The importance to ethics of the free-will question is a subject upon which there has existed almost as much diversity of opinion as on the free-will question itself. It has been urged by advocates of free-will that its denial involves the denial of merit and demerit, and that, with the denial of these, ethics collapses. It has been urged on the other side that, unless we can forsee, at least partially, the consequences of our actions, it is impossible to know what course we ought to take under any given circumstances; and that if other people's actions cannot be in any degree predicted, the foresight required for rational action becomes impossible. I do not propose, in the following discussion, to go into the free-will controversy itself. The grounds in favour of determinism appear to me overwhelming, and I shall content myself with a brief indication of these grounds. The question I am concerned with is not the freewill question itself, but the question how, if at all, morals are affected by assuming determinism. (§ 24 ¶ 1)

In considering this question, as in most of the other problems of ethics, the moralist who has not had a philosophical training appears to me to go astray, and become involved in needless complications, through supposing that right and wrong in conduct are the ultimate conceptions of ethics, rather than good and bad, in the effects of conduct and in other things. The words good and bad are used both for the sort of conduct which is right or wrong, and for the sort of effects to be expected from right and wrong conduct, respectively. We speak of a good picture, a good dinner, and so on, as well as of a good action. But there is a great difference between these two meanings of good. Roughly speaking, a good action is one of which the probable effects are good in the other sense. It is confusing to have two meanings for one word, and we therefore agreed in the previous section to speak of a right action rather than a good action. In order to decide whether an action is right, it is necessary, as we have seen, to consider its probable effects. If the probable effects are, on the whole, better than those of any other action which is possible under the circumstances, then the action is right. The things that are good are things which, on their own account, and apart from any consideration of their effects, we ought to wish to see in existence: they are such things as, we may suppose, might make the world appear to the Creator worth creating. I do not wish to deny that right conduct is among the things that are good on their own account; but if it is so, it depends for its intrinsic goodness upon the goodness of those other things which it aims at producing, such as love or happiness. Thus the rightness of conduct is not the fundamental conception upon which ethics is built up. This fundamental conception is intrinsic goodness or badness. (§ 24 ¶ 2)

As the outcome of our discussions in the previous section, I shall assume the following definitions. The objectively right action, in any circumstances, is that action which, of all that are possible, gives us, when account is taken of all available data, the greatest expectation of probable good effects, or the least expectation of probable bad effects. The subjectively right or moral action is that one which will be judged by the agent to be objectively right if he devotes to the question an appropriate amount of candid thought, or, in the case of actions that ought to be impulsive, a small amount. The appropriate amount of thought depends on the importance ofthe action and the difficulty of the decision. An act is neither moral nor immoral when it is unimportant, and a small amount of reflection would not suffice to show whether it was right or wrong. After these preliminaries, we can pass to the consideration of our main topic. (§ 24 ¶ 3)

The principle of causality—that every event is determined by previous events, and can (theoretically) be predicted when enough previous events are known—appears to apply just as much to human actions as to other events. It cannot be said that its application to human actions, or to any other phenomena, is wholly beyond doubt; but a doubt extending to the principle of causality must be so fundamental as to involve all science, all everyday knowledge, and everything, or almost everything, that we believe about the actual world. If causality is doubted, morals collapse, since a right action, as we have seen, is one of which the probable effects are the best possible, so that estimates of right and wrong necessarily presuppose that our actions can have effects, and therefore that the law of causality holds. In favour of the view that human actions alone are not the effects of causes, there appears to be no ground whatever except the sense of spontaneity. But the sense of spontaneity only affirms that we can do as we choose, and choose as we please, which no determinist denies; it cannot affirm that our choice is independent of all motives, and indeed introspection tends rather to show the opposite. It is said by the advocates of free-will that determinism destroys morals, since it shows that all our actions are inevitable, and that therefore they deserve neither praise nor blame. Let us consider how far, if at all, this is the case. (§ 25 ¶ 1)

The part of ethics which is concerned, not with conduct, but with the meaning of good and bad, and the things that are intrinsically good and bad, is plainly quite independent of freewill. Causality belongs to the description ofthe existing world, and we saw that no inference can be drawn from what exists to what is good. Whether, then, causality holds always, sometimes, or never is a question wholly irrelevant in the consideration of intrinsic goods and evils. But when we come to conduct and the notion of ought, we cannot be sure that determinism makes no difference. For we saw that the objectively right action may be defined as that one which, of all that are possible under the circumstances, will probably on the whole have the best consequences. The action which is objectively right must therefore be in some sense possible. But if determinism is true, there is a sense in which no action is possible except the one actually performed. Hence, if the two senses of possibility are the same, the action actually performed is always objectively right; for it is the only possible action, and therefore there is no other possible action which would have better results. There is here, I think, a real difficulty. But let us consider the various kinds of possibility which may be meant. (§ 26 ¶ 1)

In order that an act may be a possible act, it must be physically possible to perform, it must be possible to think of, and it must be possible to choose if we think of it. Physical possibility, to begin with, is obviously necessary. There are circumstances under which I might do a great deal of good by running from Oxford to London in five minutes. But I should not be called unwise, or guilty of an objectively wrong act, for omitting to do so. We may define an act as physically possible when it will occur if I will it. Acts for which this condition fails are not to be taken account of in estimating rightness or wrongness. (§ 26 ¶ 2)

To judge whether an act is possible to think of is more difficult, but we certainly take account of it in judging what a man ought to do. There is no physical impossibility about employing one’s spare moments in writing lyric poems better than any yet written, and this would certainly be a more useful employment than most people find for their spare moments. But we do not blame people for not writing lyric poems unless, like FitzGerald, they are people that we feel could have written them. And not only do we not blame them, but we feel that their action may be objectively as well as subjectively right. if it is the wisest that they could have thought of. But what they could have thought of is not the same as what they did think of. Suppose a man in a fire or a shipwreck becomes so panic-stricken that he never for a moment thinks of the help that is due to other people, we do not on that account hold that he does right in only thinking of himself. Hence in some sense (though it is not quite clear what this sense is), some of the courses of action which a man does not think of are regarded as possible for him to think of, though others are admittedly impossible. (§ 27 ¶ 1)

There is thus a sense in which it must be possible to think of an action, if we are to hold that it is objectively wrong not to perform the action. There is also, if determinism is true, a sense in which it is not possible to think of any action except those which we do think of. But it is questionable whether these two senses of possibility are the same. A man who inds that his house is on fire may run out of it in a panic without thinking of warning the other inmates; but we feel, rightly or wrongly, that it was possible for him to think of warning them in a sense in which it is not possible or a prosaic person to think of a lyric poem. It may be that we are wrong in feeling this difference, and that what really distinguishes the two cases is dependence on past decisions. That is to say, we may recognize that no different choice among alternatives thought of at any time would have turned an ordinary man into a good lyric poet; but that most men, by suitably choosing among alternatives actually thought of, can acquire the sort of character which will lead them to remember their neighbours in a fire. And if a man engages in some useful occupation of which a natural effect is to destroy his nerve, we may conceivably hold that this excuses his panic in an emergency. In such a point, it would seem that our judgment may really be dependent on the view we take as to the existence of freewill; for the believer in freewill cannot allow any such excuse. (§ 27 ¶ 2)

If we try to state the difference we feel between the case of the lyric poems and the case of the fire, it seems to come to this: that we do not hold an act objectively wrong when it would have required that we recognize as a special aptitude in order to think of a better act, and when we believe that the agent did not possess this aptitude. But this distinction seems to imply that thereis not such a thing as a special aptitude for this or that virtue; a view which cannot, I think, be maintained. An aptitude for generosity or for kindness may be as much a natural gift as an aptitude for poetry; and an aptitude for poetry may be as much improved by practice as an aptitude for kindness or generosity. Thus it would seem that there is no sense in which it is possible to think of some actions which in fact we do not think of, but impossible to think of others, except the sense that the ones we regard as possible would have been thought of if a different choice among alternatives actually thought of had been made on some previous occasion. We shall then modify our previous definition of the objectively right action by saying that it is the probably most beneficial among these that occur to the agent at the moment of choice. But we shall hold that, in certain cases, the fact that a more beneficial alternative does not occur to him is evidence of a wrong choice on some previous occasion. (§ 27 ¶ 3)

But since occasions of choice do often arise, and since there certainly is a sense in which it is possible to choose any one of a number of different actions as right and some as wrong. Our previous definitions of objectively right actions and of moral actions still hold, with the modification that, among physically possible actions, only those which we actually think of are to be regarded as possible. When several alternative actions present themselves, it is certain that we can both do which we choose, and chose which we will. In this sense all the alternatives are possible. What determinism maintains is, that our will to choose this or that alternative is the effect of antecedents; but this does not prevent our will from being itself a cause of other effects. And the sense in which different decisions are possible seems sufficient to distinguish some actions as right and some as wrong, some as moral and some as immoral. (§ 28 ¶ 1)

Connected with this is another sense in which, when we deliberate, either decision is possible. The fact that we judge one course objectively right may be the cause of our choosing this course: thus, before we have decided as to which course we think right, either is possible in the sense that either ill result from our decision as to which we think right. This sense of possibility is important to the moralist, and illustrates the fact that determinism does not make moral deliberation futile. (§ 28 ¶ 2)

Determinism does not, therefore, destroy the distinction of right and wrong; and we saw before that it does not destroy the distinction of good and bad: we shall still be able to regard some people as better than others, and some actions as more right than others. But it is said that praise and blame and responsibility are destroyed by determinism. When a madman commits what in a sane man we should call a crime, we do not blame him, partly because he probably cannot judge rightly as to consequences, but partly also because we feel that he could not have done otherwise: if all men are really in the position of the madman, it would seem that all ought to escape blame. But I think the question of choice really decides as to praise and blame. The madman, we believe (excluding the case of wrong judgment as to consequences), did not choose beteen different courses, but was impelled by a blind impulse. The sane man who (say) commits a murder has, on the contrary, either at the time of the murder or at some earlier time, chosen the worst of two or more alternatives that occurred to him; and it is for this we blame him. It is true that the two cases merge into each other, and the madman may be blamed if he has become mad in consequence of self-indulgence. But it is right that the two cases should not be too sharply distinguished, for we know how hard it often is in practice to decide whether people are what is called responsible for their actions. It is sufficient that there is a distinction, and that it can be applied easily in most cases, though there are marginal cases which present difficulties. We apply praise or blame, then, and we attribute responsibility, where a man, having to exercise choice, has chosen wrongly; and this sense of praise or blame is not destroyed by determinism. (§ 29 ¶ 1)

Determinism, then, does not in any way interfere with morals. It is worth noticing that freewill, on the contrary, would interfere most seriously, if anybody really believed in it. People never do, as a matter of fact, believe that anyone else’s actions are not determined by motives, however much they may think themselves free. Bradshaw consists entirely of predictions as to the actions of engine-drivers; but no one doubts Bradshaw on the ground that the volition of engine-drivers are not governed by motives. If we really believed that other people’s actions did not have causes, we could never try to influence other people’s actions; for such influence can only result if we know, more or less, what causes will produce the actions we desire. If we could never try to influence other people’s actions, no man could try to get elected to Parliament, or ask a woman to marry him: argument, exhortation, and command would become mere idle breath. Thus almost all the actions with which morality is concerned would become irrational, rational action would be wholly precluded from trying to influence people’s volitions, and right and wrong would be interfered with in a way in which determinism certainly does not interfere with them. Most morality absolutely depends upon the assumption that volitions have causes, and nothing in morals is destroyed by this assumption. (§ 30 ¶ 1)

Most people, it is true, do not hold the freewill doctrine in so extreme a form as that against which we have been arguing. They would hold that most of a man’s actions have causes, but that some few, say one per cent, are uncaused spontaneous assertions of will. If this view is taken, unless we can mark off the one per cent of volitions which are uncaused, every inference as to human actions is infected with what we may call one per cent of doubt. This, it must be admitted, would not matter much in practice, because, on other grounds, there will usually be at least one per cent of doubt in predictions as to human actions. But from the standpoint of theory there is a wide difference: the sort of doubt that must be admitted in any case is a sort which is capable of indefinite diminution, while the sort derived from the possible intervention of freewill is absolute and ultimate. In so far, therefore, as the possibility of uncaused volitions comes in, all the consequences above pointed out follow; and in so far as it does not come in, determinism holds. Thus one per cent of free will has one per cent of the objectionableness of absolute freewill; and has also only one per cent of the ethical consequences. (§ 30 ¶ 2)

In fact, however, no one really holds that right acts are uncaused. It wouldbe a monstrous paradox to say that a man’s decision ought not to be influenced by his belief as to what is his duty; yet, if he allows himself to decide on an act because he believes it to be his duty, his decision has a motive, i.e. a cause, and is not free in the only sense in which the determinist must deny freedom. It would seem, therefore, that the objections to determinism are mainly attributable to misunderstanding of its purport. Hence, finally it is not determinism but freewill that has subversive consequences. There is therefore no reason to regret that the grounds in favour of determinism are overwhelmingly strong. (§ 30 ¶ 3)