As I said at our first meeting, I am a biologist, really an experimental biologist; I'm not a philosopher and I have little experience of dealing with the mind. Therefore this lecture will be rather a biologist's down-to-earth lecture, before we go back to more refined philosophical considerations of the body.
I want to ask whether questions we normally ask about the mind, free-will, determinism and so on, are really phrased in terms of an adequate model. I'm going to suggest that they very often are not; that very often people think that we are dealing with a system in which the details of the output correspond precisely to the details of the input; and I'm going to say that that is not the way many biological things work, and I don't think the mind does so either. And then; many of the questions about free-will—can you make up your mind to do something—are usually discussed as though you were a man in a padded cell, totally isolated; and of course most people, most of the time at least, are not in padded cells, but are in situations in which they are continually showered with all sorts of inputs from their environment. And in particular from other people.
I shall approach by asking what we know of the character of other biological systems, simpler than mind.
This can be ‘accused’ of being a sort of ‘reductionism’. I agree, but I think it legitimate, so long as we recognise that appeal to (‘reduction to’) a lower level entity should never be allowed to force us to reject anything discernible in the higher entity. What we may have to do is to change our concept of the lower entity to accommodate the new properties we find that it can, under some circumstances, exhibit. When I said this, the other night, using an example from chemistry, Christopher Longuet-Higgins rebuked me, saying that the Schrodinger equations contain everything that there is to know about the interactions between atoms and molecules. Conceivably; but are they not a trifle on the inscrutable side? No one can solve them for anything more complex than the hydrogen atom, and, if they do contain everything necessary to understand protein molecules, no one can get it out of them. So when, some time ago, Jacob and Monod in Paris discovered that protein molecules, highly complex molecules, could exhibit a property called allostery, they got the Nobel Prize for it. Allostery is a name I don't suppose many of you have heard before. It's the phenomon that if you act on one sensitive site in a complex body, sometimes another sensitive site somewhere else gets ready to be active or becomes active; a phenomenon, one might, say, that is not outside the ordinary run of sexual experiences. But it was quite novel about protein molecules, and its discovery there did, and quite justly, obtain the Nobel Prize. So I think that reductionism—attempting to reduce things to the simpler elements—may be OK if you do it carefully enough. I'm going to suggest that many of the discussions one reads about the mind and about determinism don't do it carefully enough, and actually reduce too far, to things like Newtonian billiard ball atoms, when they ought to have stopped at some intermediate stage.
So, let's consider some of these intermediate stages. One of the main characteristics of biological systems is a sort of inherent cussedness. When I was at school we learnt something called La Chatelier's rule, which was about the world in general, to the effect that if you do anything to something, it will react in such a way as to cancel out what you've done. It's not actually absolutely true, I think, but it's a fairly widely experienced situation, and it certainly applies very well to many biological systems. If you run very fast and increase the CO2 content of your blood, the body does something that reduces the CO2 and brings it back to normal. Similarly if you change the pH of your blood, its acidity, again there's a very efficient mechanism for restoring it to normality. This is a tendency to restore the status quo. It's usually called ‘homeostasis’, keeping the thing stationary, at the same situation. A more general phenomenon is one where the things are changing all the time. Being in my professional experience an embryologist and evolutionist, I'm always used to things that change, and don't stay still. Embryos never stay still, they go on developing. But they go on developing along definite pathways, and if you chop a bit out, or add a bit on, or do something like that to an embryo, unless you go too far, it usually succeeds in getting itself back to normality and producing the normal end result. This is not a stabilised situation; it's a stabilised pathway of change; what I call ‘homeorhesis’.
To take a very simple example of this sort of phenomenon; if you drop the soap in the bath, and pull the plug, the soap goes down the drain, pretty well independently of where you dropped it. This is, of course, due to there being a physical constraint on the system, the bath with the drain at its lowest point.
In biological systems, there is not, in general, anything which can be compared to the bath. The constraints on the system have to be generated by the system itself. This requires, firstly, that the system must be complex, containing many component processes; and secondly, that these must interact in ways which ‘stabilise’ (better, ‘canalise’) a certain pathway of change, by feedback devices which monitor whether one component is going too far off track, and if so operate to bring it back. Such a pathway of change is a ‘chreod’. In a chreodic system, so long as external factors do not push the system beyond certain limits, the end result is independent of what the inputs precisely are—the system will finish up at its inbuilt end-point, just as any stream within a certain valley will finish up flowing out by the main river. Unless, of course, it is pushed across a watershed, when it will flow down another valley into another river.
Many biological systems, for instance embryos, have a character which cannot be described by a single chreod, but only by an ‘epigenetic landscape’, containing many chreods separated by ‘watersheds’ (which technically have the perhaps unfortunate name of ‘catastrophe regions’). For instance there are ‘watersheds’ separating developmental pathways leading from the early embryonic condition to adult nerve, muscle, epidermis, various glands etc.
The mind has to be regarded as a system of chreodic systems. Each of its components is complex, with many sub-components, and these interact in ways which tend to stabilise the patterns of activity. Let me give some examples:
First, biochemical. We know at least a score of abnormal genes in man, which cause the appearance of abnormal enzymes, which in turn bring about abnormalities in mental functioning and thus in behaviour. The normal forms of all these enzymes must be involved in normal brain functioning.
Next anatomical. The actual course of the nerves, and their interconnections, is best known for one of the evolutionarily oldest and most primitive parts of the brain, the cerebellum, which is concerned with learning and then controlling (unconsciously) skilled movements, such as running downstairs. Even though our knowledge of the ‘wiring’ of the cerebellum is probably only roughly accurate, it is certain that many of the cells function by monitoring the activity of other cells, and signalling the result on to still another group, so as to adjust their activity accordingly. That is to say, many of the cells are acting as feedbacks, to keep the whole activity within certain limits.
Then language. A word (or concept) is a chreod covering a certain range of values. (An orange may be a tangerine, a mandarin, a Navel, a Seville etc.)
Finally, consider computer intelligence. People who are trying to program computers to carry out mindlike activities have had to build chreodic properties into them, that is, the ability to deal with individual details in connection with larger systems. When Winograd was constructing a program to allow a computer to understand normal language, he had to ensure that when presented with the input ‘Is there a block on a green table? What colour is it?’ the last word, ‘it’, was not taken alone, but in its place in the whole context, which makes it clear that the question is about the colour of the block, not of the table. Unless such comparisons can be made between different parts of an utterance, and a balance struck between them, natural language would not remain within the chreod of ‘making sense’, but would seem chaotic and impossible to interpret.
The paradox of the mind is that although it is the most flexible instrument which has been produced by evolution, making an enormous range of reactions, each very appropriate to a larger variety of inputs than any other system can take cognisance of, it does this, not by letting ‘specifically complex input’ be translated directly into a correspondingly specific output, but by inserting an intermediate stage, in which the ‘demands’ of the input are matched by a suitable combination of pre-existing lower-level (less complex) multi-purpose chreods (programmes, symbols). For instance, we usually formulate a new thought by a new combination of existing words, rather than by just defining a new word.
Now any system which has a chreodic character, that is, a tendency to change in some definite direction, or to reach some definite end-point, could be said to exhibit an unconscious will, with the end-point as its goal. Thus I am claiming that all mental events have something of a will-like character.
Of course, in pre-Freudian days, will was nearly always taken to mean conscious will—except by a very few sophisticated novelists, such as Lady Murasaki in 10th century Japan. Even now, the problem which troubles most people is the effectiveness or otherwise of conscious free will. What, if anything, does the concept of chreods allow us to say about it? Consciousness undoubtedly has a largely private character, and is therefore difficult for science to get to grips with. (Kenny will probably discuss further just how private it is). I can only offer some rather disconnected remarks.
First, (a parenthesis), I would like to ask whether consciousness is so very important after all. Some Eastern religions have attributed the highest values to meditation, not action; for them free will, or even consciousness of the external world in general, are not so very important. Even if we place greater value on Good Action, rather than Good Being, might it not be argued that the really Good Man just acts rightly as a straightforward expression of what he is, without having to think consciously, or argue with himself, about it? Is our emphasis on the problem of free will part of the Western World's insistence—still further emphasised in the Puritan tradition—that no action can be of value unless it is difficult and unpleasant?
It has not been too difficult to discover something about the machinery which decides whether conscious experience will occur or not. We know about brain centres which control sleep, or arousal, or even rather more specific experiences of pleasure, discomfort and so on.
This is far from telling us exactly what we shall be conscious of. We clearly are never conscious of everything which is available for us to be conscious of at any moment. We tend to be conscious, firstly, of anything unexpected—if any of the audience had a bright blue face, I should probably consciously notice it, even in a general glance round. Secondly, we tend to be conscious of anything relevant to some contest to which we are directing attention. This ‘direction of attention’ is something much more specific than mere general arousal—I think it is one of the main mysteries in the operations of the mind. It seems to me that the important questions about Free Will are closely allied to the problem of how we can direct or concentrate attention on some particular context.
In the first place, I want to make the point that the subjective feeling of free will is not at all connected with the logical problem with which it is usually associated in discussion, namely whether we are dealing with a system which is not subject to the deterministic laws of causation. It is only in the most trivial instances, when we are least conscious of ‘making a choice’—for instance when we have to pick up either a cup of tea or of coffee while engaged in a conversation about something much more interesting—that we can be satisfied to ask whether our action is determined by cause and effect or whether there is some Stochastic Demon throwing dice inside us. In more important choices—shall I get married to Joan and start earning a salary, or stay on as a student and get Honours?—the feeling is not just of the operations of Chance—though that may also be involved—but rather of being pulled in several directions.
One can describe the situation by saying that the state of one's mental apparatus is such that it is poised on a watershed between two or more chreods. It is in a ‘catastrophe region’; and a slight diversion one way or the other will take it into one or another valley of the epigenetic landscape, which may lead to very different end-results. Perhaps we are particularly likely to be conscious of the existence of choices between chreods when we are posed on the watersheds between them.
One might take the view that the feeling of ‘exerting will power’ is merely a symptom which arises when our internal state is somewhere in the watershed or catastrophe region; that the strength of the feeling, or the demands on our will, simply reflect the strength of the opposing forces which are tending to push us in one or other direction; but that the outcome of the opposition of these forces is fully determined in advance, and that however much we feel we are exerting will power, we are not influencing the situation at all, merely registering it by a certain emotional response. I think the gist of the debate about free will is to decide whether this is an adequate description.
In the first place, I should argue that it is certainly inadequate to suggest that the outcome of a conflict between two or more internal drives which we have at any given moment is determined by the strengths which those drives have built up over any substantial period. To suppose so leaves out of account that we normally live in surroundings which continually bombard us with unforeseeable inputs—and the chreods between which we are poised are systems which ‘canalise’, towards distant goals, not only the internal forces which arise within us from our genetic endowment and our previous experience, but also the immediate input with which we are showered by our environment. And our environment usually includes other persons, and we cannot tell in advance just which ones we will run into and what they will communicate to us, thus contributing a push off the crest of the watershed in one or other direction. This liability to unforeseeable inputs gives our actions some degree of the ‘indeterminateness’ or rather unpredictability of the classical example of ‘chance operations’, the tossing of a coin. No one supposes that, when you toss a penny, the deterministic laws of classical physics are in any way violated—but the inputs into the systems of movements are so many that the outcome cannot be predicted.
My own strongest experiences of free will—or the exertion of will power—have probably been in the last stages of quarter mile and half mile races. When you are within fifty yards of the tape, you feel: just how far can you concentrate on putting on a sprint at the expense of pushing your oxygen debt, and the lactic acid of your blood, to levels which feel most unpleasant? The point I made above is that some unforeseeable input from the external world can make a real difference. The stop watch will show that you did better when out of the corner of your eye you saw some other man's beastly legs pounding away just half a yard ahead of you, than when you didn't. And you felt you ‘tried’ harder: was this just a reflection of what the visual evidence of his presence deterministically did to you, or did it make any difference?
Well, the evidence that he was there, and winning, only spurred you on to cliff-hang a bit more perilously on the ‘run’ side of the watershed because you went on concentrating attention on a particular objective, winning the race. But supposing you had diverted your attention to something else. All sorts of odd things happen at the extremes of physical endurance—you may feel you are really three feet up in the air above your head, watching; you may feel, you have got your nostrils and mouth fully open to take in the air you need, but there is no muscle power to suck it in, it is all being used in your legs. If you did divert attention to consciously experiencing these things, you wouldn't run faster. But could you control your attention in this way?
This sort of question—and not any professional expertise about Heisenberg and quantum indeterminacy—is, in my opinion, the crucial issue about the effectiveness of willing.
I will put the question in more general terms:
(i) the effectiveness of an external or internal input, in influencing the output from the brain, is, usually at least, greater when that input is consciously experienced than when it is not;
(ii) it is more likely to be consciously experienced when ‘attention’ is directed towards some context with which it is connected;
(iii) if, therefore, we can ‘will’ to concentrate attention on some particular context, we could control the effectiveness of the various inputs impinging on our mental activities at that moment.
I do not think the question is answerable in terms of existing science. The problem is one of the content of consciousness. You—or those of you who have not got bored—are conscious of my voice and what I am saying; but there are many other things you might be conscious of; you are unlucky if there is not a pretty girl, or a handsome boy, within a few yards, and you might be concentrating your attention on them. What directs your attention? That is the question. But how could science approach it? There is nothing in fundamental physics—the inscrutable Schrodinger equations—which opens even a possibility that a system of atoms might not merely react to its surroundings but consciously experience them. Nothing even in such more complex scientific entities as genes, or cells, opens any door out of the realm of ‘input stimuli and output reactions’, into the realm of conscious experience. Until we have something to say about what it is which differentiates a conscious experience from an unconscious reaction to a stimulus, we cannot throw any light on the basic question—what decides, and how, which of our mental and perceptual activities is enabled to ‘register’ as ‘information’ which plays its part in the deciding whether we slip off the watershed down into one chreodic valley or the other? But one thing is, I think, clear. In discussing free will, we are not inhabiting a world to which Newtonian or Laplacian determinism, or Quantum Indeterminacy, have any relevance. We are concerned with a chreodic world, in which outputs are largely independent of inputs; and we are asking questions about the distribution of something—Conscious Attention—which we can experience but cannot describe.
Waddington was rather defensive at the outset—he feared that he might be charged with reductionism but I thought the charge which was much more likely to be levelled against him was that of vitalism. The chreods are strongly reminiscent of the entelechies (another Greek word used in the last century by biologists) which were found by many biologists to be necessary and by many other scientists to be impossible, as a means of describing the phenomena of the organic living world. What I would like to extract from this part of Waddington's lecture as part of a programme for the whole of this course of lectures is to try and see what is true about vitalism and how this can be reconciled with what is true in mechanism—as it were, how the controversies of the last century can now be resolved at a higher plane. We know in part that the homeostatic mechanisms which the cyberneticians have invented and which we are fairly familiar with enable us to go some of the way, but it is fairly evident that Waddington does not want to say that his chreods are merely homeostats, otherwise he wouldn't have introduced a new word. For myself, I recognise some of what he is trying to put forward (although I am unhappy about some of the examples—the oranges, the Seville oranges and mandarins—this did not seem to be a very happy chreod, but no doubt Waddington will be able to explain it rather more fully). More generally what I think we have to recognise is that some sort of teleological explanation, explanations in terms of purposes (we ask ‘What is that organ for?’, we ask of certain patterns of animal behaviour ‘Why is the robin billing and cooing?’) is essential if we are to get an adequate understanding of biological phenomena and, a fortiori, of specifically human affairs. But we have to raise a sort of neo-Kantian question ‘How are teleological explanations possible?’ That is to say, ‘How are they compatible with the kind of explanation which Longuet-Higgins is always going to trot out his Schrodinger equations to prove as being absolutely basic and essential?’ And this is a question which we have not yet begun to answer.
I now want to turn to the latter part of Waddington's lecture where he was talking about freewill and consciousness, and where he was putting forward the suggestion that consciousness was perhaps dispensable from an elucidation of the problem of freewill, although noticeably not dispensable from his account of the universe as a whole, and I think we should be prepared to go quite a long way with Waddington. That is to say, what is crucial to whether a person is responsible for his actions or not, is not that he is very conscious of acting rightly: we are right to be rather suspicious of people who judge the moral worth of their actions from how hard they can persuade themselves that they at that time were resisting the temptations which beset them. Rather we should be prepared to count to a man's credit the unremembered acts that constitute the greater part of a good man's life and seize as the crucial way into the concept of responsibility that which the word evidently originally meant—‘respondeo; I answer’—answerability to the question ‘why did you do this?’ The person needn't at the time be answering the question; it is only the sort of answer he would have given. This goes right back to the Greeks—Socrates was very keen on ‘logon echein te kai didonai’: ‘to have and to be able to give reasons for what you are doing’, and the question which we need to consider is whether there is a place in our account of the mind, and possibly in our account of other phenomena, for this question or whether it is, as has often been suggested and as mechanism seems to suggest, no longer possible to ask this question and expect it to be answered in view of what we already know about the way that we are made.
On these two points we have yet much discussion—and for the moment I shall just give it back to Waddington—perhaps to befriend the oranges or perhaps to answer one of these deeper questions.
Well no, I think I will leave the deeper question to the others and perhaps, later, all of us. I just want to say about the idea that chreods are a vitalistic conception, that is really not so at all. The idea of a chreod is simply of the same kind as the idea of a homeostatic system, as used by the cyberneticians. Really biologists invented the idea of homeostasis long before electrical engineers got around to inventing the word ‘cybernetics’; it is a well-known notion that goes back to the late nineteenth century. An embryologist like myself enlarged this idea to include a time dimension, and a chreod is very like a homeostatic system, only it is going along a pathway in time. Most engineers, and most of the early biologists were only concerned with holding something steady, but I think this is inadequate for many things in biology. I don't think you can leave out that change in time and still have an adequately flexible notion. But the notion of a chreod, or a stabilised pathway or a canalised pathway, or whatever you like to call it, is not vitalistic at all but is just the same sort of thing as a homeostatic system.
Again I think we may come back to the question of teleology. All I will say at the moment is that the reason why most animals behave as they do is that they find it advantageous in competition with others, and natural selection has tailored them to do it.
Well, I want first of all to disown a view which is sufficiently different from my own for me to disagree with it, and that is ‘it's all in the Schrodinger equation’. If I ever said that I certainly didn't mean the Schrodinger equation is enough for us in answering interesting and important questions at higher levels. To say that it is all implicit in the Schrodinger equation is quite a different thing from saying that with the aid of the Schrodinger equation we can solve all problems in chemistry, biology, sociology, and economics. And I hoped I had made it clear last time that that was my general view.
Well, having said that, may I come to John Lucas’ question which I think is a fair question in relation to Waddington's paper. ‘How are reasons possible; how can we discuss giving reasons for doing something?’ I would have thought that that was not an insuperably difficult problem if one thinks about examining the output—if I may refer again to computer programs—examining the output of a computer program, finding a somewhat puzzling set of symbols, and then enquiring, as one can, ‘Why did it print these out?’ One answers the question by referring back to the program and seeing the various choices that were made at the various stages, according to various circumstances, and examining why the circumstances were that way at that time, and referring to the program to note that if x was equal to 2 at that time then y would have to be given the value of 3—or something like that. One can in fact discuss the reasons why something worked out the way it did in relation to computer programs. I wouldn't feel insulted if somebody said the reason why I used the word—some word that I might have just picked up from Waddington in the course of this conversation—was that I had just heard it from Waddington. I would not be insulted. I would think that was a perfectly good reason, a penetrating observation on what was going on in my mind.
One further point I would like to raise with Wad is something that he might well have said himself if he had had time. I very much like his idea of there being various valleys in this so-called epigenetic landscape, a valley being essentially one of these chreods down which the river flows, changing its course within limits—limits set by the banks of that valley. But I would like to reinforce his idea that ways of behaving are essentially chreods between which we have got to choose and that a very small influence from outside or inside may make all the difference. One might, for example, catch sight of something out of the corner of one's eye and say something entirely different from what one would otherwise have said. The immense sensitivity of the human being to minute perturbations from outside is enough of a reason for asserting that any attempt to predict people's behaviour in detail is doomed to failure; and of course the idea of doing so doesn't really make sense anyway when we come to discuss human affairs.
I'd like to make two points about Waddington's paper. Firstly I don't think that freewill is the kind of thing he thinks it is and secondly I don't think that consciousness is the kind of thing that he thinks it is.
Waddington believes that to have freewill is to have certain kinds of experiences—pretty unpleasant experiences to judge by his reference to the puritan tradition and those last few seconds before breasting the tape. Now I think that to have freewill is to be able to do things when you want to and to be able not to do them if you don't want to. It is a matter of a certain sort of capacity or ability or power; it isn't a matter of a particular experience. Sometimes we feel that we can do things when we want to and we are wrong, other times we can do things when we want to and we don't have any special feeling about it. Whether one feels one has freewill is quite a different question from whether one has it or not.
Now, secondly, I think that consciousness is not the private thing which is unamenable to scientific treatment that Waddington thinks it is. When I say that consciousness is not particularly private I don't mean to be propounding anything very difficult or metaphysical. I simply mean that I can tell by watching other people behave whether or not they are conscious, and I can tell by looking at this mug that it isn't conscious and I know that this mug is not conscious as certainly as I know anything. I was slightly disappointed that Waddington came to the conclusion that atoms and molecules were probably not conscious. I thought from his initial remarks about the sex-life of proteins that we were perhaps going to come to a conclusion, which he seemed to defend the other day, that consciousness is widely spread throughout the universe. It certainly seems to me that it is quite consistent with the view that consciousness is something private which we can never really make contact with except in our own case that this mug should be conscious. What on Waddington's view is wrong with the idea that, for all we know, this mug may be in excruciating pain? Of course we have to combine with this the supposition that the excruciatingness of its pain is only matched by the stoic quality of its fortitude. But it seems to me that on the view that Waddington has put forward there would be nothing nonsensical about that supposition. Against this, I think that consciousness, like freewill, is a matter of having certain sorts of ability. To be conscious is, for instance, to see and hear. Whether somebody can see or hear is a matter of whether he can discriminate between certain things, and whether he can discriminate between certain things is something that we can test both in simple everyday ways and in complicated experimental ways. Consciousness is therefore something which can be tested; it isn't something that is private and unamenable to scientific treatment.
I want to make a reply to some of the points just made by Kenny. Now, I am not a bit certain that the mug is not a little conscious. I said that I think you have to add to the definition of atoms something to do with consciousness, but I added that this consciousness is not going to be as highly evolved as ours; I feel that it would be somehow akin to it, some kind of beginnings of it, as much simplified compared to our consciousness as our structure is more complicated than that of an atom. I am definitely not ruling out that there is some sort of thing allied to consciousness all through the world.
Then this question of the privacy of consciousness. I think that one of the routes which may enable us to approach this is now becoming a fashionable experience, though not so much in scientific circles: namely, experimentation with what tend to be called novel forms of consciousness. If everybody's got the same sort of consciousness, there is not much you can say about it—except have you got it or haven't you got it? But if people start getting funny states of consciousness, by breathing exercises and yoga positions, or LSD if you like, and if they can communicate—which they do seem to be able to do, so I gather—to communicate something about these to other people, and in fact train people to get into specific states of consciousness, then I think this phenomenon is becoming more amenable to scientific study, I carefully said—I am always hedging my bets, I'm afraid—that consciousness was pretty private; but I'm not excluding that it may also be somewhat public.
On this point that freewill is the ability to act when you want to or to do what you like, I just think we differ on this. I may will to go to the Ritz, but if I haven't a five pound note in my pocket I can't do it; the ability to do something doesn't seem to me to be at all the same thing as consciously willing to do it. Again, you have to remember the terrific number of things you do when you are not consciously willing to do them; you do them as a spontaneous reaction to an input; or you want to answer the telephone and you just run down the stairs—a terrifically complex performance carried out by your cerebellum, but you aren't really conscious of doing it. I don't consider that as something that poses the problem of freewill. That merely poses the problem of the mechanisms of the wiring circuits in your cerebellum, which is quite a different kettle of fish.