The Philosophical Review, Vol. 66, No. 4 (Oct., 1957), pp. 439-450

THE idea that free will can be reconciled with the strictest determinism is now very widely accepted. To say that a man acted freely is, it is often suggested, to say that he was not constrained, or that he could have done otherwise if he had chosen, or something else of that kind; and since these things could be true even if his action was determined it seems that there could be room for free will even within a universe completely subject to causal laws. Hume put forward a view of this kind in contrasting the "liberty of spontaneity ... which is oppos'd to violence" with the nonexistent "liberty of indifference . . . which means a negation of necessity and causes."' A.J. Ayer, in his essay "Freedom and Necessity"2 was summing up such a position when he said, "from the fact that my action is causally determined ... it does not necessarily follow that I am not free"3 and "it is not when my action has any cause at all, but only when it has a special sort of cause, that it is reckoned not to be free."4 I am not here concerned with the merits of this view but only with a theory which appears more or less incidentally in the writings of those who defend it. This is the argument that so far from being incompatible with determinism, free will actually requires it. It appears briefly in Hume's Treatise and was set out in full in an article by R. E. Hobart.5 P. H. Nowell-Smith was expressing a similar view when he said of the idea that determin- ism is opposed to free will that "the clearest proof that it is mistaken or at least muddled lies in showing that I could not be free to choose what I do unless determinism is correct. . . . Free- dom, so far from being incompatible with causality implies it."6 Ayer has taken up a similar position, arguing that the absence 1 Treatise, bk. II, pt. III, sec. 2. 2 Polemic, no. 5 (I946); reprinted in his Philosophical Essays (London, I954). 3 Philosophical Essays, p. 278. 4 Ibid., p. 28I. 5 "Freewill as Involving Determinism," Mind, XLIII (1934), I-27. 6 "Freewill and Moral Responsibility," Mind, LVII (I948), 46. 439 I PHILIPPA FO O T of causal laws governing action "does not give the moralist what he wants. For he is anxious to show that men are capable of acting freely in order to infer that they can be morally re- sponsible for what they do. But if it is a matter of pure chance that a man should act in one way rather than another, he may be free but he can hardly be responsible."7 This argument is not essential to the main thesis of those who use it; their own account of free will in such terms as the absence of constraining causes might be correct even though there were no inconsistencies in the suggestion put forward by their libertar- ian opponents. But if valid the argument would be a strong argument, disposing of the position of anyone who argued both that free will required the absence of determining causes and that free will was a possibility. That the argument is not valid, and indeed that it is singularly implausible, I shall now try to show. It is, I think, surprising that it should have survived so long; this is perhaps because it has not had to bear much weight. In any case the weapons which can be used against it are ones which are in general use elsewhere. In discussing determinism and free will it is important to be clear about the sense which is given in this context to words such as "determined" and "caused." Russell gave this account: The law of universal causation . . . may be enunciated as follows: There are such invariable relations between different events at the same or different times that, given the state of the whole universe throughout any finite time, however short, every previous and subse- quent event can theoretically be determined as a function of the given events during that time.8 This seems to be the kind of determinism which worries the defender of free will, for if human action is subject to a universal law of causation of this type, there will be for any action a set of sufficient conditions which can be traced back to factors outside the control of the agent. We cannot of course take it for granted that whenever the word "determined" or the word "cause" is used this is what 7Philosophical Essays, p. 275. 8 "On the Notion of Cause," in Our Knowledge of the External World (London, I914), p. 22I. 440 FREE WILL AND DETERMINISM is implied, and what is intended may be in no way relevant to the question of free will. For instance, an action said to be determined by the desires of the man who does it is not necessarily an action for which there is supposed to be a sufficient condition. In saying that it is determined by his desires we may mean merely that he is doing something that he wants to do, or that he is doing it for the sake of something else that he wants. There is nothing in this to suggest determinism in Russell's sense. On the whole it is wise to be suspicious of expressions such as "determined by desire" unless these have been given a clear sense, and this is particularly true of the phrase "determined by the agent's character." Philosophers often talk about actions being deter- mined by a man's character, but it is not certain that anyone else does, or that the words are given any definite sense. One might suppose that an action was so determined if it was in character, for instance the generous action of a generous man; but if this is so we will not have the kind of determinism traditionally supposed to raise difficulties for a doctrine of free will. For nothing has been said to suggest that where the character trait can be predicated the action will invariably follow; it has not been supposed that a man who can truly be said to be generous never acts ungenerously even under given conditions. Keeping the relevant sense of "determinism" in mind, we may now start to discuss the view that free will requires determinism. The first version which I shall consider is that put forward by Hobart, who suggests that an action which is not determined cannot properly be called an action at all, being something that happened to the agent rather than something he did. Hobart says, "In proportion as it [the action] is undetermined, it is just as if his legs should suddenly spring up and carry him off where he did not prefer to go." To see how odd this suggestion is we have only to ask when we would say that a man's legs were carrying him where he did not prefer to go. One can imagine the scene: he is sitting quietly in his chair and has said that he is going to go on reading his book; suddenly he cries, "Good heavens, I can't control my legs!" and as he moves across the room, he hangs on to the furniture or asks someone else to hold him. Here indeed his legs are carrying him where he does not 44I PHILIPPA FOOT want to go, but what has this to do with indeterminism, and what has the ordinary case, where he walks across the room, to do with determinism? Perhaps Hobart thinks that when a man does something meaning to do it, he does what he wants to do, and so his action is determined by his desire. But to do something meaning to do it is to do it in a certain way, not to do it as the result of the operation of a causal law. When one means to do something, one does not call out for help in preventing the movement of one's limbs; on the contrary, one is likely to stop other people from interfering, saying, "I want to do this." It is by such factors that walking across the room is distinguished from being carried off by one's legs. It is to be explained in terms of the things said and done by the agent, not in terms of some force, "the desire," present before the action was done and still less in terms of some law telling us that whenever this "desire" is found it will be followed by the action. The indeterminist has no difficulty in distinguishing an action from something that happens to the agent; he can say exactly the same as anyone else. Nowell-Smith seems to be thinking along somewhat the same lines as Hobart when he attacks C. A. Campbell for saying that free will requires indeterminism: The essence of Campbell's account is that the action should not be predictable from a knowledge of the agent's character. But, if this is so, can what he does be called his action at all? Is it not rather a lusus naturae, an Act of God or a miracle? If a hardened criminal, bent on robbing the poor-box, suddenly and inexplicably fails to do so, we should not say that he chose to resist or deserves credit for resisting the tempta- tion; we should say, if we were religious, that he was the recipient of a sudden outpouring of Divine Grace or, if we were irreligious, that his "action" was due to chance, which is another way of saying that it was inexplicable. In either case we should refuse to use the active voice.9 It is hard to see why a man who does something inexplicably does not really do it. Let us suppose that the hardened criminal's action really is inexplicable; we can only say, "He just turned away," and not why he did so; this does not mean that he did it by accident, or unintentionally, or not of his own free will, and 9 Ethics (London, I954), pp. 28i-282. 442 FREE WILL AND DETERMINISM I see no reason for refusing to use the active voice. In any case, to explain an action is not necessarily to show that it could have been predicted from some fact about the agent's character-that he is weak, greedy, sentimental, and so forth. We may if we like say that an action is neverfully explained unless it has been shown to be covered by a law which connects it to such a character trait; but then it becomes even more implausible to say that an action must be explicable if we are to admit it as something genuinely done. In the ordinary sense we explain the criminal's action if we say, for instance, that a particular thought came into his mind; we do not also have to find a law about the way such thoughts do come into the minds of such men. A rather different version of this argument runs as follows. We hold responsible only a man who is a rational agent; if someone were always to do things out of the blue, without having any reason to do them, we should be inclined to count him as a lunatic, one who could not be held responsible for his actions, so that even if he did things he would do things for which he could not be held responsible. And is it not through being determined by motives that actions are those of a rational agent whom we can praise or blame? It certainly would be odd to suppose that free will required the absence of motives for action. We do not of course expect that everything that the rational man does should be done with a motive; if he gets up and walks about the room he need not be doing so in order to take exercise; it is quite usual for people to do this kind of thing without any particular purpose in view, and no one is counted irrational for doing so. And yet we do expect a man to have a motive for a great number of the things that he does, and we would count anyone who constantly per- formed troublesome actions without a motive as irrational. So it looks as if a moral agent is a man whose actions are in general determined, if determinism is involved in "having a motive" for what he does. What does it mean to say that someone had a motive for doing what he did? Often this particular expression means that he did it with a particular intention, so we should first say something about intentions and the sense in which they can be said to 443 PHILIPPA FOOT determine action. We say that a man had a certain intention in acting when he aimed at a certain thing, and "his motive for such and such" often means "his aim in doing such and such," for instance, "His motive for going to the station was to take a train to London." But where motives are intentions it is clear that they cannot be determining causes; for intending to do x and being ready to take the steps thought necessary to do x are connected not empirically but analytically. A man cannot be said to have an intention unless he is reconciled to what he believes to be the intermediate steps. We cannot speak as if the intention were something which could be determined first, and "being ready to take the necessary steps" were a second stage following on the first. It might be objected that this does not cover the case of "doing y because one wants x" where "wanting x" does not imply trying to get x. In one sense of "want" it is possible to say, "He wants x" without knowing whether he is prepared to take steps to get it. (One might, for instance, want to go to London but not be prepared to spend the money to get there.) So that wanting seems here to be a separate condition, which might in certain cases be connected by an empirical law to the adoption of particular courses of action. Certainly wanting is not an event, but one gets rid of wanting as a determining factor too easily if one merely says that desires are not causes because they are not occurrences. We say "He wants" in this sense where he would adopt certain policies if there were no reasons for not doing so. We can say, "He wants to get to London," even when he is not prepared to take the necessary steps to get to London, provided he can say, "Trains are too expensive," or "Hitchhiking is too uncom- fortable." If we offered him a spare railway ticket or otherwise disposed of his reasons against going, and he still did not go, we would have to say, "He didn't really want to go after all." So wanting in this sense is being prepared to act under certain conditions, though not being prepared to act under the given conditions. It is a description which could be applied to a man before we knew whether he was ready to act in a given situation, and it seems that there might then be a causal relation between 444 FREE WILL AND DETERMINISM the wanting and the acting where the latter took place. This is quite true; there could be a law to the effect that when the de- scription "He wants x" applied at t1, the description "He is taking the necessary steps to get x" applied at t2. It would be possible to say this without making a mistake about what it is to want and inventing a hidden condition of body or mind. One could say, "Wanting in this sense just is being prepared to act under some conditions," and still maintain that there could be an empirical law connecting wanting with acting under a particular set of conditions. The mistake lies not in the idea that such laws are possible but in the thought that there is a reference to them in the statement that a man did one thing because he wanted some- thing else. So far we have been dealing only with cases in which a question about a motive was answered by specifying something aimed at or wanted. Now we should turn to the cases in which the motive is said to be kindness, vanity, ambition, meanness, jealousy, and so on, to see whether determinism is involved. It is easy to show that a motive is not a cause in Russell's sense, for it is clearly not an antecedent cause. Professor Gilbert Ryle has pointed out that a man who acts out of vanity is not a man who had a feeling of vanity immediately before he acted, and if it is objected that the vanity which preceded the action need not have manifested itself in a feeling, one may ask what else would count as the vanity which was causing him to act. A man's motives are not given by what was happening to him immediately before he started to act. Nor do we discover some independent condition contemporaneous with the action and a law linking the two, for again there is nothing which would count as vanity except the tendency to do this kind of thing. So much is implied in what Ryle says about acting out of vanity, but his own account of what it is to do so still uses something which is objectionably like a causal model. The analogy which he thinks apt is that between saying a man acted out of vanity and saying a piece of glass broke because it was brittle: "To explain an act as done from a certain motive is not analogous to saying that the glass broke because a stone hit it, but to the quite different type of statement that the glass broke, when the stone 445 PHILIPPA FOOT hit it, because the glass was brittle."10 The positive part of this statement seems to me mistaken. Acting out of vanity is not so closely connected with being vain as Ryle must suppose it to be. Let us suppose that his account of what it is to be vain is entirely correct; to say that a man is vain is to say that he tends to behave in certain ways, to feel aggrieved in particular situations, and so on.1" It does not follow that ascribing vanity as a motive for an action is bringing this action under the "lawlike" proposition that the agent is a man who tends to do these things. For it makes sense to say that a man acts out of vanity on a particular occasion although he is not in general vain, or even vain about this kind of thing. It cannot therefore be true that when we speak of an agent's motive for a particular action we are explaining it in terms of his character, as Ryle suggests; we are not saying "he would do that." It is of course possible to give a motive and to say that the agent has the character trait concerned, but the latter cannot be included in an account of what it is to assign a motive to a particular action. The explanation of why Ryle says what he does seems to lie in the fact that he has taken a false example of explaining an action by giving a motive. He considers as his example the expla- nation, "He boasted because he is vain," which is not in fact an explanation of the right type; considered as a statement assigning a motive to a particular action it would be uninformative, for except in very special cases boasting is acting out of vanity. It is not surprising that this particular sentence has a different function-that of relating this act of vanity to the character trait. What Ryle says about the example is correct, but it is not an example of the kind of thing he is trying to describe. It might seem as if we could reformulate the theory to meet the objection about the man who acts out of vanity on one occasion by saying that a man's acting out of vanity is like glass breaking because of a brittleness which could be temporary. "He acted out of vanity" would then be explained as meaning that at that particular time he tended to react in the ways described by 10 Concept of Mind (London, I949), pp. 86-87. 11 Ibid., p. 86. 446 FREE WILL AND DETERMINISM Ryle. (If he finds a chance of securing the admiration and envy of others, he does whatever he thinks will produce this admira- tion and envy.) This is wrong because, whereas glass which is even temporarily brittle has all the reactions which go by this name, a man who is temporarily acting out of vanity is not liable to do other things of this kind. To find concepts which this model would fit one must turn to such descriptions as "a boastful mood," "a savage frame of mind," or "a fit of bad temper." Assigning a motive to an action is not bringing it under any law; it is rather saying something about the kind of action it was, the direction in which it was tending, or what it was done as. A possible comparision would be with the explanation of a movement in a dance which consisted in saying what was being danced. Often in diagnosing motives we should look to purposes- to what the action was done for. This we should discover if we found out what the agent was prepared to go without and what he insisted on having; the fact that visitors are made to admire a garden even in the rain is strong evidence that they were invited out of vanity rather than kindness. In other cases finding the motive will be better described as finding what was being done-finding, for instance, that someone was taking revenge. We should take it that a man's motive was revenge if we dis- covered that he was intentionally harming someone and that his doing so was conditional on his believing that that person had injured him. In the same way we should take it that someone was acting out of gratitude if he (i) intended to confer a benefit and (2) saw this as called for by a past kindness. The fact that it is only the character of the particular action which is involved shows how far we are from anything which could involve motives as determining causes. We have now considered two suggestions: (i) that an unde- termined action would not be one which could properly be attributed to an agent as something that he did and (2) that an undetermined action would not be the action of a rational agent. A third version, the one put forward by Hume, suggests that an undetermined action would be one for which it would be impos- sible to praise or blame, punish or reward a man, because it would be connected with nothing permanent in his nature. 447 PHILIPPA FO O T 'Tis only [Hume says] upon the principles of necessity, that a person acquires any merit or demerit from his actions. . . . Actions are by their very nature temporary and perishing; and where they proceed not from some cause in the characters and disposition of the person, who perform'd them, they infix not themselves upon him, and can neither redound to his honour, if good, nor infamy, if evil. The action in itself may be blameable.... But the person is not responsible for it; and as it proceeded from nothing in him, that is durable and constant, and leaves nothing of that nature behind it, 'tis impossible he can, upon its account, become the object of punishment or vengeance.'2 Hume is surely wrong in saying that we could not praise or blame, punish or reward, a person in whose character there was nothing "permanent or durable." As he was the first to point out, we do not need any unchanging element in order to say that a person is the same person throughout a period of time, and our concept of merit is framed to fit our concept of personal identity. We honor people as well as nations for what they have done in the past and do not consider what has been done merely as an indication of what may be expected in the future. Moreover, it is perfectly rational to punish people for what they have done, even if there is no reason to think that they would be likely to do it again. The argument that it will be a different me who will be beaten tomorrow carries no weight, for "different" or not the back which will be beaten is the one about which I am concerned today. So we have no reason to invent something durable and constant underlying the actions which we punish or reward. And it is not in fact our practice to pick out for praise or blame only those actions for which something of the kind can be found. It would be possible, of course, that we should do this, punishing the cruel action of the cruel man but not that of one usually kind. But even in such a situation there would be no argument against the man who said that moral responsibility depended upon indeterminism; for a motive is not a determining cause, nor is an habitual motive. If we say that a man constantly acts out of cruelty, we no more say that his actions are determined than if we say that he acts out of cruelty on a particular occasion. There could of course be a law to the effect that no one who has 12 Treatise, bk. II, pt. III, sec. 2. 448 FREE WILL AND DETERMINISM been cruel for thirty years can turn to kindness after that, and this would throw responsibility back from the later to the earlier acts. But it is clear that this is a special assumption in no way involved in the statement that cruelty is a "durable and constant" element in someone's character. I have already mentioned Ayer's argument that moral re- sponsibility cannot be defended on the basis of indeterminism and will now consider his version in detail. Ayer says that the absence of a cause will not give the moralist what he wants, because "if it is a matter of pure chance that a man should act in one way rather than another, he may be free but he can hardly be re- sponsible."13 To the suggestion that "my actions are the result of my own free choice," Ayer will reply with a question about how I came to make my choice: Either it is an accident that I choose to act as I do or it is not. If it is an accident, then it is merely a matter of chance that I did not choose otherwise; and if it is merely a matter of chance that I did not choose otherwise, it is surely irrational to hold me morally responsible for choosing as I did. But if it is not an accident that I chose to do one thing rather than another, then presumably there is some causal explanation of my choice: and in that case we are led back to deter- minism.14 The "presumably" seems to be the weak link in the argument, which assumes a straightforward opposition between causality and chance that does not in general exist. It is not at all clear that when actions or choices are called "chance" or "accidental" this has anything to do with the absence of causes, and if it has not we will not be saying that they are in the ordinary sense a matter of chance if we say that they are undetermined. When should we say that it was a matter of chance that a man did what he did? A typical example would be the case in which a man killed someone with a bullet which glanced off some object in a totally unforseeable way; here he could disclaim responsibility for the act. But in this instance, and that of something done "by accident," we are dealing with what is done unintentionally, and this is not the case which Ayer has in mind. We may turn, 13 Philosophical Essays, p. 275. 14 Ibid. 449 FREE WILL AND DETERMINISM as he does, to the actions which could be said to have been "chosen" and ask how the words "chance" and "accident" apply to choices. Ayer says, "Either it is an accident that I choose to act as I do, or it is not." The notion of choosing by accident to do something is on the face of it puzzling; for usually choosing to do something is opposed to doing it by accident. What does it mean to say that the choice itself was accidental? The only application I can think of for the words "I chose by acci- dent" is in a case such as the following. I choose a firm with which to have dealings without meaning to pick on one run by an international crook. I can now rebut the charge of choosing a firm run by an international crook by saying that I chose it by accident. I cannot be held responsible for this but only for any carelessness which may have been involved. But this is because the relevant action-the one with which I am being charged-was uninten- tional; it is for this reason and not because my action was un- caused that I can rebut the charge. Nothing is said about my action being uncaused, and if it were, this could not be argued on my behalf; the absence of causes would not give me the same right to make the excuse. Nor does it make any difference if we substitute "chance" for "accident." If I say that it was a matter of chance that I chose to do something, I rebut the suggestion that I chose it for this reason or for that, and this can be a plea against an accusation which has to do with my reasons. But I do not imply that there was no reason for my doing what I did, and I say nothing what- soever about my choice being undetermined. If we use "chance" and "accident" as Ayer wants to use them, to signify the absence of causes, we shall have moved over to a totally different sense of the words, and "I chose it by chance" can no longer be used to disclaim responsibility. PHILIPPA FOOT Somerville College, Oxford 450