Throughout the last three chapters we have been considering various objections which might be urged against the theory stated in Chapters I and II. And the very last objection which we considered was one which consisted in asserting that the question whether an action is right or wrong does not depend upon its actual consequences, because whenever the consequences, so far as the agent can forsee, are likely to be the best possible, the action is always right, even if they are not actually the best possible. In other words, this objection rested on the view that right and wrong depend, in a sense, upon what the agent can know. And in the present chapter I propose to consider objections, which rest instead of this, on the view that right and wrong depend upon what the agent can do.(Ch. 6 ¶ 1)

Now it must be remembered that, in a sense, our original theory does hold and even insists that this is the case. We have, for instance, frequently referred to it in the last chapter as holding that an action is only right, if it produces the best possible consequences; and by the best possible consequences was meant consequences at least as good as would have followed from any action which the agent could have done instead. It does, therefore, hold that the question whether an action is right or wrong does always depend upon a comparison of its consequences with those of all the other actions which the agent could have done instead. It assumes, therefore, that wherever a voluntary action is right or wrong (and we have throughout only been talking of voluntary actions), it is true that the agent could, in a sense, have done something else instead. This is an absolutely essential part of the theory.(Ch. 6 ¶ 2)

But the reader must now be reminded that all along we have been using the words can, could, and possible in a special sense. It was explained in Chapter I (¶¶ 17-18), that we proposed, purely for the sake of brevity, to say that an agent could have done a given action, which he didn’t do, wherever it is true that he could have done it, if he had chosen; and similarly by what he can do, or what is possible, we have always meant merely what is possible, if he chooses. Our theory, therefore, has not been maintaining, after all, that right and wrong depend upon what the agent absolutely can do, but only on what he can do, if he chooses. And this makes an immense difference. For, by confining itself in this way, our theory avoids a controversy, which cannot be avoided by those who assert that right and wrong depend upon what the agent absolutely can do. There are few, if any, people who will expressly deny that we very often really could, if we had chosen, have done something different from what we actually did do. But the moment it is asserted that any man ever absolutely could have done anything other than what he did do, there are many people who would deny this. The view, therefore, which we are to consider in this chapter—the view that right and wrong depend upon what the agent absolutely can do—at once involves us in an extremely difficult controversy—the controversy concerning Free Will. There are many people who strenuously deny that any man ever could have done anything other than what he actually did do, or ever can do anything other than what he will do; and there are others who assert the opposite equally strenuously. And whichever view be held is, if combined with the view that right and wrong depend upon what the agent absolutely can do, liable to contradict our theory very seriously. Those who hold that no man ever could have done anything other than what he did do, are, if they also hold that right and wrong depend upon what we can do, logically bound to hold that no action of ours is ever right and none is ever wrong; and this is a view which is, I think, often actually held, and which, of course, constitutes an extremely serious and fundamental objection to our theory: since our theory implies, on the contrary, that we very often do act wrongly, if never quite rightly. Those, on the other hand, who hold that we absolutely can do things, which we don’t do, and that right and wrong depend upon what we thus can do, are also liable to be led to contradict our theory, though for a different reason. Our theory holds that, provided a man could have done something else, if he had chosen, that it is sufficient to entitle us to say that his action really is either right or wrong. But those who hold the view we are considering will be liable to reply that this is by no means sufficient: that to say that it is sufficient, is entirely to misconceive the nature of right and wrong. They will say that, in order that an action may be really either right or wrong, it is absolutely essential that the agent should have been really able to act differently, able in some sense quite other than that of merely being able, if he had chosen. If all that were really ever true of us were merely that we could have acted differently, if we had chosen, then, these people would say, it really would be true that none of our actions are ever right and that none are ever wrong. They will say, therefore, that our theory entirely misses out one absolutely essential condition of right and wrong—the condition that, for an action to be right or wrong, it must be freely done. And moreover, many of them will hold also that the class of actions which we absolutely can do is often not identical with those which we can do, if we choose. They may say, for instance, that very often an action, which we could have done, if we had chosen, is nevertheless an action which we could not have done; and that an action is always right, if it produces as good consequences as any other action which we really could have done instead. From which it will follow that many actions which our theory declares to be wrong, will, according to them, be right, because these actions really are the best of all that we could have done, though not the best of all that we could have done, if we had chosen.(Ch. 6 ¶ 3)

Now these objections seem to me to be the most serious which we have yet had to consider. They seem to me to be serious because (1) it is very difficult to be sure that right and wrong do not really depend, as they assert, upon what we can do and not merely on what we can do, if we choose; and because (2) it is very difficult to be sure in what sense it is true that we ever could have done anything different from what we actually did do. I do not profess to be sure about either of these points. And all that I can hope to do is to point out certain facts which do seem to me to be clear, though they are often overlooked; and thus to isolate clearly for the reader’s decision, those questions which seem to me to be really doubtful and difficult.(Ch. 6 ¶ 4)

Let us begin with the question: Is it ever true that a man could have done anything else, except what he actually did do? And, first of all, I think I had better explain exactly how this question seems to me to be related to the question of Free Will. For it is a fact that, in many discussions about Free Will, this precise question is never mentioned at all; so that it might be thought that the two have really nothing whatever to do with one another. and indeed some philosophers do, I think, definitely imply that they have nothing to do with one another: they seem to hold that our wills can properly said to be free even if we never can, in any sense at all, do anything else except what, in the end, we actually do do. But this view, if it is held, seems to me to be plainly a mere abuse of language. The statement that we have Free Will is certainly ordinarily understood to imply that we really sometimes have the power of acting differently from the way in which we actually do act; and hence, if anybody tells us that we have Free Will, while at the same time he means to deny that we ever have such a power, he is simply misleading us. We certainly have not got Free Will, in the ordinary sense of the word, if we never really could, in any sense at all, have done anything else than what what we did do; so that, in this respect, the two questions certainly are connected. But, on the other hand, the mere fact (if it is a fact) that we sometimes can, in some sense, do what we don’t do, does not necessarily entitle us to say that we have Free Will. We certainly haven’t got it, unless we can; but it doesn’t follow that we have got it, even if we can. Whether we have or not will depend upon the precise sense in which it is true that we can. So that even if we do decide that we really can often, in some sense, do what we don’t do, this decision by itself does not entitle us to say that we have Free Will.(Ch. 6 ¶ 5)

And the first point about which we can and should be quite clear is, I think, this: namely that we certainly often can, in some sense, do what we don’t do. It is, I think, quite clear that this is so; and also very important that we should realise that it is so. For many people are inclined to assert, quite without qualification: No man ever could, on any occasion, have done anything else than what he actually did do on that occasion. By asserting this quite simply, without qualification, they imply, of course, (even if they do not mean to imply), that there is no proper sense of the word could, in which it is true that a man could have acted differently. And it is this implication which is, I think, quite certainly absolutely false. For this reason, anybody who asserts Nothing ever could have happened, except what actually did happen, is making an assertion which is quite unjustifiable, and which he himself cannot help constantly contradicting. And it is important to insist on this, because many people do make this unqualified assertion, without seeing how violently it contradicts what they themselves, and all of us, believe, and rightly believe, at other times. If, indeed, they insert a qualification—if they merely say, In one sense of the word could nothing ever could have happened, except what did happen, then, they may perhaps be perfectly right: we are not disputing that they may. All that we are maintaining is that, in one perfectly proper and legitimate sense of the word could, and that one of the very commonest senses in which it is used, it is quite certain that some things which didn’t happen could have happened. And the proof that this is so, is simply as follows.(Ch. 6 ¶ 6)

It is impossible to exaggerate the frequency of the occasions on which we all of us make a distinction between two things, neither of which did happen,—a distinction which we express by saying, that whereas the one could have happened, the other could not. No distinction is commoner than this. And no one, I think, who fairly examines the instances in which we make it, can doubt about three things: namely (1) that very often there really is some distinction between the two things, corresponding to the language which we use; (2) that this distinction, which really does subsist between the things, is the one which we mean to express by saying that the one was possible and the other impossible; and (3) that this way of expressing it is a perfectly proper and legitimate way. But if so, it absolutely follows that one of the commonest and most legitimate usages of the phrases could and could not is to express a difference, which often really does hold between two things neither of which did actually happen. Only a few instances need be given. I could have walked a mile in twenty minutes this morning, but I certainly could not have run two miles in five minutes. I did not, in fact, do either of these two things; but it is pure nonsense to say that the mere fact that I did not, does away with the distinction between them, which I express by saying that the one was within my powers, whereas the other was not. Although I did neither, yet the one was certainly possible to me in a sense in which the other was totally impossible. Or, to take another instance: It is true, as a rule, that cats can climb trees, whereas dogs can’t. Suppose that on a particular afternoon neither A’s cat nor B’s dog do climb a tree. It is quite absurd to say that this mere fact proves that we must be wrong if we say (as we certainly often should say) that the cat could have climbed a tree, though she didn’t, whereas the dog couldn’t. Or, to take an instance which concerns an inanimate object. Some ships can steam 20 knots, whereas others can’t steam more than 15. And the mere fact that, on a particular occasion, a 20-knot steamer did not actually run at this speed certainly does not entitle us to say that she could not have done so, in the sense in which a 15-knot one could not. On the contrary, we all can and should distinguish between cases in which (as, for instance, owing to an accident to her propeller) she did not, because she could not, and cases in which she did not, although she could. Instances of this sort might be multiplied quite indefinitely; and it is surely quite plain that we all of us do continually use such language: we continually, when considering two events, neither of which did happen, distinguish between them by saying that whereas the one was possible, though it didn’t happen, the other was impossible. And it is surely quite plain that what we mean by this (whatever it may be) is something which is often perfectly true. But, if so, then anybody who asserts, without qualification, Nothing ever could have happened, except what did happen, is simply asserting what is false.(Ch. 6 ¶ 7)

It is, therefore, quite certain that we often could (in some sense) have done what we did not do. And now let us see how this fact is related to the argument by which people try to persuade us that it is not a fact.(Ch. 6 ¶ 8)

The argument is well known; it is simply this. It is assumed (for reasons which I need not discuss) that absolutely everything that happens has a cause in what precedes it. But to say this is to say that it follows necessarily from something that preceded it; or, in other words, that, once the preceding events which are its cause had happened, it was absolutely bound to happen. But to say that it was bound to happen, is to say that nothing else could have happened instead; so that, if everything has a cause, nothing ever could have happened except what did happen.(Ch. 6 ¶ 9)

And now let us assume that the premise of this argument is correct: that everything really has a cause. What really follows from it? Obviously all that follows is that, in one sense of the word could, nothing ever could have happened, except what did happen. This really does follow. But, if the word could is ambiguous—if, that is to say, it is used in different senses on different occasions—it is obviously quite possible that though, in one sense, nothing ever could have happened except what did happen, yet in another sense, it may at the same time be perfectly true that some things which did not happen could have happened. And can anybody undertake to assert with certainty that the word could is not ambiguous? that it may not have more than one legitimate sense? Possibly it is not ambiguous; and, if it is not, then the fact that some things, which did not happen, could have happened, really would contradict the principle that everything has a cause; and, in that case, we should, I think, have to give up this principle, because the fact that we often could have done what we did not do, is so certain. But the assumption that the word could is not ambiguous is an assumption which certainly should not be made without the clearest proof. And yet I think it often is made, without any proof at all; simply because it does not occur to people that words often are ambiguous. It is, for instance, often assumed, in the Free Will controversy, that the question at issue is solely as to whether everything is caused, or whether acts of will are sometimes uncaused. Those who hold that we have Free Will, think themselves bound to maintain that acts of will sometimes have no cause; and those who hold that everything is caused think that this proves completely that we have not Free Will. But, in fact, it is extremely doubtful whether Free Will is at all inconsistent with the principle that everything is caused. Whether it is or not, all depends on a very difficult question as to the meaning of the word could. All that is certain about the matter is (1) that, if we have Free Will, it must be true, in some sense, that we sometimes could have done, what we did not do; and (2) that, if everything is caused, it must be true, in some sense, that we never could have done, what we did not do. What is very uncertain, and what certainly needs to be investigated, is whether these two meanings of the word could are the same.(Ch. 6 ¶ 10)

Let us begin by asking: What is the sense of the word could, in which it is so certain that we often could have done, what we did not do? What, for instance, is the sense in which I could have walked a mile in twenty minutes this morning, though I did not? There is one suggestion, which is very obvious: namely, that what I mean is simply after all that I could, if I had chosen; or (to avoid a possible complication) perhaps we had better say that I should, if I had chosen. In other words, the suggestion is that we often use the phrase I could simply and solely as a short way of saying I should, if I had chosen. And in all cases, where it is certainly true that we could have done, what we did not do, it is, I think, very difficult to be quite sure that this (or something similar) is not what we mean by the word could. The case of the ship may seem to be an exception, because it is certainly not true that she would have steamed twenty knots if she had chosen; but even here it seems possible that what we mean is simply that she would, if the men on board of her had chosen. There are certainly good reasons for thinking that we very often mean by could merely would, if so and so had chosen. And if so, then we have a sense of the word could in which the fact that we often could have done what we did not do, is perfectly compatible with the principle that everything has a cause: for to say that, if I had performed a certain act of will, I should have done something which I did not do, in no way contradicts this principle.(Ch. 6 ¶ 11)

And an additional reason for supposing that this is what we often mean by could, and one which is also a reason why it is important to insist on the obvious fact that we very often really should have acted differently, if we had willed differently, is that those who deny that we ever could have done anything, which we did not do, often speak and think as if this really did involve the conclusion that we never should have acted differently, even if we had willed differently. This occurs, I think, in two chief instances—one in reference to the future, the other in reference to the past. The first occurs when, because they hold that nothing can happen, except what will happen, people are led to adopt the view called Fatalism—the view that whatever we will, the result will always be the same; that it is, therefore, never any use to make one choice rather than another. And this conclusion will really follow if by can we mean would happen, even if we were to will it. But it is certainly untrue, and it certainly does not follow from the principle of causality. On the contrary, reasons of exactly the same sort and exactly as strong as those which lead us to suppose that everything has a cause, lead us to the conclusion that if we choose one course, the result will always be different in some respect from what it would have been, if we had chosen another; and we know also that the difference would sometimes consist in the fact that what we chose would come to pass. It is certainly often true of the future, therefore, that whichever of two actions we were to choose, would actually be done, although it is quite certain that only one of the two will be done.(Ch. 6 ¶ 12)

And the second instance, in which people are apt to speak and think, as if, because no man ever could have done anything but what he did do, it follows that he would not, even if he had chosen, is as follows. Many people seem, in fact, to conclude directly from the first of these two propositions, that we can never be justified in praising or blaming a man for anything that he does, or indeed for making any distinction between what is right or wrong, on the one hand, and what is lucky or unfortunate on the other. They conclude, for instance, that there is never any reason to treat or to regard the voluntary commission of a crime in any different way from that in which we treat or regard the involuntary catching of a disease. The man who committed the crime could not, they say, have helped committing it any more than the other man could have helped catching the disease; both events were equally inevitable; and though both may of course be great misfortunes, though both may have very bad consequences and equally bad ones—there is no justification whatever, they say, for the distinction we make between them when we say that the commission of the crime was wrong, or that the man was morally to blame for it. And this conclusion, again, will really follow if by could not, we mean would not, even if he had willed to avoid it. But the point I want to make is, that it follows only if we make this assumption. That is to say, the mere fact that the man would have succeeded in avoiding the crime, if he had chosen (which is certainly often true), whereas the other man would not have succeeded in avoiding the disease, even if he had chosen (which is certainly also often true) gives an ample justification for regarding and treating the two cases differently. It gives such a justification, because, where the occurrence of an event did depend upon the will, there, by acting on the will (as we may do by blame or punishment) we have often a reasonable chance of preventing similar events from recurring in the future; whereas, where it did not depend upon the will, we have no such chance. We may, therefore, fairly say that those who speak and think, as if a man who brings about a misfortune voluntarily ought to be treated and regarded in exactly the same way as one who brings about an equally great misfortune involuntarily, are speaking and thinking as if it were not true that we ever should have acted differently, even if we had willed to do so. And that is why it is extremely important to insist on the absolute certainty of the fact that we often really should have acted differently, if we had willed differently.(Ch. 6 ¶ 13)

There is, therefore, much reason to think that when we say that we could have done a thing which we did not do, we often mean merely that we should have done it, if we had chosen. And if so, then it is quite certain that, in this sense, we often really could have done what we did not do, and that this fact is in no way inconsistent with the principle that everything has a cause. And for my part I must confess that I cannot feel certain that this may not be all that we usually mean and understand by the assertion that we have Free Will; so that those who deny that we have it are really denying (though, no doubt, often unconsciously) that we ever should have acted differently, even if we had willed differently. It has been sometimes held that this is what we mean; and I cannot find any conclusive argument to the contrary. And if it is what we mean, then it absolutely follows that we really have Free Will, and also that this fact is quite consistent with the principle that everything has a cause; and it follows also that our theory will be perfectly right, when it makes right and wrong depend on what we could have done, if we had chosen.(Ch. 6 ¶ 14)

But, no doubt, there are many people who will say that this is not sufficient to entitle us to say that we have Free Will; and they will say this for a reason, which certainly has some plausibility, though I cannot satisfy myself that it is conclusive. They will say, namely: Granted that we often should have acted differently, if we had chosen differently, yet it is not true that we have Free Will, unless it is also often true in such cases that we could have chosen differently. The question of Free Will has thus been represented as merely the question whether we ever could have chosen, what we did not choose, or ever can choose, what, in fact, we shall not choose. And since there is some plausibility in this contention, it is, I think, worth while to point out that here again it is absolutely certain that, in two different senses, at least, we often could have chosen, what, in fact, we did not choose; and that in neither sense does this fact contradict the principle of causality.(Ch. 6 ¶ 15)

The first is simply the old sense over again. If by saying that we could have done, what we did not do, we often mean merely that we should have done it, if we had chosen to do it, then obviously, by saying that we could have chosen to do it, we may mean merely that we should have so chosen, if we had chosen to make the choice. And I think there is no doubt it is often true that we should have chosen to do a particular thing if we had chosen to make the choice; and that this is a very important sense in which it is often in our power to make a choice. There certainly is such a thing as making an effort to induce ourselves to choose a particular course; and I think there is no doubt that often if we had made such an effort, we should have made a choice, which we did not in fact make.(Ch. 6 ¶ 16)

And besides this, there is another sense in which, whenever we have several different courses of action in view, it is possible for us to choose any one of them; and a sense which is certainly of some practical importance, even if it goes no way to justify us in saying that we have Free Will. This sense arises from the fact that in such cases we hardly ever know for certain beforehand, which choice we actually shall make; and one of the commonest senses of the word possible is that in which we call an event possible when no man can know for certain that it will not happen. It follows that almost, if not quite always, when we make a choice, after considering alternatives, it was possible that we should have chosen one of these alternatives, which we did not actually choose; and often, of course, it was not only possible, but highly probable, that we should have done so. And this fact is certainly of practical importance, because many people are apt much too easily to assume that it is quite certain that they will not make a given choice, which they know they ought to make, if it were possible; and their belief that they will not make it tends, of course, to prevent them from making it. For this reason it is important to insist that they can hardly ever know for certain with regard to any given choice that they will not make it.(Ch. 6 ¶ 17)

It is, therefore, quite certain (1) that we often should have acted differently, if we had chosen to; (2) that similarly we often should have chosen differently, if we had chosen so to choose; and (3) that it was almost always possible that we should have chosen differently, in the sense that no man could know for certain that we should not so choose. All these three things are facts, and all of them are quite consistent with the principle of causality. Can anybody undertake to say for certain that none of these three facts and no combination of them will justify us in saying that we have Free Will? Or, suppose it is granted that we have not Free Will, unless it is often true that we could have chosen, what we did not choose:—Can any defender of Free Will, or any opponent of it, show conclusively that what he means by could have chosen in this proposition, is anything different from the two certain facts, which I have numbered (2) and (3), or some combination of the two? Many people, no doubt, will still insist that these two facts alone are by no means sufficient to entitle us to say that we have Free Will; that it must be true that we were able to choose, in some quite other sense. But nobody, so far as I know, has ever been able to tell us exactly what that sense is. For my part, I can find no conclusive argument to show either that some such other sense of can is necessary, or that it is not. And, therefore, this chapter must conclude with a doubt. It is, I think, possible that, instead of saying, as our theory said, that an action is only right, when it produces consequences as good as any which would have followed from any other action which the agent would have done, if he had chosen, we should say instead that it is right whenever and only when the agent could not have done anything which would have produced better consequences; and that this could not have done is not equivalent to would not have done, if he had chosen, but is to be understood in the sense, whatever it may be, which is sufficient to entitle us to say that we have Free Will. If so, then our theory would be wrong, just to this extent.(Ch. 6 ¶ 18)