I TURN now to consider more precisely the general nature of what it is that moral laws bid us to do. Let me start by supposing that we accept the general account of rightness that has been offered, viz. that ‘right’ means ‘suitable, in a unique and indefinable way which we may express by the phrase “morally suitable”, to the situation in which an agent finds himself’. This situation contains two elements, what we may call the objective and the subjective element. The objective element consists of the facts about the various persons and things involved in the situation, in virtue of which a certain act would in fact be the best possible fulfilment of the various prima facie obligations resting on the agent. Suppose, for instance, that the situation is one in which none of the special obligations such as that of keeping a promise or of making reparation for an injury rests upon the agent, but only the responsibility for bringing as much good as possible into being. Then the act which would in fact produce the maximum good will be that which best fits the objective element of the situation, and will be in this respect the right act.

The subjective element consists of the agent's thoughts about the situation. These are as much parts of the total situation as are the objective facts. And the act which is morally suitable to them, i.e. the act which the agent, in view of his opinion about the situation, thinks will be the maximum fulfilment of obligation, will be in that respect right; while in order to be completely right an act will have to be suitable both to the objective and to the subjective element, which it can be only if the agent's opinions correspond to the realities of the situation.

It is clear that when we call an act right we sometimes mean that it suits the objective features of the situation, and sometimes that it suits the subjective features. And when people express different opinions about the rightness or wrongness of an act, the difference is often due to the fact that one of them is thinking of objective and the other of subjective rightness. The recognition of the difference between the two is therefore in itself important as tending to reconcile what might otherwise seem irreconcilable differences of opinion. But the question remains, which of the characteristics—objective or subjective rightness—is ethically the more important, which of the two acts is that which we ought to do. There are various considerations that tell in favour of the objective view. When we are in a difficult moral situation, what we want to know is not what act we think likely to produce certain results, but what act will produce certain results. And we are often driven to admit that we do not know what we ought to do, while if what we ought to do depended on what we think, we could always ascertain by reflection what we think, and therefore what we ought to do.

Again, moral laws are often expressed in a form which implies the objective view. It will be enough to take two instances. The moral law about promise-keeping is usually expressed in the form ‘keep your promises’, and a promise is usually expressed as a promise to effect a certain change in the situation, e.g. to restore a book to a friend. If we accept the moral principle as it is usually expressed, it follows that the act we ought to do is that which will in fact lead to our friend's reception of the book, and that if we so deal with the book that it reaches him we shall have done our duty, while if we so deal with it that it does not reach him we shall not have done our duty, even if in the first case we have dealt carelessly with it and in the second case carefully. This no doubt seems paradoxical, but I formerly thought that the paradox could be removed by saying that while the first act is the right act, the latter is the morally better act, since it is done with greater conscientiousness.

Again, the moral principle relating to the production of good is usually couched in the form ‘do that which will produce most good’, and this, like the usual formulation of the duty of promise-keeping, implies the objective view. It implies that not the act which the agent thinks will produce the most good, but that which will in fact produce the most good, is the act that ought to be done. And, as before, we might hold that we had removed the apparent paradox involved in calling a careless act which in fact produced most good the right act, by admitting that it is morally inferior to a more careful and conscientious act which in fact produced less good.

Nevertheless, I have come to hold the opposite opinion, that it is the subjectively right act that is obligatory. I owe my conversion to Professor Prichard's Lecture on ‘Duty and Ignorance of Fact’. His reasoning is so conclusive that I cannot do better than reproduce the main features of his argument. He starts by pointing out that the moral rules which are the generalization of our thought as to the characteristics which make particular acts which are our duties to be duties, habitually take the form of saying we ought to do so-and-so, to speak the truth, to carry out the rules of one's government, &c. He points out that the doing something which is implied in such formulations always means causing a change of state of some existing thing; that e.g. telling the truth means causing some one else to have a true opinion about the state of one's own belief about something. Further, when we reflect on the changes which we think we ought to bring about, we find that they are always changes which we can only bring about indirectly, by bringing about something else directly—e.g. we can cause another man to know our thoughts only by causing certain sounds, or making certain marks on paper, or the like. Thus the general form of a moral rule is ‘a man ought, or ought not, to bring about a thing of a certain kind indirectly’. But when we bring about something indirectly, the result is not wholly due to us. The only changes we can be said quite strictly to bring about are those which we bring about directly. Thus if a moral rule is stated in terms of ‘doing something’ and of ‘bringing about something’ in the strict sense, it will take the form, ‘A man ought to do such an act or acts as will cause a thing of the kind A to assume a state of the kind x’.

But this expression is elliptical in two respects. In saying that a man ought to support his indigent parents, we do not mean that he should support them whether he has or has not such parents, or that he should support them whether he can or not. Thus the full form of a moral rule will be: ‘When the situation in which a man is contains a thing of the kind A capable of having a state of the kind x effected in it, and when also it is such that some state or combination of states which the man can bring about directly will cause a state of the kind x in A, the man ought to bring about that state or combination of states’.

The formulation of any moral rule in this way, which we have seen to be the proper expression of our normal formulation of moral laws, implies the objective view of the basis of moral obligation. And this view, Professor Prichard points out, we find in two ways attractive. (1) We tend naturally to think that obligation does not depend on our thought about the situation, but on the nature of the situation itself; and when we try to resolve our doubt about what our duty is, we often try to do so by resolving our doubt about the facts. And (2) this view implies that if some action is a duty, it would actually bring about some state referred to in a moral rule, such as the recovery of a sick man, and would not merely be an act which we think likely to do so; and we value this implication because we should like to think that if we have done a duty, we have actually achieved some such change.

Yet this view has awkward consequences. (1) In order to know that some moral rule is applicable to me here and now, I must know (a) that the situation contains a thing of the kind A capable of having a state of the kind x effected in it, and (b) that it is such that some act that I can do would cause this A to assume a state of the kind x. Now (a) is not always fulfilled. I may not know whether my parents are in difficulties, or whether a man I meet is ill. And (b) is never fulfilled; I never know nor can come to know that some state which I can bring about will produce an effect of the kind x, though I may have reason to think it. Thus if duty be such as the objective theory conceives it to be, I can never know that I have any particular duty, or even that any one has ever had or will ever have a duty.

It is worth while to note in passing just what is proved and what is not proved by this argument. In constructing or in following a geometrical proof, we never know that we have before us a triangle, for instance; but we treat the figure before us as if it were a triangle, and we come to know that if it were, it would have certain properties. Similarly, we never know that an act we could do would produce a certain effect, but we may think that it would, and may know that if it would, it would be our duty to do it; and we might proceed from this to generalized moral rules, which would be hypothetical in character; e.g. ‘if you can ever produce a true opinion in the mind of some one else as to what you think, you ought to do so.’ Thus the objective view is not fatal to the possibility of knowing moral rules. But it is fatal to the possibility of recognizing particular duties incumbent on us here and now, since we can never know, for instance, that we can produce a true opinion as to our thought in any one else's mind.

Further consequences of the objective view are (2) that we can never do a duty because it is a duty, since this must mean ‘because we know it to be a duty’, (3) that some past act of mine may have been morally obligatory though I believed it was one I ought not to do, and (4) that I may do some act which is obligatory, though I do not even suspect that it will have the effect which renders it a duty.

These difficulties all arise from supposing that an act is made my duty by the objective facts of the situation. The only alternative is to suppose that it is made a duty by the subjective facts of the situation, viz. by my state of knowledge or opinion about the facts of the case. The most obvious way of describing the state of mind which makes an act my duty is to say that it is my thinking certain things likely; e.g. that a man near me has fainted and that my shouting would revive him. This view has the advantage of making it possible to discover our duties; for we can always or almost always know what it is that we think likely. This view also has the advantage of making it possible to do a duty knowing it to be a duty. It is, however, open to this objection, that since the question whether a certain act is a duty depends not merely on our thinking that there is some probability of its having a certain effect, but on our thinking it at least in a certain degree likely that it will, ‘there will be border-line cases in which I shall be unable to discover whether the degree to which I think the act likely to confer a certain benefit is sufficient to render it a duty’.1 Thus even on this view I may have a duty without being able to discover that I have it—although there are other duties which we can discover ourselves to have.

The only way to choose between the objective and the subjective view is to ask ourselves which corresponds better with what we actually think to be duties. Professor Prichard points out two ways in which our thought seems to imply the objective view. (1) We often think without question both that the situation contains something in a certain state, and that some action we could do would produce a change in it of a certain kind, and then we think without question that we ought to do the action. Here we seem to be implying that what makes us bound to do the act is not our opinion, but the fact, that the situation is of a certain kind and that the act would have a certain effect in that situation. (2) We often seek to change the mind of some one else about a duty by trying to convince him that he is wrong about the facts. ‘Thus, where A thinks he ought to vote for X rather than Y, B may try to convince A that he ought to vote for Y by arguing that X and Y will, if elected, act otherwise than as A expects.’2 Here we seem to imply that what A ought to do depends not on how he thinks X and Y would behave, but on how they would in fact behave.

On the other hand, much of our ordinary thought is in conflict with the objective view. (A) One instance will suffice to show this. Suppose one is driving a car from a side-road into a main road; the question arises, ought one to slow down before entering the main road.

‘If the objective view be right, (1) there will be a duty to slow down only if in fact there is traffic; (2) we shall be entitled only to think it likely… that we are bound to slow down; and (3) if afterwards we find no traffic, we ought to conclude that our opinion that we were bound to slow down was mistaken.… Indeed the objective view is in direct conflict with all the numerous cases in which we think without question that we ought to do something which we are thinking of as of the nature of an insurance in the interest of some one else.’3

(B) ‘The extent to which our ordinary thought involves the subjective view is usually obscured for us by our tendency to think that the terms “likely” and “probable” refer to facts in nature. For we are apt, for instance, to express our thought that some one has probably fainted, and that shouting would probably revive him, by the statements: “He has probably fainted” and “Shouting would probably revive him”. We are then apt to think that these statements state the existence of certain facts in nature called probabilities.’4

But there cannot be probabilities in nature. Whatever the precise nature of the fact expressed by the statement ‘X has probably fainted’ may be, the fact must consist in our mind's being in a certain state. Once this is realized it becomes clear that most of our ordinary thought involves the subjective view.

(C) Even when we try to change some one else's mind about a duty, we do not really imply the objective view. ‘This is shown by our thinking that when our attempt to change his opinion about the facts is over, then, whether we have or have not succeeded, the question whether he is bound to do the action will turn on the nature of his opinion about the facts.’5 We are not really trying to convince him that his duty is not what he thinks it is, but, thinking that his doing of his duty would result in very bad consequences, we try to put him into a different state of opinion on the facts, a state of opinion in which an act which we think will have good consequences will have become his duty, because his opinion about the facts has changed.

Thus on the whole the subjective view is more in agreement with our ordinary thought than the objective. Yet it is exposed to various difficulties, of which the chief are (1) that on this view knowledge of the existence of border-line cases precludes us from thinking that we can always discern our duties, and (2) (a more fundamental difficulty) that this view represents the duty of doing some action as depending not on the fact that the action would have a certain character, but on our thinking it likely that it would. To maintain this seems impossible, and we seem therefore to be in an impasse.

Professor Prichard now turns to consider a difficulty which is common to both views, and which if well founded will lead us to modify both. We have hitherto assumed that an obligation is an obligation to do some action, i.e. to produce some change in something. But we must ask whether this is true. An obligation must be an obligation to be active, and not to be affected, in a particular way. To say that an obligation is always an obligation to do some action implies that there is a particular kind of activity, distinct from other activities such as thinking or imagining, whose nature is to be the bringing about of something. But there is no type of mental activity of which the general nature is to be the producing of a change in some physical object, such as the moving of a hand or a foot. On the contrary, if we ask how we move a hand or a foot, the natural answer is that we do so by setting ourselves to do so. There is a type of mental activity of which the generic nature is to be the setting oneself to effect a change in a physical object, and of which setting oneself to move a hand or a foot is an instance. The change in the physical object, when it follows, is merely the result—the intended result, of course—of the mental activity.

Again, if we ask what we mean when we say ‘I can make a loud noise’, we find that what we mean is not that there is a special kind of activity of which we are capable consisting in bringing about a loud noise, but rather that a special kind of activity of which we are capable, consisting of setting ourselves to make a loud noise, would have a loud noise as an effect.

Two conclusions follow: (1) that the true answer to any question of the form ‘can I do so-and-so?’ must be ‘I don't know’. This is obvious in certain cases. Obviously I cannot know whether I can succeed in threading a needle. But even where we usually assume that we can effect certain changes, such as the movement of a hand or a foot, we can never know that we can, since we may have become paralysed since we last tried. And (2) whatever we are setting ourselves to do, we never know that we are doing what we are setting ourselves to do. The mental activity may be of exactly the same kind whether we do or do not by performing it bring about the bodily change intended.

‘As regards an obligation, the moral is obvious. It is simply that, contrary to the implication of ordinary language and of moral rules in particular, an obligation must be an obligation, not to do something, but to perform an activity of a totally different kind, that of setting or exerting ourselves to do something, i.e. to bring something about.’6

The question now arises whether the substitution of ‘setting ourselves to bring about some result’ for ‘bringing about some result’ makes it easier to decide between the objective and the subjective view. Professor Prichard points out that in one respect it does. ‘For once it has become common ground that the kind of activity which an obligation is an obligation to perform is one which may bring about nothing at all, viz. setting ourselves to bring about something, we are less inclined to think that, for there to be an obligation to perform some particular activity, it must have a certain indirect effect. To this extent the modification diminishes the force of the objective view without in any way impairing that of its rival’.7 But the main difficulty of the subjective view remains, that it represents ‘the obligation to do some action as depending not on the fact that the action would have a certain character, if we were to do it, but on our thinking it likely that it would’.8

This difficulty Professor Prichard removes in the following way:

‘We are apt’, he says, ‘to think of an obligation to do some action as if it were, like its goodness or badness, a sort of quality or character of the action.… And this tendency is fostered by our habit of using the terms “right” and “wrong” as equivalents for “ought” and “ought not”. For when we express our thought that we ought, or ought not, to do some action by saying that the act would be right, or wrong, our language inevitably implies that the obligation or disobligation is a certain character which the act would have if we were to do it.… And when we think this, we inevitably go on to think that the obligation or disobligation must depend on some character which the act would have. But, as we recognize when we reflect, there are no such characteristics of an action as ought-to-be-doneness and ought-not-to-be-doneness. This is obvious; for, since the existence of an obligation to do some action cannot possibly depend on actual performance of the action, the obligation cannot itself be a property which the action would have, if it were done. What does exist is the fact that you, or that I, ought, or ought not, to do a certain action, or rather to set ourselves to do a certain action. And when we make an assertion containing the term “ought” or “ought not”, that to which we are attributing a certain character is not a certain activity but a certain man. If our being bound to set ourselves to do some action were a character which the activity would have, its existence would, no doubt, have to depend on the fact that the activity would have a certain character, and it could not depend on our thinking that it would. Yet since, in fact, it is a character of ourselves, there is nothing to prevent its existence depending on our having certain thoughts about the situation and, therefore, about the nature of the activity in respect of the effects. Indeed, for this reason, its existence must depend upon some fact about ourselves. And while the truth could not be expressed by saying: “My setting myself to do so-and-so would be right, because I think that it would have a certain effect”—a statement which would be as vicious in principle as the statement: “Doing so-and-so would be right because I think it would be right”—there is nothing to prevent its being expressible in the form: “I ought to set myself to do so-and-so, because I think that it would have a certain effect.”9

This last part of Professor Prichard's theory, while both true and important, is not necessary for the saving of the subjective view. For even if we think that there is a character of rightness that attaches to an activity, it will, on the subjective view as now restated, be a character which belongs to the activity not because of the activity's being thought to have a certain character but because of its actually being of a certain character, the character of being the setting oneself to bring about a certain effect. This character it actually has, and there is in principle no reason why it should not be the ground of a further character of rightness. Thus, even apart from Professor Prichard's last contention, the subjective view is safe from the objections which seemed fatal to it.

There is another mode of argument by which we may, I think, satisfy ourselves of the truth of the subjective view. It might be agreed, I believe, that the act which a man in any situation ought to do is that which it would be reasonable for him to do if he wanted to do his duty in that situation. And I think we can on reflection discover two possible but wrong answers to the question what it would be reasonable for him to do. (1) There may be some change, by setting himself to bring which about he would in fact produce the result the production of which would be objectively right, e.g. would succeed in returning to a friend a book he had promised to return; a change, however, which no human foresight could foresee to be about to have this effect—e.g. a book despatched in the most careless way may by the vigilance of the Post Office or of some individual unexpectedly reach its destination. And it might happen that owing to unforeseen circumstances the careful despatch of the book might fail to lead to its reaching its destination. Yet if no human foresight can foresee these facts, no one would say that it was reasonable for a man who wanted to do his duty by his friend to despatch the book carelessly, since the successful result of this neither is nor could be foreseen by the sender. (2) There may be circumstances which the agent does not foresee, but which a wiser or better-informed person might foresee, which would in fact cause a certain activity of the agent's to produce a certain result, the production of which would be objectively right. Yet it would not be reasonable for the agent, if he wished to do his duty, to perform such an activity, since ex hypothesi he neither knows nor thinks the activity would have this result. The fact that other people might know or think this has no tendency to make it reasonable for him to act thus. What he ought to set himself to do, then, is neither that which will in fact produce the result in question, nor that which in the judgement of better-informed people is likely to produce it, but that which he thinks likely to produce it.

Yet we do not think that an agent should necessarily forthwith perform that activity of self-exertion which in his present state of opinion about the facts seems to him likely to produce the objectively right result. We often raise the problem what we ought to do, some time before the time at which whatever action is to be taken must be taken. In such a case the agent should have before his mind, as the ideal, that self-exertion which would in fact produce the right result. This, however, he cannot know; and so he must fall back on a secondary ideal, viz. that self-exertion which on the fullest consideration that he can give to the matter within the time at his disposal would seem most likely to produce the result. And he should set himself to act only when either the time-limit is on the verge of arriving, or he has reached the point of thinking that no further consideration would enable him to judge better of the circumstances and of the probable effects of alternative exertions. What he does after such consideration may reasonably be expected to be nearer to the objectively right act than what he would do after a first hasty consideration.

At first sight it might seem that in substituting ‘setting himself to bring about a certain result’ as that which the agent ought to do rather than ‘bringing about a certain result’, we have, contrary to our earlier conclusion, introduced motive into the structure of that which we ought to do. For ‘to set oneself to bring about a certain result’ seems to be perilously near to being actuated to action by the desire to bring about that result; i.e. by a certain motive. Yet the two things are quite different. Suppose we imagine, for instance, that in some situation none of the special responsibilities such as that of keeping a promise is in question, and the only responsibility that arises is that of setting ourselves to produce as much good as possible. What we ought to do, then, strictly speaking, is just to set ourselves to produce this. And that is different from doing so from any special motive, such as sense of duty or benevolence. For we may set ourselves to produce the result from any one of a variety of motives. We may think, for instance, that in setting ourselves to produce the greatest good we are also likely to acquire a good reputation for ourselves, or to get in a high degree the pleasure of having a good conscience; and either of these may be our motive. But a self-exertion which may proceed from any one of several motives cannot be identified with self-exertion from any one motive; and if it is the self-exertion that is our duty, it is not the self-exertion from any particular motive that is our duty. Or again, suppose that the main responsibility in some situation is that of fulfilling a promise, e.g. of paying a debt. We may set ourselves to do this either from the motive of sense of duty, or from the wish to avoid a legal action against us, or from the wish to injure our creditor by putting him in possession of more money than is good for him. A self-exertion which may arise from any one of these motives is not identical with a self-exertion from the sense of duty, and it is the former and not the latter that is our duty.

We are now in a position to see that two views which we have rejected owe their plausibility to an ambiguity in the notion of right action. A right action means in general one that is morally suitable to the situation. But an action may be described as morally suitable to the situation either because it is suitable to the objective elements in the situation, i.e. because it is that which would in fact produce the result which we think we ought to aim at; or because it is suitable to the subjective elements in the situation, i.e. to our thoughts about the situation and about the probable results of alternative actions. Both actions are undeniably, in different respects, right; and because the former is right in a respect in which the latter is not, it is easy to fall into the supposition that it is it that is obligatory. But it is also true that the latter is right in a respect in which the former is not, and we have seen good reasons for holding that it is it that is in fact obligatory.

Again, an action done from a certain motive is undeniably right, or morally suitable to a situation, in a sense in which a mere action, irrespective of its motive, is not. Where, for instance, the only responsibility is that of producing a maximum of good, it is more completely fitting that we should set ourselves to produce a maximum of good, from the sense of duty to do so, than that we should barely set ourselves to produce a maximum of good, it may be from some unworthy motive. And where the fulfilling of a promise is the main responsibility, it is more completely fitting that we should set ourselves to fulfil it from the sense of duty, than that we should barely set ourselves to fulfil it. And since the action from a certain motive is more fully fitting, morally, than the bare action, more completely right, it is easy to fall into the supposition that it is our duty. But we have seen good reasons for holding that this view, although it is one into which we easily fall, is not the true view. Both the view that it is our duty to produce certain results, and the view that it is our duty to act from certain motives, are natural enough perversions of what seems to be the true view, that it is our duty to set ourselves to produce certain results. It is sometimes said that it is neither results nor motives but intentions that make actions right or wrong, and this is almost true. There is a certain danger in laying the stress on intention, since intentions may remain idle; but it would be true to say that the nature of what is intended in an act is what makes the act right or wrong.

The most important point, I think, which emerges from Professor Prichard's discussion is that the only thing to which a man can be morally obliged is what I will call a self-exertion, a setting oneself to effect this or that change or set of changes. He cannot be obliged to perform an ‘act’, in the ordinary sense. For the noun ‘act’, as we ordinarily use it, stands for a complex thing; viz. the causing of a certain change by setting oneself to cause it; and this includes as an element in it the occurrence of the change. It would be absurd to say ‘I killed him, and in consequence he died’; to say ‘I killed him’ includes the statement that he died. It would not, indeed, be absurd to say ‘I hit him, and in consequence he died’, but it would be absurd to say ‘I hit him, and in consequence he suffered a blow’. Now the occurrence of the bodily change involved in the use of such words or phrases as ‘kill’, ‘hit’, ‘tell the truth’ cannot even be part of what is right or of what is wrong. This follows directly from the fact that if a man had, without knowing it, become paralysed since the last time he had tried to effect the given type of change, his self-exertion, though it would not produce the effect, would obviously be of exactly the same character as it would have been if he had remained unparalysed and it had therefore produced the effect. The exertion is all that is his and therefore all that he can be morally obliged to; whether the result follows is due to certain causal laws which he can perhaps know but certainly cannot control, and to a circumstance, viz. his being or not being paralysed, which he cannot control, and cannot know until he performs the exertion.

Now, assuming that the only thing that can be obligatory or disobligatory is a self-exertion, it can be seen that the only thing to which there can be a prima facie obligation, or to which some one else can have a claim, is also a self-exertion. For only those things are prima facie obligatory which, if there are no more pressing prima facie obligations, are actually obligatory. No one, for instance, can, have a claim to have his life saved by me; the most that any one can have is a claim to my self-exertion to that end.

At the same time, it is very natural that in our ordinary thought we should think that it is actions and not self-exertions that are right or wrong. For (1) the most direct results of the self-exertion, those within the agent's own body, have followed so constantly, within his experience, upon the self-exertion, that he not unnaturally thinks of the self-exertion+its most direct results as if they formed one single event; and (2) where the self-exertion produces its desired result, there is a further connexion between it and its result, over and above that which there usually is between a cause and its result, viz. that the one is just the attempt to produce the other.

Now we may distinguish several different self-exertions which might have some claim to be considered right, or what the agent ought to do:

(A) The self-exertion which is morally most suitable to the objective circumstances, in the sense of ‘the circumstances other than the agent's own state of knowledge or opinion’; e.g., in a case where only beneficence is in question, the self-exertion which would in fact benefit humanity most.

(B) The self-exertion which is morally most suitable to the agent's state of mind about the circumstances, in which there may be included ignorance and false opinion as well as knowledge and true opinion; i.e. the self-exertion which would be morally most suitable if the circumstances were such as he supposes them to be.

(C) The self-exertion which he thinks to be morally most suitable in the circumstances as he takes them to be.

(B) may differ from (A), in consequence of a divergence from the truth in the agent's opinion about the circumstances. (C) may differ from (B), in virtue of a divergence from truth in the agent's opinion as to what is morally suitable to the supposed circumstances. The one difference is due to a divergence from truth on a non-moral question, the other to a divergence from truth on a moral question.

All these acts are in different senses right or morally suitable—the first suitable to the objective circumstances, the second to the agent's opinion on the non-moral question, the third to his opinion on the moral question. Which of them is the action that the agent ought to do? Professor Prichard's argument seems to me to have shown that it is not the first. No one would say that the driver of a car had done right in driving fast round a corner if he thought there might quite probably be a car meeting him but in fact there were none. But the question may be asked, should we not go a stage farther and say that it is rather the third than the second that is the right act, since that alone is suitable to the agent's complete state of opinion, including his opinion on the moral as well as on the non-moral question. The suggestion is at first sight open to the objection that we should be saying that act (C) is the right act for the agent to do, simply because he thinks it is the right act. It is clear on epistemological grounds that nothing can have a character simply by being thought to have it; but we are not suggesting that act (C) has a certain character by being thought to have that character. The agent thinks it is the act suitable to, or harmonious with, his opinion on the question ‘what are the circumstances?’; and in consequence it is the act suitable to his opinion on the question ‘what is the act suitable to the circumstances?’. Thus the act has one suitability by being thought to have another. If our suggestion thus escapes the epistemological objection, there seems to be no objection to saying that it is the suitability that act (C) has, and not that which act (B) has, that makes an act one's duty. For, just as it was felt to be paradoxical to say ‘you ought to do the act which will produce certain consequences, because it will produce these consequences, and though you think it will not’, so it is paradoxical to say ‘you ought to do the act which is most suitable to your opinion about the circumstances, because it is the most suitable to your opinion, and though you think it is not’.

It is only by thus distinguishing different rightnesses or suitabilities and by making duty depend on the last of the three, that we can do justice to a thought which is inseparable from the thought of duty. This is the thought that anything that we ought to do must be something that we not only can do, but can do with the knowledge or at least the opinion that it is our duty. Suppose that we blame a man for not doing his duty, and he replies ‘but I did not know or even think it to be my duty, and therefore could not do it with the knowledge or even the opinion that it was my duty’; it would be a poor response to say ‘no, but you might have done it from some quite different motive’; for clearly a man who had acted from a different motive would have been more blameworthy than the man who did what he honestly thought was his duty. Now, when a man's opinion about the circumstances is mistaken, he cannot do act A with the knowledge or even the opinion that it is right in sense A. And when his moral insight is at fault, he cannot do act B with the knowledge or even the opinion that it is right in sense B. But even if both his opinion about the facts and his moral insight are at fault, he can always do act C with the knowledge that it is right in sense C, i.e. that it is the act which is most suitable to his opinion on the question what act is most suitable to the circumstances as he takes them to be.

There is another consideration which tends to show that it is what is right in this sense that we think an agent is obliged to do. The notion of obligation carries with it very strongly the notion that the non-discharge of an obligation is blameworthy. Now suppose that of two men one does that which he mistakenly believes to be his objective duty, and the other does that which is his objective duty, believing it not to be so, we should regard the former as at least less blameworthy than the latter; and in fact we should not regard the former as directly blameable for the act, but only, if at all, for previous acts by which he has blunted his sense of what is objectively right.

It may at first sight seem dangerous to admit this double dose of subjectivity into the answer to the question ‘what is my duty?’, by making it depend on my opinion as to what is morally suitable to what is in my opinion the state of the facts; and that is, I think, the strongest apparent objection to which this account is exposed. But ‘subjective’ is notoriously one of the vaguest of philosophical terms, and we must ask ourselves whether the account given above is a subjective account in any objectionable sense. The kind of subjective account which it seems to me important to avoid is one which says that acts are made to have some moral characteristic by being thought to have it, or (which comes to the same thing) that the opinion that an act has a certain characteristic is no more true and no more false than the opinion that it has not. Now, we are not giving such an account of ‘right’ in any of the three senses we have distinguished. In any particular situation in which a particular man is placed, there is one act which, if he had complete knowledge about the circumstances and a completely correct moral insight, he would see to be right in the first sense. There is no suspicion, even, of subjectivity in what is right in this sense. Secondly, suppose him to be mistaken about the circumstances; there is an act which is right in the second sense, in the sense of being appropriate to his opinion about the circumstances. That act is not made right in this second sense by being thought to be so; it bears the same sort of relation to the supposed situation as the first act does to the actual situation; the same kind of harmony exists in the one case as in the other; the harmony is not created by being thought to exist, it exists independently of the agent's thought about it. Thirdly, the agent may be mistaken in his moral judgement of his duty in the supposed situation; but so long as he thinks as he does, the act in which he acts on his conviction has the same sort of harmony with his conviction as an act in which a man acts on a correct conviction has with that conviction, a harmony which is not created by his opinion but is there for all to apprehend.

Error in this region arises, it would seem, only if we confuse, as we often do, one kind of rightness with another. Although, if one acts with imperfect moral insight or in accordance with insight that is morally correct but based on an incorrect view of the facts, one does what is right in one sense, and in what is from one point of view the most important sense, since it is that to which praise is appropriate (for a man is more to be blamed for acting against his convictions than for doing contrary to his convictions an act that is right in the first sense), no one should be content to have done so. He should be rather ashamed of having done an act which owes its rightness to its harmony with incorrect moral insight or incorrect opinion about the facts, and should realize that it would have been better if he could have amended his moral insight or his opinion of the circumstances, or both, so that in doing what was right in the third sense he would have also been doing what was right in the second or even in the first. If to act in accordance with one's conviction is always, in one sense, to do one's duty, it remains true that one's conscience may be very much mistaken and in need of improvement.

We may now, in the light of this discussion, consider the question of the relation of the morally good act to the right act. Is a morally good act necessarily right? Is a right act necessarily morally good? Or are the two characteristics quite independent?

If by a right act we mean an objectively right act, i.e. the act which out of all those open to a particular agent in particular circumstances will in fact produce the maximum fulfilment of the claims that exist against him, we must maintain the complete non-dependence of moral goodness and rightness upon one another. For an action's being morally good depends mainly10 on the motive from which it is done, and the goodness of the motive neither guarantees nor is guaranteed by the nature of the results that the act actually produces. Take, for instance, the case in which the motive is the sense of duty, i.e. the desire to do one's duty+the thought that a certain act is one's duty. This thought in turn rests on the thought that the act will produce certain results; and the thought that it will do so furnishes no guarantee that it actually will do so. And conversely, of course, the fact that it will do so furnishes no guarantee that it was done from the thought that it would do so and the thought that therefore it was our duty to do it. Thus a morally good act may be objectively wrong, and an objectively right act may be morally bad, or indifferent.

It is important to maintain this, as a corrective of the view that, so long as we act conscientiously, all is well. Conscience, when not accompanied by clear insight into the situation, and by foresight of the effects which acts are likely to have, has often led to acts which objectively considered were deplorably wrong, which failed lamentably to fulfil the prima facie obligations of the agent. On the other hand, we are not bound to think that there is no connexion between moral goodness and objective rightness, that a morally good act is no more likely to be objectively right than a morally bad or indifferent act, or an objectively right act no more likely to be morally good than an objectively wrong act. For the motive of a morally good act is either the sense of duty or the desire to bring some particular good thing into being, as being good, and an act so motived is far more likely to conform to objective duty than one of which the motive is either self-interest or malevolence.

Again, the act which is right in the first of the two subjective senses, the act which would be right if the situation were as the agent supposes it to be, is not necessarily morally good, nor vice versa. For on the one hand, such an act may be done with a bad motive, and will not then be morally good; and on the other hand, an act done with a good motive, and therefore morally good, may through failure of moral insight not be the act which would actually be right in the circumstances as the agent supposes them to be.

The relation between rightness in the third sense, conformity to the agent's thought on the question as to what is right in the circumstances, and moral goodness, cannot be stated so simply. The motive of a morally good act may be either the sense of duty or the wish to bring some good thing into being, as being good. In the first case the morally good act is necessarily right in the third sense; it is the act which harmonizes with the agent's thought about his duty. In the second case, it is not so. The act may be done without the agent's thinking about his duty, and then the act cannot be said to harmonize with the agent's thought about his duty, since he has no such thought. Or again he may think that act A is his duty, but do act B from some other good motive (e.g. that of kindness to an individual), and in such a case his act has some moral goodness, but does not harmonize with his thought about his duty, and is not right in the third sense.11

And conversely, rightness in this sense never guarantees moral goodness. For an act may be the act which the agent thinks to be his duty, and yet be done from an indifferent or bad motive, and therefore be morally indifferent or bad.