From the many topics in Moore's ethical writings which might profitably be discussed I am going to choose two in the present paper. They are (1) his attempted refutation of Ethical Egoism, and (2) his distinction between "Natural" and "Non-natural" characteristics, and his doctrine that the word "good" (in one very important use of it) is a name for a certain non-natural characteristic.

(I) ETHICAL EGOISM

I shall begin by defining three opposed terms, viz., "Ethical Egoism," "Ethical Neutralism," and "Ethical Altruism." The second of these is the doctrine which Moore accepts in Principia Ethica; the other two are extreme deviations from it in opposite directions. It will therefore be best to start with ethical neutralism.

The neutralist theory is that no-one has any special duty to himself as such, and that no-one has any special duty to others as such. The fundamental duty of each of us is simply to maximise, so far as he can, the balance of good over bad experiences throughout the whole aggregate of contemporary and future conscious beings. Suppose that A, by giving to B a good experience at the cost of foregoing a good experience or incurring a bad one himself, can increase this balance more than by any other means; then it is A's duty to do so. Suppose, on the other hand, that A, by getting a good experience for himself at the cost of depriving B of a good experience or giving him a bad one, can increase this balance more than by any other means; then it is equally A's duty to do so.

Ethical Egoism is the doctrine that each man has a predominant obligation towards himself as such. Ethical Altruism is the doctrine that each man has a predominant obligation towards others as such. These doctrines might be held in milder or more rigid forms according to the degree of predominance which they ascribe to the egoistic or the altruistic obligation respectively. The extreme form of Ethical Egoism would hold that each man has an ultimate obligation only towards himself. The extreme form of Ethical Altruism would hold that each man has an ultimate obligation only towards others. According to the former extreme each man's only duty is to develop his own nature and dispositions to the utmost and to give himself the most favourable balance that he can of good over bad experiences. He will be concerned with the development and the experiences of other persons only in so far as these may affect, favourably or unfavourably, his own development and his own experience. This doctrine seems to have been held by Spinoza. The extreme form of Ethical Altruism would hold that each man's only duty is to develop to the utmost the natures and the dispositions of all other persons whom he can affect and to give them the most favourable balance that he can of good over bad experiences. He will be concerned with his own experiences only in so far as they may affect, favourably or unfavourably, the development and experiences of other persons.

Now Moore professes to show in Principia Ethica (96-105) that Ethical Egoism is self-contradictory; and, if his argument were valid, a very similar argument could be used to refute Ethical Altruism. He alleges that Ethical Egoism involves the absurdity that each man's good is the sole good, although each man's good is different from any other man's good. In my opinion it involves nothing of the kind; and I will now try to justify that opinion.

First of all, what do we mean by such phrases as "Smith's good" or "Jones' evil?" The good of Smith is just those good experiences, dispositions, etc., which are Smith's; and the evil of Jones is just those bad experiences, dispositions, etc., which are Jones'.

Suppose now that A is an ethical egoist. He can admit that, if a certain experience or disposition of his is good, a precisely similar experience or disposition of B's will be also and equally good. But he will assert that it is not his duty to produce good experiences and dispositions as such, without regard to the question of who will have them. A has an obligation to produce good experiences and dispositions in himself and no such direct obligation to produce them in B or in anyone else. A recognises that B has no such direct obligation to produce them in A or in anyone else. This doctrine does not contradict itself in any way.

What it does contradict is Sidgwick's second axiom about goodness and our obligations in respect of producing it. This is stated as follows in Book III, Chapter XIII of Sidgwick's Methods of Ethics (382, in the sixth edition):—" . . . as a rational being I am bound to aim at good generally—so far as it is attainable by my efforts—not merely at a particular part of it." Since Sidgwick was an ethical hedonist, he held that nothing is intrinsically good or bad except experiences. Therefore, to "aim at good" will mean to try to produce good experiences and to avert or diminish bad experiences. Therefore this axiom means that it is each person's duty to try to produce the greatest possible net balance of good over bad experiences among all the conscious beings whom he can affect, and that he has no direct obligation to produce such experiences in one rather than in another person or set of persons, e.g., in himself as such or in others as such. Suppose he confines his efforts to himself or to his family or to his social class or to his countrymen, or even to that very extended but still restricted group which consists of everyone but himself; then he will always need some positive justification for this restriction. And the only admissible justification is that, owing to his special limitations or their special relations to him, he can produce more good on the whole by confining his efforts to a certain restricted set of conscious beings. Any restriction in the range of one's beneficent activities needs ethical justification, and this is the only valid ethical justification for it.

It is evident, then, that Sidgwick's axiom is equivalent to the assertion of ethical neutralism, and that ethical egoism is inconsistent with it. But this does not make ethical egoism self-contradictory; and, unless Sidgwick's axiom be self-evidently true, the inconsistency of ethical egoism with it does not refute that doctrine.

Precisely similar remarks would apply to any argument against ethical altruism on the same lines as Moore's argument against ethical egoism. Suppose that A is an ethical altruist. He can admit that, if a certain experience or disposition of B's is good, a precisely similar experience or disposition of his own will be also and equally good. But he asserts that it is not his duty to produce good experiences and dispositions as such, without regard to the question of who will have them. A has an obligation to produce good experiences and dispositions in others, and no such direct obligation to produce them in himself. A recognises that B has an obligation to produce good experiences and dispositions in A and in everyone except B, and that B has no obligation to produce them in B. This doctrine, again, contradicts Sidgwick's second axiom about goodness and our obligations in producing it. But it is not self-contradictory, and, unless Sidgwick's axiom be self-evident, the inconsistency of ethical altruism with it does not refute the latter doctrine.

An illuminating way of putting the difference between ethical neutralism and the other two theories is the following. Ethical neutralism assumes that there is a certain one state of affairs—"the sole good"—at which everyone ought to aim as an ultimate end. Differences in the proximate ends of different persons can be justified only in so far as the one ultimate end is best secured in practice by different persons aiming, not directly at it, but at different proximate ends of a more limited kind. The other two theories deny that there is any one state of affairs at which everyone ought to aim even as an ultimate end. In fact each theory holds that there are as many ultimate ends as there are agents. On the egoistic theory the ultimate end at which A should aim is the maximum balance of good over evil in A's experiences and dispositions. The ultimate end at which B should aim is the maximum balance of good over evil in B's experiences and dispositions. And so on for C, D, etc. On the altruistic theory the ultimate end at which A should aim is the maximum balance of good over evil in the experiences and dispositions of all others-than-A. The ultimate end at which B should aim is the maximum balance of good over evil in the experiences and dispositions of all others-than-B. And so on for C, D, etc. The main difference between the two theories is that for egoism the various ultimate ends are mutually exclusive, whilst for altruism any two of them have a very large field in common.

Now there is nothing self-contradictory in the doctrine that, corresponding to each different person, there is a different state of affairs at which he and he only ought to aim as an ultimate end. And there is nothing self-contradictory in the doctrine, which is entailed by this, that there is no one state of affairs at which everyone ought to aim as an ultimate end. Moore simply assumes, in common with Sidgwick, that there must be a certain state of affairs which is the ultimate end at which everyone ought to aim; shows that ethical egoism is inconsistent with this assumption; and then unjustifiably accuses ethical egoism of being self-contradictory.

Granted that even the extreme forms of ethical egoism and ethical altruism are internally consistent, is there any reason to accept or to reject either of them?

(1) If ethical neutralism were true, they could both be rejected. Now the following argument can be produced in favour of ethical neutralism. On any theory except this it will sometimes be right for a person to do an act which will obviously produce less good or more evil than some other alternative act which is open to him at the time. E.g., it is often the case that A could either (i) do an act which would add something to his well-being at the cost of diminishing B 's by a certain amount, or (ii) do another act which would increase his own well-being rather less at the cost of diminishing B's very much less. Plainly A would in general be producing more good by doing the latter act than by doing the former. But, if ethical egoism be true, it would be his duty to do the former and to avoid doing the latter. Again, it is often the case that A could either (i) do an act which would add something to B's well-being at the cost of diminishing his own by a certain amount, or (ii) do another act which would increase B's well-being rather less and diminish his own very much less. Plainly A would in general be producing more good by doing the latter act than by doing the former. But, if ethical altruism be true, it would be his duty to do the former and to avoid doing the latter. I think, therefore, that ethical neutralism is the only one of the three types of theory which can be combined with the doctrine that the right act in any situation will always coincide with the optimific act in that situation. Since Utilitarians hold the latter view, they ought to hold the former; and so Sidgwick was right, as a Utilitarian, to lay down an axiom which is equivalent to ethical neutralism.

It is possible, however, to distinguish between what I call an "optimising" act, and what I have called an "optimific" act; and it might be possible to combine ethical altruism with the doctrine that the right act always coincides with the optimising act. I will now explain this distinction and justify this assertion. Suppose that a certain act á is done in a certain situation S, and that the amount of value in the universe is thereby changed. It is possible to distinguish two quite different contributions which this act may make to the amount of value in the world. They may be called its direct and its consequential contributions. The act may have qualities which make it intrinsically good or bad. Even if it has not, it will have non-causal relations to other factors in the contemporary and the past situation which forms its context; and in virtue of these the whole composed of the act and the situation may be better or worse than the situation by itself or the situation combined with a different act. Any value, positive or negative, which accrues in this way to the universe through the occurrence of this act may be called its "direct" contribution. Again, the act co-operates as a cause-factor with other factors in the contemporary situation and thus leads to a train of consequences which are different from those which would have followed if no act or a different act had been done. Any value or disvalue which accrues to the universe through the values or disvalues of the consequences of an act may be called its "consequential" contribution. The "total" contribution of an act consists of its direct and its consequential contributions. An "optimific" act in a given situation may be defined as one whose consequential contribution to the value in the universe is at least as great as that of any alternative act open to the agent. An "optimising" act in a given situation is one whose total contribution to the value in the universe is at least as great as that of any alternative act open to the agent.

Let us now apply these notions. Suppose that there is a situation in which A can either (i) do an act which will increase B's well-being at the cost of considerably diminishing his own, or (ii) do an act which would add rather less to B's well-being and diminish his own very much less. Suppose further that an act of self-sacrifice has, as such, a certain amount of moral value. Then it might be that the direct contribution which the former act would make, as an act of self-sacrifice, more than counter-balances the consequential diminution which it causes by decreasing the agent's own well-being. So an altruistic act might be the optimising act even when it is not the optimific act.

Now common-sense does ascribe considerable positive value to acts of voluntary self-sacrifice as such. It is therefore conceivable that the right act, on the extreme altruistic view, might always coincide with the optimising act. But it is not necessary that it should, and it seems most unlikely that it always would. For it seems easy to conceive situations in which the most altruistic act possible would increase the well-being of others very slightly and would diminish that of the agent very much, whilst some other possible act would increase the well-being of others only a little less and would positively increase that of the agent. In such a situation it is most unlikely that the most altruistic act would be an optimising act, even when its direct contribution to the goodness in the universe, as an act of self-sacrifice, was taken into account.

It is quite plain that no attempt on these lines to reconcile ethical egoism with the doctrine that the right act coincides with the optimising act would be plausible. For common-sense attaches no positive value to an act of sacrificing others for one's own benefit, as such. Therefore, when what would be the right act on the extreme egoistic view fails to coincide with the optimific act, it is impossible that it should coincide with the optimising act.

The upshot of this discussion is as follows. Many people find it self-evident that the right act in any situation must coincide with the optimific act. Anyone who does so can safely reject pure ethical egoism and pure ethical altruism, and will almost be forced to accept ethical neutralism. When the distinction between an optimising and an optimific act is pointed out to them many of those who thought it evident that the right act must coincide with the optimific act would be inclined to amend their doctrine by substituting "optimising" for "optimific." Such people could safely reject pure egoism. It is not impossible that the most altruistic act should coincide with the optimising act even when it fails to coincide with the optimific act. But, even if it always did so, this would be a merely contingent fact; whilst they hold that the coincidence between the right act and the optimising act is necessary, since they find it self-evident. And it is very unlikely that the most altruistic act would in fact always coincide with the optimising act. Therefore the substitution of "optimising" for "optimific" would make no difference in the end. Those who find the coincidence of the right act with the optimising act self-evident could safely reject both pure altruism and pure egoism, and would have to accept ethical neutralism as the only principle of distribution which is compatible with their axiom.

(2) So far we have considered the grounds for accepting ethical neutralism, and therefore the indirect reasons for rejecting ethical egoism and ethical altruism. We will now consider the attitude of common-sense towards each of the three alternatives when judged on its own merits, (i) Common-sense would reject pure ethical egoism out of hand as grossly immoral. It is, I think, doubtful whether anyone would accept ethical egoism unless, like Spinoza, he had already accepted psychological egoism. If a person is persuaded that it is psychologically impossible for anyone to act non-egoistically, he will have to hold that each man's duties are confined within the sphere which this psychological impossibility marks out. But common-sense does not accept psychological egoism, though many philosophers have done so; and on this point common-sense is right and these philosophers are tricked by certain rather subtle ambiguities of language.

(ii) The attitude of common-sense (at any rate in countries where there is a Christian tradition) towards pure ethical altruism is different. It would, with an uncomfortable recollection of some rather disturbing passages in the Sermon on the Mount, be inclined to describe the doctrine as quixotic or impracticable but hardly as immoral. Apart from the embarrassment which persons in a Christian country feel at saying or implying that Christ sometimes talked nonsense, there is a sound practical reason for this attitude. We realise that most people are far more likely to err on the egoistic than on the altruistic side; that in a world where so many people are too egoistic it is as well that some people should be too altruistic; and that there is something heroic in the power to sacrifice one's own happiness for the good of others. We therefore hesitate to condemn publicly even exhibitions of altruism which we privately regard as excessive.

(iii) Although common-sense rejects pure egoism and does not really accept pure altruism, I do not think that it is prepared to admit neutralism without a struggle. It would regard neutralism as in some directions immorally selfish and in other directions immorally universalistic. It undoubtedly holds that each of us has more urgent obligations to benefit certain persons who are specially related to him, e.g., his parents, his children, his fellow-countrymen, his benefactors, etc., than to benefit others who are not thus related to him. And it would hold that the special urgency of these obligations is founded directly on these special relations.

(iv) The ideal of common-sense then is neither pure egoism nor pure altruism nor neutralism. I think it may be best described as "Self-referential Altruism." I will now explain what I mean by this. Each of us is born as a member of a family, a citizen of a country, and so on. In the course of his life he voluntarily or involuntarily becomes a member of many other social groups, e.g., a school, a college, a church, a trades-union, etc. Again, he gets into special relations of love and friendship with certain individuals who are not blood-relations of his. Now the view of common-sense is roughly as follows.

Each of us has a certain obligation to himself, as such. I do not think that common-sense considers that a person is under any obligation to make himself happy, i.e., to "give himself a good time." Possibly this is because most people have so strong a natural tendency to aim at getting the experiences which they expect to like and at avoiding those which they expect to dislike. On the other hand, the obligation to develop one's own powers and capabilities to the utmost, and to organise one's various dispositions into a good pattern is considered to be a strong one. This kind of action often goes very much against the grain, since it conflicts with natural laziness and a natural tendency to aim at the easier and more passive kinds of good experience. The obligation to make others happy and to prevent or alleviate their unhappiness is held to vary in urgency according to the nature of their relation to oneself. It is weakest when the others stand in no relation to the agent except that of being fellow sentient beings. It is strongest when the others are the agent's parents or his children, or non-relatives whom he loves and by whom he is loved, or persons from whom he has received special benefits deliberately bestowed at some cost to the giver.

A person's obligation towards A is more urgent than his obligation towards B if it would be right for him to aim at the well-being of A before considering that of B, and to begin to consider that of B only after he had secured a certain minimum for A. The greater this minimum is, the greater is the relative urgency of his obligation towards A as compared with his obligation towards B.

Now common-sense holds that it is one's duty to be prepared to sacrifice a considerable amount of one's own well-being to secure a quite moderate addition to the net well-being of one's parents or children or benefactors, if this be the only way in which one can secure it. But it does not consider that a person has a duty to sacrifice much of his own well-being in order to secure even a substantial addition to the net well-being of others who stand in no specially intimate relations to him.

Common-sense draws a sharp distinction between making oneself happy, and developing one's own powers and capacities to the utmost and organising one's disposition into a good pattern. The obligation to make oneself happy is held to be vanishingly feeble, whilst the obligation to develop and organise oneself is held to be very urgent. Hence it is felt to be doubtful how far a person ought to sacrifice self-development and self-culture, as distinct from his own happiness, in order to add to the happiness of others. It is only when the claim is very strong, as in the case of a child on its parents or of aged and infirm parents on a grown-up son or daughter, that common-sense approves of this kind of self-sacrifice. Even here it feels considerable hesitation; and, apart from such cases, it is extremely embarrassed. It realises that it is a good thing on the whole that a certain proportion of people should voluntarily forego the development of a great many valuable aspects of their personality in order to live in the slums and add to the well-being of other persons who have no specially urgent claims on them. But, whilst it admires those who make the sacrifice, it regrets the waste of talent; and it is relieved to think that there is no great danger of too many gifted persons following their example. On the whole it favours a kind of ethical "division of labour." A certain minimum of self-sacrifice and of self-culture is demanded of everyone; but, when this minimum has been reached, commonsense approves of certain persons specialising in self-culture and others in beneficent self-sacrifice.

Lastly, common-sense considers that each of us has direct obligations to certain groups of persons of which he is a member, considered as collective wholes. The most obvious case is one's nation, taken as a collective unity. It is held that an Englishman, as such, is under an urgent obligation in certain circumstances to sacrifice his happiness, his development, and his life for England, and is under no such obligation to Germany; and that a German is under an obligation in similar circumstances to make a similar sacrifice for Germany, and is under no such obligation to England. And so on. It should be noticed that Germans, as well as Englishmen, admit that Englishmen have this peculiar obligation towards England; and that Englishmen, as well as Germans, admit that Germans have this peculiar obligation towards Germany. This is clearly recognised by the saner citizens of both countries even when they are at war with each other.

It seems to me that the fact that an Englishman considers that a German should sacrifice himself for Germany, even when his doing so is detrimental to England, and that a German considers that an Englishman should sacrifice himself for England, even when his doing so is detrimental to Germany, is of some theoretical importance. It certainly suggests that we are concerned here with a genuine and objective, though limited, obligation; and not with a mere psychological prejudice in favour of one's own nation. Opinion has varied from time to time and from place to place as to what kind of group has the most urgent obligation on its members. At present, among most Western peoples and in Japan, the nation is put in this supreme position. Among the Greeks and Romans it was the city. It may be that in future it will be a class rather than a nation or a city. But common-sense has always held that there is some group for which all its members, and only they, were bound in certain circumstances to sacrifice their happiness, their chances of culture and development, and their lives.

I said that common-sense accepts a kind of "Self-referential Altruism." My meaning should now be clear. Common-sense is altruistic in so far as it considers that each of us is often under an obligation to sacrifice his own happiness, and sometimes to sacrifice his own development and life, for the benefit of certain other persons and groups, even when it is doubtful whether more good will be produced by doing so than by not doing so. It tends to admire these acts, as such, even when it regrets the necessity for them, and even when it thinks that on the whole they had better not have been done. It has no such admiration for the attempt to make oneself happy, as such, even when it does no harm to others. And, although it admires acts directed towards self-development and self-culture, as such, especially when they are done in face of external obstacles and internal hindrances, its admiration for them is not usually very intense.

On the other hand, the altruism of common-sense is always limited in scope. It does not hold that any of us has an equally strong obligation to benefit everyone whom he can affect by his actions. According to it, each of us has specially strong obligations to benefit certain persons and groups of persons who stand in certain special relations to himself. And these special relations to himself are the ultimate and sufficient ground of these specially urgent obligations. Each person may be regarded as a centre of a number of concentric circles. The persons and the groups to whom he has the most urgent obligations may be regarded as forming the innermost circle. Then comes a circle of persons and groups to whom his obligations are moderately urgent. Finally there is the outermost circle of persons (and animals) to whom he has only the obligation of "common humanity." This is what I mean by saying that the altruism which common-sense accepts is "self-referential."

If this be a fair account of the beliefs of common-sense, what line could a person take who found neutralism self-evident? And, again, what line could a person take who found it self-evident that the right act must coincide with the optimising act and was therefore committed at the second move to neutralism? The problem is the same for both. He would have to do three things.

(i) He would have to hold that common-sense is mistaken in thinking that these special obligations are founded directly on these special relations, (ii) He would have to show that all these special obligations, so far as they are valid at all, are derivable from the one fundamental obligation to maximise the balance of good over evil among contemporary and subsequent conscious beings as a whole. He will try to do this by pointing out that each of us is limited in his power of helping or harming others, in the range of his natural sympathies and affections, and in his knowledge of the needs of others. He will argue that, in consequence of this, the maximum balance of good over evil among conscious beings as a whole is most likely to be secured if people do not aim directly at it. It is most likely to be secured if each aims primarily at the maximum balance of good over evil in the members of a limited group consisting of himself and those who stand in more or less intimate relations to him. The best that the neutralist could hope to achieve on these lines would be to reach a system of derived obligations which agreed roughly, both in scope and in relative urgency, with that system of obligations which commonsense mistakenly thinks to be founded directly upon various special relationships. In so far as this result was achieved he might claim to accept in outline the same set of obligations as common-sense does; to correct common-sense morality in matters of detail; and to substitute a single coherent system of obligations, deduced from a single self-evident ethical principle and a number of admitted psychological facts, for a mere heap of unrelated obligations, (iii) To complete his case he would have to try to explain, by reference to admitted psychological facts and plausible historical hypotheses, how commonsense came to make the fundamental mistakes which, according to him, it does make. For common-sense rejects the neutralistic principle which he finds self-evident, and it regards as ultimate those special obligations of an agent towards certain persons and groups which he regards as derivative.

How, if at all, could this last desideratum be fulfilled? It seems to me that it might be attempted along the following lines. It must be admitted that any society in which each member was prepared to make sacrifices for the benefit of the group as a collective whole would be more likely to flourish and persist than a society whose members were not prepared to make such sacrifices. It must also be admitted that egoistic and anti-social motives are extremely strong in everyone. Suppose, then, that there were a society in which, by any means, there had arisen a strong additional motive (however mistaken and superstitious) in support of self-sacrifice, and that this motive were conveyed from generation to generation by example and precept and were supported by the sanctions of social praise and blame. Such a society would be likely to flourish and to overcome other societies in which no such additional motive existed. Its ways of thinking on these subjects and its sentiments of approval and disapproval would tend to spread, both directly through conquest and indirectly through the prestige which its success would give to it.

On the other hand, a society in which each member was prepared to sacrifice himself just as much for other societies and for their members as for his own society and its members would be most unlikely to flourish and persist. So, if there were a society in which, by any means, a strong additional motive for unlimited altruism or for neutralism had arisen and been propagated, that society would be likely to go under in conflict with one in which a more restricted self-referential altruism was practised. It therefore seems likely that the societies which would still persist and flourish after a long period of conflict would be those in which there had somehow arisen, in the remote past, a superstitious approval of altruism within certain limits and a superstitious disapproval of extending it beyond those limits. These are exactly the kind of societies which we do find. If one were asked what is the most formidable society which now exists and the one which seems to have the best chance of spreading its ideals throughout the world directly by conquest and indirectly by the prestige of its success, the answer would be Germany. And, if one were asked what is the nation whose citizens combine the most intense spirit of self-sacrifice within the group with the most rigid refusal to extend that spirit beyond the group, the answer would be the same. Non est potestas super terrain quae comparetur ei.

It seems then that, even if neutralism be true and be self-evident to the philosopher in his study, there are powerful causes which would make it likely that certain forms of self-referential altruism would appear to be true and self-evident to most unreflective persons at all times and even to reflective persons at most times. Therefore the fact that common-sense rejects neutralism and accepts certain forms of self-referential altruism as self-evident is not a conclusive objection to the truth, or even to the necessary truth, of neutralism.

(2) "NATURAL" AND "NON-NATURAL" CHARACTERISTICS¹

It is a fundamental doctrine of Moore's ethical theory that the word "good," in its most fundamental sense, is a name for a characteristic which is simple and "non-natural." He compares it in the first respect, and contrasts it in the second, with the word "yellow."

A complete discussion of this doctrine would have to begin by raising a question which Moore never did raise but which has become acute in recent years. Is "good" a name of a characteristic at all? Or do sentences like "This is good," though grammatically similar to sentences like "This is yellow" which undoubtedly ascribe a certain characteristic to a subject, really need an entirely different kind of analysis? Is it not possible that the function of such sentences is to express or to stimulate certain kinds of emotion, or to command or forbid certain kinds of action, and not to state certain kinds of fact? There is a vast amount to be said for and against this suggestion, but I do not propose to discuss it here. I shall assume, for the sake of argument, that "good" is a name for a characteristic, and that sentences like "This is good" ascribe this characteristic to a subject.

On this assumption the next topic which would have to be discussed in any full treatment of the theory would be Moore's contention that the characteristic of which "good" is a name is simple and unanalysable. This, of course, implies that it is a pure quality and not a relational property. I do not consider that Moore has produced any conclusive reasons for this opinion, and I am very doubtful whether such an opinion could be established. But I propose to waive this point also, and to assume that "good" is a name for a simple characteristic.

If we make these assumptions two questions remain. (1) What exactly is meant by the distinction between a "natural" and a "non-natural" characteristic? (2) What connexion, if any, is there between the doctrine that "good," in its primary sense, denotes a characteristic which is simple and unanalysable, and the doctrine that it denotes a characteristic which is non-natural? We will take these two questions in turn.

(1) Let us begin with complex characteristics. A complex characteristic is natural if it can be analysed without remainder into a set of simple characteristics each of which is natural. A complex characteristic is non-natural if its analysis involves at least one simple characteristic which is non-natural. Suppose, e.g., that "This is good" where "good" is used in the primary sense, could be analysed into "This is something which it would be right to desire as an end." And suppose that "right," as applied to desires, were a name for a non-natural characteristic. Then goodness, in this sense, would be a complex non-natural characteristic. So we are eventually faced with the question: "What is meant by calling a simple characteristic natural or non-natural?."

Unfortunately we shall get very little light on this question from Moore's published works. The only place, so far as I know, in which it is explicitly discussed is Principia Ethica, 40 to 41. We are told there that a "natural object" is any object that is capable of existing in time, e.g., a stone, a mind, an explosion, an experience, etc. All natural objects have natural characteristics, and some of them have also non-natural characteristics. We are told that each natural characteristic of a natural object could be conceived as existing in time all by itself, and that every natural object is a whole whose parts are its natural characteristics. We are told that a non-natural characteristic of a natural object is one which cannot be conceived as existing in time all by itself. It can be conceived as existing only as the property of some natural object.

Now it seems to me that every characteristic of a natural object answers to Moore's criterion of non-naturalness, and that no characteristic could possibly be natural in his sense. I do not believe for a moment that a penny is a whole of which brownness and roundness are parts, nor do I believe that the brownness or roundness of a penny could exist in time all by itself. Hence, if I accepted Moore's account, I should have to reckon brownness, roundness, pleasantness, etc., as non-natural characteristics. Yet he certainly counts them as natural characteristics.

I think that Moore is intending to explain the distinction between natural and non-natural characteristics in the very difficult essay in his Philosophical Studies which is entitled "The Conception of Intrinsic Value." So far as I can understand his doctrine in that essay it may be summarised as follows.

(i) The characteristics of any thing T may be first divided into two great classes, viz., those which do, and those which do not, "depend solely on the intrinsic nature of T. (ii) Those characteristics of a thing T which depend solely on its intrinsic nature may then be subdivided into those which are, and those which are not, "intrinsic characteristics" of T. Consider, e.g., an experience which has a certain perfectly determinate kind and degree of pleasantness. Suppose that it also has a certain perfectly determinate kind and degree of goodness. Then, if I understand him aright, Moore would say that both its pleasantness and its goodness are characteristics which depend solely on its intrinsic nature. He would say that its pleasantness is an intrinsic characteristic of it. And he would say that its goodness is not an intrinsic characteristic of it. (iii) Although he does not explicitly say so, I think that he would identify the non-natural characteristics of a thing with those which are determined solely by its intrinsic nature and yet are not intrinsic. The natural characteristics of a thing would be those which are either (a) intrinsic, or (b) not determined solely by its intrinsic nature.

Unfortunately Moore gives no clear account of this distinction between the intrinsic and the non-intrinsic varieties of the characteristics which depend on the intrinsic nature of a thing. All that he says is this. A complete enumeration of the intrinsic characteristics of a thing would constitute a complete description of it. But a description of a thing can be complete even if it omits those characteristics of it which, though determined solely by its intrinsic nature, are not intrinsic. E.g., a pleasant experience, which is also good, could not be completely described unless its pleasantness was mentioned. But it could be completely described without its goodness being mentioned.

I find it most difficult to follow or to accept this. I am inclined to think that the fact which Moore has in mind here is that goodness, in the primary sense, is always dependent on the presence of certain non-ethical characteristics which I should call "good-making." If an experience is good (or if it is bad), this is never an ultimate fact. It is always reasonable to ask: "What makes it good?" or "What makes it bad?" as the case may be. And the sort of answer that we should expect to get is that it is made good by its pleasantness, or by the fact that it is a sorrowfully toned awareness of another's distress, or by some other such non-ethical characteristic of it. We might, therefore, distinguish the characteristics of a thing into the following two classes, viz., ultimate and derivative. Goodness would certainly fall into the class of derivative characteristics.

Now there is a sense in which one might say that a thing could not be completely described if any of its ultimate characteristics were omitted, but that it could be completely described without mentioning all its derivative characteristics. In describing a circle, e.g., it is not necessary to mention explicitly any of the innumerable properties of circles which follow of necessity from their definition together with the axioms of Euclidean geometry.

But, although this analogy may throw some light on what Moore had in mind, it certainly does not help us much towards understanding what he means by calling goodness a non-natural characteristic, and pleasantness, e.g., a natural characteristic. In the first place, the way in which the ethical properties of a thing depend on its non-ethical properties seems to be quite unlike the way in which the remaining properties of a circle depend on its defining properties. In the latter case the dependence is equivalent to the fact that the possession of the remaining properties can be inferred deductively from the axioms of Euclid and the presence of the defining properties. But the connexion between the non-ethical bad-making characteristic of being an emotion of delight at another's pain and the ethical characteristic of being morally evil is certainly not of this nature.

Moreover, it is surely quite as evident that pleasantness and unpleasantness are derivative characteristics of an experience as that goodness and badness are. If an experience is pleasant, it is always reasonable to ask: "What makes it pleasant?" And the answer will always be to mention some non-hedonic "pleasant-making" characteristic of the experience. E.g., if it is a sensation of taste, the answer might be that it is made pleasant by its sweetness; if it is an auditory experience, the answer might be that it is made pleasant by the way in which various simultaneous and closely successive sounds are combined in it; and so on. Now Moore counts pleasantness as a natural characteristic. If he is right in doing so, it is impossible to identify the non-natural characteristics of a thing with the derivative sub-class of those of its characteristics which depend solely on its intrinsic nature. For by that criterion pleasantness would be a non-natural characteristic just as much as goodness.

It seems impossible then to extract from Moore's writings any satisfactory account of the distinction between "natural" and "non-natural" characteristics. Yet one seems to recognize fairly well what is the extension of these two terms, even if the attempts to define them are not successful. I propose now to try a different line of approach, and to suggest an epistemological description, as distinct from a logical definition, of the term "natural characteristic." That is to say, I shall try to delimit the class of natural characteristics, not by stating their intrinsic peculiarities, but by stating how we come to form our ideas of them.

I propose to describe a "natural characteristic" as any characteristic which either (a) we become aware of by inspecting our sense-data or introspecting our experiences, or (b) is definable wholly in terms of characteristics of the former kind together with the notions of cause and substance. I think that this covers every characteristic which would be universally admitted to be natural. It would, e.g., cover yellowness, both in the primary nondispositional sense in which we use the word when we say, e.g., "That looks yellow to me from here now," and in the secondary dispositional sense in which we use it when we say, e.g., "Gold is yellow." It would also cover psychological characteristics, whether non-dispositional or dispositional. We know, e.g., what is meant by the fear-quality, the anger-quality, etc., through having felt afraid, angry, etc., and having introspected such experiences. And we know, e.g., what is meant by timidity, irascibility, etc., because these dispositional properties are definable in terms of the fear-quality, the anger-quality, etc., together with the notions of cause and substance.

A "non-natural" characteristic would be described epistemologically in negative terms derived from the abovementioned epistemological description of "natural" characteristics. A characteristic would be non-natural if (a) no-one could become aware of it by inspecting his sense-data or introspecting his experiences, and (b) it is not definable in terms of characteristics of which one could become aware in those ways together with the notions of cause and substance. These epistemological descriptions of the two kinds of characteristic leave open the question whether goodness is of the natural or the non-natural kind and provide us with a criterion for answering it.

(2) We are now in a position to deal with our second question. What connexion, if any, is there between the doctrine that "good," in its primary sense, denotes a characteristic which is simple, and the doctrine that it denotes one which is non-naturall

It is plain that our epistemological criteria at once plunge us into certain questions which are important in themselves and are never, so far as I am aware, raised in Moore's writings. If "goodness" is a name for a characteristic, how do we become aware of the characteristic for which it is a name? There is lamentably little discussion on this point in works on ethics.

It seems to me evident that goodness is not a characteristic which we can become aware of by inspecting any of our sense-data; so that in this respect it is utterly unlike yellowness, sweetness, squeakiness, etc., when used in the non-dispositional sense. It is plain that, when "good" is used in its primary sense, it does not denote a characteristic whose presence is revealed to us by sight or touch or taste or hearing or smell, or any other sense that we do have or conceivably might have. It is doubtful whether goodness, in this sense, could belong to sense-data or to physical objects. And, even if it can, it is certain that we do not sense or perceive with our senses the goodness of such objects. At most we perceive with our senses certain natural characteristics which are good-making, e.g., certain combinations of colour, of sound, etc., which make the object which possesses them intrinsically good.

It seems equally clear that no simple psychological characteristic, such as we could discover by introspecting our experiences, can be identified with goodness. By introspection we become aware of experiences which are pleasant or unpleasant, toned with desire or with aversion, fearful, hopeful, and so on. Now it is true that goodness, in the primary sense, can belong to experiences. Indeed, some people would hold that, in this sense, it can belong to nothing else. Yet I think that a moment's reflexion will show that by calling an experience "good" we do not mean that it is pleasant, or that it is an experience of desire or of fear or of hope, or that it has any of the simple psychological qualities which we become aware of through introspecting our experiences.

If anyone were tempted to identify goodness with one of these simple psychological characteristics, I think that he would be doing so through the following confusion. What he really believes is that there is one and only one good-making quality of experiences, e.g., pleasantness. He then fails to notice, the distinction between goodness itself and the one and only good-making quality which he recognizes. And so he thinks he believes that "good" and "pleasant," e.g., are just two names for a single characteristic. Since pleasantness certainly is a natural characteristic, he will thus think he believes that "good" is a name for a natural characteristic. But I do not think that the belief that "good" and "pleasant" (e.g.) are two names for one characteristic would survive after the distinction between goodness itself and a good-making characteristic had been pointed out. And similar remarks would apply to any other simple psychological quality which one might be tempted to identify with the characteristic denoted by "good."

So we come to the following conditional conclusion. If the word "good," when used in its primary sense, denotes a simple quality, then that quality is almost certainly not one which a person could become aware of either by inspecting his sense-data or introspecting his experiences. And, if the characteristic which it denotes be simple, it will not be definable at all, and therefore will not be definable in terms of characteristics which a person might have become aware of in one or other of those ways. Therefore it will be a non-natural characteristic, according to our criterion. So, with our criteria, there is an important connexion between the doctrine that "good" is the name of a simple quality and the doctrine that it is the name of a non-natural characteristic.

This, however, does not settle the question whether "good" is in fact the name of a non-natural characteristic. For it is by no means certain that it is the name of a characteristic at all; and, even if it is so, it is by no means certain that this characteristic is simple. Suppose "good," in the primary sense, were a name for a characteristic which is complex. Would there then be any reason to think it non-natural? I believe that there would. I know of no proposed definition of goodness in purely natural terms which is in the least plausible. But there are definitions of goodness, containing terms which appear to be non-natural, which are not without plausibility. E.g., it would not be unplausible to suggest that "x is intrinsically good" means that x is something which it would be right or fitting to desire as an end. And it would appear that "right" or "fitting," when used in this sense, answer to our criterion of being non-natural.

The legitimate conclusion then would seem to be that, if "good" in its primary sense be the name of a characteristic, that characteristic, whether simple or complex, is non-natural. Anyone who saw reason to doubt the existence of characteristics answering to our description of "non-natural" might fairly use this conclusion as the basis of an argument to show that "good" is not a name of a characteristic at all.

It will be worth while to develop this line of argument a little further. Is there any way of becoming aware of a simple quality belonging to particulars except by inspecting sense-data or introspecting experiences which have that quality? Many people would say that there plainly is no other way. If they are right, it follows that we could not possibly have an idea of goodness if goodness were a non-natural characteristic. For, if "good" denotes a characteristic at all, it certainly denotes one that belongs to particulars and only to them. If goodness were a simple characteristic and were non-natural, the conclusion that we could not have an idea of it would follow at once from the epistemological principle which these people find self-evident. If goodness were a complex characteristic and were non-natural, the same conclusion would follow, at the second move, from the same principle. We could not have an idea of such a complex characteristic unless we had ideas of all its simple components. By hypothesis one at least of these would be non-natural; and, if the epistemological principle be accepted, we could have no ideas of any such components.

Now, although this epistemological principle does seem to me highly plausible, I am not prepared to accept it (or any other) as self-evident. Therefore I am not prepared to assert dogmatically that no characteristic of which anyone could have an idea could be non-natural. But I do think it important to point out the following conditional conclusion, If goodness be a non-natural characteristic, then anyone's idea of it must be an a priori notion or contain a priori notions as elements. For an a priori notion just is an idea of a characteristic which is not manifested to us in sensation or introspection and is not definable wholly in terms of characteristics which are so manifested. Anyone who holds that goodness is a non-natural characteristic and that he has an idea of it is therefore committed to the belief that there are a priori notions and that his idea of goodness is one of them. Now as we have seen, if "good" denotes a characteristic at all, the characteristic which it denotes is almost certainly non-natural. Therefore anyone who holds that "good" denotes a characteristic, and that he has an idea of the characteristic which it denotes, will be almost compelled to hold that there are a priori notions, and that his idea of goodness is or contains one of them.

There is one other epistemological point to be noticed in conclusion. Suppose a person regards goodness as a non-natural characteristic, and admits that its presence is always dependent on the presence of certain natural characteristics which are good-making. Suppose, further, that he holds that the connexion between a goodmaking characteristic and the goodness which it confers is necessary. Then he will be obliged to hold that there are synthetically necessary facts, and that he knows some of them. He will therefore be obliged to admit that he can make synthetically a priori judgments. The necessary connexion between those natural characteristics of a thing which are good-making and the goodness which they confer on it could not possibly be analytic. For this would involve the absurdity that the non-natural characteristic of goodness is contained as a factor in the analysis of the purely natural good-making characteristics.

Now it is fashionable at present to hold that all necessary connexion must be analytic and that there can be no synthetic a priori judgments. I do not find this principle in the least self-evident myself, though I think that it has enough plausibility and interest to justify a strenuous attempt to see whether it can be successfully applied in detail. Anyone who does accept it, and also holds that "good" is a name for a characteristic, will be compelled to draw one or other of the following conclusions. Either (a) goodness is a natural characteristic, or (b) the connexion between the good-making characteristics of a thing and the goodness which they confer on it is purely contingent and can be known only empirically. He might, of course, consistently combine both these conclusions, as Hume did. For Hume's doctrine, stated very roughly, was (a) that to be good means to be an object of emotions of approval in all or most men, and (b) that it is a contingent and empirically known fact that such emotions are called forth by what is believed to be pleasant or useful and only by objects which have that property.

NOTES

¹ This (second) part of this paper is extracted, with some additions and modifications, from a paper by the author in the Aristotelian Society's Proceedings for 1934, entitled "Is 'Goodness' a Name of a Simple Non-natural Quality?" C.D.B.