1. OUR discussion in Chapter III brought us into touch with the notion of the good of a being. This in its turn was found to be related to the notions of pain and pleasure i.e. to a further form of goodness which is here called the hedonic good.

Our term ‘hedonic goodness’ is supposed to cover roughly the same ground as the word ‘pleasure’ in ordinary language. But as we shall soon see this ground is very heterogeneous and the use of one word to cover it may produce an appearance of conceptual homogeneity by which we must not let ourselves become deluded.

To realize the heterogeneity of the conceptual field in which we are moving in this chapter some observations on language may be helpful. In English one is used to speaking of pleasure and pain as a pair of contraries or opposites. In other languages this contrast is not so clearly marked. In German for example the nearest parallel to the pair ‘pleasure-pain’ in ordinary parlance is ‘Lust-Unlust’. But the German word for ‘pain’ is not ‘Unlust’. It is ‘Schmerz’. The German pair of substantives ‘Lust-Unlust’ answers in meaning more closely to the English pair of adjectives ‘pleasant-unpleasant’ than to the substantive-pair ‘pleasure-pain’. But this correspondence too is not perfect. The words ‘pleasant’ and ‘unpleasant’ in English would most naturally be translated by ‘angenehm’ and ‘unangenehm’ in German.

Considering the important rôle which the concept of pleasure has played in ethics all through the history of the subject it is surprising how little this concept has been made the object of special investigation. Neither Hume nor the British utilitarians nor Moore and the critics of ethical naturalism in this century seem to have been aware of the problematic character of this key-notion of their own writings.1 Most writers in the past regard pleasure as either some kind of sensation or as something between sensation and emotion. Moore Broad and the non-naturalists in general take it for granted that pleasantness is a ‘naturalistic’ attribute of things and states and not an axiological term.2 This I think is a bad mistake. Some of the orthodox views of pleasure were challenged by Professor Gilbert Ryle in an important essay a few years ago.3 Since then there is noticeable a new interest in the concept for its own sake—and not merely as an item in the ethicists’ discussions of moral value.

Our discussion here of the concept of pleasure can claim neither to be deep-searching nor even very systematic. My own feeling is that I am only scratching a surface under which important problems lie hidden.

I think it is useful at least for purposes of a first approximation to distinguish three main forms—as I shall call them—of pleasure.4

The first I call passive pleasure. It is the pleasure or better: the pleasantness which we attribute primarily to sensations and other so-called states of consciousness and secondarily also to their causes in the physical world. Pleasantness as an attribute of sensations can also be spoken of as ‘the pleasures of the senses’ or as ‘sensuous pleasure’. It seems to me that this sub-form of passive pleasure is largely regarded as the prototype of all pleasure whatsoever and that this one-sided view has been much to the detriment of the philosophic discussion of these topics.

The second form of pleasure I shall call by contrast active pleasure. It is the pleasure which a man derives from doing things which he is keen on doing enjoys doing or likes to do. Active pleasure can also be called ‘the pleasures of an active life’. To the discussion of the ethical relevance of pleasure the pleasures of the active life seem to me to be at least as important as the pleasures of the senses. Yet there are few moralists apart from Aristotle who have paid much attention to active pleasure.

In addition to passive and active pleasure there is that which I shall call the pleasure of satisfaction or contentedness. It is the pleasure which we feel at getting that which we desire or need or want—irrespective of whether the desired thing by itself gives us pleasure. The pleasure of satisfaction has played implicitly if not explicitly a great rôle in the formation and discussion of the doctrine known as psychological hedonism.

2. As specimens of the use of ‘good’ to refer to the sub-form of passive pleasure which we call sensual one may offer the phrases ‘a good wine’ or ‘a good apple’. Let us here consider the case of the good apple in some detail.

It should first be noted that there are many points of view from which the goodness of apples may become assessed. Apples are food. When we say that it is good to eat apples or that apples are good for the children we are probably thinking of the nourishing value and wholesomeness of apples. This goodness of the fruit is of the form we have called the beneficial. When the beneficial nature of apples is concerned the attribution of goodness is usually not of an individual apple but of apples as such or of some kind of apple.

When the cultivator or producer of apples judges of the goodness or badness of a kind of apple he may be thinking of such questions as whether this kind is easy to cultivate or—in a cold climate—is a hardy sort of apple. From the consumer's point of view some apples are particularly good for storing others for making jam others again for eating. When judged from the producer's and the consumer's specific points of view the goodness of apples is often instrumental or utilitarian goodness for some purpose. When these forms of goodness in apples are concerned the judgment is usually about a kind of apple and not about individual apples.

Calling an individual apple good is often another way of saying that it is not damaged or decayed or diseased. The apple is then being treated as quasi a being of whose good it makes sense to talk. An apple can be ‘healthy’ as distinct from ‘wholesome’. We need not here stop to consider whether and when such talk is ‘reducible’ to talk of instrumental and utilitarian goodness.

But calling an individual apple good can also be but another way of saying that we like its taste that it is good-tasting. Then the goodness of the apple is hedonic. When it is hedonic it is moreover of the form we called passive and the sub-form we called sensuous.

The hedonic judgment need not be about an individual apple. A person who says that apples are good may mean that he likes the taste of apples and from this it would not follow that he will like the taste of all individual apples. A person again who says of an individual apple which he is not actually tasting ‘this is a good apple’ would almost certainly be pronouncing on a kind of apple of which this individual is a specimen. He finds the taste of apples of this sort good but he perhaps dislikes the taste of some other kind of apple. I shall here disregard such general or generalized hedonic judgments.

Consider the particular judgment expressed in the words ‘this apple is good’ or ‘this is a good apple’ when ‘good’ is meant hedonically. The question may be raised: Of what is goodness here really predicated? On the face of it goodness is predicated of the apple. It could however be suggested that the verbal formulation conceals a primary judgment the overt formulation of which would run ‘the taste of this apple is good’ or alternatively ‘this apple is good-tasting’ or ‘this apple has a good taste’. According to this suggestion hedonic goodness belongs primarily to the taste of the apple and secondarily only to the apple itself. The taste is a sensation or bundle of sensations; the apple is a physical object. It is a causal or dispositional characteristic of the physical object that it evokes or produces under specific circumstances taste-sensations in a sensing subject. These sensations are the primary logical subject of the hedonic value-judgment. The physical thing ‘partakes’ so to speak in the goodness of the sensations only by being their cause.

Against this idea of a primary and a secondary attribution of goodness the following objection may be raised: To call a good-tasting apple good is both common and natural. It could hardly be maintained that it were uncommon or unnatural to call its taste good too this simply is one of the uses of ‘good’. But instead of calling the good-tasting apple's taste good we could also call it agreeable or pleasant whereas the apple itself would not commonly and naturally be called by those attributes. And now someone might wish to make a subtle distinction and say that ‘good’ is primarily an attribute of the thing and secondarily of the sensations it produces ‘pleasant’ again primarily of the sensations and secondarily if at all of the physical object.

There are other senses of ‘good’ which apply only to the object and not to the sensations it produces. For example: The apple can be good for storing or good for providing us with Vitamin C but its taste cannot possess such goodness. There are thus it would seem a great many more senses in which the apple can be good or bad than in which its taste can be good or bad. But to argue from this that there is no sense of ‘good’ which applies genuinely or primarily to the taste of an apple (or to a sensation in general) seems to me to be quite wrong. The lesson taught by the use of the word ‘pleasant’ in the context is not that sensations could not be good in a primary sense but that the word ‘pleasant’ is a synonym for the word ‘good’ in one of the latter's primary uses.

I shall accept the view that hedonic goodness of the sub-form which I have called the pleasure of the senses is primarily an attribute of sensations and secondarily of the objects which produce those sensations. The sentence ‘this taste is good’ I shall say expresses a primary hedonic value-judgment and the sentence ‘this apple is good’ when ‘good’ is used in the hedonic sense a secondary hedonic value-judgment.

3. In the sensation which a thing as we say ‘produces’ when it affects our senses we sense one or several qualities of the thing. In tasting sugar for example we sense its taste-quality which is sweetness. In tasting an apple we sense many qualities: a certain juiciness sourness maybe sweetness too. We could say that the taste—quality of the apple is a bundle or mixture of several qualities. (Not all of these ingredient qualities by the way are what we would ordinarily call taste-qualities. Some are olfactory qualities. Is juiciness a taste-quality?) In a similar manner we could say that the taste—sensation which we have of the apple is a bundle or mixture of several sensations. The several sensations themselves might be called ingredients or ingredient sensations of the total sensation.

Assume that a sensation which contains several ingredient sensations of different qualities is judged good or pleasant. Then it may happen that we can point to some of those ingredients and say that we judge the sensation pleasant because of the presence in it of those very ingredients. The ingredients thus pointed to we could call good-making ingredients (ingredient sensations); and the qualities of the thing thus sensed we could call good-making qualities or properties of the thing. For example: someone may wish to maintain that what makes him like the taste of a certain apple is the presence in it of a certain juiciness and sourness. Juiciness and sourness would then be good-making qualities of the apple and the sensations of juiciness and sourness good-making ingredients of the total taste-sensation which the person has of the apple.

In a sensation judged pleasant several ingredient sensations may thus become distinguished i.e. several sense-qualities sensed. Now it may perhaps be thought that pleasantness itself is one such quality just as for example sweetness is.5

The idea that pleasure or pleasantness were a sense-quality i.e. something which we sense is I think a bad confusion. I shall briefly indicate why I think so.

In the sensations we sense qualities; sensations are of certain qualities we also say. Thus for example we may have a sensation of redness or sweetness. One sometimes calls a sensation of redness a red sensation. One can do so for the sake of verbal convenience. But it is highly misleading. For it suggests a view of sensations as a kind of thing of the sensible qualities of which: colour shape smell etc. we can talk. (The talk of the fake-entities called sense-data has I am afraid much encouraged this view among philosophers of an earlier generation.) But a sensation of red is not the sort of entity which smells or has a colour or can be tasted. One can sense a colour or a smell or a taste but one cannot sense a sensation. What sort of properties then do sensations have? A sensation can for example be intense or vivid or dim or vague it lasts for some time and then passes away and it can be pleasant or unpleasant. A pleasant sensation is I believe sometimes called a sensation of pleasure. Perhaps there is some convenience in this mode of speech. But it should be remembered that ‘a sensation of pleasure’ is misleading in much the same way as ‘a red sensation’ is. That is: just as ‘a red sensation’ may be regarded as a logically distorted form of ‘a sensation of redness’ similarly ‘a sensation of pleasure’ may be regarded as a logical distortion of ‘a pleasant sensation’.

But is not a sensation of sweetness a sweet sensation? I would answer: Properly speaking only when ‘sweet’ is used in an analogical sense to mean ‘pleasant’ or something near it. ‘Sweet’ in English has clear analogical uses as a value attribute. ‘How sweet of you’ ‘How nice of you’ ‘How good of you’ say roughly the same. It is the existence of such analogical uses of ‘sweet’ I would suggest which makes it appear more natural to speak of sweet sensations than of red or round sensations. Of the other adjectives for taste-qualities ‘bitter’ and ‘sour’ are also used analogically as value attributes. But ‘salt’ is not quite in the same way and to the same extent used analogically. This perhaps explains why—at least in my ears—it sounds less natural to call a sensation of something salt a salt sensation than to call a sensation of something sweet or bitter or sour a sweet or bitter or sour sensation.

4. As naming the opposite to the thing named by the substantive ‘pleasure’ language—i.e. the English language (cf. sect. 1) suggests the substantive ‘pain’. ‘Unpleasure’ is not a word in common use. ‘Displeasure’ again has to do with anger and trouble. It does not name an opposite to that which we have here called passive pleasure and of which the sensuous pleasures are a sub-form. If it names an opposite of pleasure at all then it would be of that form which we have called the pleasure of satisfaction or of some sub-form of it. The displeased man feels dissatisfaction at or disapproves of something.

The adjective ‘pleasant’ may be said to have two linguistic opposites ‘unpleasant’ and ‘painful’.

The same arguments which may be advanced for showing that pleasure and pleasantness are not sense-qualities also apply to the unpleasant. But which is the status of the painful in this regard?

We speak of painful sensations. But the phrase ‘a sensation of painfulness’ sounds unnatural. This would indicate that ‘painful’ too does not refer to a sense-quality.

On the other hand we speak of a sensation of pain. We have pain in a tooth or pain in the stomach. But we do not commonly say that we have pleasure in the mouth when eating an apple. Pain as has often been observed is much more sensation-like than pleasure. The word ‘pain’ has analogical uses which resemble the use of ‘pleasure’ in that they make the word a value-attribute. But it seems to me right to say that in its primary sense ‘a pain’ refers to a kind of sensation and that ‘pain’ names a sense-quality of which however there are many shades. In this respect ‘pain’ is on a different logical level both as compared with the substantive ‘pleasure’ and as compared with the adjectives ‘pleasant’ ‘unpleasant’ and ‘painful’.

Pains which are sensations are ‘bodily pains’. So-called ‘mental pains’ are not sensations. They are therefore ‘pains’ by analogy only.

That pleasure and pain are not contradictories is trivial. Not trivial however is that the two because of their logical ‘asymmetry’ are not even contraries in any of the senses of ‘contraries’ which logicians distinguish. ‘Pleasant’ and ‘unpleasant’ denote contraries likewise ‘pleasant’ and ‘painful’ and when used in the hedonic sense ‘good’ and ‘bad’. If furthermore we regard the painful as a sub-form of the unpleasant we could perhaps say that ‘painful’ names a stronger contrary to the pleasant than ‘unpleasant’.

Neither ‘painful’ nor ‘pleasant’ are privative terms. Between the pleasant and the unpleasant there is a zone of genuinely value-indifferent states and not merely of states which are left unclassified because of vagueness. This is a feature in which the hedonically good and bad differ from the instrumentally good and poor and from the technically and medically good and bad but agrees with utilitarian goodness and badness.

Are pain-sensations ipso facto painful or at least unpleasant? I find the question difficult and puzzling. Since sensations are ‘naturalistic’ things and unpleasantness is a value-attribute an affirmative answer would provide us with an interesting example of an intrinsic connexion between a section of the world of facts and a section of the realm of values.

It seems to be a logical feature of the concept of pain that most pain is also painful or unpleasant. This feature is probably logically connected with the facts that bodily injury usually is painful and that severe pain in itself often is injurious in the sense that it has adverse effects on the possibilities of the individual concerned to satisfy his needs and wants.

Yet though necessarily most pain is painful (unpleasant) all pain it seems is not. ‘A pleasant pain’ is not a contradiction in terms we have said before. (See Ch. III sect. 9.) Some pain-sensations moreover we actually like judge pleasant. An example would be when a father or mother pinches their child in a playful attitude of love or tenderness.

There can also be sensations which are both pleasant and painful to the same subject on the same occasion. But they must as far as I can see be mixed sensations—like a bitter sweetness. Thus ‘painful pleasures’ are sensations or other experiences with pleasant and painful ingredients. A ‘pleasant pain’ is different. It is not a sensation which is both pleasant and painful. It is not a painful sensation at all. It is a pain-sensation which we happen to like judge pleasant and not painful.

Of the painful pleasures it seems always true to say that their painful parts are something which we would as such rather be without than suffer. We endure the pain because the pleasure outbalances it is greater than it. It so to speak pays to suffer the pain for the sake of the pleasure. But when the child welcomes a pinch with a laugh this is not because the pain though in itself disagreeable were outbalanced by a greater pleasure. A pleasant pain is not a price we pay for some greater pleasure but is itself pleasant.

5. We distinguished (in sect. 2) between primary hedonic judgments the logical subjects of which are sensations or other states of consciousness and secondary hedonic judgments the logical subjects of which are events or things in the physical world. The secondary judgments are capable of analysis in terms of primary judgments and causal statements. The pattern of analysis is as follows:

The secondary judgment ‘this X is good’ has roughly speaking the same meaning as ‘this X produces or has a disposition or tendency to produce pleasant (agreeable) sensations of such and such a kind’. An instantiation of the pattern would be: ‘This apple is good’ means ‘this apple produces pleasant gustatory sensations’. In ordinary life we should not express ourselves thus but instead of the last sentence say ‘this apple has a good (pleasant) taste’. Let us however not now mind the suggested piece of a philosopher's jargon. Let us also for the moment forget about the fact that the analysans does not specify about whose sensations we are talking and thus is in an important respect an incomplete statement.

This analysis of secondary hedonic judgments shows a conspicuous resemblance to a well-known attempt at analysing moral judgments. I am thinking of the theory or a variant of it commonly known as the emotive theory of ethics.6 According to this theory broadly speaking the sentence ‘this X is morally good’ means the same as the sentence ‘this X produces or has a disposition or tendency to produce a feeling of moral approval’. The ‘X’ usually stands for an arbitrary human act. Some authors speak simply of ‘a feeling of approval’ omitting the adjective ‘moral’ from the analysans. This makes their theory simpler though hardly more plausible.

Our analysis of secondary hedonic judgments may be regarded as a simplified model of the corresponding ethical theory. Some of the logical features of the ethical theory which have caused dispute can I think be conveniently studied in the simplified model. One such disputed point is whether moral judgments on the emotivist analysis are true or false or whether they merely are verbalized expressions of emotion and therefore lack truth-value. The form of emotive theory which holds the first opinion could be called naturalistic subjectivism. The form which holds the second we shall call non-cognitivist subjectivism.

Are hedonic judgments true or false? The secondary judgments involve a causal component. It traces certain sensations back to a physical thing as their cause. This causal component we shall here ignore. If we ignore it our question reduces to the question whether primary hedonic judgments are true or false. To get a firmer grasp of this second question it will be necessary to distinguish two types of primary hedonic judgments. I shall call them in a technical sense first person judgments and third person judgments.

In a first person hedonic judgment the subject is judging of a sensation which he is himself now experiencing or having that it is agreeable or pleasant that he likes experiencing or having it. In a third person hedonic judgment the subject is judging of the past present or future sensations of another subject that this other subject found or finds or will find them pleasant. Also the case when a subject judges of the hedonic quality of his own past or future sensations will here count as third person judgments. The subject is then as it were speaking of himself from outside in the perspective of time.

Third person hedonic judgments obviously are true or false. That Mr So-and-so likes or does not like or dislikes the taste of an apple which he is now eating is true or false. So are the statements that most people like the taste of this or that sort of apple or that they would if they tasted it like this particular apple. The difficulties of coming to know the truth-value may be considerable: when the apple is eaten how can we know whether somebody who never tasted it would have liked it? Perhaps the answer is that we cannot know this. But from this does not follow that the statement is not true or false.

When ascertaining the truth of a third person hedonic judgment we largely rely on first person hedonic judgments. Does N. N. like the apple which he is now eating? We ask him and he replies ‘Yes’ or ‘I like it’. His words express a first person hedonic judgment. It is used for assessing the truth of the third person hedonic judgment that N. N. likes the taste of the apple he is eating or which means the same that he finds the taste of the apple he is eating good. We may regard the evidence provided by the first person judgment as being so strong that all doubts about the truth-value of the third person judgment are expelled. But sometimes we do not attach much weight to the evidence e.g. because N. N. is a very polite man and is therefore likely to say of an apple which we have offered him that he likes it even when in fact he does not. If this were the case we should probably in forming our opinion as to whether he liked the fruit or not rely more on N. N.'s facial expression when eating the apple than on his words.

Do first person hedonic judgments have a truth-value? This is a very difficult question and part of a much larger question pertaining to the logical status of first person present-tense statements in general and first person statements about sensations in particular. One may argue the view—successfully I think—that in the first person hedonic judgments no statements are made at all and that the judgments therefore cannot properly be called true or false. When the words ‘the taste of this apple is good’ are used as a first person judgment they express (‘give vent to’) my pleasure at the taste and do not state that I am pleased or describe myself as a being who approves of the taste.

The same distinction between first person and third person judgments can of course also be made for judgments about the occurrence of those more subtle phenomena called ‘feelings of moral approval or indignation’. In the case of such feelings too it is fairly obvious that the third person judgments are objectively true or false whereas it is at least arguable that the first person judgments are not true or false statements about feelings but neither true nor false expressions of feeling.

One could try to do distributive justice to the claims both of naturalistic and non-cognitivist subjectivism as theories of hedonic value-judgments by apportioning to each theory a due share in the truth—to the first because of the propositional character of the third person judgments and to the second because of the interjectional character of the first person judgments. But by practising such impartiality one runs risk of obscuring an important point. This point is that the third person judgments just because of the feature of theirs which makes them true or false viz. that they are about the valuations of other subjects (or about the judging subject's own valuations viewed in the perspective of time) are no genuine value-judgments at all. They are no value-judgments since they do not value but report or conjecture about human reactions i.e. such reactions which we call valuations. The only genuine value-judgments in the context are the first person judgments. In them the judging subject values his sensations. They are not true or false and therefore in a sense of the word no ‘judgments’ even. For this reason it seems to me fair to say that non-cognitivist subjectivism represents the correct view of hedonic value-judgments whereas naturalistic subjectivism is not a theory of value-judgments at all.

In their relation to truth (primary) hedonic value-judgments unless I am badly mistaken differ importantly from judgments of instrumental technical or utilitarian goodness. These latter judgments are true or false; one can always be mistaken in them. The primary hedonic value-judgments are neither true nor false there is no room for mistake in them. They are in this peculiar sense ‘subjective’. In their sphere one cannot distinguish between an apparent and a real good; ‘to be good’ and ‘to be judged (or considered or thought) good’ are here one and the same. But judgments in which we affirm or anticipate the occurrence of such and such hedonic valuations in ourselves or in other subjects are of course true or false judgments—though not value-judgments.

6. Hedonic goodness of the form which we called passive pleasure is a value-attribute of sensations and other states of consciousness. To call this form of goodness an attribute or characteristic or property of sensations is useful when we want to explain why pleasure itself is not a sensation or a sense-quality (of a thing which is sensed in a sensation). But to call hedonic goodness a property can also be misleading.

Sensations are tied to a subject. They are somebody's sensations. If the secondary hedonic judgment expressed in the words ‘this apple is good’ is analysed in terms of primary judgments about the goodness or pleasantness of some gustatory sensations the question will instantly arise: Of whose sensations are we here talking? The overt form of the sentence which expresses the secondary judgment does not give us any guidance since it does not mention an apple-taster. Nor does the overt form of sentences which express primary hedonic judgments always mention a subject. ‘The taste of this apple is good’ and ‘this is a pleasant taste’ are complete sentences. But I think it is correct to say that the sense of the sentences is incomplete unless it is understood from the context or otherwise whose taste-sensations are meant. If this is true of sentences expressing a primary hedonic judgment the same will a fortiori be true of any sentence expressing a secondary hedonic judgment into the analysis of which the primary judgment may enter.

How is mention of a sensing subject to be worked into the overt form of a sentence expressing a hedonic judgment about the taste of an apple? Let us assume for the sake of argument that the subject is the speaker himself. He might then instead of ‘the taste of this apple is good’ say ‘I find the taste of this apple good’. Instead of ‘find’ he may also say ‘think’ or ‘judge’ or ‘consider’. Perhaps he would use the form of words ‘I think that the taste of this apple is good’. But this sounds artificial at least as an expression of his judgment when he is actually tasting the apple. (It sounds more like a conjecture that he will like the taste.) It would also sound rather artificial to say ‘the taste of this apple is good to me’. But to say ‘I like the taste of this apple’—thus not using the word ‘good’ at all—would sound perfectly natural.

It will strike us that a majority of these sentences which mention a subject are not of the ordinary subject-predicate form in which a property is predicated of a thing. It may be questioned whether any of them really is of this form. It may also be questioned whether not the overt subject-predicate form of sentences expressing hedonic judgments with suppressed reference to a sensing subject is really spurious viz. as a reflexion of the judgment's logical form. Among the sentences which we gave as examples the one which comes nearest to being of the ordinary subject-predicate form is ‘the taste of this apple is good to me’. If ‘good to me’ could be said to name a property then this sentence could be safely said to be a subject-predicate sentence. But ‘good to me’ rather suggests a relation between the apple-taster and the apple-taste than a property of the taste-sensation. If it be asked what this relation is a plausible answer would be that it is the liking-relation or some relation closely akin to it. The sensing subject likes enjoys approves of the taste of the apple. This is what could be reasonably meant by the unnatural-sounding form of words ‘the taste of this apple is good to me’.

The sentence ‘I think that the taste of this apple is good’ contains a subject-predicate sentence as a part within a ‘that’—clause. It is not itself a subject-predicate sentence. As already noted it has a certain artificiality about itself. It suggests to us that a distinction could be made between the taste's being good—as it were ‘in itself’—and somebody's thinking or judging or opining of the taste that it is good. I can of course think or judge or opine that I shall like the taste of this apple and perhaps find on tasting it that I was mistaken. But as we had already occasion to stress when we were speaking of the subjectivity of hedonic value-judgments I cannot judge the taste of the apple which I am now tasting good and be mistaken. Therefore to think that the taste is good is not to think that something is thus and thus but might be otherwise. The taste which is the logical subject of the first person judgment cannot from the point of view of its hedonic value be anything else but what I think it is. The use of ‘think that’ is here quite unlike the normal use of this phrase.

It was this slightly unnatural use of ‘that’—clauses in value-sentences which suggested to Moore7 the following refutation of subjectivism in value-theory: If ‘this is good’ means the same as ‘I think that this is good’ then by raising the same question over again for the ‘this is good’ which occurs inside the ‘that’—clause we get as an answer that ‘this is good’ means the same as ‘I think that I think that this is good’ and so on ad infinitum. This is a clever point as it stands but not of much consequence for the task of refuting subjectivism. It only establishes that if the words ‘I think that this is good’ are offered as an equivalent of ‘this is good’ then the ‘I think that…’—form functions in a peculiar way which should make its very use here sound suspect to our logical ear. If the subjectivist says that ‘this is good’ means ‘I think this good’ or ‘I consider this good’ omitting the word ‘that’ he is already better off and Moore's objection cannot be raised against him.8

These considerations in my opinion tend to show that hedonic goodness is not a property and that the subject-predicate form of sentences which express hedonic judgments is a spurious reflexion of the logical form of such judgments which is the relational form. As the logically most satisfactory formulations of such judgments in language I should regard their formulation with the aid of the verb ‘to like’ as for example in the sentence ‘I like the taste of this apple’. But this needless to say does not make the subject-predicate form either useless or incorrect as a shorthand formulation in all those innumerous cases when mention of a sensing subject is omitted from the sentence. The sentences ‘this is a good apple’ or ‘this is a good taste’ are all right as they stand. But they are apt to mislead the philosopher by concealing a logical form.

7. To think of pleasure exclusively in terms of predications of pleasantness to various logical subjects is thus philosophically misleading. To think of pleasure in terms of liking is philosophically enlightening and helpful. Beside ‘like’ also ‘approve’ and ‘enjoy’ are relational verbs which may be used for expressing hedonic judgments. That of which we approve is normally said to please us but is not ordinarily said to give us pleasure. ‘Enjoy’ is more obviously hedonic than ‘approve’; any source of enjoyment can also be called a source of pleasure and vice versa.

Among our likings an important position are held by things we like to do. One man likes to watch cricket another to play chess a third likes to get up early in the morning.

The way in which liking to do is connected with pleasure is rather different in these three examples. The pleasure of watching a game is mainly I should think of the form which we have called passive pleasure. Watching a game means the acquisition of experiences which the man who likes watching the game finds pleasant. Is his pleasure that of the senses? There are reasons for saying that it is since our man is enjoying a sight. There are perhaps even stronger reasons for saying that it is not since enjoying this sight normally requires both knowledge of the rules of the game and some familiarity with the practice of playing it. The border between sensuous pleasure and other forms of passive pleasure is very elastic.

The pleasure of playing a game has many aspects. Sometimes one plays a game just for amusement as a pastime. Then the pleasure which one derives from playing the game is substantially passive pleasure i.e. the pleasantness of certain experiences. But often we like playing a game not ‘for the sake of amusement’ but because we are interested in the game keen on the art of playing it. Then the delight we take in the game is of the form which I called active pleasure. The same is the case with any pleasure derived from the practising of any activity of which it is true to say that we are keen on it or that we like it ‘for its own sake’. It will occur to us that there is a connexion between technical goodness and the form of hedonic goodness called active pleasure. This connexion may be intrinsic. The two forms of goodness are nevertheless distinct.

Consider next the man who likes to get up early in the morning. Must he find early rising pleasant? Some men may rise early for ‘hedonic’ reasons i.e. in order to enjoy the morning: the freshness of the morning air the beauty of the sunrise etc. But rather few I think of those who say they ‘like’ to get up early would give such reasons for their liking. Someone may like to get up early because he has so many things to do that if he stays in bed till late he will not have time to do them at all or his afternoon will be badly rushed or he will have to work at night. But he may be completely indifferent to the peculiar pleasures of the early morning hours. Should this man not rather then say that he wants to get up early than say that he likes to get up early? Or perhaps the suggestion will be that our man should say that he has to or must get up early considering that this is not anything which he likes or wants to do ‘for its own sake’ but something that is forced upon him by the ‘practical necessities’ of life (Cf. Ch. VIII.) The answer is that the uses of ‘like to do’ and ‘want to do’ and ‘have to do’ shade into one another and that we sometimes say that we like to do things the doing of which is a source neither of passive nor of active pleasure to us.

But is there not at least a remote connexion with pleasure also in this third case which we have been discussing? The man who rises early may want to do so in order to avoid having to rush his day's work which is an unpleasant thing. Or he may be anxious to finish his set work as early in the day as possible so that he can relax and do in the afternoon what he ‘really likes’ i.e. that which affords him (active or passive) pleasure. These possibilities are not unrealistic. Now it may be argued that unless our man has some such desire either to avoid something unpleasant or to secure for himself some pleasure then he could not say truly of himself even that he wants to get up early in the morning. To argue thus about the man would be to apply to his case a general philosophic thesis about the nature of man viz. that all action is in the last resort necessarily prompted by a desire to secure some pleasure or avoid something unpleasant. This is the thesis or a version of the thesis known as psychological hedonism. We must here try to form some opinion of its truth.

8. Not everything which a man can be said to do is voluntary action. For example: getting fat or sleeping. And not everything which a man does and which is voluntary action can he also on any ordinary understanding of the words be said to want to do. Most things which a man does ‘because this is the custom’ e.g. taking off his hat when greeting a lady or ‘because this is the rule or law’ e.g. driving to the left or halting in front of a major road ahead are not things he ‘wants to do’ whenever he does them. But some of the things which a man voluntarily does he also wants to do and some of the things he wants to do he also likes to do.

To maintain psychological hedonism for all voluntary action is hardly feasible. To maintain it for those acts and activities only which a man enjoys doing or likes to do may seem too narrow to be of much interest. The exact scope of the claim of psychological hedonism is seldom made clear. I shall here regard it as a thesis concerning at least all those things which a man can be properly said to want to do. What does this thesis say? That too is seldom made sufficiently clear. I shall here understand it as saying something which can for purposes of a first approximation be stated as follows: If in an individual case we raise the question why a man wants that which he wants to do the answer will if not immediately then after a chain of questions and answers have to mention something which this man likes finds pleasant or dislikes finds unpleasant or painful. This statement is admittedly vague but I hope the subsequent discussion of it will make its intended meaning clearer. (See also Ch. V sect. 2.)

A man says he wants to get up early to-morrow morning. Why? Because he wants to see the sunrise. Why? Because he enjoys the sight likes it. The answer to the further question why he likes sunrises is not that their sight gives him pleasure. For his liking of the sight of a sunrise and the sight's giving him pleasure are one and the same.

The question may of course be raised why the spectacle pleases him. It is not certain that it can be answered. It would be answered by pointing to some fact in the man's life-history which tells us what made him like (‘caused’ him to like) sunrises. This question ‘Why?’ does not ask for a reason or motive but for a (kind of) causal explanation. The question ‘Why?’ which is relevant to the thesis of psychological hedonism is a question concerning reasons for doing i.e. ends in acting.

Our early riser can rightly be called a pleasure-seeker. His action is motivated by a desire to secure for himself a certain pleasant experience. He wants to get up early not because early rising is as such or in itself pleasant to him but because the act is conducive to pleasure i.e. is a necessary condition for his attaining a pleasant experience in the end.

Consider now the following case: A man wants to get up early some morning to see the sunrise. He has never seen one before but he has heard many people praise the beauty of the sight. He does not however expect that he will particularly like it and he is not anxious to secure for himself a new pleasure. He simply is curious to know what the sight which so many people praise looks like. Inquisitiveness can be a thoroughly self-sufficient motive of action. The chain of questions and answers could run as follows: Why does he want to get up early? Because he wants to see the sunrise. Why does he want to see the sunrise? Because he is curious about the sight. Here is no mention of liking or pleasure. Perhaps we can also answer the question: What made him curious? and that the answer is: The fact that so many poets whom he reads have written enthusiastically about sunrises. This question is causal. The fact that its answer can be said to hint at the likings and pleasures of poets is here irrelevant. Our man could equally well be curious to see something which people notoriously dislike. His inquisitiveness not his desire for some pleasant experience prompts his action.

I think that the case of the curious man refutes hedonism but the difficulty is to see that it really does so decisively. For could one not argue as follows:

Our inquisitive man is anxious to satisfy his inquisitiveness. The satisfaction of this desire gives him pleasure. If any case of doing that which one wants to do can be correctly described as a case of satisfying some desire—and let us not here query the correctness of this view—and if to have one's desires satisfied is a pleasure is then not psychological hedonism after all right?

Before we answer this question we must say some words about the relation of pleasure to satisfaction of desire.

Desire is sometimes called a dissatisfaction or discontentedness with a prevailing state of things; it is an impulse or longing to change this state to another. The change at which a desire aims could therefore also be characterized as a transition from a present ‘unpleasure’ to a future pleasure. Here however great caution is needed. The phrase ‘transition from a present unpleasure to a future pleasure’ can mean many things. It can mean that the state in which we are now is judged unpleasant and the one at which we aim is thought of as pleasant. For example: we feel uncomfortable or bored and want to do something which will make us feel comfortable or which will amuse us cheer us up. Here unpleasantness and pleasantness are clearly hedonic features of the states. This is a common type of desire-situation. But it is not the only type. The case e.g. of the man who wants to satisfy his curiosity is different. To have an unsatisfied curiosity can be agitating exciting vexing. It can be unpleasant too. But it need not be so. Someone may even think that curiosity is more of a pleasant than an unpleasant feeling. Yet on the other hand: if attempts to satisfy our curiosity fail we shall probably feel annoyed or grieved or even outraged and these states we should ordinarily judge unpleasant. Something corresponding it seems holds good of desire in general: frustration is unpleasant sometimes to the degree of being painful.

I think it is correct to say that frustration of desire is intrinsically unpleasant. We should not say that a desire had become ‘frustrated’ unless one or several unsuccessful attempts to gratify the desire had not had some hedonically bad consequences such as anger annoyance grief impatience hurt vanity or the like.

Similarly it seems to me right to think that satisfaction of desire is intrinsically pleasant. But the connexion between satisfaction and pleasure is more complicated than the connexion between frustration and hedonic badness. This is due to the fact that the description of a desire involves mention of an object of desire. The object is that which we desire. The attainment of the object intrinsically ‘by definition’ satisfies the desire. Thus attainment of the object is in a sense intrinsically satisfying. But from this does not follow that the object itself were intrinsically pleasant or otherwise hedonically good. The hedonically good consequences which are intrinsically connected with satisfaction of desire consist in feelings of contentedness or joy or power or relief or something similar. If they are not consequent upon the attainment of the object we should doubt whether there was any desire at all or whether it had been correctly described as a desire for so-and-so.

It is on the evidence of such hedonically tinged consequences both of frustration and of satisfaction that we often judge the strength of the desire. ‘He cannot have wanted it very eagerly since he was not very glad when he got it’; ‘He must have desired it strongly since failure to get it depressed him so much’ we say.

For the pleasure which is intrinsically connected with satisfaction of desire we already coined (in section 1) the name pleasure of satisfaction. It is the existence of an intrinsic tie between satisfaction and pleasure I think which is above all responsible for the strong appearance of truth which the thesis of psychological hedonism undoubtedly possesses. The refutation of hedonism must therefore not consist in an attempt to deny the existence of this connexion.

Wherein then does the refutation consist? As far as I can see it consists solely in this:

The pleasure of satisfying a desire can never be an object of that same desire. For satisfaction presupposes a desire and a desire in its turn presupposes something which we desire an object. Therefore the object of desire must necessarily be different from the pleasure of satisfying that desire. But the pleasure drawn from the satisfaction of a desire can itself become the object of a new desire. This is perhaps not very common but it is not an entirely unrealistic possibility. Consider again the man who wants to see the sunrise because he is curious about the sight. It is conceivable that experience had taught him that satisfying a curiosity is something very pleasant. He is of a peculiar inquisitive disposition or temper. Each time he is curious about something he is extremely tense and when his curiosity has become satisfied he has an immensely exhilarating and joyful feeling of relief. This man welcomes every opportunity for satisfying a curiosity because of the pleasure this gives him. This pleasure is the passive pleasure of some peculiar experience following upon the satisfaction of a curiosity. There is nothing in the logic of things to prevent it from becoming the object of a new desire viz. the desire for the pleasure of having satisfied a curiosity. But whether the original pleasure of satisfaction will or will not itself thus turn into an object of desire is an entirely contingent matter. It is the contingent nature of this fact which in my opinion constitutes the refutation of psychological hedonism. The error of hedonism is that it mistakes the necessary connexion which holds between the satisfaction of desire and pleasure for a necessary connexion between desire and pleasure as its object. This mistake is easy to make but not quite easy to expose. This is why I have spent so much time here on the refutation of psychological hedonism although this doctrine is said to have been refuted over and over again in the past I am not certain that any ‘refutation’ does full justice to the complications of the theory and the problems raised by it still continue to vex me.

9. Something can be an object of desire although it is thought unpleasant. The question may be raised whether anything can be an object of desire because it is considered unpleasant. Can in other words the contrary of pleasure be an object of desire?

A pain can be an object of desire. A man can want to inflict pain upon himself as a chastisement. But he can also want this simply because he likes the pain. This is perhaps perverse but it is not contrary to logic. It is not illogical since—as we have said before—a pleasant pain is not a contradiction in terms.

A man can also want something for himself which he finds unpleasant—for example to undergo a surgical operation because he considers it necessary or good for his bodily health or general well-being. A man can desire the unpleasant out of sheer curiosity. ‘What will be my reaction? Shall I faint or vomit?’ One can be intensely curious about such things. A man could take a perverse pleasure in vomiting and for the sake of this pleasure want to eat something extremely distasteful. In none of these cases however in which something unpleasant is wanted do we want the unpleasant for the sake of its being unpleasant but for the sake of something else—some pleasure maybe—to which it is conducive.

The fact that the contrary of pleasure if I am right thus necessarily is not an object of desire may have contributed to the illusion of psychological hedonism that pleasure necessarily is the ultimate end after which people aspire whenever they want something.

10. A few brief remarks will here be made on the doctrine known as Ethical Hedonism.

We shall distinguish between two principal forms which this doctrine may assume:

Firstly Ethical Hedonism can be a theory about the concept of goodness or the meaning of the word ‘good’. In its crudest form this theory maintains that any context where the word ‘good’ is used (not mentioned) either is one from which ‘good’ can be eliminated by simply substituting for it the word ‘pleasant’—as in the phrase ‘a good-smelling flower’—or is one from which ‘good’ can be eliminated by means of an analysis in terms of ‘pleasure’—as on our suggested view of secondary hedonic judgments it may become eliminated from the phrase ‘a good apple’.

Secondly Ethical Hedonism can be an axiological theory about the character of good things. In its crudest form this theory defends some such view as that those and only those things are good which either are judged pleasant in themselves or are (somehow) instrumental or useful for the production of pleasure i.e. causally responsible for the coming into being of pleasant things.

Moore in his well-known criticism of hedonism thought that the axiological theory could be refuted on the ground that it conflicts with our value-intuitions and the conceptual theory on the ground that it commits ‘the Naturalistic Fallacy’. ‘Good’ means good Moore says and not anything else e.g. pleasant. About this I shall only say that it seems to me just as obvious that ‘good’ sometimes means ‘pleasant’ or can otherwise become translated into hedonic terms as it seems to me obvious that ‘good’ does not always mean ‘pleasant’ or can become thus translated. The pleasant pleasure we have called a form of the good or of goodness. It is equally futile to try to reduce this form to one or several others as it is to try to reduce all other forms to it. But there may exist logical connexions of a more complex and subtle nature between the forms.

If one is aware of the multiform nature of goodness one will realize that the general question ‘Is pleasure good?’ is unintelligible unless the form of goodness is specified. In one sense of ‘good’ the question is just as empty of content or logically defect as the questions ‘Is pleasure pleasant?’ or ‘Is goodness good?’. When correctly stated the question must mean something along the following lines: Are the things which are good hedonically also good in some other respect? And here this other respect must be specified. The question may be well worth discussing. So may the converse question be: Are the things which are in such-and-such respect good also good hedonically? In the case of neither question however would an affirmative answer establish that pleasure is the ‘sole and ultimate good’ in any reasonable sense of those unprecise words.

• 1. Hume calls pleasure and pain alternatingly ‘impressions’ and ‘perceptions’, sometimes also ‘sensations’. Bentham calls them ‘interesting perceptions’.
• 2. See e.g., the discussion by Broad in ‘Certain Features in Moore's Ethical Doctrines’ in The Philosophy of G. E. Moore (1942), pp. 57–67 and the reply by Moore, op. cit., especially p. 587.
• 3. Gilbert Ryle, Dilemmas (1954), pp. 54–67.
• 4. Substantially the same distinction between three forms of pleasure is made by Broad in Five Types of Ethical Theory (1930), pp. 187 and 191 f.
• 5. Passow's well-known Handwörterbuch der griechischen Sprache gives for the Greek word ηδυς, of which ‘hedonic’ and ‘hedonism’ are linguistic off-springs, the German equivalents süss (‘sweet’) and angenehm (‘pleasant’) and suggests that the word was originally used for taste. Even if this suggestion be true, it does not follow, however, that the word was used as a name of a sense-quality as distinct from a value-attribute of sensations of this quality. The two uses, when speaking of sweet-tasting things, probably merged into one. These observations on the Greek ηδυς may be said to show how deep-rooted is the tendency to view pleasure as quasi a sensible quality.
• 6. As a prototype may be taken the theory of Edvard Westermarck, first presented in his monumental work The Origin and Development of the Moral Ideas I–II (1906–08) and further developed and defended in Ethical Relativity (1932).
• 7. See Ethics (1912), p. 76. Moore, when giving the ‘refutation’, is immediately dealing with the subjectivist suggestion that ‘this is tight’ meant the same as ‘somebody believes that this is right’. But the same argument applies, he maintains, to the suggestion that the tightness of an act meant that the act is thought to be right (p. 80) and also to the corresponding suggestion about goodness (p. 98).
• 8. It is significant that Moore, when stating the subjectivist view, often uses the formulation ‘thinks it right (good)’, but when trying to refute it by means of the argument ad infinitum switches to the formulation ‘thinks (believes) that it is right (good)’.