I Want to talk about a strong sense of the terms ‘predict’, ‘prediction’, ‘predictability’, and so on; to take up the question whether a social science may aspire to predict in the strong sense; and to consider some bearings of this question on the free-will problem.

Moore in Ethics (1912), ch. vii, writes – if I understand him right – that our basic experience of free-will resides in our certain feeling, in regard to our past actions, that we could have acted differently if we had so chosen; more exactly, that we should have acted differently if we had so chosen, which precisely means that we could have acted differently. He adds the qualifying adverb ‘sometimes’; I think we may well say always, keeping in mind, of course, (i) that we are only concerned here with ‘actions’ proper, as contrasted with involuntary movements, twitchings, starts, fits, etc., and also with wholly habitual, routine-like movements or manipulations performed without any attention and without the slightest deliberation, and (ii) that only such actions are meant here as are within our range of physical or psycho-physical power, which is anyhow implicit in the concept of choice: there is no sense in my saying that I might here and now ‘choose’ to carry this building on my back to Paris, to dismiss the present Government or to continue writing in impeccable Japanese. However, Moore is aware of the clearly meaningful objection that ‘acting as we choose’ (within those obvious limits) would not establish the fact of free-will unless it were also the case that our choice itself is (often) free, in other words that we not only choose to act so instead of acting differently but also choose to choose so instead of choosing differently. He argues that this too is indeed (often) the case, emphasising in this context our experience of the unpredictability, even by ourselves, of our future choices (or of many or most of them). He concludes, though, on a note of doubt which shows that he is not quite sure in his own mind to have definitively disposed of the objection.

I Want to examine how far the question ‘What is it for a man to act morally?’ can be answered in terms of the sociological concept of a role. Is, for example, acting as a moral agent consistent with acting in the capacity of one's role, or even identical with it if being a moral agent is acting a role? In the first part of my paper I shall examine some of the relations between morality and roles, especially from the point of view of that freedom which is always regarded as a necessary condition of a man's being a moral agent. In the second part I shall examine the concept of a role itself, not from the point of view of sociology to which it notionally belongs, but from the point of view of the philosophy of mind; this is the point of connection between my own theme and the general theme of this series of lectures. The role concept will thus be a bridging concept between ethics and the philosphy of mind. Finally I shall bring to bear on the idea of the moral agent any insights that may be yielded by the psychology of role-acting.

There are various ways in which a person's freedom of action may be diminished or restricted. Instead of acting, he may be acted upon; or he may be unable to help doing what he does. He may suffer from some disability, have a duty imposed on him or do something because he is obliged to do it. In this essay, I wish to examine the notion of being obliged to do something. I shall investigate the differences between ‘being obliged by A to do X’ and ‘being obliged to A for X’; between ‘being obliged to do X’ and ‘having (or being under) an obligation to do X’; between being physically obliged and being morally, legally and logically obliged; and between saying ‘X is something I am obliged to do’, ‘X is something I choose to do’ and ‘X is something I ought to do’.

It is not easy to explain how people know what they are going to do. The phenomenon occurs: obviously we often do know what we are going to do; but its explanation is less obvious. When I say this, I do not mean that we have some mysterious method by which we discover what we are going to do, like forecasting tomorrow's weather. Usually we know without any investigation, and without the use of any method of discovery. You know some of the things that you are going to do tomorrow immediately – i.e. you know them without the mediation of any evidence. This is a rather mysterious phenomenon, and it is not at all easy to classify it. How is it possible for there to be any immediate knowledge of something which lies in the future? Is it like the kind of immediate knowledge of the past which is supplied by memory? If so, does this analogy help us to understand our knowledge of our own future actions? Or is memory equally mysterious?

The question I shall consider is whether there are any mental causes, that is, whether there is anything which is both a state of mind and a cause of other mental or physical happenings. The obvious common-sense answer to this question is Yes. In ordinary discourse (that is, outside philosophy and psychology) we constantly refer to human actions and experiences in what appears to be causal language; we seem to be saying that some states of mind are causes of other states of mind, and of some bodily activities. Sometimes we do this by using the word ‘cause’ itself— ‘His driving into the ditch was caused by his seeing a child run into the road’, ‘The cause of his silence was his wish to protect his friend’. More often we use expressions which in physical contexts are admittedly causal— ‘A glimpse of the look on her face made him hesitate’, ‘Ambition was the driving force of his career’. Most philosophers have taken for granted the genuineness of these causal attributions. But recently it has been frequently denied that any state of mind can be properly described as the cause either of other states of mind or of physical occurrences, at any rate if by a state of mind one means a state of consciousness. Some philosophers who deny this are materialists, advocates of the Unity of Science. They believe that everything that happens in the universe can be causally accounted for by reference to a single set of fundamental laws, the laws of physics, and since a state of consciousness is not, as such, a physical state of any kind, it cannot have a place in the causal explanation of any event, and must be an epiphenomenon. Others are believers in an indeterminist interpretation of human freedom, and hold that o t introduce the category of cause and effect into the explanation of human actions is to deny that there is any genuine free will. I shall not be concerned with either of these general metaphysical objections, neither of which convinces me; but I shall consider some more specific objections to the admission of particular sorts of mental causes.

Time's arrow is not a property of time but of events: the way in which they succeed one another in our experience and, as many believe, in a reality independent of our experience. I hope to throw a little light on one aspect of this difficult matter by treating it from the standpoint of logic, topology and information theory. If I succeed in my hope I shall still be leaving many other matters unresolved. Let me state briefly what these other matters are in order not to exaggerate the scope of any contribution I hope to make.

One of the theories defined in Baldwin's Dictionary of Philosophy and Psychology, published in 1901, is ‘The Double Aspect Theory’. It is ‘the theory of the relation of mind and body, which teaches that mental and bodily facts are parallel manifestations of a single underlying reality’. It ‘professes to overcome the onesidedness of materialism and idealism by regarding both series as only different aspects of the same reality, like the convex and the concave views of a curve; or, according to another favourite metaphor, the bodily and the mental facts are really the same facts expressed in different language’.

Life Science Library now claims to examine ‘the most complex of all biological organs: the human mind’, and scientists quite commonly make no distinction between mind and brain — they delight in talking about the brain classifying, decoding, perceiving, deciding or giving orders. And while resisting the conceptual muddle involved in talking of the brain doing what persons do, the identity hypothesis tries to provide a philosophically respectable basis for the equation of mind and brain, maintaining that ‘mind’ is just a term for a group of activities and dispositions, and that these in turn are in some sense to be identified with brain activities or traces. On the other hand, from the point of view of religion and traditional philosophy the suggestion is completely unplausible — creative or inventive thought, and aesthetic, moral or religious experiences seem so far removed from mechanical or physiological processes that a good deal of softening up is necessary if any kind of identity theory is to get a fair hearing. This softening up is best carried out by considering the difficulties in the main rival philosophical view, interactionism.

In recent years philosophers have given a good deal of attention to imperatives. They have concerned themselves mainly with the logical grammar of sentences of this kind, that is to say their relations to each other and to interrogative and indicative sentences. Very often this topic has been raised in terms of the problem ‘Is imperative inference possible, and if so, what kind of inference is it?’. Many philosophers have contended that there are logically valid inferences that involve imperative sentences. Against this it has been argued that no such inferences are possible. It has even been held that there are no such things as imperatives at all – regarded, that is, as types of expression logically sui generis and independent of indicative sentences.

In a broadcast talk delivered in 1956, the late J. L. Austin began by outlining to his listeners his now well-known concept of ‘the performative utterance and its infelicities’; and at the end of that first section of his talk he made this comment: ‘That equips us, we may suppose, with two shining new tools to crack the crib of reality maybe. It also equips us – it always does – with two shining new skids under our metaphysical feet’. In this talk I intend to illustrate a particular respect in which, in moral philosophy, the partial pessimism of Austin's comment has proved abundantly justified. I shall try to show how in this field one shining new tool has led in fact to the skidding of moral philosophers' feet – how one bright idea (an idea, as it happens, very closely akin to that which Austin himself was discussing) has led some influential theorists off in the wrong direction, and the rest of us back in the end, with rather little gained, to a position not far from that of our predecessors of about a hundred years ago. This dismal story, I should in justice make clear at the outset, is not that of the whole of moral philosophy, not even of moral philosophy in English; I shall be tracing only one out of several concurrent lines of thought, but the line which I shall trace will be, I think, readily recognised as having been, and perhaps as still being, more conspicuous than most.

Just as there was a vogue at one time for identifying thinking either with mere processions or with more or less organised processions of images, so there is a vogue now for identifying thinking with something oddly called ‘language’, namely with more or less organised processions of bits of French or English, etc.

Whether the word ‘passion’, as indicating the suffering or affection from without of a soul, is by now no more than a dead metaphor, surviving from an antique conception of the mind; whether, indeed, there is any way open to us of determining the passivity or otherwise of our inner life, apart, that is, from how it strikes us, from how we are prompted to describe it, are not questions that I can take up this evening. It is enough for my purpose that for much of the time our feelings, our emotions, our inclinations are as fluctuating or as imperious as if they were not totally under our control. We are elated: we are dejected: we get angry, and then our anger gives place to a feeling of absurdity: we remain in love with someone who is lost to us but whom we cannot renounce: we are interested in something, and suddenly we are bored, or frightened that we will be bored: we see a stranger, someone who is nothing to us, who is poor or crippled, and we feel guilt: someone does something wrong or foolish, and we are unaccountably transported by laughter, by ‘sudden glory’ as Hobbes called it, knowing what it was about, and then, as unaccountably, we are thrown down. Man is, in Montaigne's famous phrase, une chose ondoyante, a creature of inner change and fickleness.