C. D. Broad, Five Types of Ethical Theory, 1930 (F) DETAILED DISCUSSION OF EACH OF THE THREE METHODS.
(1) Intuitionism. We may divide Sidgwick's discussion of this subject into two main parts, viz.: (1, 1) a general treatment of the subject, and (1, 2) a detailed analysis and criticism of the alleged moral intuitions of common-sense. The former is contained in Book 1, Chap. 2 VIII, and Book III, Chaps. I and XIII. The latter is contained in Book III, Chaps. III to XI inclusive.
(1, 1) We must begin by stating more definitely what is the subject-matter of moral judgments. So far we have said that Intuitionism, in the narrower sense to which we are now confining the term, holds that certain types of action are intrinsically right or wrong without regard to their consequences. This statement must now be made [209] more accurate. In order to do this let us take a concrete example. It is held by many people that it is always wrong to tell a lie, no matter how disastrous the consequences of telling the truth might be. We are not at present concerned with the correctness of this doctrine, but only with its precise meaning. Now it is, plain that an action cannot be called a "lie" without reference to certain of the consequences which the agent expects that it will have. He must expect that his action will produce certain beliefs, and he must hold that these beliefs will be false.
The action, then, is judged to be wrong because the agent expects it to have consequences of a certain kind. But, if so, it may be asked, how does Intuitionism differ from a teleological type of ethical theory, such as Utilitarianism? Does not Utilitarianism also condemn lying because it is likely to have consequences of a certain kind?
To deal with this question let us begin by defining a "lie" as a statement made by an agent with the intention of producing a false belief. This definition would presumably be accepted both by Intuitionists and by Utilitarians. It will be seen that the definition includes a reference to certain consequences (viz., the production of a belief and to a certain characteristic of these consequences (viz., the falsity of this belief). In any particular case both the Utilitarian and the Intuitionist will know, or reasonably suspect, that there will be other consequences beside the production of a belief, and that the belief will have other characteristics beside falsity. These, however, form no part of the definition of a "lie", though they do form part of the intention of the person who tells a lie. Now the fundamental difference between the Teleologist and the Intuitionist [210] in this case seems to be the following: The Teleologist is interested only in the goodness or badness of the intended consequences. For he recognises only what I have called earlier in this chapter the teleological and the logical senses of "ought". When he says: "I ought to do X", he always means simply and solely:
The Teleologist will therefore take into account all the intended consequences, whether they be included in the definition of the action as a "lie" or not. And he will take into account all those characteristics, and only those characteristics, of the intended consequences which he holds to be relevant to their goodness or badness. These may include other characteristics beside those involved in the definition of the act as a "lie"; and they may not include that particular characteristic at all. Thus, e.g., a Utilitarian would not consider the characteristic of falsity, which is involved in the definition, to be directly relevant; whilst he would consider that another characteristic which is not involved in the definition, viz., the tendency to diminish human happiness, is directly relevant. It is of course quite possible to imagine a non-hedonistic Teleologist who held that true belief is intrinsically good, and that it is therefore our duty to produce as much true belief as possible. Such a Teleologist would still differ fundamentally from an Intuitionist about lying. An Intuitionist need not hold that true belief is intrinsically good and false belief intrinsically bad, and he certainly will not hold that truth-telling is right and lying wrong simply because the former tends to [211] increase and the latter to diminish the amount of true belief in the world. Cases might easily arise in which it would be almost certain that more true belief would be produced by telling a lie than by telling the truth. In such cases a Teleologist of the kind just described would consider it his duty to tell a lie, whilst an Intuitionist about lying would still hold that it is wrong to do so.
- "I ought (in the teleological sense) to desire Y";
- "So far as I can see X is the most suitable means open to me for producing Y"; and
- "I ought (in the logical sense) to choose the most suitable means open to me for producing what I ought (in the teleological sense) to desire."
The fundamental difference between the Intuitionist and the Teleologist is that the former does, and the latter does not, recognise a sense of "right" which applies to actions and intentions and is not analysable into "conducive to good consequences". It is not true to say that the Intuitionist takes no account of intended consequences when judging of the rightness or wrongness of an action. What is true is that he takes no account of the goodness or badness of the intended consequences. For him a lie is wrong simply and solely because it is intended to produce a false belief, and not because a false belief is an intrinsically bad state of mind. For the Teleologist the other characteristics of the consequences are relevant only in so far as they make the consequences intrinsically good or bad, and to say that a lie is wrong simply means that its consequences will on the whole be bad.
There is a further difference, which is not, I think, so fundamental, but which certainly has existed between most Intuitionists and most Teleologists. A Teleologlst cannot reasonably take into account anything less than the whole of the consequences intended by the agent. For he has to consider the conduciveness of the action to good or evil results, and it would seem quite arbitrary to exclude from his survey any part of the consequences which the agent foresaw and desired or tolerated. But many Intuitionists [212] have held that the rightness or wrongness of an action was completely determined by certain characteristics of a certain restricted part of its total intended consequences. If, e.g., its immediate consequences had a certain characteristic, then it would be right (or wrong) no matter what might be its remoter consequences and no matter what might be the other characteristics of its immediate consequences. E.g., some Intuitionists would hold that, if I were asked a question about a certain matter, it would be my duty to make such a statement as would produce a true belief on that matter, I even though I knew that its remoter consequences would be false beliefs on other matters, and even though the belief which I produce would be intensely painful to my hearer in addition to being true.
I do not think that there is any logical necessity for a person who admits that there is a sense of "right" which applies to actions and intentions and is not analysable into "conducive to good consequences" to go to these extremes. He might reasonably hold that the rightness or wrongness of an action was determined by certain of the characteristics of all its intended consequences. But I think it is true that certain forms of Intuitionism could hardly be held if this view were taken. The point is this. Teleological theories would make all statements about the rightness or wrongness of classes of action into empirical propositions about general tendencies. We might be able to conclude by induction from past experience that lying generally has bad consequences; but we could not be sure that every lie, under existing circumstances, or any lie, under certain conceivable circumstances, would have such consequences. Hence, on a teleological theory, there could be no propositions of the form: "Such and such a type of action would always be [213] right (or wrong)." Now most Intuitionists have claimed to know some propositions of this kind by direct insight into the terms. Now it is difficult to see how they could possibly do this unless they knew that all but the immediate consequences, and all but a few of the characteristics of these, were irrelevant to the rightness or wrongness of the action. To define a type of action, such as "lying", we must take a few outstanding features of the act or of its immediate consequences. In any concrete instance of lying the act will have many remote consequences which the agent can foresee; and all its consequences, immediate and remote, will have many characteristics beside the one which makes it a lie by definition. Any Intuitionist who claims to be able to see that every lie as such must be wrong is claiming to see that all the remoter consequences of a lie, and all the other characteristics of the consequences except those involved in the definition of the act as a "lie", are irrelevant to the wrongness of the act.
It is now easy to see what is the connexion between the epistemological division of ethical theories into Intuitionist and Empirical and the ontological division of them into Deontological and Teleological.
- As we have seen, on a Teleological theory, both singular judgments of the form: "This act is right (or wrong)" and universal judgments of the form: "All acts of a certain kind are right (or wrong)" essentially involve judgments about all the consequences of the act or class of acts so far as these can be foreseen at all. Such judgments are of course purely empirical, like all judgments which involve particular causal laws. So a teleological theory is, to this extent, necessarily an empirical or inductive theory.
- Nevertheless, every Teleological theory does involve at least one a priori judgment. For it [214] will always involve some judgment of the form: "Anything that had a certain non-ethical characteristic (e.g., pleasantness) would necessarily be intrinsically good." Such judgments have nothing to do with causation. They claim to express a necessary connexion between a certain non-ethical characteristic and the ethical characteristic of goodness. The only kind of induction on which they are based is what Mr Johnson calls "intuitive induction", such as we use in coming to see that shape and size are necessarily connected, and not what he calls "problematic induction," such as we use in making the probable generalisation that all cloven-footed animals chew the cud.
- Any Deontological theory which claims to make universal judgments of the form: "All acts of such and such a kind are right (or wrong)" does claim to make a priori judgments in a sense in which teleological theories deny that they can be made. For it defines the kind of action under consideration by one or a few characteristics of its immediate consequences; and it claims to see that these suffice to make all such actions right (or wrong), and that the more remote consequences and the other characteristics of the consequences will always be irrelevant to the rightness (or wrongness) of the action. It is plain that, if such judgments can be made at all, they must be a priori. They may be compared with the judgment that, if the sides of a triangle be equal, this suffices to make it equiangular, and that the size, position, colour, etc., of the triangle are irrelevant. A Deontologist of this kind is called by Sidgwick either a dogmatic or a philosophic Intuitionist. The distinction between these two subdivisions corresponds to the distinction which we have drawn between pluralistic and monistic Deontologists. For a Dogmatic Intuitionist is one who holds that there are many independent [215] intuitively certain judgments asserting that such and such kinds of action are necessarily right (or wrong). And a Philosophic Intuitionist is one who holds that all the more concrete judgments of this kind can be subsumed under one or a few supreme moral principles which are intuitively certain. It is worth while to remark at this point that, although it is theoretically possible for a teleological theory to be pluralistic (since it may hold that there are several independent characteristics, each of which would suffice to make a thing intrinsically good), and although it is theoretically possible for a deontological theory to be monistic, yet in fact teleological theories have tended to be monistic and deontological theories to be pluralistic. No one has produced a plausible monistic deontological theory whilst universalistic hedonism is a fairly plausible form of monistic teleological theory. And this fact has often made people prefer teleological theories, since monism in such matters is more satisfactory to the intellect than pluralism.
(d) As I have said, a Deontologist might hold that it was necessary to consider all the foreseen consequences of an action before one could decide whether it was right or wrong. If such a Deontologist made universal judgments about the rightness or wrongness of certain types of action he would have to confine them to statements about general tendencies, just as the Teleologist has to do. He could not say: "Every lie is as such necessarily wrong," though, in the case of any particular lie, he might be able to say that this be is certainly right or certainly wrong. He could, no doubt, make the generalisation that any lie told in exactly similar circumstances with exactly similar foreseen consequences would necessarily be right, if this lie be right, or wrong, if this lie be wrong. But such generalisations are [216] hardly worth making. This kind of Deontologist seems to be what Sidgwick means by an Aesthetic Intuitionist.
(1, 2) This is perhaps as much as we need say on the general topic of Intuitionism. We can now pass to Sidgwick's criticism of the morality of common-sense. Into the details of this I shall not enter. The essence of the matter is this. Sidgwick holds that common-sense does claim to be able to see by inspection that certain types of action are necessarily right (or wrong) without regard to the goodness or badness of their consequences. And, although it does not ignore intended consequences, since it defines many types of action by reference to some of the characteristics of some of their intended consequences, yet it holds that certain characteristics of the immediate consequences suffice to make such actions right (or wrong). Common-sense then is dogmatically intuitive, though this does not necessarily imply that it does not use other and incompatible criteria of right and wrong. This seems to me to be true.
The upshot of his very elaborate discussion of common-sense morality is as follows. If there be genuine moral axioms they must fulfil the following conditions. Their terms must be clear and distinct; the propositions themselves must continue to seem self-evident no matter how carefully they be examined, and no matter with what difficulties we may confront them; and they must be mutually consistent. Moreover, it is important that there should be a clear consensus of opinion in their favour. If something seems self-evident to me and does not seem so to someone else who, so far as I can see, is as competent as I and is really contemplating the same situation as I, I am reduced to a state of hesitation, There are two special dangers about alleged ethical axioms. In the first place, [217] we are liable to confuse strong impulses with genuine intellectual insight, and to judge as wrong what we impulsively dislike. Secondly, rules which really rest on custom and the opinion of the society in which we have been brought up may gain the appearance of moral axioms. A grown man seems to himself to know intuitively what politeness or honour or fashion forbids. Yet such codes certainly have been imposed on him from without, and are largely lacking in rational justification. It is quite certain that common-sense morality contains a great deal of material of this kind. Now a careful discussion of the alleged axioms of common-sense morality shows that they do not answer to the required conditions. Agreement exists only so long as we keep to vague generalities and simple cases. As soon as we go into detail doubts and difficulties arise, both as to the meaning of the terms and as to the range of application of the principles. The central part of each duty seems clear, but it is surrounded with a margin of uncertainty. And, when the duties which it has laid down as absolute and unexceptionable conflict, common-sense either suggests no principle of reconciliation, or one so complex and qualified as to be no longer self-evident, or else it falls back on some teleological principle such as Utilitarianism.
I think that anyone who reads the relevant chapters in Sidgwick will agree that the extreme form of Intuitionism which he ascribes to common-sense cannot be maintained. And he is no doubt right in thinking that common-sense wants to hold something like this, and retreats from it only at the point of the bayonet. Sidgwick's conclusion is that we are forced to a mainly teleological view, eked out by a few very abstract intuitions about right and wrong modes of distributing good and evil. This does not seem to me to [218] be certain; and I propose as briefly as possible, and therefore somewhat dogmatically, to state a form of Intuitionism which is not open to Sidgwick's objections and is not flagrantly in conflict with reflective common-sense.
(a) Whenever a man is called upon either to act or to abstain from action he is in presence of a highly complex situation, composed of pre-existing persons, institutions and things, in various relations to each other and to himself. Let us call this the "initial phase". Whether he acts or abstains from action this phase will be succeeded by others. The initial phase, together with its subsequent developments, may be called a "total course of events". If the agent abstains from action there will be what I will call an "unmodified course of events". If he acts he will introduce an additional cause-factor into the initial phase, and this will make the subsequent phases, and therefore the total course of events, different from what they would otherwise have been. We then have a "modified course of events". According to what action he performs we shall have a correspondingly different modified course of events. Now of course each phase will itself be highly complex. If we denote the unmodified course by F1, F2 . . . Fn, then any phase such as Fr will consist of factors which we might symbolise by fry fr1, fr2 . . . frm. Suppose that, instead of abstaining from action, the man does a certain act x. The initial phase will then consist of all the factors in F1, together with the additional factor x, which of course will not simply be added to the rest but will stand in perfectly definite relations to them. The subsequent phases will be modified in a characteristic way by the addition of this cause-factor to the initial phase and will become Fx1, Fx2 . . . Fxn.
(b) Now it seems to me that we have to distinguish two [219] 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. I will now try to explain what I mean by these two notions. 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 Fx2, 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 m 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 [220] situation F1 than y is" means that x is more fitting to the whole course of events Fx1, Fx2 . . . Fxn than y is to the whole course of events F1Fy2 . . . Fyn. 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. By analogy with mechanics we may talk of the "resultant fittingness" and the, various "component fittingnesses". But, unfortunately, there is no simple general rule, like the parallelogram of forces, by which the resultant fittingness can be calculated from the various component fittingnesses.
(c) Having now tried to explain what I mean by the "fittingness" of an action, I will next consider its "utility". We have now to leave out of account the relations of fittingness or unfittingness between an action and the modified course of events which it initiates, and to consider simply the intrinsic goodness or badness of such a course of events. This will be determined by the qualities of the component events and their relations to each other. 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, Fx2 . . . Fxn is on the whole intrinsically better than the course of events F1, Fy2 . . . Fyn. The intrinsic goodness of a whole course of events is a function of the intrinsic goodness or badness of its successive phases, and the intrinsic goodness or badness of any phase is a function of the Intrinsic goodness or badness of its various factors or aspects. If Hedonism were true our calculations would be simplified, because no characteristic but pleasure or pain would have to be considered in our estimate. But the notion of utility is wider than this, [221] and it would still have application even if the Hedonistic view of what constitutes intrinsic goodness were found to be inadequate or false. We have to take account of the consequences of an act both in considering its fittingness and in considering its utility in a given initial situation. For the act has not to fit merely the initial phase or the total course of events which would have occurred if the initial phase had been allowed to develop without interference from the agent. It has to fit the total course of events which will occur if the initial phase is modified by it as an additional cause-factor. And of course it is still more obvious that utility in a given situation involves a reference to consequences. For it just is the causal characteristic of producing a course of events which have such qualities or such relations among themselves as to be intrinsically good.
(d) Now it seems to me 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. As against the pure Teleologist I would say that we all do recognise relations of fittingness and unfittingness. And, as against the pure Deontologist, I would say that we do think it reasonable to weigh utility against fittingness; and that a sane person judges it right, though regrettable, to do an act which is unfitting if this be the only means open to him of avoiding a course of events which, from [222] their qualities and their mutual relations, would be intrinsically very evil.
"It hath been the wisdom of the Church of England, ever since the first compiling of her Publick Liturgy, to keep the mean between the two extreams, of too much stiffness in refusing, and of too much easiness in admitting any variation from it."And I intend to follow the excellent example of my national Church.(e) If I am right, the kind of Intuitionist with whom Sidgwick contends in his discussion of the morality of common-sense makes two fundamental mistakes. In the first place, he identifies rightness with fittingness, and fails to see that utility is also a factor in determining rightness. Secondly, he takes far too simple-minded a view of fittingness. He thinks that the fittingness of an action is completely determined by its relations to the initial situation or the phase that immediately succeeds it. And he forgets that even these phases may be very complex, and that the fittingness of the action to each factor must be considered.
It seems quite clear that the Intuitionist will have to moderate his claims very greatly. He will be confined to statements about tendencies to be right and tendencies to be wrong. He can say that a he has a very strong tendency to be wrong, and that it will be wrong unless telling the truth would have very great disutility or unless the situation be of a certain special kind in which it is a matter of honour to shield a third person. And it is very doubtful whether any general rules can be given for balancing one kind of fittingness against another or for balancing fittingness on the whole against utility on the whole. When it comes to estimating resultant fittingness from component fittingnesses and unfittingnesses, and to estimating total rightness from total fittingness and total utility, we are soon reduced to [223] something analogous to those perceptual judgments on very complex situations which we have constantly to make in playing games of skill. No doubt this is an unsatisfactory conclusion, and at first sight it compares ill with the sweet simplicity of Utilitarianism. But, if it is so, it is so. And perhaps we may say that Utilitarianism is at once too simple in theory and too difficult in practice to satisfy either the philosopher or the plain man for very long.
It remains to say something about the few highly abstract principles which Sidgwick does regard as intuitively certain. They are the following.
- If an action would be right when done by A and wrong when done by B in precisely the same circumstances, there must be some qualitative dissimilarity between A and B which accounts for this. The mere fact that B is numerically other than A is irrelevant.
- If an action would be right when done by A to B and would be wrong when done in precisely similar circumstances by A to C, there must be some qualitative dissimilarity between B and C to account for this. The mere fact that B is numerically other than C is irrelevant.
- Any general rule ought to be applied impartially to all persons who come within the scope of the rule.
I will comment on these three principles before I mention the others which Sidgwick accepts. The first two, though not absolutely verbal, are extraordinarily trivial. Any pair of individuals always do differ qualitatively from each other in innumerable ways. Some of these qualitative differences are, and some are not, ethically relevant. And qualitative dissimilarities which are ethically relevant to certain types of action will be ethically irrelevant to others. If A admires red hair and B does not, this may make it right for A and [224] wrong for B to propose marriage to the red-haired C; and it may make it right for A to propose to C but wrong for him to propose to the otherwise similar but yellow-haired D. But if A had to rescue either C or D from drowning, and could not rescue both, the difference in the colour of their hair would not be an adequate ground for saving one and letting the other drown. What we want are some self-evident principles as to precisely what kinds of qualitative differences are relevant and what are irrelevant grounds for two people to act differently in similar circumstances or for the same person to act differently in similar circumstances towards two people. Sidgwick's principles are rather like the famous Principle of Indifference in Probability. Two alternatives are equally probable if there be no relevant dissimilarities between them; but what kinds of dissimilarity are relevant and what are not? If I had the chance of saving the life of one, but not of both, of two persons would the fact that one was my mother and that the other was my second cousin be a relevant ground for saving the former? As regards the third principle it is difficult to see that it states an absolutely unexceptionable duty. Certainly, if I have to administer a rule inflicting a penalty on all members of a certain defined class, it will be both unfitting and contrary to utility if I inflict the penalty on some members of the class and not on others. And I shall be inexcusable if I break it in favour of someone who does not differ relevantly from those on whom I inflict the penalty. But this is merely a particular case of the second rule. Suppose, however, that I see that there is a relevant difference between certain members of the class contemplated by the rule and others, am I never to break the rule in their favour? May not the unfittingness of ignoring these relevant [225] differences in some cases outweigh the unfittingness and the disutility of making exceptions to the rule which it is my duty to administer? Such conflicts plainly can arise where a man has to administer an obviously unjust and inadequate rule; and, when they reach a certain degree of acuteness, it is very hard to be sure about the duty of the officer.. We might be inclined to say that it was his duty to break the rule, but that his employers would have a right to punish him for doing so.
I pass now to Sidgwick's three remaining principles.
This, as he points out, is quite compatible with its being reasonable to prefer a nearer to a remoter good on the groumds of the greater likelihood of getting it, of greater keenness of appreciation in youth, and so on. The only doubt that I feel about this principle is concerned with order in time. Most people would be inclined to think that a life which began unhappily and ended happily was to be preferred to one, containing the same balance of happiness, which began happily and ended unhappily. It is difficult to be sure whether they really think that mere order is relevant, as their language would suggest. For there are the secondary pleasures and pains of anticipation and memory to be considered. The anticipation of happiness is always pleasant, and is perhaps more so if one is now unhappy. The anticipation of unhappiness is always unpleasant, and is perhaps more so if one is now happy. The memory of past happiness tends to be painful if one is now unhappy; whilst the memory of past unhappiness is on the whole not unpleasant if one is now happy. Now the primary happiness of the earlier half of the one life may be [226] reduced by the secondary unhappiness of anticipating the primary unhappiness of its later half; and the primary unhappiness of its later half will certainly be increased by the secondary pain of remembering the lost primary happiness of the first half. In the case of the other life the primary unhappiness of the first half may be reduced by the secondary pleasure of anticipating the primary happiness of the second half; and the primary happiness of the second half will not be reduced, and may be increased, by the memory of the primary unhappiness of the first half. So perhaps the truth of the matter is simply this. Of two lives which contain the same amounts of primary happiness and unhappiness, occurring in opposite order in time, the life in which the primary unhappiness precedes the primary happiness will contain more secondary happiness and less secondary unhappiness than that in which the primary happiness precedes the primary unhappiness. If this be the whole truth, the case under discussion is no real exception to Sidgwick's principle. But I do not feel completely certain that it is the whole truth, and to that extent I feel a faint doubt about the principle.
- Mere difference in the date in one's life in which any good is to be enjoyed makes no difference to its value.
The two remaining principles are of extreme importance in connexion with the controversy between Egoistic and Non-egoistic types of ethical theory. They are as follows.
From these two principles he deduces what he calls the Principle of Rational Benevolence, viz., that I ought to try to produce good states in any other individual as much as in myself, except in so far as I am less certain of [227] being able to produce them in him, or less certain that such states in him would be good, or can see that more good would be sacrificed in me than would be produced in him.
- The good of any one individual is of no more importance, from the point of view of the Universe, than the equal good of any other. And
- it is my duty to aim at good generally, so far as I can bring it about, and not merely at a particular part of it.
It will be best to defer the discussion of these two principles and of Sidgwick's inference from them till we deal with the question of Egoism. In the meanwhile I think we can say that, on their negative side, Sidgwick's principles are principles of indifference or impartiality. They tell us that certain kinds of difference, viz., the numerical difference between one individual and another, and the difference in temporal position between one event in a man's life and another, are not ethically relevant grounds for a difference of action or treatment or valuation. On their positive side they assert that a difference in action or treatment or valuation always does need justification, and that it must be justified by some kind of dissimilarity of quality or relation.
(2) Hedonism. We can now pass to the type of theory called "Hedonism", which is a form of teleological theory. As I have said, the discussion is best subdivided into
(2, 1) Hedonism in General;
(2, 2) Egoistic Hedonism; and
(2, 3) Universalistic Hedonism or Utilitarianism.(2, 1) We divided this into
(2, 11) The Ethical Problem andI will now say something about each of these in turn.
(2, 12) The Factual Problem.(2, 11) Since Hedonism, in its most rigid form, would be a purely teleological theory, a complete discussion of it would have to begin by considering whether any purely teleological theory of ethics could possibly be adequate. This question we have already discussed in connexion with Intuitionism, and we need say nothing further about it. In any case Sidgwick, though a Hedonist, is not a pure [228] teleologist, since his six ethical intuitions are deontological propositions. The ethical question that remains is this. Is it the case that nothing is intrinsically good or bad except experiences, that no characteristic of an experience has any bearing on its intrinsic value except its pleasantness or painfulness, and that the measure of its intrinsic value is the nett balance of pleasantness over painfulness which characterises it? Sidgwick discusses this question in Book III, Chap. XIV.
It seems to me important to begin by trying to get a dear idea of what we mean by "a pleasure" and "a pain"; for, on this point psychologists, to my mind, are very confused. The old tripartite division subdivides all mental events into Cognitions, Conations, and Feelings. And it seems to identify "Feelings" with pleasures and pains. Now this seems to me to be a radically unsatisfactory and unscientific division. I would first divide mental events into those which are and those which are not directed to objects. If there be any members of the second class, and I think it is plausible to maintain that there are, I confine the name "Feelings" to them. In the first class would certainly come Cognitions, Conations, and Emotions. You cannot cognise without cognising something, or will without willing something, or have an emotion without having it towards something. As regards those mental events which are called "Sensations," it seems to me that some, e.g., visual and auditory sensations, are plainly Cognitions, and therefore fall into the first class. With regard to others it is difficult in practice to decide whether they ought to go into the first or the second class, though it is certain that any one of them must in fact go into one class or the other. There are some "Sensations", e.g., those which we get from [229] processes in our bodies, which are often called "Feelings", and which it seems highly plausible, though not absolutely necessary, to put in the second class. Now I am very much inclined to agree with M'Taggart that really all members of the first class are Cognitions. It is plain that Emotion and Conation presuppose cognition, and that it is cognition which provides them with their objects. Now it seems plausible to suggest that, e.g., to fear something just is to cognise that thing and to have this cognition "toned" or qualified in a certain characteristic way. In fact to fear an object is to cognise it "fearfully"; to desire an object is to cognise it "desiringly"; and so on. Of course these qualitative differences among cognitions carry with them all kinds of causal differences. If I cognise an object fearfully my subsequent mental states and bodily actions will tend to be characteristically different from what they would be if I cognised it desiringly. If this be so, the fundamental division of mental events is into Cognitions and Feelings. And a cognition is called an "Emotion" if it has any one of the innumerable specific kinds of emotional quality; it is called a "Conation" if it has the "desire-aversion" quality; and so on. It seems plain that these qualities are not mutually exclusive, like determinates under the same determinable. The very same cognition may have several different emotional qualities and also the conative quality. It will then count both as a conation and as a mixed emotion.
We are now in a position to deal with pleasures and pains. It seems to me that there is a quality, which we cannot define but are perfectly well acquainted with, which may be called "Hedonic Tone". It has the two determinate forms of Pleasantness and Unpleasantness. And, so far as I can see, it can belong both to Feelings and to those [230] Cognitions which are also Emotions or Conations. Whether it can belong to Cognitions which have neither an emotional nor a conative quality, if such there be, is more doubtful. "A pleasure" then is simply any mental event which has the pleasant form of hedonic tone, and "a pain" is simply any kind of mental event which has the unpleasant form of hedonic tone. There is not a special kind of mental events, called "pleasures and pains"; and to think that there is is as if one should solemnly divide human beings into men, women, and blondes. It is of course true that the commonest, and some of the most intense, pleasures and pains are feelings, in my sense of the word. But remorse, which is memory of certain events, having a certain emotional tone, is plainly a pain as much as toothache. And hope, which is expectation of certain events, having a certain emotional tone, is plainly as much a pleasure as the sensation of smell which we get from a rose or a violet.
Now any mental event which has hedonic quality will always have other qualities as well, and its specific hedonic quality will often be causally determined by its specific non-hedonic qualities. Thus the painfulness of remorse and the pleasantness of hope are determined respectively by the specific kinds of emotional quality which these two cognition have. And this is even more obvious in the case of bodily feelings. Headaches and toothaches are both pains, for they both have unpleasant hedonic tone. But each has its own specific sensible quality of "headachiness" and "toothachiness", beside further modifications, such as "stabbingness", "throbbingness", etc., which may be common to both. And the painfulness of these feelings seems to be causally determined by their non-hedonic sensible qualities. At this point I cannot refrain from throwing out an interesting [231] question which I must not pursue further. Is the connexion between such and such non-hedonic qualities and such and such a form of hedonic quality merely causal and logically contingent, or Is it intrinsically necessary? Is it, e.g., logically possible that there should have been minds which had experiences exactly like our experiences of acute toothache in all their sensible qualities, but in whom these sensations were pleasantly toned? The reader may find it amusing to speculate on this question for himself.
We can now deal with the question of pleasures and pains of different quality, which Mill raised, but which he and his critics have so lamentably failed to state clearly. We must first divide the characteristics of any experience into Pure Qualities and Relational Properties. We must then further subdivide the Pure Qualities into Hedonic and Non-hedonic, and the Relational Properties into Causal and Non-causal. Take, e.g., remorse. Its hedonic quality is unpleasantness. It has, beside, that characteristic emotional quality in virtue of which we call it "remorse". It has the non-causal relational property of being A cognition of one's own past misdeeds. And it may have the causal property of tending to make us avoid in future such actions AS we are now regretting. Now it is perfectly plain that there are "differences of quality" among pleasures and pains in the sense that two experiences which were exactly alike in hedonic quality might differ in non-hedonic quality (as a headache and a toothache do), or in non-causal relational property, or in causal property. The pure Hedonist holds that differences of non-hedonic quality and non-causal relational property make no difference to the intrinsic value of an experience. Nothing is relevant to the value of the experience except its hedonic quality and a certain one [232] of its causal properties, viz., what Bentham called its "fecundity". Fecundity is the causal property of tending to produce other experiences which are pleasant or painful. Mill presumably held that, although no experience would have any intrinsic value, positive or negative, unless it were pleasant or painful, yet of two experiences which had precisely the same hedonic quality and precisely the same fecundity one might be better than the other in virtue of some difference in non-hedonic quality, or in non-causal relational property, or in some causal property other than fecundity. This view appears to be perfectly consistent logically, whether it be in fact true or not.
There is, however, another and more subtle sense in which it is conceivable that pleasures or pains might "differ in quality". It is commonly assumed that hedonic tone is a determinable quality having two and only two determinate forms under it, viz., pleasantness and unpleasantness, though of course each can be present in various degrees of intensity. This may very well be true; but there is another possibility which is at least worth mentioning. Is it not possible that there may be several different determinate forms of pleasantness and unpleasantness, just as there are several different shades of redness and several different shades of blueness? If this were admitted, it might be held that nothing is relevant to the goodness or badness of an experience except its hedonic quality and its fecundity, and yet that two experiences which had exactly the same degree of pleasantness and the same fecundity might differ in value because they had this pleasantness in different determinate forms. It is just conceivable that Mill may have meant this. He was so confused that he probably did not himself know precisely what he meant; very likely he was thinking in [233] a vague way of both these entirely different senses of "qualities of pleasure", without ever clearly distinglushing them. A person who took the present view might be called a "pure hedonist" but not a "purely quantitative hedonist".
As regards the characteristics which make an experience intrinsically good or bad, Sidgwick is definitely a pure quantitative hedonist. He seems not to have envisaged the possibility which I have described as pure, but not purely quantitative, hedonism. And his discussion is to some extent confused by the assumption that pleasures and pains are a specific kind of experience, instead of being any kind of experience which happens to have pleasantness or painfulness.
I do not propose to go into the details of Sidgwick's argument. In the end, as he is well aware, each man must decide such questions for himself by direct inspection. All that the philosopher can do is to make sure that no relevant facts have been ignored, that no logical fallacies are committed, and that the issue is not confused by verbal ambiguities. I will therefore put the matter as briefly and clearly as I can in my own way. The contention which we have to examine is that no relational property of an experience, and no quality of it except its hedonic quality, has any bearing on its intrinsic goodness or badness. If this were so, it would follow that no causal characteristic of it can have any bearing on its goodness or badness as a means except its fecundity, i.e., its tendency to produce pleasant or painful experiences. I shall first try to convince the reader that this is not in fact true. And I shall then try to point out the kind of fallacy which is, I think, committed by those persons who profess to show that it is true.
(i) Since this is a general proposition, it can be refuted [234] if we can produce a single convincing contrary instance. Now consider the state of mind which is called "malice". Suppose that I perceive or think of the undeserved misfortunes of another man with pleasure. Is it not perfectly plain that this is an intrinsically bad state of mind, not merely in spite of, but because of, its pleasantness? Is it not plain that any cognition which has the relational property of being a cognition of another's undeserved misfortunes and the hedonic quality of pleasantness will be worse in proportion as the pleasantness is more intense? No doubt malice is a state of mind which on the whole tends to increase human misery. But surely it is clear that we do not regard it as evil, simply as a means. Even if we were quite sure that all malice would be impotent, it seems clear to me that we should condemn it as intrinsically bad.
This example, if it be accepted, not only refutes the general contention of the pure hedonist, but also brings out an important positive fact. Malice is not intrinsically bad simply because it is pleasant; many pleasant states are intrinsically good. And it is not intrinsically bad simply because it has the relational property of being a cognition of another's undeserved happiness; the sorrowful cognition of such an object would not be intrinsically bad. The intrinsic badness of malice depends on the combination of being pleasant with having this particular kind of object. We must therefore be prepared for the possibility that there is no single simple characteristic which is necessary and sufficient to make an experience intrinsically good or bad. It may be that intrinsic goodness or badness always depends on the combination of certain characteristics in the same experience. Any experience which combined the characteristics c1 and c2 might be intrinsically good; any [235] that combined c2 and c3 might be intrinsically bad; whilst experiences which combined c3, and c1 might be neutral.
(ii) Let us now consider what seems to me to be the fallacy in the arguments of pure hedonists. We must begin by remarking that it is logically impossible that an experience should have no characteristic except hedonic quality. It is as clear that no experience could be merely pleasant or painful as that nothing could be black or white without also having some shape and some size. Consequently the hedonist can neither produce nor conceive an instance of an experience which was just pleasant or painful and nothing more; and so he cannot judge by direct inspection that hedonic quality is necessary and sufficient to determine intrinsic value. He is therefore reduced to reflecting on instances in which hedonic quality is combined with non-hedonic characteristics. Now the utmost that he can do is this. He can take in turn each of the non-hedonic characteristics of experiences which could with any plausibility be thought to affect their intrinsic value. These can occur, or be conceived to occur, without hedonic quality, or with various degrees of pleasantness and various degrees of painfulness. He will claim to establish by inspection propositions of the following kind with regard to each of these non-hedonic characteristics.
Variations in the determinate form or in the degree of this non-hedonic characteristic [236] make no difference to the determinate form or the degree of value of the experience.
- When this characteristic is present and hedonic quality is absent the experience has no intrinsic value.
- When this characteristic is present and hedonic quality is also present the experience has intrinsic value.
- The determinate kind of value (goodness or badness) varies with the determinate kind of hedonic quality (pleasantness or unpleasantness), and its degree varies with the degree of the hedonic quality.
I do not think that any hedonist could possibly claim more than to establish these propositions in turn about each non-hedonic characteristic of an experience which seemed worth considering. I have tried to show by a contrary instance that the third of them, at any rate, is not true. But suppose, for the sake of argument that they were all true, what could legitimately be inferred? You could legitimately infer that hedonic quality is necessary to give intrinsic value to an experience. You could legitimately infer that none of these other characteristics is necessary to give intrinsic value to an experience; i.e., that, if you take any one of them, an experience could be intrinsically good or bad without possessing that one. But it would not be legitimate to infer that any experience could have intrinsic value if it had none of these characteristics. For it might be that, although an experience which had hedonic quality could have intrinsic value without c1 being present, and could have it without c3 being present, . . . and could have it without cn being present, yet it could not have intrinsic value unless one or other of the non-hedonic characteristics c1, c2 . . . cn were present m addition to the hedonic quality. To take a parallel case; there is no area which a thing must have in order to be round, but it cannot be round without having some area or other. Thus, even if all the premises which the most optimistic hedonist could demand were granted to him, he would have no right to conclude that the hedonic quality of an experience is sufficient as well as necessary to give it intrinsic value. Even if the variations in intrinsic value wave dependent on variations in hedonic quality and totally independent of variations in any non-hedonic [237] characteristic, it might still be the case that intrinsic value would not be present at all unless there were some non-hedonic characteristic in addition to the hedonic quality. To take a parallel case; the variations in the time of swing of a pendulum are independent of variations in the mass of the pendulum-bob. But it would not swing at all if the bob had no mass.
All arguments for pure quantitative hedonism, including Sidgwick's, with which I am acquainted overlook these elementary logical points. I conclude then that the arguments for this doctrine are certainly fallacious, and that the doctrine itself is almost certainly false.
Here, if I were wise, I should leave the matter. But I cannot resist the temptation of starting one more hare before I turn to another topic. We have so far talked of pleasantness and painfulness as two determinate forms of a certain determinable quality (hedonic tone) which may belong to any kind of experience. We have noted that it is a priori impossible that any experience should have only hedonic quality; it must always have some non-hedonic quality (such as toothachiness, throbbingness, etc.), and this will determine its hedonic quality. Now this suggests the following possibility. Is it not possible that what we have called "hedonic quality" is really a relational property and not a quality at all? Is it not possible that the statement: "This experience of mine is pleasant" just means: "I like this experience for its non-hedonic qualities"? I may dislike the experience as a whole, because it will have causal and non-causal relational properties in addition to its non-hedonic qualities. I like the experience of malice for its emotional quality; but I cannot confine my attention to this. I have to consider also its relational property of [238] having for its object the undeserved misfortunes of another; and my dislike for the combination of this emotional quality with this relational property overbalances my liking for the experience regarded simply as having the emotional quality. On this view we should no longer divide the qualities of an experience into hedonic and non-hedonic. All its qualities would be non-hedonic. But, if its qualities were such that I liked it for them it would be pleasant, and if its qualities were such that I disliked it for them it would be painful. And it would remain pleasant in the first case even though I disliked it on the whole, and painful in the second case even though I liked it on the whole. I think it is worth while to throw out this suggestion; but I do not wish to attach much weight to it. My argument against pure quantitative hedonism is independent of its truth or falsity. I am inclined to think that Sidgwick is taking a somewhat similar view in the very difficult discussion in Book II, Chap. II, Sect. 2, and in Book III, Chap. XIV, Sect. 4.
(2, 12) We can now pass to what I have called the Factual Problem of Hedonism. This is simply the question whether approximately accurate estimates can be made of the relative balance of pleasure and pain in alternative future possible states of affairs. This is discussed in Book II, Chaps. III to VI inclusive. Naturally every difficulty which there is in estimating the relative hedonic value of alternative future states of oneself is intensified when one tries to make such estimates about the states of other men, as Utilitarianism needs to do. I have only one comment to make. The admitted difficulties of forming such estimates are often alleged as a conclusive objection to Hedonism in general and to Utilitarianism in particular. This is no doubt legitimate as an argumentum ad hominem against any [239] Hedonist who gives himself airs and maintains that it would be easy to know what our duty is if Hedonism were true. But the important fact for those of us who have no particular ethical axe to grind is this. Hedonism has to be rejected, not because it is too complicated, but because it is far too simple. On any ethical theory which attempts to do justice to all the facts, estimates will have to be made in comparison with which those demanded by Hedonism would be child's play. In the first place, in judging the rightness of an action we shall have to balance its fittingness to the total course of events which it will modify against the intrinsic goodness and badness of these modified events. Secondly, as we have seen, the estimation of this total fittingness involves an elaborate balancing and composition of partial fittingnesses and unfittingnesses. Lastly, we have now seen reason to think that the intrinsic goodness or badness of any state of affairs will depend on many different factors, of which hedonic quality is only one. Living in such a glass-house, we shall be most unwise to cast stones at Utilitarianism on the ground of the impossible complexity of the estimates which it demands of us.
This completes what I have to say about (2, 1) Hedonism in General. I do not think that it would be profitable to comment separately on (2, 2) Egoistic Hedonism and (2, 3) Universalistic Hedonism. The only point that I wish to make is that there would seem to be no need for an egoistic teleological theory to be hedonistic. Green's theory of Self-realisation as the ultimate end to be aimed at is plainly a form of Egoism, and equally plainly not a form of Hedonism. Sidgwick's view appears to be (cf. Book I, Chap. VII) that all other forms of Egoism are so vague as to be hardly worth discussing. On investigation they prove either not to be egoistic or to be hedonistic. This may in fact be historically correct, but it does not seem clear that there could not be a perfectly definite form of non-hedonistic Egoism. However this may be, the important point which remains to be discussed is the controversy of Ethical Egoism with Nonegoistic theories of ethics. This comes under the heading:
THE RELATION BETWEEN THE THREE METHODS.