition. First, I would like to put another nail in the coffin of this identification of generativity in chreods with the generativity of grammars. I don't think even generative grammars actually generate anything at all; as Tony implies, a generative grammar is more like a set of rules for playing a game: ‘you can do this if you want to’. You and I generate sentences within the constraints of English grammar, and English grammar merely tells us what sorts of sentences we are allowed to generate if we want to speak a particular dialect of English. This takes me to a further point, which is that one can mislead oneself into imagining that one is explaining something by giving it a name. It's certainly very illuminating to study these epigenetic processes and to notice the distinctions between the different kinds of way in which living things organise themselves. The trouble is, one's so used to biological phenomena in ordinary life that it isn't until one sees them under rather unusual conditions like time-lapse photography that one recaptures one's sense of wonder. But I do think it's important not to imagine that one has in any sense accounted for any of these phenomena by calling them either generative or assimilative; what we are trying to do is to understand why and how these things happen.

In my first lecture I was trying to indicate the sort of statements which might commend themselves to us as explanations of living processes. If we did have a complete understanding of molecular biology we might be able to trace the unfolding of the programs which are embodied in our chromosomal DNA, and to see how those little blueprints in fact succeed in directing operations all round them and making things go this way or that, instructing this piece of tissue that it's got to get on with the job of being a wing and leaving it to the bit of tissue as to exactly how it does it. I personally think that this is by far the best type of explanation we could hope to have for the things that happen. Although, of course, we have such explanations for no more than a very, very few of the very, very simplest biological phenomena. But as we all agreed, I think, there are other ways of looking at the unfolding of the evolutionary process which brought our species into being, and these also can be illuminating because we know that there is a relation, for which we can see good reasons, between ontogeny and phylogeny; so all I would want to say here—and I'd invite Wad's comments—is that I don't think we actually explain things by naming them. I am sure that the concept of a chreod is a valuable concept, but it doesn't seem to me to constitute in any sense an explanation of the puzzles and the phenomena that we are faced with, and I only say this because I think we have an awful long way to go before we can understand the remarkable imperturbability of biological systems.


If it is true that generative grammars do not generate sentences, then don't blame me for introducing the word ‘generative’. It was introduced by Chomsky, and I have only adopted it from him. And, of course, in some sense, the generative chreods I spoke about don't generate the nervous system, or anything else; they are stability constraints on the way the nervous system can generate itself, if you like. I think this discussion has got into a semantic tangle about whether ‘generative’ means what it says, or something else; if necessary we can return to this later.

In the few minutes left I would like to say something about the second point that Christopher made: that giving things names does not explain them. Of course not. In my next lecture I don't promise to explain the appearance of chreods, but at least I intend to talk about it. For instance, it is clear that natural selection during the processes of evolution has played a large part in building into the genotype the ability to express itself as a generative chreod, controlling the development of the nervous system, or as an assimilative one, bringing into co-ordination the pituitary, the gonads and so on. All these examples of stabilisation have been produced during evolution, and the evolutionary forces must have played a considerable part in their genesis. But one might ask the question: ‘Does evolution have to do the whole job?’, or are there perhaps some general rules about complex interacting systems which result in them exhibiting some inherent stability properties? I shall point out in my next lecture that there are in fact some hints in support of this second alternative. Now it is when one considers questions of this kind that I think one is going beyond the assignment of names towards the discovery of explanations. The point of giving names is to get a clearer idea of what it is you want to explain; but I entirely agree with Christopher that we should not be tempted to think that we can stop there.

From the book: