FROM the question whether the characteristics right and obligatory are definable I turn to the question what are the grounds of rightness and of obligatoriness. As I have already pointed out,1 to any theory which says ‘so-and-so is the essence of rightness’ there will correspond a possible theory that that same thing is the ground of rightness, rightness itself being treated as indefinable, or definable in some other way. And the example of Bentham2 is enough to show that it is very easy to fail to distinguish between the two views. Logically, the two views are entirely different; but in their ethical consequences they will be the same. Whether you say ‘so-and-so is the essence of rightness’ or ‘so-and-so is the ground of rightness’, you will be led to the same ethical judgements on any act or type of acts; it will be in virtue of their possession or non-possession of the characteristic ‘so-and-so’ that you will judge of the rightness or wrongness of acts.
I remarked before3 that the theories which specify this or that as the ground of rightness are in general more plausible than those which specify this or that as the essence of rightness. And in one respect this is so. For when such a characteristic as ‘conducing to life’ or ‘being approved by the individual judger’ or ‘being approved by the majority of some body of men’ is put forward as the essence of rightness, we have only to examine what is in our mind when we say such and such an act is right, to see that, however closely the characteristic in question may be connected with rightness, it is not the very meaning we have in mind when we assert the rightness of an act. From another point of view, some at least of the ground-theories are less plausible than the essence-theories. For, while, until we begin to reflect carefully on our meaning when we predicate rightness, it may seem plausible to say that rightness is just the being generally approved, for instance, it is very unplausible to say that the being approved is the ground of an action's having a quite different characteristic, a characteristic of its own, that of being right.
But we had better consider the ground-theories methodically, as we considered the essence-theories. And first we may consider the evolutionary or biological theories. An easily detected characteristic of them is their instability, their tendency to turn, on examination, into theories of a different type from that to which they appear to belong. Take, for instance, Spencer's chapter on good and bad conduct, which he does not distinguish from right and wrong conduct. In fact I think it is clear that he means rather right and wrong than good and bad conduct. For he entirely ignores motive as a source of goodness and badness; but I think almost every one must agree that motive is at least the main factor in making action morally good or bad, while opinions differ on whether it has anything to do with making action right or wrong. Now Spencer's first answer to the question what makes action right is that it is its being relatively more evolved,4 and ‘most evolved’ he explains as meaning simultaneously achieving ‘the greatest totality of life in self, in offspring, and in fellow men’.5 This, so far, is a purely biological view. ‘Right’ is defined by reference not to any psychological state but simply to life. But Spencer immediately goes on to ask, ‘Is there any postulate involved in these judgements on conduct? Is there any assumption made in calling good the acts conducive to life, in self or others, and bad those which directly or indirectly tend towards death, special or general? Yes; an assumption of extreme significance has been made—an assumption underlying all moral estimates.… Yes, there is one postulate in which pessimists and optimists agree. Both their arguments assume it to be self-evident that life is good or bad, according as it does, or does not, bring a surplus of agreeable feeling.’6 He expressly says that ‘by those who think life is not a benefit but a misfortune, conduct which prolongs it is to be blamed rather than praised’.7 And he sums up by saying ‘if we call good the conduct conducive to life, we can do so only with the implication that it is conducive to a surplus of pleasures over pains’.8 Spencer's position, then, is this: action which conduced as much as possible to the increase of life would not be right unless it conduced as much as possible to the increase of pleasure; action which conduced as much as possible to the increase of pleasure would be right whether or not it conduced as much as possible to the increase of life. Clearly it is conduciveness to pleasure that is for him the real ground of rightness. This is his fundamental ethical theory. But he holds, on grounds with which we need not concern ourselves, that life always contains a surplus of pleasure over pain, and that conduciveness to life and conduciveness to pleasure always go together, so that he can say right action is always that which conduces to life, though he does not really think that it is this that makes it right. His fundamental theory turns out to be universalistic Hedonism, or Utilitarianism; the apparently biological theory turns out to be really a psychological theory. And I believe this to be in the long run true of evolutionary ethics in general, so that it need not be examined as a separate form of theory regarding the ground of rightness.
We turn then to the psychological theories. The psychological theories about the essence of rightness we divided into the reaction (or attitude) theories and the causal theories; we may consider the psychological theories about the ground of rightness under the same two heads. The reaction theories we may divide into those that rest rightness on the reaction of the individual judger, those that rest it on the reaction of a majority of men or of some class of men, and those that rest it on the reaction of the agent (who may or may not be identical with the judger). The first of these theories will be the theory that because an individual contemplating an act reacts to it with the emotion of approval, therefore the act in itself has the characteristic of rightness. This theory is open to at least three objections, any one of which is fatal. (1) In the first place, suppose that an act is contemplated by two observers one of whom reacts with approval and the other with disapproval. Then, since the act itself is the same act, we can only suppose either that they are not, both of them, grasping the whole nature of the act (e.g. one may be contemplating it simply as an act of promise-breaking and the other as an act productive of great pleasure), or that while both are grasping its whole nature, idiosyncrasies of the two men cause them to react in different ways to it. In other words, their reaction is due not directly to the nature of the act, but to two things in them, their opinions about the constitutive nature of the act, and the idiosyncrasies which lead them as a result of these opinions to react with approval or disapproval. And it is surely impossible that any quality of the act itself can be founded on a reaction which is itself founded not directly on the nature of the act but on the opinions and idiosyncrasies of individual contemplators of it.
(2) If an act is right because it is approved by A and wrong because it is disapproved by B, the same act will be in fact right and wrong. But while we might agree that the same act may be in some respects right and in others wrong, we do not suppose that the same act can be in fact right on the whole and wrong on the whole. To think this would be to put an end to all ethical judgement. The corresponding essence-theory of rightness put an end to ethical discussion because it implied that two men who respectively call an act right and wrong are not contradicting one another.9 The ground-theory puts an end to discussion because it implies that the two men are contradicting one another but nevertheless both are right.
(3) I have not so far urged that the emotion of approval presupposes a judgement that the act is right, but merely that it presupposes an opinion about it in its constitutive character, e.g. that it is an act of promise-keeping or that it is an act productive of great pleasure. But I think that in fact the emotion of approval presupposes a judgement that the act is right. About this there is perhaps room for difference of opinion. Some may think that the emotion comes first and the judgement second. I am willing to admit that there may be cases in which our first reaction to an act is not an ethical opinion about it, but a non-ethical emotion of disgust, perhaps simply due to its foreignness to our habitual ways of acting and thinking, and that this may through lack of reflection lead to the opinion that the act is wrong. But it seems to me clear that a genuine emotion of ethical disapproval presupposes a judgement that the act is wrong, and not the other way about. If this be so, the situation we are asked to believe in is this: ‘A spectator forms a certain view of the constitutive character of the act. In consequence of that he judges it to be wrong. In consequence of that he feels the emotion of disapproval. And in consequence of his doing this the act really is wrong.’ Thus his opinion that it is wrong is made indirectly the ground of its being wrong. But it is surely clear that if his opinion is incorrect the act is not wrong; and if his opinion is correct, it is correct because the act is wrong already; the act is not wrong because he has the opinion that it is.
If, on the other hand, it be suggested that the vital thing is not an emotion of ethical disapproval presupposing the opinion that the act is wrong, but merely an unfavourable emotion of disgust, it would be absurd to hold that in consequence of this the act has the ethical quality of wrongness. The only natural conclusion would be that acts are not right or wrong, but that some of them happen to disgust us and others do not. And if any one is willing to adopt this view, I do not think that he can be reasoned out of it. But apparently very few people are willing to adopt this view. Those who do not think wrong the things that most people think wrong, at least think wrong the things that most people think right.
The next theory to be examined is that an act is right because the majority of men, or of some class of men, feel the emotion of approval towards it. This ‘public’ theory is exposed to the first and the third objection which I raised to the corresponding ‘private’ theory—the objections arising from the difficulty that there is in supposing that any act can have an objective character of rightness in consequence of the reaction of individuals to it; to this difficulty the fact that many individuals and not one are involved makes no difference. The reaction of the many is just as much coloured by their idiosyncrasies as is the reaction of one individual by his. But in addition this view is exposed to the further difficulty that we often judge an action to be right when we do not for a moment suppose that there is a majority which either is actually approving of it, or would approve of it if it contemplated it. A moral pioneer, or a man who is being generally blamed for an act which he regards as justifiable, no doubt thinks that most people, if they knew all the circumstances and judged truly of them, would agree with him. But if he does, he does not think that his act is right because they would do so, but that they would do so because his act is in fact right.
We come now to the third reaction-theory—that an act is right because the doer of it approves of it. Those who hold this view are, I think, more likely to mean by approval the doer's thinking the act right, than his having a mere emotion of satisfaction not presupposing this thought. The question then comes to be, Is an act made right by the agent's thinking it right? On general grounds there would seem to be a fatal objection to this suggestion; it appears perfectly impossible that anything can be necessitated to have any attribute merely by being thought to have it;10 it does not seem possible that an act could have the characteristic of rightness by being thought to have it, any more than anything else could have any other characteristic by being thought to have it. We may, however, be faced with the indignant protest ‘Can it really be right for a man to do what he thinks wrong, or wrong for him to do what he thinks right?’ The view underlying such a protest must be one or other of two views; it must either be the view that the being thought by the agent to be right is the sole condition of an act's being right, or the view that in addition to having the other conditions of being right an act must, in order to be right, be thought by the agent to be so. On reflection it can be seen that neither of these contentions can be true. We must adhere to the general principle that a thing's being thought to have a certain characteristic cannot be either the sole condition or one of the conditions of its having that characteristic. To assure ourselves of this, we need only consider the fact that opinion must be either true or false. Now if the opinion that an act is right is false, the act is plainly not right and therefore cannot be right in virtue either of being thought to be so or of anything else. And if the opinion is true, it is true because the act is already right independently of our opinion about it.
It might however reasonably be suggested that the characteristic of being thought to be right confers on an act another characteristic; that, for instance, the character of being thought to be objectively right confers on an act the characteristic of being subjectively right, to use language which Sidgwick has used before us.11 This suggestion escapes the general objection which has been drawn above from the relation between opinion and fact. A thing may perhaps have one characteristic by being thought to have another. For instance, an imagined future state of affairs may have the characteristic of attractiveness by being thought to be such that it will yield a great balance of pleasure to the agent; it is plain that its attractiveness depends not on the characteristics it will actually have, but on those which it is thought that it will have.
It is plain, however, that if the being thought to be objectively right is made the ground of an action's being subjectively right, ‘subjectively right’ must be given a meaning other than the being thought to be objectively right; whether such a sense can be found, I will inquire later.12
I used to think that the protest I have imagined to be made could be effectively met by the distinction which should be drawn between the rightness of an act and its moral goodness. If we are asked ‘can it be right to do what you think wrong?’, our answer, it seemed to me, should be ‘yes, it can be right, since that cannot be affected by your thinking it to be so or not; but it cannot be morally good, since moral goodness depends13 on the goodness of the motive, and your motive in doing what you think wrong cannot be good. And similarly it can be wrong to do what you think right, but when it is, it may nevertheless be morally good, and will be so if you not only do what you think right but do it because you think it right. To say that when we fail to do what we think right our action is not morally good, and that when we do what we think right, because we think it right, our action is morally good, covers all the truth that lies behind the loosely worded protest “It cannot be right to do what you think wrong, or wrong to do what you think right”.’
Now, however, I am inclined to think that in saying this I was not paying enough attention to the strong persuasion which we have that a man who does what he thinks is his duty, really does his duty. This persuasion cannot be right as it stands, for nothing can have any characteristic merely by being thought to have it. But we should try to come to terms more closely with this persuasion, and this I will try to do later.14
From the reaction theories I turn to the causal theories. The only naturalistic15 causal theory that has ever found much favour is the hedonistic one which says that what makes acts right is their tendency to produce either pleasure for the agent, or pleasure for mankind or for all sentient beings (the latter difference being one of detail, and not affecting the general character of the theory, while the former affects it profoundly). I have already remarked16 that this form of Hedonism, in which productivity of pleasure is made the ground of rightness, is far more plausible than the form previously considered, in which productivity of pleasure is put forward as the essence of rightness. But I do not propose to examine Hedonism in detail, and that for various reasons. In the first place, I do not think that Hedonism has much vitality to-day. Egoistic Hedonism is put out of court by the fact which stares us in the face, that it is consideration for the rights or interests of others, far more often (to state the matter very mildly) than consideration for our own interests or rights, that makes us think it our duty to behave in a certain way. Whether consideration of our own rights or interests ever gives rise to the thought that we ought to behave in a certain way, in distinction from the thought that it would be prudent or sensible to behave in a certain way, is a question to which I hope to come later.17 But that it is the sole consideration which gives rise to the thought of duty is too palpably untrue to need serious discussion. Universalistic Hedonism, again, seems to me to have been put effectively out of court, by (inter alia) Professor Moore's arguments to show that there are other things, notably virtuous action, which we regard as good in their own right, independently of their tendency to produce pleasure. To flog Hedonism is, I believe, to flog a dead or dying horse. That many people behave in a great many of their actions as if they believed in Hedonism is true enough, but as a theory of morals it has very little if any serious claim to our attention. I do not propose therefore to join in the easy game of exposing its fallacies.
There is a further reason why I may excuse myself this task. I believe that no one holds the hedonistic creed unless he believes two things: (1) that what makes acts right is their being productive of the greatest good, and (2) that pleasure is the only thing good in itself. Mill, for instance, describes the theory of life on which Utilitarianism is grounded as being the theory ‘that pleasure, and freedom from pain, are the only things desirable as ends’.18 But it is plain that this is not in itself sufficient ground for the theory that productivity of pleasure is the sole ground of rightness; there is needed also the major premiss that productivity of what is desirable as an end is the sole ground of rightness. This premiss Mill nowhere, I think, seeks to prove; he assumes it silently. But it is just as much needed for the proof of utilitarianism as is the premiss which he takes some pains to prove. However much it might be true that pleasure is the only good, it would not follow that productivity of pleasure is the sole ground of rightness, unless it were also true that productivity of good is the sole ground of rightness. The ground of the rightness of acts is rooted, according to Utilitarianism, in the goodness of their results; and goodness is a genuinely ethical notion, not a naturalistic one.
It is true that the only reason Mill gives for the view that pleasure is the only good is at first sight a naturalistic one, viz. the psychological fact, as he holds it to be, that pleasure is the only thing that is desired.19 But it is plain that here again a further premiss is needed. That pleasure is the only thing that is desirable, or good, can follow from the fact that pleasure is the only thing that is desired, only in virtue of the further premiss that the only thing which is desired must be the only thing which is desirable, or good, and in this the non-naturalistic notion of desirable or good is already present. Thus not only does Mill's view that productivity of pleasure is the only ground of rightness rest on the non-naturalistic premiss that only pleasure is good, but this in turn rests on the naturalistic premiss that only pleasure is desired + the non-naturalistic premiss that the only thing that is desired is the only thing that is desirable.
Similarly, Sidgwick devotes a whole chapter20 to showing that ultimate good consists solely of pleasant consciousness. And this would be irrelevant to his main contention, that productivity of pleasure is the ground of rightness, unless he were assuming as self-evident that productivity of ultimate good is the ground of rightness. He is clear-sighted enough to reject the reason which Mill gives for regarding pleasure alone as ultimately good, viz. that it alone is desired, and candid enough to admit, or indeed contend, that pleasure is not the only thing that is desired for its own sake. But he is at one with Mill in accepting the two premisses: (1) that only productivity of good is what makes acts right, and (2) that only pleasure is ultimately good, premisses of which the subject of the first and the predicate of the second involve the non-naturalistic notion ‘good’.
The position then is this. There is a certain widely held view, which we may call Utilitarianism, that productivity of good is the only thing that makes acts right. There is one form of this view, hedonistic Utilitarianism, which adds the premiss that only pleasure is good. There is another form of the view, with which the writings of Professor Moore and Dr. Rashdall have made us familiar, which holds that other things besides pleasure are good. This is non-hedonistic Utilitarianism. Hedonistic Utilitarianism cannot be true unless Utilitarianism is true, but may be untrue even if Utilitarianism is true. Thinking as I do, then, that Utilitarianism in general is an untrue view, I am not much interested in the question whether, if it were true, the hedonistic or the non-hedonistic variety of it would be the true one; and if I can persuade any one that Utilitarianism is untrue, he will not wish me to discuss the other question. Yet even if Utilitarianism is not true, it is still the case that it is one of our main responsibilities to produce as much good as we can, so that the question whether pleasure is the only good remains a very important question. But it will belong to a later stage of our discussion.
It is Utilitarianism in its general form, the view that our sole duty is to produce as much good as possible, that we have now to discuss. If we could persuade ourselves that right just means ‘calculated to produce the greatest good’, the matter would be simple. But we have seen, I hope,21 that that contention is not at all plausible. If productivity of good is different from rightness but is the universal ground of rightness, how do we know this? There are, I think, only three possibilities. Either it is known by an immediate intuition, or it is established deductively, or it is established inductively. I do not know of any attempt to establish it deductively, and I cannot think of any middle term which could with any plausibility be used to connect the two terms in question. The effective alternatives appear to be intuition and induction. I will first ask whether the proposition has been established inductively. I take leave to quote some sentences from The Right and the Good.
‘Such an enquiry, to be conclusive, would have to be very thorough and extensive. We should have to take a large variety of the acts which we, to the best of our ability, judge to be right. We should have to trace as far as possible their consequences, not only for the persons directly affected but also for those indirectly affected; and to these no limit can be set. To make our inquiry thoroughly conclusive, we should have to do what we cannot do, viz. trace these consequences into an unending future. And even to make it reasonably conclusive, we should have to trace them far into the future. It is clear that the most we could possibly say is that a large variety of typical acts that are judged right appear, so far as we can trace their consequences, to produce more good than any other acts possible to the agents in the circumstances. And such a result is far short of proving the constant connexion of the two attributes. But it is surely clear that no inductive inquiry justifying even this result has ever been carried through. The advocates of utilitarian systems have been so much persuaded either of the identity or of the self-evident connexion of the attributes “right” and “optimific” (or “felicific”) that they have not attempted even such an inductive inquiry as is possible.’22
It is clear, too, that even if we could establish inductively that all optimific acts are right and all right acts optimific, that would not establish that their being optimific is the ground of their rightness, which is the proposition we are inquiring into. If we have only proved that the two attributes always go together, that is not enough. We should have to show that all right acts not only are optimific but are right because they are optimific. I do not mean to insist that it should be shown that unreflective people always reach their judgement that an act is right because they first judge it to be optimific. To this demand the utilitarian would have a perfectly proper answer. He would say, ‘Certain types of act have been in practice found to be optimific, and have in consequence been judged to be right; and so, for plain men, the character of rightness has come to seem to belong to such acts directly, in virtue of their being, e.g. fulfilments of promise, and the middle term which established their rightness has come to be forgotten. Media axiomata such as “men should keep their promises” have come to be accepted as if they were self-evidently true, and people habitually judge acts to be right on the strength of the media axiomata, forgetting the method by which the media axiomata have themselves been established’. That is a fair answer. The test I would prefer to impose is a different one, viz. this: when we reflect, do we really come to the conclusion that such an act as promise-keeping owes its rightness to its tendency to produce maximum good, or to its being an act of promise-keeping?
It seems clear that Utilitarianism has not established inductively that being optimific is always the ground of rightness, and as a rule utilitarians have not attempted to do so. The reason is simple: it is because it has seemed to them self-evident that this is the only possible ground of rightness. Professor Moore definitely says that for him the principle is self-evident.23 For my part, I can find no self-evidence about it. And I think I can point to several facts which tell against its truth, and to some which tell against there being even a constant correspondence between the two attributes, optimificness and rightness.
(1) Professor Broad has pointed out one such difficulty. Utilitarians hold that pleasure is either the only good, or is at least a good; and in the latter assertion most people would be, with certain qualifications, in agreement with them. Then, if any other consequences that an act may have be abstracted from, utilitarians are bound to say that an act which produces the greatest possible amount of pleasure is the right or obligatory act. Now, Professor Broad points out,
‘among the things which we can to some extent influence by our actions is the number of minds which shall exist, or, to be more cautious, which shall be embodied at a given time. It would be possible to increase the total amount of happiness in a community by increasing the numbers of that community even though one thereby reduced the total happiness of each member of it. If Utilitarianism be true it would be one's duty to try to increase the numbers of a community, even though one reduced the average total happiness of the members, so long as the total happiness in the community would be in the least increased. It seems perfectly plain to me that this kind of action, so far from being a duty, would quite certainly be wrong.’24
His criticism appears to be clearly justified. We should not merely not judge that such action was right because it was optimific; we should judge that it was wrong although it was optimific. It already begins to become clear that it is not our duty to increase to the utmost the total happiness, irrespective of how the happiness is distributed.
Professor Broad does not apply his argument to any other good than pleasure; for it is hedonistic Utilitarianism that he is criticizing. But the same argument will apply to any other form of good, say virtuous action or intelligent thought. The utilitarian doctrine involves that all goods are commensurable—that, for instance, in any two virtuous acts there must be different quantities of good which are in a certain ratio to each other, even if we cannot detect the ratio. And a utilitarian should maintain that it is self-evident that if we had to choose between promoting the existence of a certain amount of virtue and intelligence spread out very thin among a certain population, and a slightly smaller amount concentrated in a much smaller population (whose average virtue and intelligence would therefore be greater), we ought to choose the former. But it is clear to me that this is far from self-evident.
Thus we have already a principle which theoretically at any rate is capable of coming into conflict with the principle of producing the greatest total amount of good, viz. the principle which bids us concentrate good in a population of high average virtue and intelligence, rather than spread it out over a population of low average virtue and intelligence, if the choice ever lay between these alternatives.
(2) Consider now a case in which the size of the population is not assumed to be alterable by anything we can do. If the essential utilitarian principle is true, that productivity of maximum good is the sole ground of rightness, it ought to be quite indifferent how an ‘extra dose’25 of happiness should be distributed among the population, provided the total amount of the dose is unaltered. It would be morally just the same whether A is made very happy and B only very slightly happy, or whether A and B are both made rather happy, provided that the net gain in happiness for A and B taken together were equal in both cases. Now if A and B are people of equal moral worth, we do not really think that it would be right to distribute happiness unequally between them. Sidgwick, while criticizing some of our supposed intuitions of justice, has the candour to admit that there is one principle of justice that is axiomatic, viz. that of impartiality in the application of general rules.26 This, in its application to the case we are considering, can only mean that it is not morally indifferent how we divide an extra dose of happiness between two individuals, but that in the absence of some relevant difference between them it should be equally divided between them. But though Sidgwick recognizes the ‘principle of justice’ alongside of the ‘principle of rational benevolence’ (that which commands us to produce the maximum of good), he seems to assign to it a subordinate position. He would still, I think, say that if we can produce a greater total extra dose of happiness by giving much to A and little to B, than by giving the same amount to both, we ought to do so. This, however, is a half-way house at which we cannot stop. The principle of justice in the distribution of happiness can in no way be derived from the principle bidding us produce the greatest total of happiness. If it is true, as Sidgwick holds, then it is independent of the greatest happiness principle; and if it is independent of it, it is capable of coming into conflict with it. And where it does, I believe we should all judge that it would be rather our duty to produce a smaller increase of total happiness, fairly divided between individuals, than a slightly larger increase, very unfairly divided. Furthermore, I think we should judge not only that there is an independent moral principle bidding us divide happiness equally between people of equal moral worth, but also that the same principle bids us divide it, so far as we can, unequally between people of unequal moral worth. This appears to me just as axiomatic as the principle which bids us promote the general happiness, or (more widely) the general good.
(3) A further difficulty for utilitarians arises when we consider the distribution of pleasure between the agent and any one else. For utilitarians, it is always a duty for me to produce a greater pleasure for myself rather than a smaller pleasure for another (except of course where the ulterior consequences of the two acts would weigh the balance in favour of the latter act—but we can ignore this complication). Now the plain truth seems to be that we never judge so in fact. It seems to me that if we are honest with ourselves, which in a matter affecting us so closely it is hard to be, we shall find that we never really think ourselves morally bound to do an act which will increase our own pleasure, except for some ulterior reason, e.g. where we think that the pleasurable experience will fit us to do our work better, or that the relinquishing it to another person will tend to have a bad effect on his character.
Of these three difficulties for utilitarianism, arising out of the distribution of good, the first two may be dealt with in either of two ways. We may say (a) that quite apart from the duty to produce as much good as possible, there is an independent duty to produce a concentration of good in a smaller number of persons rather than a distribution of an equal amount of it among a larger number, and another independent duty to distribute happiness in proportion to merit. Or we may say (b) that the concentration of good in a smaller number of persons is itself a good, a good of higher order,27 as it were, than the good (consisting, say, of virtuous action, intelligent thought, and pleasure) which is thus concentrated; and similarly that the enjoyment of happiness in proportion to merit is itself a good of higher order than the happiness and the merit themselves. In this case the duty to produce such concentrations or such distributions will fall under the general duty of producing good; and our criticism of Utilitarianism will be, so far, less radical than in the other case. We shall not have established a duty other than the duty of producing good. We shall simply have shown that Utilitarianism in naming virtuous action, intelligent thought, and pleasure as the things that are good has overlooked two important goods of higher order.
It is difficult to choose between these two views. On the whole I incline towards the latter. It seems to me that the existence of a greater concentration of good is not only something in which we should in fact take greater satisfaction than in the wider and thinner distribution of the same total amount of good, but something in which it is reasonable to take satisfaction, i.e. is a greater good.28 And similarly I think it is reasonable to take satisfaction in a distribution of happiness in proportion to merit rather than in a distribution not in proportion to merit. If we had before us in imagination two communities in which the total amounts of virtue and of happiness were equal, but in one the good were happy and the bad wretched, and in the other the bad were happy and the good were wretched, I think it would be reasonable to say that the state of the first community is a better state than that of the second, and one which on that ground we ought to do our best to bring about rather than the other.
In answer to the third objection also, a utilitarian might be tempted to say that a good of higher order is involved. He might say that the enjoyment of pleasure by a man as a result of another man's action is a good of higher order, while the enjoyment of pleasure by a man as a result of his own action is not such a good. It is clear, however, that this is not true. Suppose that A, desiring to produce pleasure for B, produces it for himself, and that B, desiring to produce pleasure for himself, produces it for A. No one thinks A's enjoyment in the first case less of a good than his enjoyment in the second, though in the first it has been produced by himself and in the second by B.
The utilitarian might then seek to amend his suggestion by saying ‘the enjoyment of pleasure by a man as a result of another man's action directed to that end is a good of higher order, while the enjoyment of pleasure by a man as a result of his own action directed to that end is not such a good, and that is why it is a duty to produce pleasure for others and not a duty to produce it for oneself’. But he is not entitled to make this amendment. For on his own showing the duty of doing an act depends on the results produced, or (according to a different form of the theory) on the results intended; and he is not entitled to reckon a difference between the two motives as a difference in the results produced or intended.
Yet here also our argument does not necessarily point to a duty quite distinct from that of producing a maximum of good. For, while for a third person the enjoyment of pleasure by A is the same kind of thing as its enjoyment by B, A's own pleasure stands in quite a different relation to A from that in which B's pleasure stands to A. There is at least some ground for thinking that for A they may be good only in quite different senses of ‘good’, B's pleasure being for A a morally suitable object of satisfaction, and A's pleasure being for A only an inevitable object of satisfaction, having nothing morally suitable or unsuitable about it.29 If this be the true account, the hard fact (one of the most certain facts in morals) that we have a duty to produce pleasure for others, and have not a duty to produce it for ourselves, will involve us in admitting that it is only things that are good in the sense of being morally suitable objects of satisfaction, and not those that are good in the sense30 of being inevitable but morally neutral objects of satisfaction, that we have a duty to produce.
I pass now to an objection connected not with the distribution of pleasure but with the fact that we may by our action produce pleasure for some people and pain for others. On the utilitarian view, to each dose of pleasure there is some dose of pain that is exactly equal. The one may be represented by +x, the other by −x. Now for a utilitarian it is morally indifferent whether by your act you produce x units of pleasure for A and inflict y units of pain on B, or confer x−y units of pleasure on one of them, since in each case you produce a net increment of x−y units of pleasure. But we should in fact, I think, always judge that the infliction of pain on any person is justified only by the conferment not of an equal but of a substantially greater amount of pleasure on some one else (assuming the persons to be of equal worth). We do not, in fact, think that persons other than ourselves are simply so many pawns in the game of producing the maximum of pleasure, or good. We think they have definite rights, or at least claims, not to be made means to the giving of pleasure to others; and claims that ought to be respected unless the net pleasure, or good, to be gained for the community by other action is very considerable. We think the principle ‘do evil to no one’ more pressing than the principle ‘do good to every one’, except when the evil is very substantially outweighed by the good. This consideration seems to be perfectly clear, and it is strange that it has been overlooked by the utilitarians.
I pass next to a group of difficulties for Utilitarianism arising from our sense of special duties towards individuals, based on special relations between them and the agent. These seem to fall under three general heads. There is first the sense which we all possess that we have a special duty to make compensation to any one for any wrong we have done him. When I have wronged some one, he has ceased to be merely what Utilitarianism regards him as being, one out of many possible recipients or receptacles of good, between whom the choice is to be made simply on the basis of the question how the maximum good is to be achieved. He has become some one with a special claim on my effort, over and above the claim which all men have to my beneficence.
There is similarly the claim which those have from whom we have accepted benefits in the past. This again is a claim which, in fact, I believe every one recognizes, and it is evident that it is on it that our special duty to parents and friends in the main depends.
These two responsibilities—the responsibility for compensation and for rendering good for good—arise incidentally from past actions having another purpose. But, thirdly, there are obligations arising from acts whose express object was to create them. Our name for these acts is ‘promises’; a promise is just the voluntary making of something obligatory on us which would not, or need not, have been obligatory before. To make this clear, we must in the first place distinguish (as we do, more or less clearly, in ordinary life) between the making of a promise and the announcement of an intention. There are cases in which it is difficult to know whether some one is making a promise or is merely announcing an intention; but that does not affect the fact that the two things are in principle quite different. As Sidgwick remarks, ‘If I merely assert my intention of abstaining from alcohol for a year, and then after a week take some, I am (at worst) ridiculed as inconsistent; but if I have pledged myself to abstain, I am blamed as untrustworthy.’31 The announcement of an intention is merely a statement about one's present state of mind; a promise is a statement about the future. But, secondly, not every statement about one's own behaviour in the future is a promise. If I merely say incidentally in conversation with some one that I shall be at a certain place at a certain time, that does not constitute a promise to be there. To make a promise, there must be a more or less clear intimation to another person that he can rely upon me to do something which he, at least, regards as a service to him. The difference between this and a mere statement about the future can be seen from the fact that when I have merely made a statement about the future, what he relies upon, if he expects me to fulfil it, is my unchangeability, while, when I make a promise, what he relies upon is my sense of duty.
A promise being this, an intentional intimation to some one else that he can rely upon me to behave in a certain way, it appears to me perfectly clear, that, quite apart from any question of the greatness of the benefits to be produced for him or for society by the fulfilment of the promise, a promise gives rise to a moral claim on his part that the promise be fulfilled. This claim will be enhanced if there are great benefits that will arise from the fulfilment of the promise in contrast to its violation; or it may be overridden if the fulfilment of the promise is likely to do much more harm than good. But through all such variations it remains as a solid fact in the moral situation; and it arises solely from the fact that a promise has been made, and not from the consequences of its fulfilment. I would go so far as to say that the existence of an obligation arising from the making of a promise is so axiomatic that no moral universe can be imagined in which it would not exist.
These seem to me to be the main difficulties in the way of accepting Utilitarianism as a complete ethical creed; these are the principles of duty which seem to emerge as distinct from the principle ‘promote the maximum good’.
I may be allowed to reinforce these criticisms of Utilitarianism by quoting some words from the most sagacious, if not the most consistent or systematic, of the British Moralists. In his ripest work on ethics, the Dissertation on the Nature of Virtue,32 Butler indicates more clearly than in the Sermons his distrust of the view which treats zeal for the general good as the only virtue.
‘Without inquiring’, he says, ‘how far, and in what sense, virtue is resolvable into benevolence, and vice into the want of it; it may be proper to observe, that benevolence, and the want of it, singly considered, are in no sort the whole of virtue and vice. For if this were the case, in the review of one's own character, or that of others, our moral understanding and moral sense would be indifferent to every thing, but the degrees in which benevolence prevailed, and the degrees in which it was wanting. That is, we should neither approve of benevolence to some persons rather than to others, nor disapprove injustice and falsehood upon any other account, than merely as an overbalance of happiness was foreseen likely to be produced by the first, and of misery by the second. But now, on the contrary, suppose two men competitors for any thing whatever, which would be of equal advantage to each of them: though nothing indeed would be more impertinent, than for a stranger to busy himself to get one of them preferred to the other; yet such endeavour would be virtue, in behalf of a friend or benefactor, abstracted from all consideration of distant consequences: as that examples of gratitude, and the cultivation of friendship, would be of general good to the world. Again, suppose one man should, by fraud or violence, take from another the fruit of his labour, with intent to give it to a third, who he thought would have as much pleasure from it as would balance the pleasure which the first possessor would have had in the enjoyment, and his vexation in the loss of it; suppose also that no bad consequences would follow; yet such an action would surely be vicious. Nay, farther, were treachery, violence and injustice, no otherwise vicious, than as foreseen likely to produce an overbalance of misery to society; then, if in any case a man could procure to himself as great advantage by an act of injustice, as the whole foreseen inconvenience, likely to be brought upon others by it, would amount to; such a piece of injustice would not be faulty or vicious at all: because it would be no more than, in any other case, for a man to prefer his own satisfaction to another's in equal degrees.
‘The fact then appears to be, that we are constituted so as to condemn falsehood, unprovoked violence, injustice, and to approve of benevolence to some preferably to others, abstracted from all consideration, which conduct is likeliest to produce an overbalance of happiness or misery. And therefore, were the Author of nature to propose nothing to himself as an end but the production of happiness, were his moral character merely that of benevolence; yet ours is not so.… The happiness of the world is the concern of him, who is the Lord and the Proprietor of it: nor do we know what we are about, when we endeavour to promote the good of mankind in any ways, but those which he has directed; that is indeed in all ways not contrary to veracity and justice.’
These weighty words of Butler's answer better to what we really think on moral questions, than a theory which makes the production of good at all costs the only duty.
The fact is that Utilitarianism is a product of the craving for a simple creed, and that the facts of the moral life are too complex to fit into its scheme. If the root idea of rightness is suitability to the situation, there is not the slightest reason to anticipate that the only way in which an act can be right, i.e. fit a situation, is by being likely to amend it to the greatest possible extent; it may fit it no less by harmonizing with existing features of the situation, such as the existence of a claim to the fulfilment of a promise.
I must discuss at this point Professor Broad's view on this matter. He is dissatisfied, as I am, both with Utilitarianism (whether hedonistic or agathistic) and with out-and-out Intuitionism. He sees that such an act as the fulfilling of a promise has a tendency to be right which does not arise from a tendency to promote the general good, but from the fact that a promise has been made. He sees, on the other hand, that it cannot be maintained, with out-and-out Intuitionism, that all promises should be kept irrespectively of the consequences. He therefore puts forward the following analysis:
‘We have to distinguish two quite different ethical features of the action x, viz., its fittingness or unfittingness to the total course of events as modified by it, and its utility or disutility.… Fittingness or unfittingness is a direct ethical relation between an action or emotion and the total course of events in which it takes place. As this course of events consists of a number of successive phases, it is possible that a certain action may be fitting to some of the phases and unfitting to others. In particular it might be “immediately fitting”, i.e. it might be appropriate to the initial phase F1, but it might be unfitting to some or all of the later modified phases F2x, etc. Again, since each phase is itself complex, the action might be fitting to certain factors of a certain phase but unfitting to other factors of that phase. It is quite easy to give examples. If I am asked a certain question and answer it in a certain way I may be answering that question truly but my answer may lead to subsequent false inferences. It might then be said that this answer was fitting to the initial phase, but was unfitting to subsequent phases in the course of events as modified by it. It would then become a question whether a true answer, or a lie, or silence was the most fitting action on the whole, given the initial phase. The second complication may be illustrated as follows. I may be an elector to an office, and one of the candidates may have done me a service. To prefer him to a better qualified candidate would fit one aspect of the situation, since it would be rewarding a benefactor; but it would be unfitting to other factors in the situation, since it would be an act of bad faith to the institution which was employing me as an elector and an act of injustice to the other candidates. The statement that “x is more fitting to be done in the situation F1 than y is” means that x is more fitting to the whole course of events F1 F2x… Fxn than y is to the whole course of events F1 F2y… F2y. The fittingness of an act to a whole course of events will be a function of its fittingness or unfittingness to each phase in the series, and its fittingness to any phase in the series will be a function of its fittingness or unfittingness to each factor or aspect of that phase.’33
Having explained the fittingness of an action, he proceeds to consider its utility, and defines it thus:
‘The statement that “x is more useful to be done than y in the situation F1” means that, apart from all reference to fittingness and unfittingness, the course of events F1 F2x… Fnx is on the whole intrinsically better than the course of events F1 F2y… Fny’34
He adds that
‘the rightness or wrongness of an action in a given initial situation is a function of its fittingness in that situation and its utility in that situation. The pure Deontologist would deny that its utility or disutility was relevant to its rightness or wrongness. The pure Teleologist would deny that there is such a relation as direct fittingness or unfittingness, and would make its rightness or wrongness depend entirely on its utility or disutility. Both these extremes seem to me to be wrong, and to be in flagrant conflict with common sense.’35
On this I have two comments to make, (1) Professor Broad has already described rightness as a certain unique mode of fittingness of an act to a situation36—what we may call moral fittingness. Now if this be what rightness is, we cannot make rightness depend on a combination or balance of fittingness and utility. Unless the utility has a tendency to make the act fitting, it cannot have a tendency to make it right, if rightness is a kind of fittingness. It seems to me then that he should make rightness depend not on a joint consideration of fittingness and utility, but on a joint consideration of fittingness arising from utility and fittingness arising from other sources, such as that a promise has been made. I feel, myself, no difficulty in recognizing, in the tendency which an act has to amend the situation in the best possible way, i.e. to produce the maximum good, something in virtue of which that act tends to be fitting to the situation.
(2) I find a difficulty in Professor Broad's conception of an act as fitting the later phases in a process modified by its own occurrence. It seems to me clear that the situation which an act must fit if it is to be right is the situation that exists when, or just before, the act is done, not the situation as it will develop if modified by the act. Take the example that he takes, a true statement which may lead in the future to the formation of false opinions by the person to whom I speak or by some one else. It seems to me clear that any tendency that my statement may have, to lead to the formation of false opinions later, must be considered under the heading of disutility and not under the heading of what Professor Broad calls direct unfittingness. For if the duty to tell the truth be one of the duties that stand outside the utilitarian scheme, if it springs not from the badness of the total consequences of a lying statement but from the special nature of one special consequence, viz. that the statement leads directly to the formation of a false opinion on the subject-matter of my statement, then so far as that goes any opinion that any one may in the future form on other subject-matters falls outside of the ‘direct unfittingness’ of the act and must come under the heading of unfittingness arising from disutility.
These are differences of detail. In the main, Professor Broad's view is just that which I wish to advocate, viz. that among the features of a situation which tend to make an act right there are some which are independent of the tendency of the act to bring about a maximum of good. To say this is to hold an intuitionistic view of one kind. It has, of course, often been pointed out that every ethical system admits intuition at some point. Utilitarianism in the general form represented by Professor Moore's ethical writings admits the supposed intuition that only what is productive of the greatest good is right. Hedonistic Utilitarianism adds to this the supposed intuition that only pleasure is intrinsically good. Sidgwick's form of hedonistic Utilitarianism adds to these intuitions two that contradict the essential principle of Utilitarianism, the ‘axiom of rational self-love’ and the ‘axiom of justice’. The objection that many people feel to Intuitionism can hardly be an objection to the admission of intuition; for without that no theory can get going. The objection rather is that Intuitionism admits too many intuitions, and further that it admits intuitions that in practice contradict one another. These objections must be considered separately.
(1) The view which admits only one intuition—that only the production of maximum good is right—gratifies our natural wish to reach unity and simplicity in our moral theory. We have a natural wish to reach a single principle from which the rightness or wrongness of all actions can be deduced. But it is more important that a theory be true than that it be simple; and I have tried to show that a system which admits only this one intuition is false to what we all really think about what makes acts right or wrong. After all, there is no more justification for expecting a single ground of rightness than for expecting a single ground of goodness, and agathistic or generalized Utilitarianism recognizes a variety of goods without succeeding in finding, or even feeling any need to find, a single ground of the goodness of them all. It is, to my mind, a mistake in principle to think that there is any presumption in favour of the truth of a monistic against a pluralistic theory in morals, or, for that matter, in metaphysics either. When we are faced with two or more ostensible grounds of rightness, it is proper to examine them to find whether they have a single character in common; but if we cannot find one we have no reason to assume that our failure is due to the weakness of our thought and not to the nature of the facts. Just so in metaphysics; where we find two types of entity that are prima facie quite different, as bodies and minds are, it is proper to ask whether they are not two forms of one kind of entity; but there is no reason for assuming that they necessarily are; and if on examination we can find no unity of nature in them, it is wiser to accept this result than to assume that there must be a unity that we have not discovered. There is no reason why all the substances in the world should be modifications of a single pattern.
(2) But it may be argued that the plurality of moral intuitions is disproved by the fact that the supposed intuitions in practice contradict one another; that often we cannot obey one without disobeying another; that sometimes we cannot obey the principle of telling the truth without disobeying the principle of not causing needless pain, or the principle of keeping promises without disobeying the principle of producing the maximum good. This objection is to be met by care in stating the content of the principles for which we claim an axiomatic character. Moral intuitions are not principles by the immediate application of which our duty in particular circumstances can be deduced. They state what I have elsewhere37 called prima facie obligations. This way of describing them is, I think, in two ways useful. In the first place, it brings out the fact that when we approach the question, what should I do in these particular circumstances, it is the fitness or unfitness of an imagined act in certain respects that first catches our attention. Any possible act has many sides to it which are relevant to its rightness or wrongness; it will bring pleasure to some people, pain to others; it will be the keeping of a promise to one person, at the cost of being a breach of confidence to another, and so on. We are quite incapable of pronouncing straight off on its rightness or wrongness in the totality of these aspects; it is only by recognizing these different features one by one that we can approach the forming of a judgement on the totality of its nature; our first look reveals these features in isolation, one by one; they are what appears prima facie. And secondly, they are, prima facie, obligations. It is easy to be so impressed by the rightness of an act in one respect that we suppose it to be therefore necessarily the act that we are bound to do. But an act may be right in one respect and wrong in more important respects, and therefore not, in the totality of its aspects, the most right of the acts open to us, and then we are not obliged to do it; and another act may be wrong in some respect and yet in its totality the most right of all the acts open to us, and then we are bound to do it. Prima facie obligation depends on some one aspect of the act; obligation or disobligation attaches to it in virtue of the totality of its aspects.
Yet the phrase ‘prima facie obligation’ does not do full justice to the facts. It says at the same time too much and too little. It says too much; it seems to say that prima facie obligations are one kind of obligation, while they are in fact something different; for we are not obliged to do that which is only prima facie obligatory. We are only bound to do that act whose prima facie obligatoriness in those respects in which it is prima facie obligatory most outweighs its prima facie disobligatoriness in those respects in which it is prima facie disobligatory.
In another way it says too little. If we dismiss the wrong suggestion just pointed out, the phrase may then suggest that the whole character of the things we are speaking of is to appear to be obligations. But on general grounds we may say that it cannot be the whole character of anything to seem to be something. There must be something which it is, as well as something which it seems to be. And in particular, if we are mistaken when we suppose these things to be obligations, that is not the whole truth about them. They are not illusions which we dispense with when we view the act in its totality and see it to be really obligatory or really disobligatory. It remains hard fact that an act of promise-breaking is morally unsuitable in so far as it is an act of promise-breaking, even when we decide that in spite of this it is the act that we ought to do.
We want an expression, then, which shall state what these things are which we have so far described by reference to what they seem to be. Professor Prichard has suggested that the word ‘claims’ may supply the need. But it has the inconvenience of stating the matter from the point of view of the persons to be affected by the action; we want also a word to express the matter from the point of view of the agent. And, further, the expression ‘claim’ seems to be completely suitable only to cases where our duty is duty to another person or persons. But there are duties besides these—the duties of improving our own character and our own intellect; and it is only metaphorical to describe our own character or intellect as having a claim on us. Mr. Carritt has suggested the word ‘responsibility’ as escaping both these objections;38 and I gladly accept the suggestion.
When we try to formulate laws of nature, we find that if we are to state them in a universal form which admits of no exception, we must state them not as laws of actual operation but as laws of tendency. We cannot say, for instance, that a certain force impinging on a body of a certain mass will always cause it to move with a certain velocity in the line of the force; for if the body is acted on by an equal and opposite force, it actually remains at rest; and if it is acted on by a force operating in some third direction, it will move in a line which is oblique to the lines of both forces. We can only say that any force tends to make the body move in the line of the force. Thus alone do we get a perfectly universal law. In the same way if we want to formulate universal moral laws, we can only formulate them as laws of prima facie obligation, laws stating the tendencies of actions to be obligatory in virtue of this characteristic or of that. It is the overlooking of the distinction between obligations and responsibilities, between actual obligatoriness and the tendency to be obligatory, that leads to the apparent problem of conflict of duties, and it is by drawing the distinction that we solve the problem, or rather show it to be non-existent. For while an act may well be prima facie obligatory in respect of one character and prima facie forbidden in virtue of another, it becomes obligatory or forbidden only in virtue of the totality of its ethically relevant characteristics. We are perfectly familiar with this way of thinking when we are face to face with actual problems of conduct, but in theories of ethics responsibilities have often been overstated as being absolute obligations admitting of no exception, and the unreal problem of conflict of duties has thus been supposed to exist.
Cf. p. 9.
Data of Ethics (cheap edition), 25.
Ibid. 26, 27.
Cf. p. 24.
The opinion that A has the attribute B may in certain circumstances cause it to have that attribute; e.g. possunt quia posse videntur. But simultaneous necessitation is of course something quite different from causation.
Methods of Ethics, ed. 7, 207.
Mainly, at least; but cf. pp. 306–8.
Or rather, apparently naturalistic. See pp. 65–7.
See pp. 272–4.
Utilitarianism, copyright eds., 10.
Methods of Ethics, ed. 7, iii. 14.
Cf. p. 42.
The Right and the Good, 36.
Five Types of Ethical Theory, 249–50.
Methods of Ethics, ed, 7, 380.
In a mathematical, not in a moral sense.
For this sense of good cf. pp. 275–6, 278–9.
Cf. pp. 272–9.
If this is a legitimate sense of ‘good’ at all; cf. pp. 284–5.
Methods of Ethics, 304.
The Works of Joseph Butler (Gladstone's edition), i. 334–7.
Five Types of Ethical Theory, 218–20.
The Right and the Good, 19.
Morals and Politics, 185.