I PASS now to consider whether any non-naturalistic attempt to define ‘right’ or ‘obligatory’ can be accepted. By a non-naturalistic attempt I mean one that defines ‘right’ by a definition which includes a reference to some distinctively ethical term other than ‘right’. Of such attempts I know only one, viz. Professor Moore's theory in Principia Ethica that right means ‘productive of the greatest possible amount of good’. And I am bound to say that this seems just as little a true account of the meaning we have in mind when we use the word right, as are the naturalistic attempts we have been considering. Is it not clear that when a plain man says ‘it is right to fulfil promises’ he is not necessarily thinking of the total consequences of such an act, still less thinking that the total consequences are always the best possible or are even likely to be so? And if some one says ‘it is right to do that which will produce the best consequences’, he does not think he is elucidating the meaning of the word ‘right’, but that he is stating the characteristic, the possession of which by an act entails its having the characteristic of rightness. I need not elaborate the point, because, as I have shown elsewhere,1 Professor Moore seems to have given up this view when he wrote his later book, Ethics, and to have adopted the view that tendency to produce the best consequences is not the essence of rightness but the ground of rightness. I think almost every one would admit that this is a far more plausible view.

I think, too, that, since in his later work Professor Moore makes no attempt to define rightness, he has presumably come to adopt the view that right is an indefinable notion. And this is the conclusion to which I am myself led by the break-down of the attempts to define rightness which we have considered, I believe I have passed under review all the main attempts at defining ‘right’ or ‘obligatory’. That they have broken down does not prove that every attempt must break down, but it creates a presumption that it will. And, indeed, the more we think of the term ‘right’, the more convinced we are likely to be that it is an indefinable term, and that when one attempts to define it one will either name something plainly different from it, or use a term which is a mere synonym of it.

The word ‘right’, when used in a context of moral thought, seems to me to mean very nearly, but not quite, the same as ‘obligatory’ or ‘what is my duty’. The first point of difference may, I think, be stated thus: In most situations that occur in life, there are a variety of claims upon me that I can by my action either satisfy or fail to satisfy. There are, or at least there may be, cases in which any one of two or more acts would completely satisfy these claims, or would satisfy them to an equal extent and to the greatest extent possible. Let there be two such acts, A and B. Then we should agree that in doing either of them we should have done a right act. But we should not in doing either of them have done an obligatory act; for I cannot be obliged to do act A if act B would equally well satisfy the claims upon me, nor can I be obliged to do act B if act A would equally well satisfy the claims upon me. My obligation in this case is not to do act A nor to do act B, but to do either act A or act B. In any situation in which there are any claims upon me, there is either one act which satisfies these claims more completely than any other would; then this act is both obligatory and right: or there are two or more acts which would fulfil the claims equally, and better than any other act open to me would; then all of these are right and none of them is obligatory, but it is obligatory to do one or other of them.

Thus it seems that both the question ‘What is the right act for me to do?’ and the question ‘What is the act which it is my duty to do?’ are wrong questions; the first because there may be more than one act that is right; the second because there may be none that is obligatory, the only obligation being to do one or other of certain acts.

Some would deny the correctness of this distinction. They would maintain that when there are two or more acts, one or other of which, as we say, we ought to do (it not being our duty to do one rather than any of the rest), the truth is that these are simply alternative ways of producing a single result, and that what is right is, strictly, not to do any of these acts, and what is obligatory is not to do ‘one or other’ of them; what is right and what is obligatory being to produce the result. This answer might, I think, fairly apply to many cases, in which it is the production of a single result that we think obligatory, the means being optional; e.g. to a case in which it is our duty to convey information to some one, but morally immaterial whether we do so verbally or in writing. But in principle, at any rate, there may be other cases in which it is our duty to produce one or other of two or more different states of affairs, without its being our duty to produce one of them rather than another; in such a case each of these acts will be right, and none of them will be obligatory.

In maintaining that an act may be right without being obligatory it might seem that I have reduced ‘right act’ to meaning ‘act which it is not my duty not to do’. So to reduce it would not be correct. For there may be cases in which none of the acts open to me will be in any respect a fulfilment of claims, and in such a case we should not call such acts right, but indifferent. I have not, however, so reduced the meaning of ‘right act’; for I have described it as including two moments—(a) that there is no other act that would more completely fulfil the moral claims on us, but also (b) that any act which is right is itself a fulfilment of at least one claim upon us.

I have pointed out this distinction between ‘right’ and ‘obligatory’ because it is, I think, clearly implied in the way in which we use the two terms; but it does not seem to me very important. The difference is a simple one, akin to that between ‘first’ and ‘second to none’, of which the former is applicable to any competitor who beats all his rivals, while the latter is applicable to any competitor who, while he may be equal with some, is not inferior to any. Or again, the difference might be expressed by saying that in calling an action obligatory we are implying not only that it is right, but that any other in the circumstances would be wrong.

I propose next to consider Professor Broad's discussion of the meaning of ‘ought’ and ‘right’, which I have found very suggestive. He begins his discussion2 by distinguishing a wider and a narrower sense of ‘ought’. ‘In its narrower sense’, he says, ‘it applies only to actions which an agent could do if he willed. But there is a wider sense in which there is no such implication. We can say that sorrow ought to have been felt by a certain man at the death of a certain relation, though it was not in his power to feel sorrow at will. And we can say that virtue ought to be rewarded.’

On this I would comment as follows. I should agree that we often use ‘ought’ in this wider sense. But I should maintain that such a use is not strict. Can we seriously say that sorrow ought to have been felt by some one at the death of a relation? Only, it seems to me, (a) if we think that it was possible for him (and I agree with Professor Broad in holding this to be impossible) then and there to summon up a feeling of sorrow, or (b) if we think that by acting differently in the past he could have so modified his character that he would now have felt sorrow; and in the latter case the proper application of ‘ought’ is to say ‘he ought to have so acted in the past’, not ‘he ought to have felt sorrow now’. Apart from such a thought, all we are entitled to say is, not that he ought to have felt sorrow now, but that his not feeling it is a bad thing, a manifestation of a bad character. The wider use of ‘ought’ is really an improper use of it, one which we could not seriously defend. Or again, take the saying that virtue ought to be rewarded. We can say this properly only if we think that some being or beings, God or men, can and ought so to act that virtue will be rewarded. Unless we think this, all we are justified in saying is that an arrangement of human affairs in which virtue is not rewarded is a bad one; the specific justification required for saying ‘virtue ought to be rewarded’ is absent.

Professor Broad next,3 following Sidgwick, distinguishes three applications of the word ‘ought’. Some people, he says, judge that there are certain types of action that ought to be done in all or in certain types of situation, regardless of the goodness or badness of the probable consequences. This he calls the deontological application. Secondly, there are people who deny that they ever make such judgements as these, but nevertheless judge that every one ought to aim at certain ends, without any ulterior motive, e.g. at his own greatest happiness or at the greatest happiness of all sentient beings. This he calls the teleological application. Lastly, there may be people who deny that there are any types of action that are obligatory irrespective of their consequences, and also that there are any ends which every one ought to aim at, but who would admit a third application of ‘ought’. They would say that if a man in fact takes a certain end as ultimate, he ought to adopt such means as will bring it into being, and not do things which he believes will be inconsistent with its realization. This Professor Broad calls the logical application of ‘ought’.

He next asks how these three different applications are related to the wider and the narrower meaning of ‘ought’ which he has already distinguished. He points out that ‘ought’, when used in its teleological application, is used in its wider sense. For it is plain that we cannot desire a certain end at will, any more than we can at will feel sorrow at the death of a particular person, or love a particular person. Secondly, he argues that ‘ought’, when used in its logical application, is used in its narrower sense. For, since we believe that it is within the power of any sane person to be consistent if he tries, we believe that if he desires a certain end he can, if he tries, adopt the appropriate means to it. The logical ‘ought’ is thus a special case of the deontological ‘ought’. Finally, it is obvious that the deontological application of ‘ought’ involves the use of ‘ought’ in its narrower sense. Thus the three applications involve no new sense of ‘ought’ but only the two previously recognized.

Now if we have been right in saying that the wider sense of the word ‘ought’ is a loose and improper sense of it, and one in which we should not persist in using it when the implications of such a use have been pointed out, and if, as Professor Broad correctly says, the teleological application of it involves the wider sense, it will follow that the teleological application of it is an improper application, since it is an application of the word in an improper sense. And further, the logical application is also an improper application. It is true that in this case one of the conditions involved in the proper use of the word ‘ought’ is fulfilled, namely that we can be consistent if we choose, that we can will the means if we desire the end. But the other condition of the proper use of ‘ought’ is not fulfilled. For no one really thinks that the fact that a person desires a certain end makes it obligatory on him to will the means to it; if we think the end is a bad one (or that his desiring it is bad), we think that in spite of his desiring the end he ought not to adopt the means. Thus the logical application of ‘ought’, also, is an improper application of it, and we are left with but one proper application, as we are left with but one proper sense; viz. the application to acts within the agent's power to do if he chooses, and imposed on him by the moral law. In other words the categorical imperative is the only true imperative. When some one uses ‘ought’ in the teleological application he is emptying ‘ought’ of its real meaning, and all that he has a right to say is that it would be good if people aimed at certain ends; and when he uses it in its logical application he is equally emptying it of its proper meaning, and all that he is entitled to say is that a man who desires certain ends can hope to get them only if he adopts certain means. In neither of these statements does the distinctive meaning of ‘ought’ appear at all.

Professor Broad now proceeds to the relation between ‘right’ and ‘ought’.4 He points out (1) the distinction I have already pointed out, viz. that there may be cases in which several alternative acts are right and none of them is obligatory. (2) He holds that in a further respect the meaning of ‘ought’ is more restricted than that of ‘right’. For he holds that we tend to confine the word ‘ought’ to cases where we believe that there are motives and inclinations against doing the rightest action open to the agent. He quotes with approval Sidgwick's remark that we should hardly say of an ordinary healthy man that he ought, in the narrower sense of ‘ought’, to take adequate nourishment; though we might say this of an invalid with a disinclination to take food, or of a miser.

Sidgwick's example is ill-chosen. We may know (or rather, have strong reasons for thinking) that a certain man has a natural liking for food, and that he never has an antipathy to food as such. Yet we know that in ordinary human nature, and therefore probably in his, there are many other desires which may incidentally conflict with desire for food (such as dislike of the particular food that is available, or of the company in which it would have to be eaten). And, knowing the possibilities of such conflict, I think we should not hesitate to say of such a man that he ought to take the food that he needs in order to keep him fit. And similarly, though we may know that a woman's natural love for her child is strong, we should not hesitate on that ground to say that she ought to look after her child's welfare; for we know that, however strong maternal love may be, there are many other desires with which it will often come into conflict. In principle it seems that, however much we know that an agent has a natural inclination to do a certain right act, we can never know that he has not some other inclination which might incidentally conflict with that inclination; and therefore that we never need hesitate on this ground to describe the right act as obligatory. Some one might, however, try to restate Sidgwick's view in such a way as to avoid this objection. He might say ‘Granted that we never can know this; yet suppose that in fact a man had an inclination to do the act which is right, and no inclination leading him towards any alternative act; would not this act then be (although we could not know it to be) right without being obligatory?’ In support of this he might plead the assumption which we commonly make that what a man is morally obliged to do must be something that he can either do or refrain from doing, and add that in the case supposed the agent could not refrain from doing the act which is right.

The assumption that psychical necessity excludes obligation is one that should not be lightly made. For in the long run we must admit5 that what a man does he does by a psychical necessity, and if the assumption is added to this admission, the conclusion can only be that obligation does not exist. We may perhaps get some light on the question by considering on what grounds, in a particular situation, we should reject certain things as not being our duty, however suitable they might be. We should reject (1) any act which would involve a metaphysical impossibility. When we have done a wrong, perhaps the most suitable thing, if it were possible, would be to undo our own act; but this we reject because from the nature of things it is impossible to undo the past. We should reject (2) any act which would involve a control over matter which we are convinced our mind does not possess. If the only way we could help some one we had wronged was by performing an impossible feat of endurance, we should reject that. We should reject (3) anything which would involve an impossible control over the state of our own mind. If we hate some one, then however right it would be that we should forthwith love him as intensely as we now hate him, we know that we cannot by a decision effect that result here and now, and we do not (or should not) think it our duty to love him now, though we should think it our duty to try to mould our character so as to love him in time. But (4) if there is any act which we think we could do if we desired sufficiently strongly to do it, as in fact any act of self-exertion6 can be done if we desire sufficiently strongly to do it, then we do not ask whether we in fact desire it sufficiently strongly to enable us to perform the act of self-exertion. The very fact that we recognize that the act would be right involves some attraction toward it, and we do not ask whether this is strong enough, for the excellent reason that it is only by our success or failure to do the act that we can discover whether the desire was strong enough. Even if there is a necessity to do the right act, or a necessity to do the wrong act, we never know there is, and therefore there is nothing to prevent us from thinking of the right act as obligatory.

But some one might reply, ‘Will not your failure to do the act show that there was a psychical necessity to do otherwise and therefore no obligation to do that act? and will not your success, if you succeed, show that there was a psychical necessity to do the act, and therefore no obligation?’ The fact, however, seems to be that even if the occurrence of either act implies that we were under a psychical necessity to do it, that does not prevent our continuing to recognize that we were under an obligation to do the one act and not the other. And if that be so, it implies that the sort of freedom involved in the recognition of an obligation is not freedom of indifference to choose to do or not to do the act, but only freedom in the milder sense of capacity to do the right act if we desire sufficiently strongly to do it.

Further, it seems clear that in trying to discover whether it is our duty to do a certain act, we regard as irrelevant the state of our inclination towards or against the doing of the act; and if this be so, the absence of a contrary inclination cannot prevent that from being our duty, which otherwise would be our duty. The notion of duty or obligation undoubtedly carries with it the idea of restriction; but the nature of the restriction is not that our duty is something that we ought to do though we have a contrary inclination, but that it is something that we ought to do irrespective of the state of our inclination.

Professor Broad next proceeds to state his view of the meaning of ‘right’.

‘It seems to me’, he says, ‘that, when I speak of anything as “right”, I am always thinking of it as a factor in a certain wider total situation, and that I mean that it is “appropriately” or “fittingly” related to the rest of the situation. When I speak of anything as “wrong” I am thinking of it as “inappropriately” or “unfittingly” related to the rest of the situation. This is quite explicit when we say that love is the right emotion to feel to one's parents, or that pity and help are the right kinds of emotion and action in presence of undeserved suffering. This relational character of rightness and wrongness tends to be disguised by the fact that some types of action are commonly thought to be wrong absolutely; but this, I think, means only that they are held to be unfitting to all situations.’7

This account has the great merit of connecting the ethical sense of right and wrong with other uses of the words. It is plain that when we speak of ‘the right road’ or ‘the right key’ we are thinking of the road or key as fitting a particular situation in which some one is placed. The right road is that the taking of which fits into a situation of which the other element is his desire to get from A to B; the right key is that his using of which fits into a situation of which the other element is his wish to unlock a particular lock. It is worth while in this connexion to contrast the meaning of ‘the right road’ with that of a ‘good road’. Goodness is an attribute which belongs permanently to the road or key, so long as it remains unchanged in its other characteristics; rightness is an attribute which they have only relatively to a particular situation and a particular need. A good road need not be the right road, and a bad road may be the right road, if the one does not and the other does meet the requirements of the particular situation.

A similar distinction between goodness and rightness in their moral applications may be noted. When we examine certain emotions, for instance, such as benevolence, we can merely by examining their intrinsic nature see that they are good. There are others which cannot in virtue of their intrinsic nature be called either good or bad, but can be judged to be right in certain situations, and wrong in others; e.g. sorrow is right when one contemplates the death of a friend, and wrong when one contemplates the success of a rival.

But if rightness in its ethical application shares with rightness in other connexions the characteristic of being relational, in another respect ethical rightness is quite different from any other kind of rightness. What we mean by calling a road or a key right can be explained purely in terms of desire and of causation. The right road or the right key is that the use of which by us will have a certain desired effect, that of taking us to a definite place or of opening a definite door. Moral rightness cannot, we may say in the light of our previous argument,8 be thus explained in terms of any non-moral relation. As Professor Broad remarks, ‘the kind of appropriateness and inappropriateness which is implied in the notions of “right” and “wrong”’ (i.e. in their ethical use) ‘is, so far as I can see, specific and unanalysable’.9

The thought of rightness as being fitness, in a certain specific and unanalysable way, to a certain situation, is one that plays a large part in Samuel Clarke's moral philosophy,10 and forms one of the main merits of that not sufficiently regarded philosopher.

It is possible to state more exactly the relation between moral suitability and rightness. Suppose we take a case in which a man has to choose between two actions each of which would bring some good and some evil into existence, and that action A would produce a greater balance of good than action B. Then action B will be morally suitable to a certain degree, and in a certain respect, because it will produce some good; but we should not call it a right action. Not any and every degree of moral suitability will make an action right. On the other hand, complete suitability is not needed in order to make an action right; for action A will be right although in view of the fact that it will produce some evil it is not completely suitable morally. Rightness can be identified, then, neither with any and every degree of suitability, nor with complete suitability, but only with the greatest amount of suitability possible in the circumstances.

The same result emerges if we consider a case in which a man has made two promises, and can fulfil either only by breaking the other. If we decide that he ought to keep promise A rather than promise B, each of the actions will have some suitability because each will be the fulfilling of a promise; neither will have complete suitability, because each will be the breaking of a promise. We call right that act which is the most suitable of those possible in the circumstances. The other act cannot be called right, but only right in a certain respect.

One has, of course, to consider the question whether suitability is a genuine genus of which moral rightness is one species, or whether we are being taken in by a mere ambiguity in the term ‘suitable’, the utilitarian suitability of a road or a key having nothing whatever in common with the moral suitability of an action or an emotion. Are the two suitabilities related to each other as the ‘colourness’ of red is related to the ‘colourness’ of blue, or as the ‘ploughness’ of a certain agricultural instrument is related to the ‘ploughness’ of a failure in an examination? We surely must say of the two suitabilities, as Aristotle says of the different meanings of ‘good’, that they are not an instance of mere accidental ambiguity of a word; yet it is hard to find any element of real identity. The most obvious suggestion that arises in one's mind is that moral suitability is, after all, an example of utilitarian suitability—that to say of an act that it is right is to say that it serves a human purpose, or that it serves human purposes better than any other act possible in the circumstances. Yet I think we have only to examine carefully whether that is what we mean when we call an act right, to feel assured that it is not so.11

I am inclined to think that all that is common to these two suitabilities is that both are relations to which we feel a favourable reaction. There is some faint element of likeness in the two reactions, in that both are favourable; but we err if we therefore think there is an element of identity between utilitarian suitability and moral suitability, just as we err if we think that because we never call any thing good unless we have a favourable reaction to it, there is therefore a common element in the goodness of all the things we call good.

But if there is no real identity between moral suitability and utilitarian suitability, there seems to be another form of suitability which has an affinity with moral suitability, viz. aesthetic suitability. There seems to be something not altogether different in the way in which a situation calls for a certain act, and the way in which one part of a beautiful whole calls for the other parts. Here, as in the case of a right act, there is no question of subserving an extraneous purpose; there is a direct harmony between the parts of the composition, as there is between a moral situation and the act which completes it. The harmony is not of the same kind—rightness is not beauty; but there seems to be a genuine affinity, which justified the Greeks in their application of the word kalo,n to both.

If Professor Broad's view is correct, as I think it is, moral rightness is a complex characteristic. It includes in it the generic quality of suitability, which it shares with the rightness of an element in a beautiful whole. And it includes in it the differentia which distinguishes it from every form of rightness but itself. It is a complex characteristic, just as redness is a complex characteristic, including in it a generic and a differential element. Now redness, though complex, is not definable; we can begin to define it, when we say it is a form of colour, but we cannot complete the definition, since if we try to state what distinguishes it from other forms of colour we can only say that it is the being redness that does so. In the same way we can begin to define moral rightness, because we can say it is a form of suitability; but we cannot complete the definition, since if we ask what kind of suitability it is we can only say that it is the kind of suitability that is rightness. Professor Broad seems to me to be right in considering that no further analysis of it is possible.

Now it is to be noted that, whereas we cannot seriously say of any one that he ought to have a certain emotion, because we do not think it is in his power to acquire it forthwith, there is no such limitation to the use of the word ‘right’. We can still call grief the right or fitting emotion in certain situations, for instance, even if we do not suppose the person we are thinking about has it in his power to feel grief in those circumstances. Its fittingness depends solely on the nature of the circumstances and not at all on his capacity or incapacity. Thus, while we had to reject the wider use of ‘ought’ (that in which it is used when the capacity to act or feel in the way in question is not believed to be present) as being a loose use, it is the wider use of ‘right’ that is the proper use of it; although it must be granted that when we use ‘right’ of acts, as opposed to emotions, we usually think of them as being in the agent's power to do or to forbear from doing. Our common use of the word ‘right’ is so fluid that, although what it naturally conveys is simply the notion of fitness or correctness, without implying either that there is only one act or emotion that fits the situation, or that it is in the agent's power to produce the act or emotion in question, yet by usage ‘right’ is very often treated as equivalent to ‘obligatory’. This is clearly so in the common phrases ‘the right act,’ ‘the right thing to do’, where the use of the definite article shows the first of these implications to be present, and the second is in fact also present.

I have spoken of acts as being obligatory, and this language is often convenient, for brevity. But it is not strictly correct. For consider the situation when an obligation really exists, viz. before the act in question, or any alternative act, has been done. We cannot then, strictly speaking, say ‘such and such an act is obligatory’, for the act is not there, to be either obligatory or anything else. Nor, again, can we say ‘such and such an act would be obligatory if it were done’; for clearly its obligatoriness, if it has any, does not depend on its being done. The only strict language which we can use in the circumstances is ‘so and so is obliged to act in such and such a way’. In fact, obligatoriness is not a characteristic that attaches to acts; obligation is something that attaches to persons.12