ONE of the objects specified in Lord Gifford's will as forming the purpose of his lectures is that of ‘Promoting, Advancing, Teaching, and Diffusing… the Knowledge of the Nature and Foundation of Ethics or Morals, and of all Obligations and Duties thence arising’. This is the object to which I intend to devote myself in this course of lectures. I propose to take as my starting-point the existence of what is commonly called the moral consciousness; and by this I mean the existence of a large body of beliefs and convictions to the effect that there are certain kinds of acts that ought to be done and certain kinds of things that ought to be brought into existence, so far as we can bring them into existence. It would be a mistake to assume that all of these convictions are true, or even that they are all consistent; still more, to assume that they are all clear. Our object must be to compare them with each other, and to study them in themselves, with a view to seeing which best survive such examination, and which must be rejected either because in themselves they are ill-grounded, or because they contradict other convictions that are better grounded; and to clear up, so far as we can, ambiguities that lurk in them.
This is the time-honoured method of ethics. It was the method of Socrates and of Plato; we find constantly, in Xenophon's Memorabilia or in a dialogue of Plato, some ambitious definition of a virtue rejected on the ground that it would lead us to regard some act as virtuous, or again as unvirtuous, that no one really thinks to be so. It was the method of Aristotle, and has indeed nowhere been better formulated than it is by him. ‘We must… set the observed facts before us and, after first discussing the difficulties, go on to prove, if possible, the truth of all the common opinions about these affections of the mind, or, failing this, of the greater number and the most authoritative; for if we both refute the objections and leave the common opinions undisturbed, we shall have proved the case sufficiently.’1 Again, Kant's method was the same. ‘I have adopted in this work’, he says in the Preface to the Grundlegung, ‘the method which I think most suitable, proceeding analytically from common knowledge to the determination of its ultimate principle’;2 and to this ‘common knowledge’ he again and again returns, as to that on which his own theory is based and by comparison with which it must from time to time be tested.
I would add a further remark. Aristotle habitually takes as a starting-point a consideration of the opinions not only of the many but also of the wise. He is predisposed to think that in all the main theories, no less than in the views of the plain man, there is much that is true, and that even when theories are in broad opposition to each other, each is probably erring only by overstatement or mis-statement of something that is profoundly true. It would indeed be strange if any of the main theories of ethics were completely in error; it is far more likely that each has grasped something that is both true and important but has, not through blindness to moral values but by some apparently trivial logical error, claimed as the whole truth what is only one of a set of connected truths. I may illustrate the ideal which ethical theory should aim at, by an example taken from a different field. A physical object which has a certain shape, say the circular shape, when seen from different points of view presents a variety of apparent shapes, ranging from the circular through a whole series of elliptical shapes, which are explicable as the result of different perspectives. So too, if we could reach the truth about the essential problems of ethics, we should be able to recognize the varieties of common opinion and the varieties of philosophical view as being none of them wholly false, but all of them distortions of the truth due to the different perspectives in which men have looked at the problems. Some slight contribution to this result is all that I would claim for the attempt which will follow.
The method of ethics is in this respect different from that of the physical sciences. In them it would be a great mistake to take as our starting-point either the opinions of the many or those of the wise. For in them we have a more direct avenue to truth; the appeal must always be from opinions to the facts of sense-perception; and natural science entered on its secure path of progress only when in the days of Galileo men began to make careful observations and experiments instead of relying on a priori assumptions that had hitherto prevailed. In ethics we have no such direct appeal. We must start with the opinions that are crystallized in ordinary language and ordinary ways of thinking, and our attempt must be to make these thoughts, little by little, more definite and distinct, and by comparing one with another to discover at what points each opinion must be purged of excess and mis-statement till it becomes harmonious with other opinions which have been purified in the same way.
In the complex fabric of common opinions about moral questions two main strands may be discovered. On the one hand, there is a group of opinions involving the closely connected ideas of duty, of right and wrong, of moral law or laws, of imperatives. On the other hand, there are opinions involving the idea of goods or ends to be aimed at. In the one case the ideal of human life is envisaged as obedience to laws, in the other as the progressive satisfaction of desire and attainment of ends. The one may be called the Hebrew, the other the Greek ideal. In the first way of thinking, the laws of human life were originally thought of not as grasped on their own merits by human thought but as having been authoritatively revealed on Mount Sinai as the will of God. But as ethics came to win for itself a status independent of religion, these laws or others like them came to be thought of as grasped by an intuitive act of human reason, as categorical imperatives directly apprehended as involved in the nature of the moral universe. At the same time the plurality of the imperatives in the original code tended to be pruned down, as when Christ reduced the ten commandments to two, stating our duty to God and to man respectively, or as when Kant reduced the multiplicity of imperatives to the one imperative, ‘act so that your act can be universalized’. But through all such varieties of view the thought remains in essence the same, that the essence of the good life is obedience to one or more principles. On the other hand, we have the whole variety of teleological systems of ethics, which start with the thought of certain things as good, and of the good life as essentially the attempt to bring these into existence. In the crudest form of this general type of theory, the individual's own pleasure is thought of as the supreme end. In the course of time, this view was modified in two directions. On the one hand, other elements than pleasure were recognized as elements in the end; as when Aristotle substituted for pleasure euvdaimoni,a which he thought of as including pleasure indeed but as having for its main constituent good activity. On the other hand, the general pleasure came to be thought of as being, instead of his own pleasure, the proper object of each agent's activity. And finally, by a fusion of these two corrections of the original, crude view, we have the view of ideal Utilitarianism, that the supreme end is to secure, both for oneself and for others, a life which includes in it both good activity and pleasure.
The general antithesis between ethical systems in which duty is the central theme, and those in which goods or ends are the central theme, is clear enough. Yet it would be a mistake to suppose that there has ever been an ethics of duty which did not include a recognition of intrinsic goods, or an ethics of ends which did not include a recognition of duties. Kant's ethics has been perhaps the nearest approach to a pure ethics of duty; and he claims to evolve the whole duty of man by an analysis of the implications of the notion of duty, without introducing the thought of goods to be aimed at. But it has often been shown that when it comes to the point, he has to argue for the wrongness of certain acts on the ground of the badness of the results they bring. On the other hand, Aristotle's ethics would seem at first sight to be based entirely on the notion of a good or an end to be achieved; but in his discussion of the individual virtues he does not relate the virtuous act to the final goal of human life, but treats it as simply right in its own nature. Nor are these mere lapses on the part of Kant or Aristotle, due to lack of firmness of purpose. The facts have been too strong for them; both the notion of right and the notion of good are implied in the study of moral questions, and any one who tries to work with one only will sooner or later find himself forced to introduce the other.
The question remains whether either is more fundamental than the other, in the sense that the other can be defined by reference to it. Only a very careful attention to each of the two terms will justify us in giving any answer to this question, and we may find even after careful attention that we cannot give a simple answer.
The question what is the relation between the attributes goodness and rightness is, however, only part of a larger question or series of questions which can be asked about either of them. About each of them we can, to begin with, ask the question whether it is definable or indefinable. By this question I mean the question, with regard to ‘right’ and again with regard to ‘good’, whether that which we are thinking of when we use these terms can be fully expressed by using a complex expression of such a form as ‘a which is b’ or ‘m in the relation r to n’ in which none of the terms used is a synonym of the term about which we are asking the question. The question is, in fact, whether ‘good’ or ‘right’ can be elucidated without remainder in terms other than itself.
The various theories which offer definitions of ethical terms may be classified in various ways, by using a variety of different principles. It seems to me that it is on the whole best to divide them into two main classes. In one of these the term in question is defined by reference to the attitude of some being or other. One would be holding a view of this kind if one defined a right act, or a good moral state, as one which God approves, or again as one which a majority of men approve. Here the rightness or goodness of that which is right or good is identified with God's having, or most men's having, a certain attitude towards that thing. In the other main type of view, the term in question is defined by reference to the total consequences of the act or moral state in question. One would be holding a view of this kind if one defined a right act as one which would produce a maximum of life, or as one which would produce a maximum of happiness.
This is not a logically perfect classification of the attempts to define ethical terms; for it is not based on an a priori disjunction. It is not possible to see a priori that any such attempt must be an attempt to define them either by reference to a mental attitude or by reference to total consequences, as it is possible to see that any angle must be right, acute, or obtuse. Nor do I see how this list of the types of theory could be added to so as to make a complete list of all the possible types of attempt to define ethical terms. What I think we can assure ourselves of by inspection is that in fact all or almost all the attempts to define them have conformed to one or other of these types.
There is another way of classifying them which cuts right across this classification, viz. Professor Moore's classification into naturalistic and non-naturalistic definitions. The former are definitions which claim to define an ethical term without using any other ethical term; the latter are attempts to define one ethical term by the aid of another. It is clear that mental-attitude theories may be of either of these types. If you define ‘right’ as meaning what is approved by the community, you are putting forward a naturalistic definition. If you define ‘good’ as meaning ‘such that it ought to be desired’, you are putting forward a non-naturalistic definition. Consequence theories also may be either naturalistic or non-naturalistic. If you define ‘right’ as ‘productive of the greatest pleasure’, you are putting forward a naturalistic definition. If you define it as ‘productive of the greatest amount of good’, you are putting forward a non-naturalistic definition.
We might adopt either of these classifications as our main classification, and use the other for purposes of subdivision.
We may either use the scheme:
or the scheme:
From one point of view the former classification seems the more fundamental. It seems to bring together the theories that have most in common. Thus the definition of ‘right’ as ‘productive of most pleasure’ (2a) seems to have more in common with the definition of it as ‘productive of most good’ (2b) than with the definition of it as ‘approved by society’ (1a). And historically the two former are more closely connected; for theory (2b) has in fact been produced by reflection on the shortcomings of theory (2a), which has no historical connexion with theory (1a).
Yet in their general colouring, if I may put it so, all naturalistic theories have more in common than any of them has with any non-naturalistic theory. For all naturalistic theories amount to saying that all the statements in which we use either the predicate ‘right’ or the predicate ‘good’, and think that in doing so we are dealing with a very special kind of attribute, are really statements of ordinary matters of fact which can be discovered by mere observation. Whether a certain kind of act is commanded by society, or whether it produces more pleasure than any other possible act would, is a thing to be discovered (if at all) by ordinary observation; if that is all that ‘ought’ means, there is no need and no place for a special branch of study called ethics; for there are no ineradicably ethical terms. On the other hand, all non-naturalistic theories have more in common with one another than any of them has with any naturalistic theory; for while they define some ethical terms by reference to others, they preserve at least one ethical term as irreducibly different from any term expressive of ordinary matter of fact. And if, as we have seen, there have been historical connexions between certain naturalistic and certain non-naturalistic theories, there have also been historical connexions between attitude theories and consequence theories. The whole sociological school of ethics, while it is evolutionary in its origin, and therefore began by defining ethical terms by reference to the biological consequences of certain acts or states, shows a tendency to relapse into defining ‘right’ as ‘commanded by the community’—a tendency most clear in the French sociologists, who claim to reduce ethics to the science des mæurs, the historical and comparative study of the codes current in different communities, and reject the notion that there is any absolute standard by which these codes can be judged to be higher or lower.
On the whole, then, the division into naturalistic and non-naturalistic theories is the more important, and I have had it always in the background of my mind, though I have not thought it necessary to discuss all the naturalistic theories consecutively and all the non-naturalistic theories consecutively.
It is not always clear at first sight to which type a well-known theory really belongs. At first sight Hedonism, in all its forms, whether egoistic or utilitarian, would seem to belong to the class of naturalistic theories about rightness, and it is often so described—described as ‘reducing’ rightness to the tendency to produce pleasure. But we must be careful to distinguish two possibilities. A hedonist may take the view that this is what rightness is, that this is its correct definition; and then he is offering a naturalistic theory. But he may be holding something quite different. He may be holding that rightness is something indefinable, and merely claiming that that which makes right acts right is their tendency to promote pleasure. Then he is holding that a non-ethical characteristic, a pyschological characteristic, is the ground of rightness but not its essence. And if so, the theory, whether right or wrong, is not a naturalistic theory; the specific quality of rightness is then still recognized as something not reducible to merely psychological terms.
The hedonists have not always been very clear which of the two things they meant. Bentham, for instance, seems to take the naturalistic view. I take leave to quote some sentences from an earlier book of my own.3
‘He says4 that “when thus interpreted” (i.e. as meaning “conformable to the principle of utility”), “the words ought and right… and others of that stamp, have a meaning; when otherwise, they have none”. And elsewhere5 he says “admitting (what is not true) that the word right can have a meaning without reference to utility”. Yet as Sidgwick points out,6 “when Bentham explains (Principles of Morals and Legislation, chap. i, para. 1, note) that his fundamental principle ‘states the greatest happiness of all those whose interest is in question as being the right and proper end of human action’, we cannot understand him really to mean by the word ‘right’ ‘conducive to the general happiness’; for the proposition that it is conducive to general happiness to take general happiness as an end of action, though not exactly a tautology, can hardly serve as the fundamental principle of a moral system.” Bentham has evidently not made up his mind clearly whether he thinks that “right” means “productive of the general happiness”, or that being productive of the general happiness is what makes right acts right; and would very likely have thought the difference unimportant. Mill does not, so far as I know, discuss the question whether right is definable. He states his creed in the form “actions are right in proportion as they tend to promote happiness”,7 where the claim that is made is not that this is what “right” means, but that this is the other characteristic in virtue of which actions that are right are right. And Sidgwick says8 that the meaning of “right” or “ought” “is too elementary to admit of any formal definition”, and expressly repudiates9 the view that “right” means “productive of any particular sort of result”.’
It is impossible to know what the Egoism or the Utilitarianism of any particular thinker means, and to pass judgement on it, until we have decided which of the two things the writer in question means; and since different writers have meant different things, and the same writer has sometimes not known which of the two he meant, it is not possible to say that Egoism or again that Utilitarianism is the name of any one ethical theory. Each of the two names may stand for either of two theories which logically are quite different, though they unite in the general characteristic of laying great stress on pleasure in the discussion of ethical questions.
Up to now I have referred to these various views as views that may be taken about ethical characteristics in general. It is clear, however, that it is possible to hold one type of view about one ethical characteristic and another about another, to hold for instance a relational view about ‘right’ and a non-relational view about ‘good’; and these views might be united in a single system. So that we cannot use this classification as a classification of ethical systems taken as a whole; though there will probably be a tendency for a thinker who holds one type of view about one ethical characteristic to hold a similar view about others. Again, it is possible that one type of view might be true about one characteristic and false about another. From this point onwards, we must consider the main ethical characteristics separately, and ask what kind of view is true about them. Now the two main ethical characteristics or groups of characteristics are those which are designated by such terms as ‘right’, ‘obligatory’, ‘my duty’ on the one hand, and by ‘good’, ‘noble’, ‘valuable’ on the other. A good many people are disinclined to admit any clear-cut distinction between the two groups, and inclined to use ‘good’ and ‘right’, for instance, indifferently as applied to actions. And I should be the last to claim that in all our ordinary use of language we draw a clear distinction between the two words. What I would claim, and I think the claim is very important if we want to think clearly about ethics, is that there are two quite different characteristics, which, once we have grasped the difference between them, we see it to be proper to designate by different words, even if we do not always in ordinary speech so designate them; and I suggest ‘right’ as on the whole the most convenient word for designating the one, and ‘good’ as the most convenient term for designating the other. That the characteristics are different I hope to show later.10
Eth. Nic. 1145b 2–7.
Akad. Ausgabe, iv. 392 (Abbott's translation, p. 9).
The Right and the Good, 7–8.
Principles of Morals and Legislation, chap, i, para. 10.
Ibid., para. 14, 10.
Methods of Ethics, ed. 7, 26 n.
Utilitarianism, copyright eds., 9.
Methods of Ethics, ed. 7, 32.
Cf. pp. 165–7.