APPENDIX 1: PHILOSOPHICAL OR CONCEPTUAL ANALYSIS
Many of you who have reached thus far in this book on the analysis of arguments will also be studying some philosophy, and you will almost certainly have come across philosophers offering or asking for philosophical analyses of concepts or ideas or proposals or claims. They often call this activity conceptual analysis and talk about philosophy as a second order discipline which studies, and in particular analyses, the concepts of other disciplines. I do not intend to discuss how accurate or fruitful such a view of philosophy might be, but I do want to try to suggest how some of the things we have been looking at relate to philosophical analysis as I understand it, and to give a sketch of some of the main things that are subsumed under that label. It is only fair to admit that what I shall be saying would not be endorsed by all philosophers, but that tends to be the case with most things philosophers find to talk about. It is still perhaps something of a scandal that philosophers cannot even agree on what many of them think is the central method in their subject, the central activity upon which they are engaged.
The first and preliminary point I want to make is that I have avoided talking about concepts so far, and I intend to continue in this abstention. It is not that I have anything against concepts as such, but I think what matters to people who ask about concepts in philosophy or similar discussions can usually be handled by talking about the ways people use language.
Talking about ways of using language is not simple, but it is in general nearer to hard data than whatever concepts might be. I am not claiming that using language is all that having concepts amounts to, but rather that most problems stated in terms of concepts can be more easily handled by switching to corresponding questions about language use. So if someone asked whether little Johnny has the concept "uncle" I suggest we translate his question into one about how little Johnny uses the word "uncle" (assuming Johnny speaks English) - does he use it only of male persons? If he knows Sam is Jane's mother's brother, will he agree that Sam is Jane's uncle? Can he make claims about how uncles fit into the kinship structure? etc., etc. Grasping a concept is a matter of degree; we may find Johnny can do some of the things I've mentioned, but not others. But whatever having a concept may be, at the moment there is little more we can do than investigate claims about concepts via language use. One important alternative is to investigate the ability to discriminate - if a chicken can distinguish between red triangles and red squares we might want to say it has a conceptual distinction between triangular shapes and square ones. But in most of the cases dealt with in philosophy and education we are concerned with language-users and with discriminations enshrined in the language.
A second preliminary point is that people often think about language and concepts in a highly privatized way. There is a great deal of discussion in philosophy about whether a truly private language is even possible, but we can avoid that contention by noting the simple fact that all the languages we actually speak are as public as can be. They are not our inventions, they are picked up in interaction with other humans who are already speaking them and who can only guide the learner by reference to what is open to mutual inspection. However private a pain might be, English speakers can only learn how to talk about pains and can only be guided by their elders in doing so by reference to things everyone can perceive. This simple fact puts a certain limit upon how different our uses of language can be, and so provides a basis for optimism that we can give an account of what part of the language means in general for all speakers; we are not reduced to listing what it means for me, and you, and for the next person. (It is another important corollary of this point that words cannot in general stand for "ideas".)
I have suggested that we see questions about concepts as questions about how bits of language are used. An analysis of the concept of teaching is then to be given by looking at how we use the verb "to teach". But what uses? What are we looking for here? There are, I think, several different things we could be looking for, and perhaps part of the confusion surrounding philosophical analysis arises from a failure to distinguish some of these distinct questions.
Perhaps the most obvious question in a case like this is "what does the word mean?" We want to know exactly and explicitly what we are trying to convey when we say that someone is teaching. As I said in Section 13, there are occasions when this is a worthwhile thing to ask: if we are teaching a foreigner how to speak English; or if we come across an unfamiliar word; or just for its own sake on some occasions. We have to find another way of saying exactly no more and no less than the word "teach" says. We can test our hypotheses by seeing if they work with other uses of the word, since what we are looking for is an account of what the word "teach" contributes to the sentences in which it occurs, not just of what it might be doing in one sentence. So, suppose someone suggests that to teach is to do something with the intention of bringing about learning (so "John is teaching Mary Greek" amounts to "John is doing something with the intention that Mary learns Greek"). That may sound fairly plausible - try it out in some of your own examples.
Now consider the sentence "Last year's hurricane taught me never to be without a large supply of candles." Where is the intentional action there? In such a case there are as usual various alternatives we can take. We could say the word "teach" has two senses, or perhaps we could talk about a literal and metaphorical use. Or we could try for a more general, unified account - perhaps we could drop the intention in favour of simply causing learning to take place. My own inclination is to think that we are better off going for the unified account, even if it is often pretty bare; but there is no royal road to the truth, we have to do the best we can as an overall account of the way the language works.
Discovering different senses that are not marked in the language often looks like a simple evasion (in the terms we met in Section 10), but a contrast between literal and metaphorical usage is likely to be necessary in the overall picture so that solution to our problem might be acceptable. The main point now is not to analyse the concept of teaching, but simply to see how one might go about it and how one can test suggestions that are made.
While we are often concerned with what exactly a term (or a phrase or a whole sentence) means, there are other questions we can ask. Language usage is in many ways remarkably regular: we keep meeting new people and new situations but we find that we agree on how to describe them, on how to talk about them. What do we rely on to do this? Consider the case of recognizing letters. Once a person can read, he can pick out the letter "a" from other letters and he can pick out different tokens of the letter "a". But now just think of what handwritten "a's" look like, and perhaps Gothic "a's", and even Greek alphas, and what a capital "A" looks like. Each of you can group all these shapes together as "a's", and distinguish them from "o's" and "g's" and the other letters. But how do you do it? What are the instructions you would give a computer to do it for you? Again, while philosophers and ordinary people often agonize over free will, we can usually discuss events taking place around us in terms of whether the people involved acted of their own free will or not. We may not always agree, of course, just as some letters are indecipherable; but in general there is a lot of agreement, and our present question is about the basis for such consensus in language use. Following Mackie, we can borrow a word from the psychologists (who got it from the stage) and talk here of observational cues governing our usage.
But the regularities in our language use are not all simply matters of reporting what is happening. Of tremendous importance are the regular moves from one sentence to another that we allow or reject. A fluent speaker of English can take any declarative sentence and negate it, or make one or more questions out of it. There are many such regular grammatical relations between sentences, which govern what we say. Perhaps of more importance for philosophers and for conceptual analysis are the logical relations between sentences - the fact that "A is a large cat" does not entail "A is a large animal" tells us something important about the concept "large". The issues we looked at under "Ellipsis" in Section 11 come in here, since ellipses are ellipses of logical structure, structure which helps govern what we say. These grammatical and logical structures can be called logical cues to draw attention to their parallelism with the observational cues of the previous paragraph.
One thing logical and observational cues have in common is that people often don't know what they are. Simply knowing how to speak the language does not give you any special insight into what the logical or observational cues are which govern your use of language. This is obscured because many people who use a language can begin to give a pretty accurate account of what they intend to convey by their language (though often they cannot even do that) and so it is easy to think that all questions about the meaning or analysis of what we say can be answered simply by being a fluent speaker.
I have talked of our language being governed by observational and logical cues. You shouldn't make too much out of the word "governed". One of the most significant facts about language is its freedom from the constraints of the actual situations in which language users live. We continually talk about people or happenings that are not present; we speculate about possibilities; a lot of the time we talk about things or situations that never have, and never will exist; and we refrain from talking about what is present to our senses. Thought at least seems free. My point was that to understand all this mental creation there must be some links with the rules embodied in what I have been calling observational and logical cues. But the creative freedom that we enjoy in the case of language allows room also for error and confusion. In particular, the grammatical rules of language allow us to say things that are too far from the touchstones provided by observational cues or which overlook too much elided logical structure and which are in consequence all but unintelligible, incoherent, or seriously incomplete. It was these dangers that I pointed to in stressing the need to notice the missing portions of elliptical expressions. Given the values deeply embedded in our talk about argument, mere description of logical cues often shades over into logical criticism. And, as we shall soon see, observational cues can figure in similar philosophical criticism of received notions.
So far we have seen two questions that can be asked of a piece of language. What is it intended to convey? What cues govern its regular use? A third question is: what is going on in the world when people use the language correctly? We have already met one example of an answer to this kind of question in the Section on Facts and Values (Section 14). I claim there that when we call something "murder" we are not simply describing an event, we are also indicating our attitude towards it. The difference between calling something a killing or a murder is largely that the second involves the expression of our values. Now whether this particular claim is right or wrong, we can see that there is a question of what is going on when we use language correctly, and the answer to such a question might be different from the answers given to the first or the second questions already distinguished. Of course, in many cases all three questions may have the same answer, but that doesn't make them any the less distinct questions. Let us follow Mackie in calling an answer to the third kind of question a "factual analysis" of the piece of language (or concept). I would claim that when there is a big difference between what a term is intended to convey and the observational cues upon which its usage is based there is often a philosophical problem regarding the factual analysis: can we say that the facts are as we mean them to be, or are we misleading ourselves because the facts are little more than what the cues offer us.
One of the causes of confusion in a lot of analytical philosophy is the fact that disagreement over a factual analysis is often presented as if it were a disagreement over what terms or phrases mean (i.e. as a disagreement over our first question). It would take us too far to explore this issue, but let me offer a crude example, staying with the concept of murder. One might say that the term "murder" conveys that an action is a killing and is not to be done; that is what the word means. I have already suggested a factual analysis, inspired by the thought that the observational cues for murder are only things like it being a killing; we cannot observe that it is not to be done. The factual analysis I have suggested uses this consideration about cues to cast doubt upon the legitimacy of part of the meaning of the term. Other philosophers will offer a different view which endorses the claim to objective wrongness. I think we are differing over the factual analysis, but the argument is often conducted as if it were about the meaning of the term.
I have suggested that philosophical criticism may lead us to cast doubt on what people usually mean by a term. We can see this as raising a fourth question about a piece of language: how should we use the language in future? This is not asking for any description of what we have been doing; it requires of us a decision about how to behave in future. It ought to be obvious that such a decision does not follow from any facts about what we or other people have been doing in the past. Suggesting that it does is one of the main reasons a lot of analytical philosophy is conservative, since these philosophers suppose that their job is finished once they have described existing usage. But in education and elsewhere one of the reasons for taking notice of concepts and claims is that we may feel it is necessary to start doing something different. While we don't have to go on doing what has become traditional, in some cases there is no pressing reason for a change. But when we get the kind of discrepancy between the answers to our different factual questions such as we have noted above, then we are almost forced to consider the evaluative one of how we will use language in future.
The first three questions above are about how we use language. We can hope to get answers to them that will cover the whole range of normal usage. Obviously, they are not always going to be easy questions to answer. As I have indicated in passing already, such general accounts of aspects of language use may often be fairly schematic; they may not be particularly informative. They may not be what we are looking for. Here we can note a couple of distinctions which may illuminate both the way language is used and what we are looking for in asking about concepts.
One such distinction may be expressed by contrasting a concept with a conception. Let us take teaching again. It may be that the concept of teaching is a pretty unspecific concept: a matter of doing things, or even of things happening, that help to cause learning. It may seem that such a concept would cover almost anything. And so it might, if we look at the multifarious ways people actually use the word "teach" and its cognates. But educators and ordinary people often have some preferred versions of teaching, favourite images or ideals of teachers and teaching. Such specific versions of teaching I shall label conceptions of teaching. One reason you may find the kind of conceptual analysis I have described above unhelpful is that what you really want is not a general account of how a set of terms is used, but rather a detailed specification of what is being referred to on some special occasions. (This is another route to talk of "real" or "true" teaching - a preferred interpretation of teaching - compare Section 10.)
One way in which this contrast between concepts and conceptions arises is from our acquiring beliefs about things. When in Section 10 I mentioned the possibility of changing the meaning of a term in the course of investigation (as for instance "sugar" started off referring to a sweet substance manufactured from certain crops and ends up labelling a kind of carbohydrate) you could say that what was discovered to be part of a reasonable conception of something becomes part of the meaning of the corresponding term. (When it was discovered that sugars were carbohydrates this fact became incorporated into at least the scientists' meaning of the word "sugar".) This example may also suggest that it is often not easy to decide how much to allocate to the meaning of a term and how much to attribute to a conception of the thing; we can often get by without having to decide. Yet it can, I think, help to have the distinction available. One small contribution it can make is to explain away the attractions of private views of language and meaning, since my conceptions of education or democracy or socialism may be very different from yours. A lot of the terms of particular interest to philosophers are indeterminate, or very general, like these, so once again what we discuss are not really the concepts but various competing conceptions that can be subsumed under them.
Another distinction that can usefully be kept in reserve can be expressed by contrasting what something means with what someone means by saying it. This leads on to fairly complicated questions that relate to the conventions of polite conversation. But the basic point can be grasped from a simple example. We can convey a great deal more when we use language than the literal meaning of our sentences. Part of this is due to the other speech acts we may be performing (we looked very briefly at this idea in Section 15). But part of this extra load of meaning can come from our exploitation of the rules for communication. Suppose I ask you about a student's essay in my role as examiner, and you tell me only that the handwriting is very elegant. I would take that to mean the essay is a bad one. You don't say that. You say, the handwriting is elegant, and so perhaps it is. That is what your words mean. But by uttering them, and them only, you mean that the essay is hopeless. How? By exploiting a rule of polite conversation that people should give pertinent answers. Here you have deliberately broken this rule, and so I infer that you've broken it because there is no encouraging pertinent answer, i.e. because the essay is hopeless. A philosophical moral of all this is that we should beware of arguments which appeal to what we don't ordinarily say. We may not ordinarily say something because it can be conveyed quite adequately by the rules of dialogue rather than because it is untrue or absurd.
The various speech acts and the rules for conversation allow us to convey a lot more by using sentences than the literal meaning of such sentences. They indicate two more ways in which what may appear to be discussions of meaning or of conceptual analysis are really about something else. Bearing them in mind may help us clear up some of the confusions of analytical philosophy.
Before closing this brief survey of some of the things that go on in the course of philosophical analysis it is worth mentioning another use of the word "analytic" in philosophy. In Section 10 we met claims like "Either it is raining or it is not raining" that I said were true but empty because their truth depended only on their form. Such tautologies and other claims which are true in virtue of their form and the definitional connections between their constituent terms (such as "All brothers are male," given that a brother is a male sibling) are often referred to as "analytic propositions" or "analytically true". There can of course be analytically false claims too (e.g. "All brothers are female"). Analytically true and false claims are contrasted with all other claims which are labelled "synthetic". Many philosophers have now given up this way of talking, but you may still meet it and it is worth noting that it is distinct from the talk about analysis we have been examining.
URL http://www.uwichill.edu.bb/bnccde/epb/analysis.html
HTML prepared June 2000, last revised June 29th, 2000.