R. M. Hare, The Language of Morals (1952)

11
'OUGHT' AND IMPERATIVES

11.1. Since a large part of my argument hinges on the assumption, hitherto not fully defended, that value-judgements, if they are action-guiding, must be held to entail imperatives, and since this assumption may very well be questioned, it is time to examine it. It might be held, for example, that I can without contradiction say 'You ought to do A, but don't', and that therefore there can be no question of entailment; entailment in any case is a very strong word, and though many might be found to agree that value-judgements are action-guiding in some sense, it might be held that they are action-guiding only in the sense in which even plain judgements of fact may be action-guiding. For example, if I say 'The train is just about to depart', this may guide a person who wants to catch the train to take his seat; or, to take a moral case, if I say to a person who is thinking of giving some money to a friend supposedly in distress, 'The story he has just told you is quite untrue', this may guide him to make a different moral decision from that which he would otherwise have made. And similarly it might be held that value-judgements are action-guiding in no stronger sense than these statements of fact. It might be urged that, just as the statement that the train is going to depart has no bearing upon the practical problems of someone who does not want to catch the train, and just as, if the man who is thinking of giving money to his friend does not recognize that the truth or otherwise of his friend's story has any bearing on the question, it may not affect his decision, so, if a man has no intention of doing what he ought, to tell him that he ought to do something may not be accepted by him as a reason for doing it. I have put as forcibly as possible this objection, which strikes at the root of my whole argument. The objection alleges, in brief, that 'ought'-sentences are not imperatives, neither do they entail imperatives without the addition of an imperative premiss. In answer to this, I have to show that 'ought'-sentences, at any rate in some of their uses, do entail imperatives.

It is necessary first to recall something that I said earlier (7.5) in discussing the evaluative and descriptive forces of value-judgements. We noticed that it is possible for people who have acquired very stable standards of values to come to treat value-judgements more and more as purely descriptive, and to let their evaluative force get weaker. The limit of this process is reached when, as we described it, the value-judgement 'gets into inverted commas', and the standard becomes completely 'ossified'. Thus it is possible to say 'You ought to go and call on the So-and-sos' meaning by it no value-judgement at all, but simply the descriptive judgement that such an action is required in order to conform to a standard which people in general, or a certain kind of people not specified but well understood, accept. And certainly, if this is the way in which an 'ought'-sentence is being used, it does not entail an imperative; we can certainly say without contradiction 'You ought to go and call on the So-and-sos, but don't'. I do not wish to claim that all 'ought'-sentences entail imperatives, but only that they do so when they are being used evaluatively. It will subsequently become apparent that I am making this true by definition; for I should not say that an 'ought'-sentence was being used evaluatively, unless imperatives were held to follow from it; but more of that later.

Thus one answer which we can make to the objection is that the cases which appear to support it are not genuine value-judgements. In the example quoted, if a man has no intention of doing what he ought, and if, therefore, telling him what he ought to do is not taken by him as entailing an imperative, that merely shows that, in so far as he accepts that he ought to do so-and-so (and of course no premiss enables a conclusion to be drawn unless it is accepted), he accepts it only in a non-evaluative, inverted-commas sense, as meaning that so-and-so falls within a class of actions which is generally held (but not by him) to be obligatory in the evaluative, imperative-entailing sense. This is an answer which disposes of some awkward cases, but which will not be accepted as a complete answer unless we extend its scope considerably. For it may be held that there are some genuine value-judgements which do not entail imperatives.

11.2. Let us recall something else that I said earlier (4.7). Practical principles, if they are accepted sufficiently long and unquestioningly, come to have the force of intuition. Thus our ultimate moral principles can become so completely accepted by us, that we treat them, not as universal imperatives, but as matters of fact; they have the same obstinate indubitability. And there is indeed a matter of fact to which it is very easy for us to take them as referring, namely, what we call our 'sense of obligation'. This is a concept that now requires investigation.

It is easy to see how, if we have been brought up from our earliest years in obedience to a principle, the thought of not obeying it becomes abhorrent to us. If we fail to obey it, we suffer remorse; when we do obey it, we feel at ease with ourselves. These feelings are reinforced by all those factors which psychologists have listed;1 and the total result is what is generally called a feeling of obligation. It is a fact that we have this feeling of obligation -- different people in different degrees, and with different contents. Judgements that I have a feeling of obligation to do X or Y are statements of empirical fact. This is not the place to argue about their interpretation; it is no doubt possible to dispute whether sentences like 'A is suffering from remorse' or 'B feels it is his duty to do Y' are reports of private mental events or are to be interpreted behaviouristically; but such controversies do not here concern us. Here it is important to point out a fact which has been singularly ignored by some moralists, that to say of someone that he has a feeling of obligation is not the same as to say that he has an obligation. To say the former is to make a statement of psychological fact; to say the latter is to make a value-judgement. A man who has been brought up in an Army family, but has become affected by pacifism, may well say 'I have a strong feeling that I ought to fight for my country, but I wonder whether I really ought'. Similarly, a Japanese brought up in accordance with Bushido might say 'I have a strong feeling that I ought to torture this prisoner in order to extract information which will be to my Emperor's advantage; but ought I really to do so?'

The confusion between psychological statements about a feeling or sense of obligation and value-judgements about obligation itself is not confined to professional philosophers. The ordinary man so very rarely questions the principles in which he has been brought up, that he is usually willing, whenever he has a feeling that he ought to do X, to say on this ground alone that he ought to do X; and therefore he often gives voice to this feeling by saying 'I ought to do X'. This sentence is not a statement that he has the feeling; it is a value-judgement made as a result of having the feeling. For those, however, who have not studied the logical behaviour of value-judgements, and have not reflected on such examples as those of the pacifist and the Japanese just given, it is easy to take this remark as a statement of fact to the effect that he has the feeling, or to confuse it in meaning with this statement. But anyone, except a professional philosopher maintaining at all costs a moral sense theory, could be got to see that the meaning is not the same, by being asked 'Wouldn't it be possible for you to feel just like that, although you really oughtn't to do X?' or 'Mightn't you feel like that and be wrong?'

The confusion, however, goes deeper than this. We have seen that there is a conscious inverted-commas use of value-words in which, for example, 'I ought to do X' becomes roughly equivalent to 'X is required in order to conform to a standard which people in general accept*. But it is also possible to use the word 'ought' and other value-words, as it were, unconsciously in inverted commas; for the standard which people in general accept may also be the standard which one has been brought up to accept oneself, and therefore not only does one refer to this standard by saying 'I ought to do X', but one has feelings of obligation to conform to the standard.

It is then possible to treat 'I ought to do X' as a confused mixture of three judgements.

  1. 'X is required in order to conform to the standard which people generally accept' (statement of sociological fact);
  2. 'I have a feeling that I ought to do X' (statement of psychological fact);
  3. 'I ought to do X' (value-judgement).
Even this tripartite division conceals the complexity of the meaning of such sentences; for each of the three elements is itself complex and can be taken in different senses. But even if we confine ourselves to the three elements just given, it is usually impossible for an ordinary person, untrained in logical subtleties, to ask or to answer the question 'Which of these three judgements are you making, just (1), or (1) and (2), or all three, or some other combination?' The situation is very similar to that of the scientist who is asked by the logician 'Is your statement that phosphorus melts at 44° C. analytic or synthetic; if you found a substance which was in other respects just like phosphorus, but which melted at another temperature, would you say "It isn't really phosphorus" or would you say "Then after all some phosphorus melts at other temperatures"?2 The scientist might well, as Mr. A. G. N. Flew has pointed out to me, answer 'I don't know; I haven't yet come across the case which would make me decide this question; I have got better things to worry about'. Similarly, the ordinary person, making moral decisions on the basis of his accepted principles, very rarely has to ask himself the question that we have just asked. So long as his value-judgements correspond with the accepted standards, and with his own feelings, he does not have to decide which he is saying, because, as we might put it, all three are as yet for him materially equivalent; that is to say, no occasion arises for saying one which is not also an occasion for saying the other two. He therefore does not ask himself
'As I am using the word "ought", are the sentences "I ought to do what I feel I ought" and "I ought to do what everybody would say I ought" analytic or synthetic?'
It is the crucial case that makes him answer such a question; and in morals the crucial case comes when we are wondering whether to make a value-decision which is in disagreement with the accepted standards or with our own moral feelings -- such cases as I have cited. It is these cases that really reveal the difference in meaning between the three judgements that I have listed.

My answer to the objection then is, that cases which are alleged to be value-judgements not entailing imperatives will always on examination be found to be cases where what is meant is not of type (3) above, but of type (1) or (2) or a mixture of both. This contention is, of course, impossible to prove or even to render plausible, unless we know when we are to count a judgement as of type (3); but I propose to get over this difficulty in the only possible way, by making it a matter of definition. I propose to say that the test, whether someone is using the judgement 'I ought to do X' as a value-judgement or not is,

'Does he or does he not recognize that if he assents to the judgement, he must also assent to the command "Let me do X"?'
Thus I am not here claiming to prove anything substantial about the way in which we use language; I am merely suggesting a terminology which, if applied to the study of moral language, will, I am satisfied, prove illuminating. The substantial part of what I am trying to show is this, that, in the sense of 'value-judgement' just defined, we do make value-judgements, and that they are the class of sentences containing value-words which is of primary interest to the logician who studies moral language. Since what we are discussing is the logic of moral language and not that tangled subject known as moral psychology, I shall not here inquire farther into the fascinating problem, discussed by Aristotle, of akrasia or 'weakness of will'3 -- the problem presented by the person who thinks, or professes to think, that he ought to do something, but does not do it. The logical distinctions which I have been making shed considerable light on this question: but much more needs to be said, chiefly by way of a more thorough analysis of the phrase 'thinks that he ought'. For if we interpret my definition strictly, and take it in conjunction with what was said earlier (2.2) about the criteria for 'sincerely assenting to a command', the familiar 'Socratic paradox' arises, in that it becomes analytic to say that everyone always4 does what he thinks he ought to (in the evaluative sense). And this, to put Aristotle's objection in modern dress, is not how we use the word 'think'. The trouble arises because our criteria, in ordinary speech, for saying 'He thinks he ought' are exceedingly elastic. If a person does not do something, but the omission is accompanied by feelings of guilt, &c, we normally say that he has not done what he thinks he ought. It is therefore necessary to qualify the criterion given above for 'sincerely assenting to a command', and to admit that there are degrees of sincere assent, not all of which involve actually obeying the command. But the detailed analysis of this problem requires much more space than I can give it here, and must wait for another occasion.

11.3. The best way of establishing the primary logical interest of the evaluative sense of 'ought' is to show that, but for the existence of this sense, none of the familiar troubles generated by the word would arise. For of the three possible paraphrases of 'I ought to do X' given on p. 167, the first two are statements of fact. This is because, if they are expanded, it will be found that the word 'ought' in them always occurs in inverted commas or inside a subordinate clause beginning with 'that'. Thus (1) might be further paraphrased

'There is a principle of conduct which people generally accept, which says "One ought to do X in circumstances of a certain kind"; and I am now in circumstances of that kind'.
Similarly, (2) might be further paraphrased
'The judgement "I ought to do X" evokes in me a feeling of conviction'
or
'I find myself unable to doubt the judgement "I ought to do X" '
(though the latter paraphrase is a good deal too strong; for not all feelings are irresistible; there is indeed an infinite gradation from vague uneasy stirrings of conscience to what are often called 'moral intuitions'). Now the fact, that when (1) and (2) are expanded the original judgement which they paraphrase occurs within them inside inverted commas, shows that there must be some sense of that original judgement which is not exhausted by (1) and (2); for if there were not, the sentence in the inverted commas would have in its turn to be paraphrased by (1) or (2), and we should be involved in an infinite regress. In the case of (1), I do not know any way of getting over this difficulty; in the case of (2), it can be got over temporarily by substituting for (2) some such paraphrase as 'I have a certain recognizable feeling'. But the device is only temporary; for if we are asked what this feeling is, or how we recognize it, the reply can only be 'It is the feeling called "a feeling of obligation"; it is the feeling you usually have when you say, and mean, "I ought to do so-and-so".'

This means that neither (1) nor (2) can give the primary sense of 'I ought to do X'. Now let us suppose (as is not the case) that (3) generates none of the logical puzzles of the kind that we have been discussing; let us suppose, that is to say, that (3) can be analysed naturalistically. If this were so, then these puzzles would not arise in the cases of (1) or (2) either; for since, besides the expression in inverted commas, there is nothing else in the expansions of (1) and (2) that cannot be analysed naturalistically, it would be possible to effect a completely naturalistic analysis of all uses of 'ought', and thus of 'good' (12.3). The fact that this is not possible is entirely due to the intractably evaluative character of (3). Itis due ultimately to the impossibility, mentioned earlier (2.5) of deriving imperatives from indicatives; for (3), by definition, entails at least one imperative; but if (3) were analysable naturalistically, this would mean that it was equivalent to a series of indicative sentences; and this would constitute a breach of the principle established. Thus it is this fact, that in some of its uses 'ought' is used evaluatively (i.e. as entailing at least one imperative) that makes a naturalistic analysis impossible, and hence generates all the difficulties that we have been considering. A logician who neglects these uses will make his task easy, at the cost of missing the essential purpose of moral language.

It is this, above all, that makes the first part of this book relevant to what is discussed in the remainder. For all the words discussed in Parts II and III have it as their distinctive function either to commend or in some other way to guide choices or actions; and it is this essential feature which defies any analysis in purely factual terms. But to guide choices or actions, a moral judgement has to be such that if a person assents to it, he must assent to some imperative sentence derivable from it; in other words, if a person does not assent to some such imperative sentence, that is knock-down evidence that he does not assent to the moral judgement in an evaluative sense -- though of course he may assent to it in some other sense (e.g. one of those I have mentioned). This is true by my definition of the word evaluative. But to say this is to say that if he professes to assent to the moral judgement, but does not assent to the imperative, he must have misunderstood the moral judgement (by taking it to be non-evaluative, though the speaker intended it to be evaluative). We are therefore clearly entitled to say that the moral judgement entails the imperative; for to say that one judgement entails another is simply to say that you cannot assent to the first and dissent from the second unless you have misunderstood one or the other; and this 'cannot' is a logical 'cannot' -- if someone assents to the first and not to the second, this is in itself a sufficient criterion for saying that he has misunderstood the meaning of one or the other. Thus to say that moral judgements guide actions, and to say that they entail imperatives, comes to much the same thing.

I do not in the least wish to deny that moral judgements are sometimes used non-evaluatively, in my sense. All I wish to assert is that they are sometimes used evaluatively, and that it is this use which gives them the special characteristics to which I have drawn attention; and that, if it were not for this use, it would be impossible to give a meaning to the other uses; and also that, if it were not for the logical difficulties connected with the evaluative use, the other uses could be analysed naturalistically. Ethics, as a special branch of logic, owes its existence to the function of moral judgements as a guide in answering questions of the form 'What shall I do?'

11.4. I am now in a position to answer an objection which may have occurred to some readers. Writers on ethics often condemn 'naturalism' or some related fallacy in others, only to commit it themselves in a subtler form. It may be alleged that I have done this. I suggested earlier (5.3) that the term 'naturalist' should be reserved for such ethical theories as are open to refutation on lines similar to those marked out by Professor Moore. We must therefore ask whether any analogous refutation of my own theory can be constructed. Now it is true that I am not suggesting that moral judgements can be deduced from any statements of fact. In particular I am not suggesting the adoption of definitions of value-terms of the sort which Moore mistakenly attributed to Kant. Moore accused Kant of saying that 'This ought to be* means 'This is commanded'.5 This definition would be naturalistic; for 'A is commanded' is a statement of fact; it is expansible into 'Someone [it is not disclosed who] has said "Do A" '. The fact that the imperative is in inverted commas prevents it affecting the mood of the complete sentence. Needless to say, I am not suggesting any such equivalence, either for 'good' or for 'ought' or for any other value-word, except perhaps when they are used in what I have called an 'inverted-commas' sense, or in some other purely descriptive way. But it might nevertheless be said that according to my treatment of moral judgements certain sentences would become analytic which in ordinary usage are not analytic -- and this would be very like Moore's refutation. For example, consider sentences like the Psalmist's

Eschew evil and do good,6
or the line of John Wesley's hymn
In duty's path go on.7
On my theory these would, it might be alleged, become analytic; for from 'A is evil' is deducible the imperative sentence 'Eschew A', and from 'Path P is the path of duty' is deducible the imperative sentence 'Go on in path P'.

Now it must be noticed that such sentences as those quoted are expansible into sentences in which a value-judgement occurs in a subordinate clause. Thus if, instead of the archaic 'Eschew evil' we write 'Do not do what is evil', this can be expanded into 'For all X, if X is evil, do not do X'. For this instruction to be applied, it is necessary that we should conjoin it with a minor premiss 'A is evil', and from the two premisses conclude 'Do not do A'. For this reasoning to be helpful, it is necessary that the minor premiss 'A is evil' should be a statement of fact; there must be a criterion for telling unambiguously whether it is true or false. This means that in this premiss the word 'evil' must have a descriptive meaning (whatever further meaning it may have). But if the reasoning is to be valid, the word 'evil' in the major premiss must have the same meaning as in the minor; there also, therefore, it must have a descriptive meaning. Now it is this descriptive content which prevents the major premiss being analytic. Sentences of the sort we are discussing are normally used by people who have firmly established value-standards, and whose value-words have, therefore, a large element of descriptive meaning. In the sentence 'Do not do what is evil', the evaluative content of 'evil' is for the moment neglected; the speaker, as it were, lets his support of the standard slip for a moment, only in order to ram it back into place with the imperative verb. This is a first-class exercise in the maintenance of our standards, and that is why it is so much in place in hymns and psalms. But it can only be performed by those who are in no doubt as to what the standard is.

Contrast with these cases others which are superficially similar. Suppose that I am asked 'What shall I do?', and answer 'Do whatever is best' or 'Do what you ought to do'. In most contexts such answers would be regarded as unhelpful. It would be as if a policeman were asked 'Where shall I park my car?' and replied 'Wherever it is legitimate to park it'. I am asked by the speaker to give definite advice as to what he is to do; he asks me, just because he does not know what standard to apply in his case. If, therefore, I reply by telling him to conform to some standard as to whose provisions he is quite in the dark, I do not give him any useful advice. Thus in such a context the sentence 'Do whatever is best' really is analytic; for because the standard is assumed to be unknown, 'best' has no descriptive meaning.

Thus my account of the meaning of value-words is not naturalistic; it does not result in sentences becoming analytic which are not so in our ordinary usage. Rather it shows, by doing full justice to both the descriptive and the evaluative elements in the meaning of value-words, how it is that they play in our ordinary usage the role that they do play. A somewhat similar difficulty is presented by Satan's famous paradox 'Evil be thou my good'. This yields to the same type of analysis, but for reasons of space I am compelled to leave the reader to unravel the problem himself.

11.5. It may be asked at this point 'Are you not assimilating moral judgements too much to the ordinary universal imperatives that exist in most languages?' It has indeed been objected to all imperative analyses of moral judgements, that they would make a moral judgement like 'You ought not to smoke (in this compartment)' the equivalent of the universal imperative 'No Smoking'. And they are clearly not equivalent, though both, according to the theory which I have been advocating, entail 'Do not smoke'. It is therefore necessary to state what it is that distinguishes 'You ought not to smoke' from 'No smoking'. I have already touched on this problem, but it requires further discussion.

The first thing to notice about 'No smoking' is that it is not a proper universal, because it implicitly refers to an individual; it is short for 'Do not ever smoke in this compartment'. The moral judgement 'You ought not to smoke in this compartment' also contains references to individuals; for the pronouns 'you' and 'this' occur in it. But, in view of what I have said above (10. 3), this is not the end of the matter. The moral judgement 'You ought not to smoke in this compartment' has to be made with some general moral principle in mind, and its purpose must be either to invoke that general principle or to point to an instance of its application. The principle might be 'One ought never to smoke in compartments in which there are young children' or 'One ought never to smoke in compartments in which there is a "No Smoking" notice'. It is not always easy to elicit just what the principle is; but it always makes sense to ask what it is. The speaker cannot deny that there is any such principle. The same point might be put another way by saying that if we make a particular moral judgement, we can always be asked to support it by reasons; the reasons consist in the general principles under which the moral judgement is to be subsumed. Thus the particular moral judgement 'You ought not to smoke in this compartment' depends on a proper universal, even though it is not itself one. But this is not true of the imperative 'Do not ever smoke in this compartment'. This invokes no more general principle; it is itself as general as it requires to be, and this is not general enough to make it a proper universal.

The difference in universality between 'Do not ever smoke in this compartment' and 'You ought not to smoke in this compartment' may be brought out in the following way. Suppose that I say to someone 'You ought not to smoke in this compartment', and there are children in the compartment. The person addressed is likely, if he wonders why I have said that he ought not to smoke, to look around, notice the children, and so understand the reason. But suppose that, having ascertained everything that is to be ascertained about the compartment, he then says 'All right; I'll go next door; there's another compartment there just as good; in fact it is exactly like this one, and there are children in it too'. I should think if he said this that he did not understand the function of the word 'ought'; for 'ought' always refers to some general principle; and if the next compartment is really exactly like this one, every principle that is applicable to this one must be applicable to the other (8.2). I might therefore reply 'But look here, if you oughtn't to smoke in this compartment, and the other compartment is just like this one, has the same sort of occupants, the same notices on the windows, &c, then obviously you oughn't to smoke in that one either'. On the other hand, when the Railway Executive is making the momentous decision, on which compartments to put notices saying 'No Smoking', nobody says 'Look here! You've put a notice on this compartment, so you must put one on the one next to it, because it's exactly like it'. This is because 'No Smoking' does not refer to a universal principle of which this compartment is an instance.

It is, in fact, almost impossible to frame a proper universal in the imperative mood. Suppose that we try to do this by generalizing the sentence 'Do not ever smoke in this compartment'. First we eliminate the implicit 'you' by writing 'No one is ever to smoke in this compartment'. We then have to eliminate the 'this'. A step towards this is taken by writing 'No one is ever to smoke in any compartment of British Railways'. But we still have left here the proper name 'British Railways'. We can only achieve a proper universal by excluding all proper names, for example by writing 'No one is ever to smoke in any railway compartment anywhere'. This is a proper universal; but it is a sentence which no one could ever have occasion to utter. Commands are always addressed to someone or to some individual set (not class) of people. It is not clear what could be meant by the sentence just quoted, unless it were a moral injunction or other value-judgement. Suppose that we imagine God issuing such a command. Then it becomes at once like the Ten Commandments in form. Historically speaking 'Honour thy father and thy mother' is supposed to have been said, not to everyone in general, but only to members of the chosen people, just as 'Render to no man evil for evil' was addressed to Christ's disciples, not to the world at large -- though He intended that all men should become His disciples. But suppose that this were not so; suppose that 'Render to no man evil for evil' were addressed literally to the unlimited class 'every man'. Should we not say that it had become equivalent in meaning to the value-judgement 'One ought to render to no man evil for evil'? Similarly, a proverbial expression like 'Let sleeping dogs lie' may without much of a jolt be paraphrased by the (prudential) value-judgement 'One ought to let sleeping dogs lie'.

On the other hand, ordinary so-called universal imperatives like 'No Smoking' are distinguished from value-judgements by not being properly universal. We are thus able to discriminate between these two kinds of sentence, without in the least abandoning anything that I have said about the relation between value-judgements and imperatives. For both the complete universal and the incomplete entail the singular: 'Do not ever smoke in this compartment' entails 'Do not (now) smoke in this compartment'; and so does 'You ought not to smoke in this compartment', if it is used evaluatively. But the latter also entails, as the former does not, 'No one ought to smoke in any compartment exactly like this one', and this in its turn entails 'Do not smoke in any compartment exactly like this one'.

These considerations alone, however, would not be sufficient to account altogether for the complete difference in 'feel' between 'You ought not' and 'Do not ever'. This is reinforced by two other factors. The first has already been alluded to; the complete universality of the moral judgement means that we cannot 'get away from it'; and therefore its acceptance is a much more serious matter than the acceptance of an imperative from whose range of application we can escape. This would explain why imperatives such as those laws of a State, which are of very general application, and therefore very difficult to escape from, have a 'feel' much more akin to moral judgements than have the regulations of the Railway Executive. But a more important additional factor is that, partly because of their complete universality, moral principles have become so entrenched in our minds -- in the ways already described -- that they have acquired a quasi-factual character, and are indeed sometimes used non-evaluatively as statements of fact and nothing else, as we have seen. None of this is true of imperatives like 'No Smoking'; and this in itself would be quite enough to explain the difference in 'feel' between the two kinds of sentence. Since, however, I do not wish to deny that there are non-evaluative uses of moral judgements, but only to assert that there are evaluative uses, this difference in 'feel' in no way destroys my argument. It would indeed be absurd to pretend that 'No Smoking' is in all respects like 'You ought not to smoke'; I have been maintaining only that it is like it in one respect, that both entail singular imperatives such as 'Do not smoke (now)'.

Notes

1 Cf. J. C. Flugel, Man, Morals and Society, especially ch. iii.

2 G. H. von Wright, Logical Problems of Induction, ch. iii.

3 Nicomachean Ethics, VII. I ff.

4 Strictly, 'always, if physically and psychologically able', cf. p. 20.

5 Principia Ethica, pp. 127-8.

6 Ps. XXXIV, V. 14.

7 Hymns Ancient and Modern (1950), no. 310.