Argument Analysis 11-17

11. ELLIPSIS

We have just been looking at one way people avoid admitting their mistakes. What I want to look at now is an aspect of language that we can too easily exploit to prevent ourselves seeing that what we are saying is crucially incomplete. We utter a lot of words, but the action is mostly off-stage.

We have seen already that we tend to leave a lot out in ordinary arguments. Many vital parts of an argument would be too tedious and too obvious to include; other parts may be too complicated to express exactly, though we may be fairly sure of them (I suggested that this was part of the utility of conditional statements in many cases); and we often leave our hearers to supply crucial parts of the argument - as a way of flattering their intelligence, or keeping their attention, perhaps. What I want to claim now is that we also leave a lot out of many of the actual statements we do make, a lot which it is necessary to replace if we are to understand what is being said.

The sentences we use can be incomplete, can fail to make explicit everything that is being said, in various ways. One obvious type of incompleteness is found in proper names. You can overhear someone talking of Mary or Graham and while you can understand what she is saying you probably won't know exactly which Mary or Graham she is talking about. The audience has to help in specifying the person referred to since the name itself doesn't tell you exactly. It is true that a few names become common currency - if I wrote about Ronald Reagan you would probably know which man I was referring to, though there are almost certainly many men with that name - but in general, names don't come so closely linked to a particular person, place or thing. Again, words like the personal pronouns or other "deictics" ("this", "yesterday", "here", etc.: pointing words) do not explicitly tell you to what they are referring; that has to be supplied from knowing the context in which the sentence has been used. But these sorts of incompleteness are so pervasive, and in a way so obvious, that I do not want to make anything of them.

The kind of incompleteness I am interested in occurs when we leave out part of a sentence which, as far as the meaning goes, rightfully belongs there. It is the phenomenon that grammarians have sometimes called "ellipsis". Grammatically, there is nothing wrong with the sentence, but it lacks a component which is essential for it to mean anything definite.

One sort of ellipsis which may display what I am getting at is very common in ordinary talk of left and right. If someone asked the way to the Cayman Islands by boat, you might tell him to turn left or veer to the left after a 100 miles. Someone sailing from Trinidad northwards has indeed to go the left, or to port, if you will, to meet with the Cayman Islands; but someone sailing from Miami will have to turn right. Again the case is obvious, and the omission is obvious - we leave out which frame of reference we are using. If you want a class to turn to your left you may have to tell them to turn right: there is no great mystery about this example of ellipsis, but what is important is that you see that we don't know what is being said, and so we don't know how to start evaluating it, when someone says "A is to the left of B" until we have supplied the appropriate frame of reference. In ordinary life we almost never bother to do this explicitly because it would be too pedantic; but we must do it implicitly.

What I want to claim is that there are many other such omissions in our regular grammatical language which are in a way more serious than that in the case of right and left. They are more serious because they are less obvious, less noticeable, but they leave what is said equally indeterminate. One reason they are not noticeable is the fact that ellipsis is not ungrammatical - "California is to the left of Michigan" is a perfectly acceptable sentence of English, it is not grammatically unfinished like "Mary is bigger than"; but in saying that it is elliptical I am saying its meaning is equally unfinished. But the examples I think important are less noticeable also, because they are rather more complicated than left and right, which are fairly simple and obviously cry out for a frame of reference. My examples are important because, whereas a frame of reference is not usually anything to fight over, what gets left out in these cases is very often vitally important, so ellipsis allows crucial issues to be overlooked or remain in a confused state.

But let us look at some examples. Consider a very famous pronouncement: "All men are created equal." It sounds very fine, but what does it mean? How can we start investigating whether it is true? If we look at new-born babies it is obvious that they are unequal in various ways - in weight, height, number of chromosomes, etc.; and if we look at adults the differences become even more striking. But are these inequalities relevant to our claim or not? How can we decide? What I have said in general about ellipsis should suggest that we cannot decide without slipping in some more content because the claim is unfinished as it stands. If we want to talk about equality or inequality, we must always supply the respects in which we are interested. Two bags of oranges may be equal in the number of oranges, or equal in the amount of vitamin C contained in the oranges, or in any number of other respects; but there is no issue of their being equal or unequal, just so. Of course, in many contexts, we can easily supply the respects that are in question - if a carpenter asks you whether two pieces of wood are equal it is a fair bet that he means equal in length, not equal in density or in smoothness. But, just as with left and right, we do have to supply the missing part of the meaning, and we can make mistakes even in simple cases (and perhaps young children who lack our particular interests in things are prone to make more such mistakes). And the same goes for the exalted sort of claim we started with - we have to supply the respects that are in question for the claim to be saying anything definite; unfortunately we too often slur over doing this and imagine that there is still something worth arguing about.

Another notion that frequently occurs together with equality, especially in political and educational debate, is that of opportunity. Many people say they want equal opportunities. At least they have begun to fill in the gap for "equal"; we know they don't want equal height, equal food intake, equal wealth; they want equal opportunities. But unfortunately in filling one gap they have brought another with them. What is an opportunity? It is surely an opportunity to do something, or to be something. UWIDITE has given you an opportunity to get a certificate; doing this course may also have given you opportunities to meet people, to get out of the house for a few hours, and so on. But it probably won't give you an opportunity to fly to the moon or to become Prime Minister. Has this course helped to equalize opportunities? How can we begin to find out until we know which opportunities, and whose opportunities. A bigger question that is often asked concerns whether it is a good thing to pursue equal opportunities. Again we can hardly start to think seriously about this question until we decide which opportunities we are talking about. One man's opportunities may be another man's degradation. Waving banners for equal opportunities, without stopping to think which opportunities and what their consequences might be, is logically absurd and may be factually suicidal; but it happens all the time.

Let us take one more example. Many teachers fuss about children's needs. They think they should cater to little Johnny's needs. But what does little Johnny need? How do we set about telling? Well, surely, we have to start by deciding what he needs things for (or what he needs things to be). If he is to read Macbeth, he probably needs a copy of the text; or he may first need to be able to read. If we are concerned about his health, he may need vaccinations, extra food, fewer household chores, or whatever. If we want him under sedation, he may need an extra dose of some drug. He has an unlimited number of needs, depending on the unlimited number of things someone may be aiming at for him. There is no sense in which we can cater for his needs, full stop. We only have something definite when we fill in what they are needed for (so formally we can say that "A needs X" is elliptical for "A needs X for/to Y"). But of course very often we don't bother to specify these, and so hope to get by with whatever vague consensus there might be, or when we do, they are hopelessly unspecific - Johnny needs to actualize himself (! what informative claims could that give rise to?). So we utter a lot of words, and people can easily murmur assent, but the thoughts have very little determinate content. It is almost as if the language prevents us from facing the real and contentious issues.

I have briefly tried to illustrate the claim I made earlier, that a great deal of our language is elliptical, it leaves unsaid things which have to be understood in order to convey a determinate meaning. If these things are not inserted then really we are not saying anything, however uplifting the appearances. If you have been persuaded by the examples, you may still be wondering how to diagnose and cure this corruption of thought. I have said that elliptical language is perfectly grammatical, so you can't appeal directly to your linguistic intuitions. What can you do? What I have been trying to suggest as a potent test is to ask yourself the question, "What more do I need to know to test this claim for truth or falsehood?" This is the question we used when looking at left and right. If you tell me that I'll find the book on the left end of the shelf, I have to fill in the frame of reference you are using to know exactly which part of the shelf to look. Does Johnny need another dose of rat poison? If we are trying to kill him, and the first dose didn't work, then yes; if we are trying to build up his strength, then no. To work out the answer, you have to fill in what the poison might be needed for.

When you think you have found a case of ellipsis you should usually find that you can sensibly and grammatically restore the missing part. We can say both that Johnny needs another dose of rat poison and, more explicitly, that Johnny needs another dose of rat poison to finish him off. The original idea of ellipsis in grammar was of part of a sentence that could optionally be left out, and so could be put back again without too much disruption or unnaturalness. My use of the term is slightly different, but I would still think that we can restore the elided elements without grammatical impropriety, though the explicit versions may well sound pedantic and trivially obvious in some cases.

It should be noted that ellipsis is different from lack of specificity. No description is complete in every detail. If I say at a particular time and place that Ronald Reagan is walking, I don't tell you with what velocity with respect to the earth he is walking, nor with what gait, nor with what intentions, nor an unlimited number of other facts about the situation. But none of these are cases of ellipsis because you don't need to know any of them in order to establish whether or not Reagan is walking. No doubt there are odd "borderline" cases of locomotion in which we might not be clear whether what he is doing is properly described as walking, but in general, we know what counts and what doesn't, and the simple expression "is walking" gives us everything we normally need, so it is not elliptical, however much it leaves unspecified. But if someone says Reagan needs to walk more, that claim is elliptical because we can't begin to establish whether it is true before we fill in somehow why he needs to walk: for his health, it is perhaps true; for justice in Central America, perhaps not.

There is a symptom of ellipsis that is worth noting. In general, as we have seen already, we cannot accept a contradiction. But sometimes we can find ourselves in what looks like such a predicament. Imagine a quarrel between husband and wife. The husband tells you, "Mary doesn't need a new dress; the wardrobe's full of her dresses"; Mary tells you "I do need a new dress; everything's out of style". On the face of it, you might find yourself agreeing with both of them: Mary needs a new dress, but then again she doesn't need one. How can this be? Although it looks like a contradiction, it cannot really be a contradiction. Grammatically it may be a contradiction, but as we know already, these sentences are elliptical and so more is being said than appears on the grammatical surface, and this extra content allows them to be true together. As we have seen, Mary's truth is that she needs a new dress to keep up with fashion; her husband's truth is that she doesn't need a new dress to be respectable. These two claims are not in conflict; obviously, they can both be true. (Notice that we have here one simple case of how to deal with so-called relative truths.) So if you seem inclined to accept both sides of what looks like a contradiction, it is a fair bet that the claims are elliptical, that more is really going on; you then have to go back to the test I gave above to uncover what this extra meaning is.

(For the sake of completeness, I should mention another use of contradiction in English. Sometimes people say things like "It is, and it isn't" to indicate borderline cases. If someone asks whether it is raining and you look out and see the odd drop of rain, you may say that it is, and it isn't. I don't know whether other languages use a grammatical contradiction in this way. For our purposes, it is enough to note that the content of what is being said is not a contradiction, but that it is not a symptom of ellipsis either.)

We have come upon at least two reasons for taking ellipsis seriously. One reason appeared just now in looking at the quarrel over Mary's new dress. People can easily start shouting "I need it. No, you don't" at each other, with increasing vehemence, and without any clue as to how to resolve the conflict, because that is what the disagreement appears to be about. But, once we explicitly fill in the gaps, we can see that the real disagreement is about the importance of keeping up with fashion. For Mary, that is an important goal. For her husband, it may not seem at all important, indeed it may never have occurred to him. If there is any chance of a rational resolution of their dispute, it must be by focussing on their priorities, and by the two antagonists coming to agree on a set of priorities. This may not be easy, but at least it keeps the discussion moving and focussed on genuine disagreements, instead of getting stuck in the rut of the original elliptical counter-claims.

The second reason cropped up in the examples of equality and opportunity, and when we first looked at needs. To make anything of talk about these topics, people have to put something in the gaps, but the way the language works allows them not to have to consider exactly and explicitly what to put there. They can forget about the problem altogether, and say virtually nothing. Or, perhaps more dangerously, they can slip in their own preconceptions and hope that other people will swallow what they say without themselves raising any awkward questions. In this way, the important questions are answered by default. What should schools be trying to do for Johnny? That is a large and controversial question; it is one that people ought not to leave unexamined. But once educators start talking about catering for Johnny's needs, they are usually assuming answers to the big questions, but neither stating nor defending these answers. They slip in their largely unexamined aims of education, when other people - Johnny or his parents, say - might have quite different aims, and they keep the discussion away from these contentious issues by talking only of the means to achieve something rather than the something we might be trying to achieve.

EXERCISE H

1. How would you make the point of the following passage in terms of ellipsis?

It is illogical to reason thus: "I am richer than you, so I am better than you. I am more eloquent than you, so I am better than you." But these reasonings are logical: "I am richer than you, so my wealth is greater than yours. I am more eloquent than you, so my speech is better than yours." [36] [Answer.]

2. Suggest an analysis of "A caused B" that incorporates the following point:

"What caused this man's skin cancer?" may mean "Why did this man develop skin cancer now when he did not develop it before?".... But the same question may mean "Why did this man develop skin cancer, whereas other men who were also exposed to radiation did not?" [37] [Answer.]

3. Are claims like "Johnny is older than Mary" or "Abeola is married" elliptical? [Answer.]

12. VERBS AND ABSTRACT NOUNS

In talking about truth I made a point that I want to take up and elaborate a little in the light of what we have seen about ellipsis. I take for granted that you grasp the differences between verb phrases and nouns in grammar. The basic point I made earlier was that when there is a pair of related words, one of them part of a verb phrase, or the whole of it, and the other a noun, usually a sort of abstract noun, it is usually easier to get to grips with the verb phrase form; one is less likely to drift into nonsense. A few examples of such pairs are: is true/truth; is triangular/triangularity; is just/justice; thinks/thought; perceives/perception; needs/need. (Language is an incredibly complex structure, and we are only scratching the surface; there are often other nouns related to such verb phrases, e.g. triangle; perceiving; etc., which I shall not be dealing with.)

The basic point rests on the claim that the fundamental sense of these items is found in the verb phrase forms. The noun forms are linguistic conveniences, which do, however, encourage the mistaken belief that there are some mysterious things (truth, triangularity or justice) existing apart from the various actual situations in the world that are just, or involve triangular objects. There is, I think, a genuine sense in which the nouns are abstract; they leave out of consideration part of the various situations to which the verb phrases truly apply. And it is in leaving things out that there arises the question of ellipsis.

Let us start with a familiar example. I have claimed already that when someone needs something it is always for something else, or to be something else. We could sum up briefly by saying that A needs X is an elliptical way of saying that A needs X for/to be Y. When we use the verb form we often leave out the Y, but we cannot grammatically leave out the X. But now when we switch to the noun form (which with typical English nonchalance is phonetically the same) and talk of A's needs, we usually omit both the X and the Y. So if talk of what A needs encourages vagueness and confusion, talk of A's needs is likely to be worse. We can still fill in the gaps, but it is now distinctly uncomfortable: "Johnny's need for an extra dose of rat poison to finish him off" is hardly a smooth noun phrase. One important consequence of this greater laxity is that talk of a person's needs is ambiguous: when we talk of Johnny's needs we may be primarily thinking of items that belong in the X slot, or we may be thinking of items that belong in the Y slot, or of course a confused medley of both. Those of you who have come across Maslow's "hierarchy of needs" would be well advised to go through it to see which slot he is dealing with on any particular occasion.

This sort of ambiguity is important because it occurs in one way or another in many of the abstract nouns I have mentioned. I would not claim that all such ambiguities can be traced back to the workings of ellipsis, but I think it important to note that there are ambiguities in the usage of these nouns. Thus "John's perception" may refer to what John perceives or to John's perceiving of it (cf. the more extreme case of "John's vision" - is it what he wrote about in the Book of Revelations, or is it something to be tested by an ophthalmologist?); "truth" sometimes points towards the various true propositions there are, other times to whatever is involved in something's being true. It must be admitted that when we take careful note of English grammar some of these ambiguities are dissolved - "Jane knows three truths" cannot (or perhaps can hardly) refer to the abstract relation, but rather to items known - but too often we don't pay that sort of attention, and anyway it won't always help us out. I should also acknowledge that "what John perceives" may not be as straightforward as it seems, but this is not a course in the philosophy of perception.

EXERCISE I

In the light of our previous studies and the above section, what would you say about the following remark made in the course of criticizing Rousseau's theory that children will develop automatically in the right direction:

If the underlying theory or assumption is false, then of course this deduction must also be false. [38] [Answer.]

13. WHAT DOES IT MEAN?

Our overall aim is to learn better how to deal with what a person says, in particular with the arguments and reasons she offers for a belief or proposal. We have looked at the structure of arguments, and we have begun to see some of the things that can go wrong with the claims we make, in particular how we try to evade falsification and how we leave vital things unspecified. But all along I have been assuming that we have a fair grasp of at least those bits that are being explicitly stated. This is not, however, always the case; you may often have wondered what I have been getting at, and I'm sure there are other cases in which you have wondered what exactly is being said, what the words actually mean.

There are different reasons for such incomprehension. Reading this text you may have come across words you didn't know - "ellipsis", for example. When you read psychology you may find people talking gaily of operant conditioning, or an Oedipus complex, and you may not know what they are getting at. In these kinds of case, it should not be too hard to find out. The words or phrases are technical terms in some subject, and you just have to read the relevant text-book, and go further into the subject, to find that you know what the words mean. In many cases, in fact, you have only to read on a little further beyond a word that you don't know to get a pretty clear idea of what it means. But there's no harm in looking in a dictionary.

Words such as these technical terms are simply unfamiliar; once we get used to them, they bring to mind something fairly definite (though notice that it need not be any kind of mental picture). But there are other words that may be very familiar, but which conjure up merely an indeterminate blur. I have already said that ellipsis can contribute to this slurring of meaning: we can talk passionately about equality without any clear idea of what we are talking about. But it is not just ellipsis that causes this indeterminacy. There are whole areas of vocabulary that seem virtually empty of meaning - "development", "self-realization", "creative", "mature", "optimally rounded person",.... In some of them one can discern a few husks of sense, but for the most part they carry only an aura, either of something good and wise, or of something outmoded and wicked, but never precise or definite. Such words might make good preaching, but they are virtually useless for serious thought.

One strategy people often adopt when faced with terms like these is to ask for a definition: "Define your terms!" Asking for a definition may be a good strategy in the first sort of case we looked at, where a word is unfamiliar, technical, foreign, etc. But in the present sort of case, it is a hopeless move. Why? Definitions are of words; they tell you how people use words, what they are trying to convey by using them. Now if I am right in suspecting that in much debate "education" or "creativity" convey practically nothing, then a good definition of these words in such contexts will simply spell out the fact that they convey practically nothing. A good definition of an empty or indeterminate term will simply replace it by a string that is equally empty or indeterminate. And you're no better off than you were before.

In fact, however, it is worth noting that in ordinary life we very rarely ask for definitions, or offer them. I didn't define my use of "ellipsis", I tried to indicate the sort of thing the word picked out. And in many other cases what we need to know is not so much a dictionary definition, but rather an account of how a word works in the language, and in particular, of its logical connections with other parts of the language. The word "large" means pretty much the same as the word "big", but that is of little use to a child learning the word - more likely he needs to know that if Fido is a large dog, and all dogs are animals, it doesn't follow that Fido is a large animal; "large" works differently from an adjective like "hairy" or "sick". Here again in fact we have a case of ellipsis, a large dog is (approximately) a dog larger than the average dog; there is an element of comparison in the meaning of "large" that is absent from most adjectives. These facts about logical connections can often be more helpful in clarifying what is meant by some piece of language than the regular dictionary definition.

If definitions won't help us in such cases of empty verbiage, what will? We have already met the question I would recommend: "How would we go about finding out whether what is said is true or false?" This question forces us to specify what has to be the case for a claim to be true, or what has to be the case for the claim to be false. If you can't answer either of these questions, you can be pretty sure that you don't understand what is being said. That may be because the words are obscure or the thought is very complex, but if the words are normal and you can unravel the grammar then I think you should conclude that really nothing is being said. It may sound good, but if it lacks determinate meaning it is not worth polluting your head with.

While my main aim is to disabuse you of the myth that definitions must always be ritually intoned at the beginning of an intellectual enquiry, I should acknowledge that definitions can occasionally be of some use. Several very unspecific words are used to cover significantly different things (e.g. "democracy"). A proper definition is not much use here, as I have said: but what people are often wanting is not a proper definition but rather an account of one particular conception of some of the things that are covered by the term. So someone might well start a discussion of democracy by telling us what particular conception (perhaps one stressing representative institutions, or again one stressing universal participation in ruling) she is using. She will tend to do this by talking of her "definition" of democracy. This way of talking is misleading, but what she is doing is perfectly sensible.

In some other cases, a definition serves to sum up the results of theoretical and empirical enquiry. To take a very crude example, a definition of "whale" saying that what it refers to is not a fish but a mammal might do this. If you want to sum up or "freeze" the present state of enquiry in a subject, offering these sorts of definition is one way of doing it. The danger is that people too often imagine that producing a definition is a way of saving themselves the trouble and tentativeness of actual enquiry. But you cannot avoid these things. The definition of "whale" is no more secure than our present biology; it isn't anything to make a fetish of.

While I have conceded that in these ways and others, some definitions, or what are offered as such, can be of use, the point to reiterate is that even here there are dangers because people confuse definitions, which are of words, with claims about the things defined. So people imagine they are defining whales, not the word "whale". But they aren't.

EXERCISE J

1. My colleagues have approved the following statement:

The seven grades are intended to represent the following standards of performance ... B. An intermediate passing standard, slightly above the average expected. A possible candidate for further studies, but not specialising in this subject area. C. An intermediate passing standard, slightly below the average expected. Not recommended for further studies...

How much can one learn from such information about what the standards are? [Answer.]

2. How determinate are the aims attributed to the "intellectual growth" approach to the curriculum here:

This approach to the curriculum begins with the assumption that intellectual development is not to aim at the objective of conveying the past to the present. The past may be interesting, and an appreciation for past accomplishment and cultural ways may have some elements of worth, but they are always subordinate to the principal aim of the curriculum. This principal aim is to open the minds of students in such a way that persons are made capable of meeting reality in all its various forms. The school's curriculum should equip students for intellectual growth and maturity. [39] [Answer.]

14. FACTS AND VALUES

In this and the following sections we shall touch upon some very large and complex issues. But as before, some of the most fundamental distinctions are among the most useful, so you should not be put off by the fact that this is only a beginning.

In the previous sections I have often contrasted describing an argument with evaluating it. This is one instance of what seems to me a vitally important distinction, the distinction between facts and values, between factual issues and evaluative issues. But it is not perhaps an easy distinction to expound, in part because it is not presented to us explicitly. A lot of our thought and language intertwines facts and values and mixes them up in various ways; but to attempt to distinguish between the two can help us see what is really going on and suggest appropriate responses to what people say. I have contrasted describing with evaluating, though one might prefer to say that it is a matter of two different sorts of description; after all, calling an argument "good" or "valid" is describing it, even if it is describing it in a different way from saying that it has three premises and a conclusion. The point may be clearer if we attend to a contrast between what is and what should be, or ought to be, or even is to be (but what the grammarians call "gerundive" rather than simply future). Issues of fact are simply issues of what is the case, whereas issues of value involve also the idea that something should be, that something ought to be, that something is to be emulated, say, or avoided. Evaluative issues in general involve a reference, explicit or implicit, to an agent or judge, over and above any factual claims that might be involved.

Let us take the claim that democratic decision-making wastes time. There is a factual component to this claim: presumably something like the fact that democratic decision-making is more time-consuming than alternative methods. But this factual element is not the only, or indeed the most important, component of the claim (though if this factual claim is in error the whole claim collapses). That is rather the evaluation that we should not so use the organization's time. In saying that something wastes time, we are not simply reporting a fact, but are also saying this fact ought not to be accepted or tolerated. Language leaves the judges unspecified. It perhaps insinuates that any impartial judge would come to such a verdict about democratic methods, and so, of course, should we.

Let us look at another example. Consider the argument: "No communists are Christians, so all non-Christians are communists." Firstly, if I say it has one premise, "No communists are Christians", I am simply describing it; I am making a simple factual claim, which in this case is true. Secondly, if I say the argument is worthless, I am making an almost purely evaluative claim; I hardly tell you anything about the argument except it is such that you ought not to be swayed by it, and it is not the sort of argument to use with a good conscience. Now thirdly, what am I doing if I say the argument is invalid? If you recall what was said earlier, you might well reply, correctly, that I am doing two things: I am describing the argument factually (it is such that its premise could be true while its conclusion was false), but also, intertwined in some way with this factual issue, I am evaluating it negatively - I am conveying much the same as I was with the less restrained claim that it is worthless. Philosophers argue a lot about the best analysis of such cases, but for our purposes it is enough to note that there are many words like "valid" and "invalid" which exemplify a tying together of simple factual issues with evaluative ones.

Since ordinary language treats values as just another kind of fact, it is particularly easy for it to roll the two together and not notice what is happening. I think that facts and values are very different in at least one important respect, but without worrying about that, it is important that you be able to distinguish the factual from the evaluative aspects of an issue. Given the intertwining of the aspects in ordinary language, it is often necessary to take particular care in the choice of the words you use to formulate a problem. Suppose you wonder whether it is right for teachers to flog small children. You may create problems for yourself if you start by asking whether teachers should punish children, since the word "punish" brings with it an aura of some act being permissible, even if unwelcome, and so may tend to prejudge the issue. It might be better to start by talking about flogging; if you think that somehow you can justify it, then you might be able to label it punishment. So sometimes what may seem like merely verbal disputes are of some importance, because of the differing evaluations linked to what, as far as descriptive content goes, may be similar words. One man's liberty may be another man's licence, and while those are the terms they use they are not likely to make much progress towards resolving their difference of opinion concerning what is permissible or to be encouraged. Similarly, calling abortion "murder" may be a way of summing up an enquiry into its moral status; it is hardly a good way of beginning one.

I have claimed that we can in principle disentangle factual from evaluative issues, even when, as often, they are knit together in our language and thought. It may be worth noting that we can also make simple factual claims about what people value. If I say John admires Picasso, or almost everyone admires Rembrandt, I am making straightforwardly factual claims; they are claims about what some people value, but they are not themselves evaluative, they don't tell you how you should think yourself. The difference comes out in the contrast between "Rembrandt is much admired" (factual claim) and "Rembrandt is an admirable painter" (evaluative claim, which suggests the audience should join in the impartial admiration). It may also be worth noting that when I say something is a factual claim, I am not saying it is a true factual claim, only that it is the sort of claim that is simply either true or false; and when I say something is evaluative, I am not saying it is acceptable.

15. SOME USES OF ARGUMENT

Our central topic has been the giving of reasons for a conclusion. When people offer a reason for a conclusion they may be doing various things, they may be performing various different "speech-acts", as philosophers sometimes say. They may be making an excuse, or making a recommendation, or putting someone down, or any number of other things. All such speech-acts are performed through using language, which is why they have that name (though in fact some such actions can be performed without actually using language).

In this section I want us to distinguish two large classes of speech-act that may be involved in putting forward an argument, in giving reasons for a conclusion. The two speech-acts go along with the distinction explained in the previous section between facts and values. One, the factual speech-act, I shall call "explaining"; the other, the evaluative, I shall call "justifying". One reason for distinguishing them is that in some contexts either might be offered in answer to a question "Why?". Thus if someone asked "Why do teachers flog pupils?" you might answer him by explaining what it is about teachers, pupils, the school system, the cultural background of the people involved, etc. that brings it about that teachers often flog their pupils - this would be to take the question as asking for a factual explanation of the occurrence. On the other hand, you might try to offer some account that seeks to show why it is acceptable for teachers so to behave, you might invoke some principles and some facts of the situation to justify the occurrence. Or again, if you are wondering why you have to do a philosophy of education course, you may be wanting a factual, explanatory answer (perhaps by reference to the history of the course and the motivations of the people organizing and now teaching the course), or more likely you may be wanting an evaluative answer, a justification, an account of why you should be asked to take such a recondite subject. Not now why you do, but why you ought to.

If you think about possible answers to these questions along the two lines I have indicated I hope you will see that however close they may get, there is still a different point to the two types of answer. In the case of this course, my point may be to explain to you the intricate history behind the present requirement that you take the course; this may involve mention of what once upon a time seemed good reasons for including the course to the people devising it. But the point of mentioning these reasons is that they contributed to the historical development, not that they are good reasons for you to do the course. On the other hand, if I am trying to justify the requirement, then the point of my remarks would be to provide good reasons for supporting the requirement, to show that it is sensible and profitable. You will also see that in many cases quite different issues will arise for these two different speech-acts, quite different sorts of claim may be needed for a successful explanation or justification. Grasping the distinction, you may be able to see what is lacking in any particular example.

It is not my intention to say anything about the structure of explanations and justifications except to focus on one point that has occurred in passing already. Justifications are built on factual claims, on a factual scaffolding, so evaluating a justification will normally require that we consider both the truth or falsity of the factual claims involved and the acceptability of the principles invoked. For example, if we are thinking of rejecting something because it wastes time, we should at least establish that it does indeed take more time than our other options, besides considering the value we are placing on saving time. Or let us suppose someone offers a justification of capital punishment in terms of deterrence. This will involve at least two crucial items: a factual claim that capital punishment does deter people from committing certain crimes; and an evaluative principle that it is acceptable to kill someone for committing these crimes in order to deter others. The basic point is that, if this is to work as a justification, these two claims have to be true or acceptable. If capital punishment does not deter these crimes, then the "justification" is useless; equally, if we think it unjustified to hang people to deter others from stealing sheep, then any deterrence is irrelevant, we cannot use a principle we don't accept to justify a policy.

It is staggering how often these simple points are overlooked - people take a statement of intention as a reliable account of what actually happens or what will happen (cf. the lists of objectives offered to defend curricula without any investigation of the predictions involved), or they simply don't think to ask whether there is any evidence at all to support the factual claims essential for a particular justification.

16. REFLECTIVE EQUILIBRIA

In this section I want to mention a vast topic, one that we cannot even begin to discuss. I have said that sometimes we use argument, we give reasons, either to explain something or to justify something (or at least to try to do these things). The issue I want to broach is the relation between these somethings explained or justified and the explanations or justifications of them.

In some cases, though by no means all, the things explained or justified are what we can call "singular" items; they are expressed in statements about a particular thing, or person, or place, etc. When this is so, we are often fairly sure of the truth of such statements, but not so sure of the claims we invoke to explain or justify. These claims will often explicitly involve what we call "generalizations" (claims about all or most or some of a certain sort), and it is arguable that they will always implicitly involve some sort of generalization. But it is by no means so easy to be sure of the truth of these generalizations. In such circumstances, then, it is very tempting to think that what we must do is base our beliefs upon the singular data, so that we see our discoveries as being built upon a lot of data collection. The generalizations we invoke to explain or justify are supported somehow by the kind of singular item we may later use these generalizations to explain or justify. Thus to take an example, we may offer to explain why Ade is not a Christian by claiming that Ade is a communist and no communists are Christians. Now this explanation invokes a universal generalization, "No communists are Christians," and we may suppose our knowledge of this claim is based on lots of examples of Ade's sort, where we have found people who are communists but not Christians, and vice versa. Again, if we invoke a principle that it is wrong to tell lies in justifying someone's action, it is tempting to see that principle as based on many cases where we have judged that a particular person's telling of a lie has been wrong.

One point I want to make about explanation and justification is that despite the big differences between facts and values, there is basically the same logical structure in arguments that explain as in arguments that justify. Again, in fact, we do not have to address the question whether there is a big difference or not between facts and values. The second point, however, is that the picture suggested in the preceding paragraph is misleading. None of our claims (except perhaps some parts of logic and mathematics) are absolutely secure; possibilities of error are pervasive. Since this is so (unfortunately you will have to take it on trust in this course) what we do in fact is to balance the things to be explained or justified against the explanations or justifications and try to come up with the best combination as far as we can judge at the time. To borrow a phrase from an eminent moral philosopher, John Rawls, we seek to find a "reflective equilibrium" between explanations and things to be explained, between justifications and things to be justified. We try to do the best we can, with what we have at hand; but tomorrow is another day, and may reveal something that upsets our best compromise. When you learn a subject it may seem fixed and immutable, but when you really know it, you see that it isn't so; and when you reflect on its history you see also that it hasn't been so; there is absolutely no reason to suppose that it has suddenly stopped.

Let us see what this abstract account comes to in real life. Take a case of explanation first. I have said we often think of the theories we use to explain facts as somehow based on similar facts, and as much more vulnerable than those facts. Sometimes this is a reasonable view, but not always. Sometimes our theories, the generalizations we use to explain data, can correct our beliefs about the data. Newton once told a notable astronomer he couldn't have seen what he claimed to have seen, because Newton's explanation didn't allow it. We aim at the most adequate combination of beliefs about general theory and beliefs about particular happenings; neither side always has the advantage.

Similarly, in evaluative cases, neither side always has the advantage. Let us take a case from moral reasoning. We can often distinguish general principles, such as that it is wrong to tell a lie, from singular judgments that John should not have told a lie on a particular occasion. In saying that neither side should always win, I am saying that responsible moral thought seeks a balance between the principles and the judgments - in some cases the principles will dictate a judgment, and may suggest you reverse the judgment you initially came to; but in other cases, judgments may reasonably qualify or totally reject principles. If, after due consideration, it seems to you quite right for John to have lied on that occasion, you are better off qualifying your principle rather than doggedly sticking to it and trying to deceive yourself.

I have raised, and assumed answers to very large questions here. Much more needs to be said, and you should not take these few remarks as a last word on anything; but I hope you can see the fruitfulness and practical importance of this approach to the questions of how our beliefs should develop in both factual and evaluative matters.

17. STANDARDS

In the last few sections I have been pointing to similarities and contrasts between factual and evaluative matters. In this final section I wish to look at one important type of case in which factual and evaluative issues get closely intertwined, but which becomes clearer, so I believe, when they are distinguished. It is also a matter of some importance for you as students and teachers.

Some of the time we grade things or people. Much of the time we classify things or people by appeal to standards or criteria which are very similar to those used in grading. We distinguish good knives or cars or ... from bad ones by reference to the aims we have for using them. In all such cases there is a shared element in the judgments we make which is simply factual. We may grade an egg as extra-large; or claim a certain action was courageous; or judge brand X is better than brand Y for cooking pots. In all such cases one of the things we are doing is putting something (an egg, or an action, or various cooking pots) up against a set of standards or criteria and noting how far it is from them. If we know how large an egg has to be to be extra-large, then it is a simple matter of fact whether this egg is extra-large or not; and similarly with the other claims. We compare something with the criteria, and this is a factual issue: does it meet the criteria, does it fit within the given boundaries?

It must be admitted that this sort of question is sometimes very precise, sometimes very vague. No doubt there are cases where the question becomes too vague; but the general point remains that where there are criteria that can be used (this does not mean that they have to be explicitly stated) then the comparison is a factual issue.

There may also be cases where people disagree about what the criteria should be. Here again, though, the basic point remains that the comparisons are factual; it is just that there is a difference of opinion about which comparisons should be made.

But while all these various judgments include a strictly factual component, there is also an evaluative issue floating around as well. A courageous action is one to emulate; an extra-large egg is one to prefer, other things being equal, to a medium-size one. One way to make this point is to say that in addition to the factual comparison of the thing in question with the criteria, there is also the evaluative endorsement of those criteria as appropriate for certain choices. Using the egg grading scheme is tantamount to recommending we choose eggs by their size. Comparing an action with one's implicit paradigm of courage is also to endorse that paradigm as worthy of emulation.

It is, I think, one of the reasons some people think grading pupils or their work is unfair that in such cases it is very unclear what further evaluation is being made. It may be straightforwardly true that Harry got 3 out of 10 (and thus an F) while Jane got 9 out of 10 (and thus an A) in a maths test; but what evaluations can reasonably be based on this? Yet the language of grading brings one along. What one has too often is a grading that has lost its point; comparing someone's performance with a criterion may be useful as a guide to what he needs next, but grading in school does not usually function in this way. But then it is as if instead of judging actions by reference to courage or pusillanimity, we judge them by how many minutes they took. Our commendations or condemnations would be pointless.

When standards do have a point, however, we find that it is not often very easy to imagine ourselves using a quite different set of standards. I have said already that calling an argument valid is evaluating it (by reference to a standard) positively. But we can see why we should be interested in applying this standard to argument and also why most other standards would be inappropriate. So whatever I might want to say about the difference between ineluctable facts and chosen values, I would not claim that we really have much choice in this area. Values are not, in general, items we can choose to be any way we happen to like; they are not in general arbitrary. But it is true some values are arbitrary and some values could be changed, or are in fact already held in different forms within a community. All I am saying is that values as such are not always so alterable, and reflection on some cases of the invocation of standards or criteria may provide examples where one type of value is clearly preferable to another.

EXERCISE K

1. In the light of the preceding remarks, consider whether the following claims are elliptical:

"John is more intelligent than George"; "George's essay is a failure; but John's should get a B." [Answer.]

2. What, if any, is the real difference between objective tests and subjective methods of assessment such as essays? [Answer.]

3. In terms of facts, values, and standards, how would you describe what is going on when a Spanish teacher changes a beginner's "Soy miedo" to "Tengo miedo"? [Answer.]

Suggested Answers

EXERCISE H

EXERCISE I

EXERCISE J

EXERCISE K