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

7
DESCRIPTION AND EVALUATION

7.1. Of all the problems raised by the preceding argument, the key problem is as follows: there are two sorts of things that we can say, for example, about strawberries; the first sort is usually called descriptive, the second sort evaluative. Examples of the first sort of remark are, 'This strawberry is sweet' and 'This strawberry is large, red, and juicy'. Examples of the second sort of remark are 'This is a good strawberry' and 'This strawberry is just as strawberries ought to be'. The first sort of remark is often given as a reason for making the second sort of remark; but the first sort does not by itself entail the second sort, nor vice versa. Yet there seems to be some close logical connexion between them. Our problem is: 'What is this connexion?'; for no light is shed by saying that there is a connexion, unless we can say what it is.

The problem may also be put in this way: if we knew all the descriptive properties which a particular strawberry had (knew, of every descriptive sentence relating to the strawberry, whether it was true or false), and if we knew also the meaning of the word 'good', then what else should we require to know, in order to be able to tell whether a strawberry was a good one? Once the question is put in this way, the answer should be apparent. We should require to know, what are the criteria in virtue of which a strawberry is to be called a good one, or what are the characteristics that make a strawberry a good one, or what is the standard of goodness in strawberries. We should require to be given the major premiss. We have already seen that we can know the meaning of 'good strawberry' without knowing any of these latter things -- though there is also a sense of the sentence 'What does it mean to call a strawberry a good one?' in which we should not know the answer to it, unless we also knew the answer to these other questions. It is now time to elucidate and distinguish these two ways in which we can be said to know what it means to call an object a good member of its class. This will help us to see more clearly both the differences and the similarities between 'good' and words like 'red' and 'sweet'.

Since we have been dwelling for some time on the differences, it will do no harm now to mention some of the similarities. For this purpose, let us consider the two sentences 'M is a red motor-car' and 'M is a good motor-car'. It will be noticed that 'motor-car', unlike 'strawberry', is a functional word, as defined in the preceding chapter. Reference to the Shorter Oxford English Dictionary shows that a motor-car is a carriage, and a carriage a means of conveyance. Thus, if a motor-car will not convey anything, we know from the definition of motor-car that it is not a good one. But when we know this, we know so little, compared with what is required in order to know the full criteria of a good motorcar, that I propose in what follows to ignore, for the sake of simplicity, this complicating factor. I shall treat 'motor-car' as if it did not have to be defined functionally: that is to say, I shall assume that we could learn the meaning of 'motor-car' (as in a sense we can) simply by being shown examples of motor-cars. It is, of course, not always easy to say whether or not a word is a functional word; it depends, like all questions of meaning, on how the word is taken by a particular speaker.

The first similarity between 'M is a red motor-car' and 'M is a good motor-car' is that both can be, and often are, used for conveying information of a purely factual or descriptive character. If I say to someone 'M is a good motor-car', and he himself has not seen, and knows nothing of M, but does on the other hand know what sorts of motor-car we are accustomed to call 'good' (knows what is the accepted standard of goodness in motor-cars), he undoubtedly receives information from my remark about what sort of motor-car it is. He will complain that I have misled him, if he subsequently discovers that M will not go over 30 m.p.h., or uses as much oil as petrol, or is covered with rust, or has large holes in the roof. His reason for complaining will be the same as it would have been if I had said that the car was red and he subsequently discovered that it was black. I should have led him to expect the motor-car to be of a certain description when in fact it was of a quite different description.

The second similarity between the two sentences is this. Sometimes we use them, not for actually conveying information, but for putting our hearer into a position subsequently to use the word 'good' or 'red' for giving or getting information. Suppose, for example, that he is utterly unfamiliar with motor-cars in the same sort of way as most of us are unfamiliar with horses nowadays, and knows no more about motor-cars than is necessary in order to distinguish a motorcar from a hansom cab. In that case, my saying to him 'M is a good motor-car' will not give him any information about M, beyond the information that it is a motor-car. But if he is able then or subsequently to examine M, he will have learnt something. He will have learnt that some of the characteristics which M has, are characteristics which make people -- or at any rate me -- call it a good motor-car. This may not be to learn very much. But suppose that I make judgements of this sort about a great many motor-cars, calling some good and some not good, and he is able to examine all or most of the motor-cars about which I am speaking; he will in the end learn quite a lot, always presuming that I observe a consistent standard in calling them good or not good. He will eventually, if he pays careful attention, get into the position in which he knows, after I have said that a motor-car is a good one, what sort of a motor-car he may expect it to be -- for example fast, stable on the road, and so on.

Now if we were dealing, not with 'good', but with 'red', we should call this process 'explaining the meaning of the word' -- and we might indeed, in a sense, say that what I have been doing is explaining what one means by 'a good motorcar'. This is a sense of 'mean' about which, as we have seen, we must be on our guard. The processes, however, are very similar. I might explain the meaning of 'red' by continually saying of various motor-cars 'M is a red motor-car', 'N is not a red motor car', and so on. If he were attentive enough, he would soon get into a position in which he was able to use the word 'red' for giving or getting information, at any rate about motor-cars. And so, both with 'good' and with 'red', there is this process, which in the case of 'red' we may call 'explaining the meaning', but in the case of 'good' may only call it so loosely and in a secondary sense; to be clear we must call it something like 'explaining or conveying or setting forth the standard of goodness in motor-cars'.

The standard of goodness, like the meaning of 'red', is normally something which is public and commonly accepted. When I explain to someone the meaning of 'red motor-car', he expects, unless I am known to be very eccentric, that he will find other people using it in the same way. And similarly, at any rate with objects like motor-cars where there is a commonly accepted standard, he will expect, having learnt from me what is the standard of goodness in motor-cars, to be able, by using the expression 'good motor-car', to give information to other people, and get it from them, without confusion.

A third respect in which 'good motor-car' resembles 'red motor-car' is the following: both 'good' and 'red' can vary as regards the exactitude or vagueness of the information which they do or can convey. We normally use the expression 'red motor-car' very loosely. Any motor-car that lies somewhere between the unmistakably purple and the unmistakably orange could without abuse of language be called a red motor-car. And similarly, the standard for calling motor-cars good is commonly very loose. There are certain characteristics, such as inability to exceed 30 m.p.h., which to anyone but an eccentric would be sufficient conditions for refusing to call it a good motor-car; but there is no precise set of accepted criteria such that we can say 'If a motor-car satisfies these conditions, it is a good one; if not, not'. And in both cases we could be precise if we wanted to. We could, for certain purposes, agree not to say that a motor-car was 'really red' unless the redness of its paint reached a certain measurable degree of purity and saturation; and similarly, we might adopt a very exact standard of goodness in motor-cars. We might refuse the name 'good motor-car' to any car that would not go round a certain race-track without mishap in a certain limited time, that did not conform to certain other rigid specifications as regards accommodation, &c. This sort of thing has not been done for the expression 'good motor-car'; but, as Mr. Urmson has pointed out, it has been done by the Ministry of Agriculture for the expression 'super apple'.1

It is important to notice that the exactness or looseness of their criteria does absolutely nothing to distinguish words like 'good' from words like 'red'. Words in both classes may be descriptively loose or exact, according to how rigidly the criteria have been laid down by custom or convention. It certainly is not true that value-words are distinguished from descriptive words in that the former are looser, descriptively, than the latter. There are loose and rigid examples of both sorts of word. Words like 'red' can be extremely loose, without becoming to the least degree evaluative; and expressions like 'good sewage effluent' can be the subject of very rigid criteria, without in the least ceasing to be evaluative.

It is important to notice also, how easy it is, in view of these resemblances between 'good' and 'red', to think that there are no differences -- to think that to set forth the standard of goodness in motor-cars is to set forth the meaning, in all senses that there are of that word, of the expression 'good motor-car'; to think that 'M is a good motor-car' means neither more nor less than 'M has certain characteristics of which "good" is the name'.

7.2. It is worth noticing here that the functions of the word 'good' which are concerned with information could be performed equally well if 'good' had no commendatory function at all. This can be made clear by substituting another word, made up for the purpose, which is to be supposed to lack the commendatory force of 'good'. Let us use 'doog' as this new word. 'Doog', like 'good', can be used for conveying information only if the criteria for its application are known; but this makes it, unlike 'good', altogether meaningless until these criteria are made known. I make the criteria known by pointing out various motor-cars, and saying 'M is a doog motor-car', 'N is not a doog motor-car', and so on. We must imagine that, although 'doog' has no commendatory force, the criteria for doogness in motor-cars which I am employing are the same as those which, in the previous example, I employed for goodness in motor-cars. And so, as in the previous example, the learner, if he is sufficiently attentive, becomes able to use the word 'doog' for giving or getting information; when I say to him 'Z is a doog motor-car', he knows what characteristics to expect it to have; and if he wants to convey to someone else that a motor-car Y has those same characteristics, he can do so by saying 'Y is a doog motor-car'.

Thus the word 'doog' does (though only in connexion with motor-cars) half the jobs that the word 'good' does -- namely, all those jobs that are concerned with the giving, or learning to give or get, information. It does not do those jobs which are concerned with commendation. Thus we might say that 'doog' functions just like a descriptive word. First my learner learns to use it by my giving him examples of its application, and then he uses it by applying it to fresh examples. It would be quite natural to say that what I was doing was teaching my learner the meaning of 'doog'; and this shows us again how natural it is to say that, when we are learning a similar lesson for the expression 'good motor-car' (i.e. learning the criteria of its application), we are learning its meaning. But with the word 'good' it is misleading to say this; for the meaning of 'good motor-car' (in another sense of 'meaning') is something that might be known by someone who did not know the criteria of its application; he would know, if someone said that a motor-car was a good one, that he was commending it; and to know that, would be to know the meaning of the expression. Further, as we saw earlier (6.4), someone might know about 'good' all the things which my learner learnt about the word 'doog' (namely, how to apply the word to the right objects, and use it for giving and getting information) and yet be said not to know its meaning; for he might not know that to call a motor-car good was to commend it.

7.3. It may be objected by some readers that to call the descriptive or informative job of 'good' its meaning in any sense is illegitimate. Such objectors might hold that the meaning of 'good' is adequately characterized by saying that it is used for commending, and that any information we get from its use is not a question of meaning at all. When I say 'M is a good motor-car', my meaning, on this view, is to commend M; if a hearer gets from my remark, together with his knowledge of the standard habitually used by me in assessing the merits of motor-cars, information about what description of motor-car it is, this is not part of my meaning; all my hearer has done is to make an inductive inference from 'Hare has usually in the past commended motor-cars of a certain description' and 'Hare has commended M' to 'M is of the same description'. I suspect that this objection is largely a verbal one, and I have no wish to take sides against it. On the one hand, we must insist that to know the criteria for applying the word 'good' to motor-cars is not to know -- at any rate in the full or primary sense -- the meaning of the expression 'good motor-car'; to this extent the objection must be agreed with. On the other hand, the relation of the expression 'good motor-car' to the criteria for its application is very like the relation of a descriptive expression to its defining characteristics, and this likeness finds an echo in our language when we ask 'What do you mean, good?', and get the answer 'I mean it'll do 80 and never breaks down'. In view of this undoubted fact of usage, I deem it best to adopt the term 'descriptive meaning'. Moreover, it is natural to say that a sentence has descriptive meaning, if the speaker intends it primarily to convey information; and when a newspaper says that X opened the batting on a good wicket, its intention is not primarily to commend the wicket, but to inform its readers what description of wicket it was.

7.4. It is time now to justify my calling the descriptive meaning of 'good' secondary to the evaluative meaning. My reasons for doing so are two. First, the evaluative meaning is constant for every class of object for which the word is used. When we call a motor-car or a chronometer or a cricket-bat or a picture good, we are commending all of them. But because we are commending all of them for different reasons, the descriptive meaning is different in all cases. We have knowledge of the evaluative meaning of 'good' from our earliest years; but we are constantly learning to use it in new descriptive meanings, as the classes of objects whose virtue we learn to distinguish grow more numerous. Sometimes we learn to use 'good' in a new descriptive meaning through being taught it by an expert in a particular field -- for example, a horseman might teach me how to recognize a good hunter. Sometimes, on the other hand, we make up a new descriptive meaning for ourselves. This happens when we start having a standard for a class of objects, certain members of which we have started needing to place in order of merit, but for which there has hitherto been no standard, as in the 'cactus' example (6.2). I shall in the next chapter discuss why we commend things.

The second reason for calling the evaluative meaning primary is, that we can use the evaluative force of the word in order to change the descriptive meaning for any class of objects. This is what the moral reformer often does in morals; but the same process occurs outside morals. It may happen that motor-cars will in the near future change considerably in design (e.g. by our seeking economy at the expense of size). It may be that then we shall cease giving the name 'a good motor-car' to a car that now would rightly and with the concurrence of all be allowed that name. How, linguistically speaking, would this have happened? At present, we are roughly agreed (though only roughly) on the necessary and sufficient criteria for calling a motor-car a good one. If what I have described takes place, we may begin to say 'No cars of the nineteen-fifties were really good; there weren't any good ones till 1960'. Now here we cannot be using 'good' with the same descriptive meaning as it is now generally used with; for some of the cars of 1950 do indubitably have those characteristics which entitle them to the name 'good motor-car' in the 1950 descriptive sense of that word. What is happening is that the evaluative meaning of the word is being used in order to shift the descriptive meaning; we are doing what would be called, if 'good' were a purely descriptive word, redefining it. But we cannot call it that, for the evaluative meaning remains constant; we are rather altering the standard. This is similar to the process called by Professor Stevenson 'persuasive definition';2 the process is not necessarily, however, highly coloured with emotion.

We may notice here that there are two chief ways in which a change in standard may be reflected in, and indeed partly effected by, a change in language. The first is the one which I have just illustrated; the evaluative meaning of 'good' is retained, and is used in order to alter the descriptive meaning and so establish a new standard. The second does not often occur with the word 'good'; for that word is so well-established as a value-word that the procedure would be practically impossible. This procedure is for the word to be gradually emptied of its evaluative meaning through being used more and more in what I shall call a conventional or 'inverted-commas' way; when it has lost all its evaluative meaning it comes to be used as a purely descriptive word for designating certain characteristics of the object, and, when it is required to commend or condemn objects in this class, some quite different value-word is imported for the purpose. The two processes may be illustrated and contrasted by a somewhat over-schematized account of what has happened in the last two centuries to the expression 'eligible bachelor'. 'Eligible' started off as a value-word, meaning 'such as should be chosen (sc. as a husband for one's daughters)'. Then, because the criteria of eligibility came to be fairly rigid, it acquired a descriptive meaning too; a person, if said to be eligible, might, in the eighteenth century, have been expected to have large landed estates and perhaps a title. By the nineteenth century, however, the criteria of eligibility have changed; what makes a bachelor eligible is no longer necessarily landed property or a title; it is substantial wealth of any kind provided that it is well-secured. We might imagine a nineteenth-century mother saying 'I know he is not of noble birth; but he's eligible all the same, because he has £3,000 a year in the Funds, and much more besides when his father dies'. This would be an example of the first method. On the other hand, in the twentieth century, partly as a reaction from the over-rigid standards of the nineteenth, which resulted in the word 'eligible' lapsing into a conventional use, the second method has been adopted. If now someone said 'He is an eligible bachelor', we could almost feel the inverted commas round the word, and even the irony; we should feel that if that was all that could be said for him, there must be something wrong with him. For commending bachelors, on the other hand, we now use quite different words; we say 'He is likely to make a very good husband for Jane', or 'She was very sensible to say "yes"'.

The close connexion of standards of values with language is illustrated by the plight of the truly bilingual. A writer equally at home in English and French relates that once, when walking in the park on a rainy day, he met a lady dressed in a way which the English would call sensible, but the French ridicule; his mental reaction to this had to be expressed bilingually, because the standards he was applying were of diverse origin; he found himself saying to himself (slipping from English into French) 'Pretty adequate armour. How uncomfortable though. Why go for a walk if you feel like this? Elle est parfaitement ridicule.' This cleavage of standards is said sometimes to produce neuroses in bilinguals, as might be expected in view of the close bearing of standards of values upon action.3

7.5. Although with 'good' the evaluative meaning is primary, there are other words in which the evaluative meaning is secondary to the descriptive. Such words are 'tidy' and 'industrious'. Both are normally used to commend; but we can say, without any hint of irony, 'too tidy' or 'too industrious'. It is the descriptive meaning of these words that is most firmly attached to them; and therefore, although we must for certain purposes class them as value-words (for if we treat them as purely descriptive, logical errors result), they are so in a less full sense than 'good'. If the evaluative meaning of a word, which was primary, comes to be secondary, that is a sign that the standard to which the word appeals has become conventional. It is, of course, impossible to say exactly when this has happened; it is a process like the coming of winter.

Although the evaluative meaning of 'good' is primary, the secondary descriptive meaning is never wholly absent. Even when we are using the word 'good' evaluatively in order to set up a new standard, the word still has a descriptive meaning, not in the sense that it is used to convey information, but in the sense that its use in setting up the new standard is an essential preliminary -- like definition in the case of a purely descriptive word -- to its subsequent use with a new descriptive meaning. It is also to be noticed that the relative prominence of the descriptive and evaluative meanings of 'good' varies according to the class of objects within which commendation is being given. We may illustrate this by taking two extreme examples. If I talk of 'a good egg', it is at once known to what description of egg I am referring -- namely, one that is not decomposed. Here the descriptive meaning predominates, because we have very fixed standards for assessing the goodness of eggs. On the other hand, if I say that a poem is a good one, very little information is given about what description of poem it is -- for there is no accepted standard of goodness in poems. But it must not be thought that 'good egg' is exclusively descriptive, or 'good poem' exclusively evaluative. If, as the Chinese are alleged to do, we chose to eat eggs that are decomposed, we should call that kind of egg good, just as, because we choose to eat game that is slightly decomposed, we call it 'well-hung' (compare also the expression 'good Stilton cheese'). And if I said that a poem was good, and was not a very eccentric person, my hearer would be justified in assuming that the poem was not 'Happy birthday to you!'

In general, the more fixed and accepted the standard, the more information is conveyed. But it must not be thought that the evaluative force of the word varies at all exactly in inverse proportion to the descriptive. The two vary independently: where a standard is firmly established and is as firmly believed in, a judgement containing 'good' may be highly informative, without being any the less commendatory. Consider the following description of the Oxford Sewage Farm:

The method employed is primitive but efficient. The farm is unsightly, obnoxious to people dwelling near it, and not very remunerative, but the effluent from it is, in the technical sense, good.4
Now here, as may be seen by consulting handbooks on the subject, there are perfectly well-recognized tests for determining whether effluent is good or bad. One manual5 gives a simple field test, and another6 gives a series of more comprehensive tests which take up seventeen pages. This might tempt us to say that the word is used in a purely descriptive sense and has no evaluative force. But, although admittedly in calling effluent good in this technical sense we are commending it as effluent and not as perfume, we are nevertheless commending it; it is not a neutral chemical or biological fact about it that it is good; to say that it was bad would be to give a very good reason for sacking the sewage-farmer or taking other steps to see that it was good in future. The proper comment on such a lapse was made by a former Archbishop of York, speaking to the Congress of the Royal Sanitary Institute, 1912:
There is now, I hope, no need of the trenchant eloquence of that noble-hearted pioneer of sanitary science, Charles Kingsley, to insist that it is not religion, but something more nearly approaching blasphemy, to say that an outbreak of disease is God's will being done, when patently it is man's duty which is being left undone.7
It is true that, if the word 'good' in a certain sentence has very little evaluative meaning, it is likely that it has a fair amount of descriptive meaning, and vice versa. That is because, if it had very little of either, it would have very little meaning at all, and would not be worth uttering. To this extent the meanings vary inversely. But this is only a tendency; we may do justice to the logical phenomena by saying that 'good' normally has at least some of both sorts of meaning; that it normally has sufficient of both sorts taken together to make it worth uttering; and that, provided that the first two conditions are satisfied, the amounts of the two sorts of meaning vary independently.

There are, however, cases in which we use the word 'good' with no commendatory meaning at all. We must distinguish several kinds of such non-commendatory uses. The first has been called the inverted-commas use. If I were not accustomed to commend any but the most modern styles of architecture, I might still say 'The new chamber of the House of Commons is very good Gothic revival'. I might mean this in several senses. The first is that in which it is equivalent to 'a good example to choose, if one is seeking to illustrate the typical features of Gothic revival' or 'a good specimen of Gothic revival'. This is a specialized evaluative sense, with which we are not here concerned. I might mean, on the other hand, 'genuinely preferable to most other examples of Gothic revival, and therefore to be commended within the class of Gothic revival buildings, though not within the class of buildings in general'. With this sense, too, we are not now concerned; it is a commendatory use, with a limited class of comparison (8.2). The sense with which we are concerned is that in which it means, roughly, 'the sort of Gothic revival building about which a certain sort of people -- you know who -- would say "that is a good building".' It is characteristic of this use of 'good' that in expanding it we often want to put the word 'good' inside inverted commas; hence the name. We are, in this use, not making a value-judgement ourselves, but alluding to the value-judgements of other people. This type of use is extremely important for the logic of moral judgements, in which it has caused some confusion.

It is to be noticed that it is easiest to use 'good' in an inverted-commas sense when a certain class of people, who are sufficiently numerous and prominent for their value-judgements to be well known (e.g. the 'best' people in any field), have a rigid standard of commendation for that class of object. In such cases, the inverted-commas use can verge into an ironic use, in which not only is no commendation being given, but rather the reverse. If I had a low opinion of Carlo Dolci, I might say 'If you want to see a really "good" Carlo Dolci, go and look at the one in . . .'.

There is another use in which the absence of evaluative content is not sufficiently obvious to the speaker for us to call it either an inverted-commas or an ironic use. This is the conventional use, in which the speaker is merely paying lip-service to a convention, by commending, or saying commendatory things about, an object just because everyone else does. I might, if I myself had no preference at all about the design of furniture, still say 'This piece of furniture is of good design', not because I wished to guide my own or anyone else's choice of furniture, but simply because I had been taught the characteristics which are generally held to be criteria of good design, and wished to show that I had 'good taste' in furniture. It would be difficult in such a case to say whether I was evaluating the furniture or not. If I were not a logician, I should not ask myself the questions which would determine whether I was. Such a question would be 'If someone (not connected in any way with the furniture trade), consistently and regardless of cost filled his house with furniture not conforming to the canons by which you judge the design of this furniture to be good, would you regard that as evidence that he did not agree with you?' If I replied 'No, I would not; for what furniture is of good design is one question, and what furniture one chooses for oneself is another', then we might conclude that I had not been really commending the design by calling it good, but only paying lip-service to a convention. We shall recur to this sort of cross-examination later (11.2).

These are only some of the many ways in which we use the word 'good'. A logician cannot do justice to the infinite subtlety of language; all he can do is to point out some of the main features of our use of a word, and thereby put people on their guard against the main dangers. A full understanding of the logic of value-terms can only be achieved by continual and sensitive attention to the way we use them.


Notes

1 Mind, lix (1950), 152 (also in Logic and Language, ii, ed. Flew, 166).

2 Ethics and Language, ch. ix.

3 P. H. J. Lagarde-Quost, 'The Bilingual Citizen', Britain Today, Dec. 1947, p. 13; Jan. 1948, p. 13.

4 Social Services in the Oxford District, p. 322.

5 Kershaw, Sewage Purification and Disposal, pp. 213-14.

6 Thresh, Beale, and Suckling, The Examination of Waters and Water Supplies, 6th ed., ch. xx.

7 Kershaw, op. cit., p. 4.