Conceptual Analysis of Education

Let us first of all address ourselves to the question of what it is to analyse a concept. What is a concept? It obviously is not the same as an image; for, to revert to our previous case, we can have a concept of ‘punishment’ without necessarily having a picture in our mind of a criminal being hung or a boy being beaten. Is to have a concept then to be able to use the word ‘punishment’ correctly? If we have the concept, it might be said, we can relate ‘punishment’ to other words like ‘guilt’ and say things like, ‘Only the guilty can be punished’. Indeed it was the understanding of this connection that probably led one of the teachers in our imaginary conversation to say that keeping the whole class in did not constitute ‘punishment’; for guilt had not been established.

This ability to relate words to each other would also go along with the ability to recognize cases to which the word applied. This looks a much more promising approach to making explicit what it is to have a concept. But it will not quite do for two reasons. In the first place we often make distinctions between things or group things together but have not got a word for marking the difference or similarity. Are we then to say that in such cases we have no concept?

This would mean denying that animals, which make quite complicated discriminations, have any concepts. It would mean that children, who behave differentially towards their mother very early in their lives, have no concept of their mother until they can use the word ‘mother’. And what is the point of being so restrictive? Would it not be better to say that our possession of a concept is our ability to make discriminations, and to classify things together if they are similar? To be able to use a word appropriately is a sophisticated and very convenient way of doing this. Indeed it could be regarded as a sufficient condition for the possession of a concept though not a necessary one.

In other words, we would probably be prepared to say that a person had a concept of ‘punishment’ if he could relate the word ‘punishment’ correctly to other words such as ‘pain’ and ‘guilt’ and apply it correctly to cases of punishment. But the absence of this ability to use the word would not straightaway lead us to say that he had not the concept. He might, for instance, get upset when he saw cases of wanton cruelty but not get upset when he saw cases of punishment; but for some reason or another, he might not have been introduced to the words which have been developed for marking these distinctions.

The second reason why it is not altogether satisfactory to equate having a concept with the possession of an ability, whether it be the specific ability to use words appropriately, or the more general one to classify and make discriminations, is that both types of ability seem to presuppose something more fundamental, namely the grasp of a principle which enables us to do these things. Locke said that an idea is ‘the object of the understanding when a man thinks’ and this is probably as near as we can get to saying what a concept is. But it is singularly unilluminating.

As, however, our understanding of what it is to have a concept covers both the experience of grasping a principle and the ability to discriminate and use words correctly, which is observable in the case of others as well as ourselves, there is, amongst philosophers generally, a tendency to rely on this publicly observable criterion of having a concept. For it is possible to say more about it than it is about the subjective side. This public criterion is necessary to identify having a concept, but having a concept is not identical with it. So much, then, for the object of our scrutiny in this branch of philosophical activity. But what do we do in philosophy when we analyse a concept?

As the concept in question is usually one the possession of which goes with the ability to use words appropriately, what we do is to examine the use of words in order to see what principle or principles govern their use. If we can make these explicit we have uncovered the concept. Historically, philosophers such as Socrates attempted to do this by trying out definitions. Now there is a strong and a weak sense of ‘definiiton’ in such cases.

The weak sense is when another word can be found which picks out a characteristic which is a logically necessary condition for the applicability of the original word. Thus, to revert to our case of ‘punishment’, a logically necessary condition for the use of this word is that something unpleasant should be done to someone. If it were not, if for instance someone who committed a crime were sent on a pleasure cruise, we would refuse to apply the word ‘punishment’. Part of our concept of ‘punishment’, therefore, is that something unpleasant is inflicted. The strong case of definition is when conditions can be produced which are logically both necessary and sufficient. In other words if one can say ‘if and only if characteristics x, y, z are present, then a person is being punished’, we would have a really strong sort of definition. In actual practice, we only have such definitions in artificially constructed symbolic systems, such as geometry, where we lay down tight conditions for the use of words such as ‘triangle’.

With words that are employed in a much looser way in ordinary language, such as ‘courage’ and ‘justice’, we would be hard put to it ever to find such a tight set of defining characteristics. In conceptual analysis we usually settle for making explicit defining characteristics in the weak sense. In attempting to make explicit the rules behind our usage of words, and thus get clearer about our concepts, it is important to distinguish logically necessary conditions from other sorts of conditions that may be present. To understand this difference is, in fact, to understand the difference between doing philosophy and doing science. It is probably the case, for instance, that acts of punishment are performed only by people with central nervous systems. But we would not have to know about that in order to understand what is meant by ‘punishment’.

Indeed countless people understand perfectly well what is meant by ‘punishment’ who have never heard of a central nervous system. The possession of a central nervous system is, therefore, only a general empirical condition of punishment rather than part of our understanding of ‘punishment’. The connection, therefore, between ‘punishment’ and the possession of a nervous system is quite different from the relationship between, say, our understanding of ‘hearing’ and the possession of ears.

For the possession of this particular part of the body is inseparable from our understanding of what it is to hear something. We could not conceivably hear without ears of some sort. Similarly, it is only an empirical fact that most learning is brought about by some form of teaching. But ‘teaching’ could not be conceived of without some reference to learning. ‘Learning’ therefore, enters into the analysis of ‘teaching’, for this connexion is not purely de facto. Now though much of what has been called conceptual analysis seems to consist in looking for logically necessary conditions for the use of a word, and hence to be concerned with ‘definition’ in a loose sense, it has become fashionable in recent times to deny that it is ever possible to produce such definitions.

Ordinary language is not static; it is a form of life. If we think that we have got a concept pinned down, we are apt to come across a case where we would naturally use the word but where the condition which we have made explicit is not established. We might think, for instance, that a necessary condition of using the word ‘punishment’ is that something unpleasant should be inflicted on the guilty. Yet we do talk of boxers taking a lot of punishment. And of what are they guilty? Wittgenstein made this general point by taking the example of ‘games’. He claimed that there is no one characteristic in terms of which roulette, golf, patience, etc., are all called ‘games’. Rather they form a ‘family’ united ‘by a complicated network of similarities overlapping and criss-crossing; sometimes overall similarities, sometimes of detail’ – rather like the similarities between faces of people belonging to the same family.

There is, he argued, no one characteristic or group of characteristics that all games possess in order to be called ‘games’. This, at least, should warn us that we may not always be successful even in our search for logically necessary conditions for the use of a word. But sometimes we may be. Is there, for instance, a use of ‘punishment’ in which there is no suggestion of something unpleasant being inflicted on someone? Actually it can be doubted whether Wittgenstein was even right about this particular concept. For how would we know which samples to lay out in order to look for the similarities?

Why did not Wittgenstein take gardening or getting married as examples of games? Does not this show that there is some more general principle which underlies calling things ‘games’ which he might have overlooked? It might be argued that gardening or getting married might be ranked by someone as games. But this brings out that whether something is a game or not does not depend on any simple observable property of the sort that makes a thing a triangle. Rather it depends on how a human being conceives an activity. A necessary condition of calling something a game is, surely, that it must be an activity which is indulged in non-seriously. Now ‘nonseriously’ does not mean that the player lacks involvement in it or that he does not give his attention to it. Rather, it means that he can conceive it as not being part of the ‘business of living’. He does not do it out of duty or prudence or for any other reason of that sort.

This example brings out two points which Wittgenstein himself made. The first is that we must not look for defining characteristics in any simple, stereotyped way, with the paradigm of just one type of word before us. The second is that concepts can only be understood in relation to other concepts. ‘Non-serious’ has to be understood in relation to a whole family of concepts included under ‘serious’. Thus if we are attempting an analysis of concepts by examining the meaning of words, we usually proceed by taking cases within their denotation and trying out suggestions about defining characteristics.

This is how Socrates, for instance, proceeded in the early part of The Republic in order to get clearer about ‘justice’. He took different cases and tried out suggestions such as ‘justice consists in giving every man his due’, ‘justice is the interest of the stronger’ and so on. In this way we gradually make explicit the links between words which reflect our conceptual structure. But we must also pay attention to what we mean by using a word in the sense of the job that we conceive of the word as doing in the context in which we employ it. For words are not just noises or marks on paper; they are more like tools.

They do specific jobs in social life. We could, for instance, only understand what ‘non-serious’ means when applied to games if we have an understanding of the form of life which renders some things ‘serious’. Obviously one of the main jobs that words do is to convey information, to describe things and situations. But this is only one of their jobs. Sometimes we use words to warn people. At other times we use words to suggest courses of action to people. At other times we use words to express what we wish. And so on. The use of words, in other words, is a form of purposeful behaviour, but it has to be understood in terms of the other non-linguistic purposes that people have in their social life.

Commands, for instance, such as ‘Halt’ have to be understood as having a specific sort of function in social situations where some are in authority over others and where they are expected to direct the behaviour of their subordinates by using words in a certain tone of voice. Usually the way in which words are put together in what are called ‘sentences’ gives a very good clue to the job that they do. Sentences can be used to make statements (or assert propositions).

When they do this they are usually couched in what grammarians call the indicative form. If we say, for instance, that the motorist was punished for exceeding the speed limit we are stating a proposition that can be either true or false. The job done by the words in this case is, therefore, to describe or indicate a state of affairs that is assumed to have occurred. The point of such an utterance, in other words, is to convey information. But when the headmaster uses the sentence ‘Punish the boy’ his words, which are now arranged in the imperative form, do not state anything that is true or false.

They have a different function, that of getting someone to do something. The grammatical form of a sentence can be, however, rather deceptive. When, for instance, it is asserted that a child needs love, it looks as if the sentence is simply stating a fact, or conveying information. But it could be argued that what the sentence is really doing is to lay down standards about how a child should be treated, that it has, in other words, a guiding function. ‘Need’ in this context, is performing a normative role.

And there are some words such as ‘right’ and ‘wrong’, ‘good’ and ‘bad’ which almost always have this general function of laying down standards of conduct. If, therefore, we are trying to analyse a concept it is important to realize that this cannot be done adequately by just examining the use of words in any self-contained way. We have to study carefully their relation to other words and their use in different types of sentences. An understanding of their use in sentences does not come just by the study of grammar; it is also necessary to understand the different sorts of purposes that lie behind the use of sentences. And this requires reflection on the different purposes, both linguistic and non-linguistic, that human beings share in their social life.
Read More : Conceptual Analysis of Education

The Point of Conceptual Analysis of Education

The question is often put to philosophers when they have done some conceptual analysis: ‘Whose concept are you analysing?’ The first answer, obviously enough, is our concept. For concepts are linked indissolubly with the social life of a group, and it would be impossible for an individual to have a purely private concept of, say, ‘punishment’. But, it might be said, there are subtle differences between groups of language users, and though there are obviously common elements in a concept there are also likely to be different emphases and differences in valuation, as, for example, in the case of the concept of ‘education’. But this type of objection really misses the point of doing conceptual analysis, which is to get clearer about the types of distinction that words have been developed to designate.

The point is to see through the words, to get a better grasp of the similarities and differences that it is possible to pick out. And these are important in the context of other questions which we cannot answer without such preliminary analysis. Ordinary language is a record of connections and distinctions that men with predominantly practical purposes have found it important to make. It is therefore a valuable guide, but it should never be treated as a repository of unquestionable wisdom. Ordinary usage, for instance, reveals in the case of ‘punishment’ a connection that practically minded people insist on between the committing of an offence against rules and the infliction of something unpleasant on the offender. In so far, therefore, as we talk about ‘punishment’ we acquiesce in this demand that ordinary language reflects.

Conceptual analysis helps us to pin-point more precisely what is implicit in our moral consciousness. But it also enables us to stand back a bit and reflect on the status of the demand to which the word bears witness. It frees us to ask a fundamental question in ethics, which is that of whether this demand is justified. In our view there is little point in doing conceptual analysis unless some further philosophical issue is thereby made more manageable. The first thing to say, therefore, about the point of doing conceptual analysis is that it is a necessary preliminary to answering some other philosophical questions.

We cannot tackle the question in ethics of whether there are any good reasons for punishing people until we are clear what we mean by ‘punishment’. Questions of analysis in other words are often linked with questions of justification. Socrates raised questions about the meaning of ‘justice’ because he was also interested in the reasons that there might be for living a just life. But there are also other more all-pervasive conceptual questions with which the analysis of a particular concept is often linked.

These are usually called metaphysical questions, those that deal with categories of thinking which structure a conceptual scheme. We employ for instance, concepts such as ‘thing-hood’, ‘causality’ and ‘time’ to make the world intelligible. In metaphysics the status of such categoreal concepts is examined. Could we, for instance, dispense with the concept of ‘consciousness’ in making human behaviour intelligible? These are general questions about the justification of our conceptual schemes. The linkage of conceptual analysis with these other types of philosophical question explains the fact that philosophers do not indulge in an undiscriminating analysis of any old concepts. They do not attempt the analysis of concepts such as ‘clock’ and ‘cabbage’ unless there are further issues with which the analysis is connected. Questions, for instance, might be raised about clocks if some philosophical issue about the status of temporal distinctions was at stake.

There is much about vegetables in Hume’s Dialogues on Natural Religion because his interest in justification, in this case in the theological argument from design, led him to examine the different types of order in the world. Why should not its order, he asked, be more like that of a vegetable than that of a house? Similarly Aristotle raised questions about the nature of vegetables because he was interested in the general nature of living things and in the role of ‘purpose’ as a categoreal concept. But without an interest in such further questions what is the point of doing conceptual analysis? It is difficult to understand, too, how all interest in any such further philosophical questions could itself be self-contained. It is difficult to conceive, for instance, how a person could be interested in philosophical questions such as ‘How do we know that we ought to punish people?’ unless he is also interested in the moral question of whether or not we ought to punish people. Philosophers often devote themselves to analysing the concepts of the particular sciences and to enquiring into the epistemological status of the methods of enquiry employed. They ask questions, for instance, about the concept of ‘the unconscious’ and about the validity of psycho-analytic methods in testing hypotheses. But it is often difficult to separate an interest in such general philosophical questions about the status of consciousness from more particular questions about the validity of psycho-analytic theory.

Philosophy, as has been explained, is concerned with second-order questions about science, morality, religion and other such human activities. But the point of asking such questions is usually provided by concrete worries at the first-order level. Lavoisier, for instance, was led to make important discoveries in chemistry through an interest in its conceptual scheme. His interests were in part philosophical even though he was professionally a scientist. But he could not have asked such questions in such a precise form without a passionate interest in and detailed knowledge of the phenomena which chemists were trying to explain.

Thus, though a philosopher could be worried about the concept of ‘the unconscious’ because, like Ryle, he is concerned only with some metaphysical thesis about the status of the mental, he could also be worried by it as a psychologist concerned with giving a theoretical explanation of concrete phenomena. A philosophical psychologist would probably be afflicted by both sorts of worry! A philosopher might be worried, similarly, about ‘the state’ because he had a general interest of a metaphysical sort in ‘the general will’; or he might be worried about it because, like Locke or Burke, he had a practical interest in rights and representation. But in all such cases the concern about concepts has point because of some further concern. To do conceptual analysis, unless something depends on getting clearer about the structure underlying how we speak, may be a fascinating pastime, but it is not philosophy. Once it is appreciated that conceptual analysis must have some point, it can also be appreciated that the inability to emerge with a neat set of logically necessary conditions for the use of a word like ‘knowledge’ or ‘education’ is not necessarily the hall-mark of failure.


For, in the process of trying to make explicit the principles which underlie our use of words, we should have become clearer both about how things are and about the sorts of decisions that have to be faced in dealing with them. We are in a better position to look through the words at the problems of explanation, justification or practical action that occasion such a reflective interest. To revert to the example from which we started: that of ‘punishment’. An analysis of this concept reveals the connection demanded by men that pain should be inflicted, usually by those in authority, on offenders. This makes explicit much that is of ethical significance. There is first of all a problem about the infliction of pain; for this is usually regarded as something that is prima facie undesirable. How is this to be justified? Supposing it can be, on the grounds that it will deter others from committing similar offences and produce less unhappiness in the long run than not inflicting it. Why then should it be inflicted on the offenders? For the concept of ‘punishment’ seems to require this as well. Surely because of some in-built notion of justice that requires discrimination against people only on relevant grounds. But how can justice be justified, in general and in this particular application of the principle?

Surely too, the operation of punishment as a deterrent presupposes a very important assumption about human beings, namely that they are responsible for their actions in the sense that they can be deterred by a consideration of foreseen consequences. And is this assumption justified? Is it not an assumption of great moral significance, because of its connection with our notion of man as a chooser? What would happen to our social life if we gave it up? Then there is the connection between ‘punishment’ and ‘authority’; for we can only distinguish punishment from some cases of revenge because those who inflict the pain, like fathers and teachers, are authorized to do this. But what is meant by ‘authority’? And call this type of institution be justified? What is the role of authority in social life? If pursued in this way the analysis of the concept of ‘punishment’ does at least two very important jobs.

Firstly, it enables us to see more clearly how a concept is connected not only with other concepts but also with a form of social life that rests on a network of interlocking assumptions – e.g. about human responsibility, rights connected with authority, and the role of pain in our lives. We thus begin to get a better understanding of the type of social life to which we seem to be committed if we admit the necessity of punishment. Secondly, however, by laying bare the structure of this concept we also show the extent to which it rests on certain moral assumptions which can be challenged. To discuss their status would take us far into moral philosophy. Now to what extent would our inability to produce a set of logically necessary conditions for the use of the word ‘punishment’ be detrimental to these further purposes that lie behind conceptual analysis? Suppose, for instance, we can produce cases of the use of the word ‘punishment’ where guilt has not been established – as in the case of the schoolmaster who keeps all the class in.

Suppose people talk of a boxer receiving ‘punishment’ when there is neither guilt nor anyone in authority who administers the pain. Surely by reflecting on such cases our understanding of our social life is increased as well as our sensitivity to the complexity of moral issues. We come to distinguish what are often called central cases of the application of ‘punishment’ from more peripheral ones. The central cases are those in which all the conditions are present which enable us to distinguish ‘punishment’ from other allied notions such as ‘revenge’, ‘spite’ and ‘coercion’. The existence of such cases explains how ‘punishment’ performs a distinctive function in the language which reflects our social life, how people come to acquire the concept, and how people come to use derivative expressions such as boxers taking a lot of punishment.

By determining which cases are central we come to learn a lot more, not just about words, but about the structure of our social life and the assumptions which underlie it. If, for instance, we challenge the moral assumptions which underlie punishment we will be led to see what else we may be challenging as well. It is only if we think that there must be some essence in the nature of things or institutions, which our concepts reflect, that we will be dismayed if we fail to produce a hard and fast set of logically necessary conditions for all uses of a word. If we do not hold such a crude view of the relationship between words and things we will not measure the success of conceptual analysis by the extent to which we can produce definitions. Rather we will measure it by the extent to which our understanding is thereby increased about how things are in the world and of the possible stances that we can adopt towards our predicament in it.
Read More : The Point of Conceptual Analysis of Education

Analytic philosophy of education

Philosophers of education are given to making pronouncements on a range of substantive educational issues: the nature of education, teaching, and learning; curriculum content and worthwhile knowledge; moral education; cognitive development; school organisation and administration; educational policy; and educational research methodology, to name just some of the major themes.

To help focus discussion, I take the important background question to be how we could ever know whether any of these pronouncements are warranted. The epistemological question is important because the approaches to philosophy of education I shall be considering offer different answers. Moreover, differences over epistemology will signal more systematic differences over philosophy of education in general. Before the rise of APE in the late 1950s, it was fashionable among philosophers of education to attempt to deduce educational claims from philosophical premises.

As the 1942 and 1955 National Society for the Study of Education yearbooks indicate, philosophy of education was something of a smorgasbord, with characteristic educational positions being associated with particular philosophical ‘isms’, such as empiricism, existentialism, rationalism, pragmatism, and so on (Brubacher, 1942, 1955). Waiving the matter of defending one ism over another, success was construed as establishing the existence of educational implications.

Ismism, as it was unkindly termed, fell on hard times in the mid 1950s after a vigorous ‘implications’ debate, conducted largely in Educational Theory and the Harvard Educational Review, prompted a consensus that the quest for educational implications had failed and that the enterprise was mistaken. (The main papers can be found in Lucas, 1969.) Although technical worries over the nature of implication were extensively canvassed, the most important determinant of the new consensus, and its subsequent development, was a new view of the nature of philosophy.

For arising out of the so-called linguistic revolution in philosophy, particularly the Ordinary Language variety associated with Oxford philosophy, came the view that philosophy should be construed as an activity rather than a body of claims from which deductions might be made. The philosophical activity most characteristic of APE, early and late, is the analysis of concepts, although latter day examples are mostly in disguise, often as exegesis of some theorist (e.g. White, 1990, on Aristotle) or some established educational practice (e.g. Wenham, 1991, on teaching).

For example, in his influential paper ‘Towards an Analytic Philosophy of Education’ Scheffler (1954: 9) recommends construing philosophy of education as ‘the rigorous logical analysis of key concepts related to the practice of education’. In answering his own question about the main concerns of philosophy of education, Peters (1966: 18) remarks: ‘There is, first of all, the analysis of concepts specific to education – such as “education”, “teaching”, “training”, and “university”, and “school”’.

And, despite subsequent criticisms of the methods of conceptual analysis, a quick check of recent papers in the field’s journals shows Barrow (1991) analysing ‘critical thinking’, Gaden (1990) analysing ‘responsibility’, McLaughlin (1990) analysing ‘belief’, and Harvey (1990) analysing ‘stereotype’, to mention just a few of the fairly conspicuous examples. Construed modestly, conceptual analysis serves the always important and useful function of distinguishing uses of terms, sorting out ambiguities, and in general clarifying what is being claimed.

Often associated with an anti-essentialist interpretation of the later Wittgenstein, it captures the idea of philosophy as linguistic therapy (see Strawson, 1992: 3–14). Except where educational theory falls into a chronic ambiguity of expression and sloppiness of usage, the impact of modest analysis has been, as one might expect, modest. Construed more ambitiously (by using the tools of modest analysis) two related approaches to conceptual analysis have been employed in APE to justify substantive educational claims. The first sees the analysis of concepts as a quest for logically necessary and sufficient conditions of usage or, where these are difficult to obtain, at least logically necessary conditions.

Thus, if an analysis of usage suggests that we would not say of someone that they were being educated unless they were learning something worthwhile, then ‘Education involves the learning of something worthwhile’ expresses a conceptual truth (see e.g. Hirst and Peters, 1970: 2–28). The aim of such a philosophy of education is to produce a comprehensive set of conceptual truths about education. These conceptual truths are meant to provide a kind of intellectual scaffolding for educational theory, around which empirical claims are added. The APE developed by Richard Peters, Paul Hirst, and other philosophers of education who plied their trade at the London Institute of Education during the 1960s and beyond, clearly displays this version of analysis. (For recent booklength versions see e.g. Barrow and Woods, 1988; Callan, 1988; and Hamm, 1989.)

The second ambitious view of analytic methodology sees analysis as a quest for the basic presuppositions of knowledge. Applied to education, this methodology looks for what substantive educational claims might be presupposed by any educational theorising. And once again it was the London Institute philosophers of education who led the way in employing the methodology most systematically and to greatest effect. Thus Hirst’s influential forms of knowledge thesis owes much to an analysis of what is presupposed in having a mind.

For example, ‘the phrase, to have “a rational mind” certainly implies experience structured under some form of conceptual scheme’ (Hirst, 1965: this volume, p. 254). Similarly, Peters’s view of what to choose as worthwhile activities in education is based on an analysis of the logical presuppositions of rational, disinterested choice (Peters, 1966: 151–66). The above ambitious versions of analysis are related, in the first instance, according to how ‘presupposition’ is construed.

Interpreting the ambiguous expression ‘S entails S’ as ‘S is a necessary condition for the truth of S’, ‘presupposition’ may mean no more than ‘is a necessary condition for the truth of S’, in which case the two become one. On Strawson’s (1952: 175) usage, ‘presupposition’ means ‘is a necessary condition for the truth or falsity of S’. Further permutations are possible according to how ‘entails’ is understood, though ironically, any further taxonomy risks outrunning the actual conduct of analysis in the literature.

In any case, it is clear that claims to have established conceptual truths about education, or to have discovered the logical presuppositions of an inclusive range of educational discourse, can (and did) exert considerable purchase on substantive educational theorising. The rewards of success were the familiar ones: acknowledged standing as a separate branch of educational studies, akin to history or sociology of education, the proliferation of professional journals, and the establishment of learned societies.
Read More : Analytic philosophy of education

Criticisms of analytic of Education methodology

Three methodological criticisms of APE have been particularly important in pointing the way to post-analytic alternatives in philosophy of education. The first, widely felt in analysis and articulated by Edel (1973: this volume, p. 40), is epistemic: ‘A central difficulty in analytic philosophy of education seems to me to reflect a soft spot in the analytic theory generally – how to judge what is a correct or adequate analysis.’ Edel locates the difficulty in APE’s heritage from positivism: an untenable distinction between analytic (or conceptual) and synthetic (or empirical) claims.

By the early 1970s, Edel was in good company, for the analytic/synthetic distinction had sustained damaging criticism from Quine (1951, 1960) and conspicuous appeal to analyticity had greatly diminished as a ground for the justification of claims in mainstream, Anglo-American, philosophy. Indeed, its lingering influence in philosophy of education was something of an anomaly. With appeal to analyticity disallowed, the structure of justification changes. Take, for example, the defence of S: ‘Teaching is an activity conducted with the intention of bringing about learning.’ Instead of engaging in the usual meaning analysis of ‘teaching’ in order to establish a case for conceptual truth of S, we can view S as providing the most immediate context for understanding the conceptual role of ‘teaching’.

But relevant context extends beyond S to other related sentences: S is embedded in some theory, T. Defending S now becomes a matter of arguing the epistemic case for T against rivals. Although theory choice has its problems it has considerably more resources for decision than meaning analysis. Note that in formulating the problem of justification in this manner a move has been made against merely assuming the epistemic privilege of ordinary language and its embedded commonsense theory of the world (see Evers, 1979).

A second problem lies with the provision of necessary conditions for the correct use of a term. While analytic philosophers of education might be willing to concede on sufficient conditions, it is important to hold the line on necessary conditions. So, to Wittgenstein’s anti-essentialist charge that, for example, there is no common feature to all games (which would undermine the condition that the presence of such a feature is a necessary condition for something to be called a game), Hirst and Peters (1970: 6) respond that he overlooked something: ‘A necessary conditon of calling something a game is, surely, that it is an activity which is indulged in non-seriously.’

One need not have played poker for oil wells, on a Mississippi riverboat, to query the ‘surely’. In response, analysts might want to distinguish two concepts: a fun one for those whose sense of propriety prevails, and a more serious one suitable for higher stakes. But now, is the sentence ‘a game is an activity indulged in non-seriously’ a conceptual truth, or even a truth? Well, once we begin to factor in the relevant social, cultural, and geographical factors that specify the surrounding theoretical context in which each use of ‘game’ figures, the epistemic sense of the question either drops out altogether, becomes relativised to prevailing local contexts, or becomes a matter of theory competition.

And the same conditions of justification will apply to distinctions between so-called central and peripheral uses of ‘game’, or ‘education’, or ‘teaching’, or ‘learning’. Since the presence of necessary conditions of usage in a sufficiently restricted domain is equivalent to their absence in a broader domain we can ask whether consideration of domain size really matters in explaining how humans can master and readily employ concepts (or use terms).

According to Stephen Stich (1992: 248–50), who appeals to empirical work in the psychology of concept acquisition, the answer is no. To appreciate the relevance of empirical conditions, consider an example. A small child is playing with a stick which at first is a gun, then a sword, then a walking stick, and finally a golf club. Analytic ways of thinking might suggest that these things all have something in common – perhaps that length exceeds breadth.

Well, this is true of every solid object in the universe that is not perfectly symmetrical. Moreover, guns can contract smoothly on their major axis without lapsing in gunhood, and likewise for swords, and even walking sticks. Perhaps the child’s imagination is not constrained by affine invariance (maintenance of proportion), or topological invariance (maintenance of boundaries). Given that language – which is quite grainy, discrete, and finite at the level of vocabulary – is being used to describe a world that is infinitely partitionable, one would expect cognitive advantages to accrue to creatures whose learning permits plasticity of feature detection.

But to meet this condition, it is better to see concepts as the result of categorisation by similarity to a prototype, or alternatively, similarity to an exemplar, where neither prototypes (the most typical members of a set) nor exemplars (the most familiar) satisfy anything like necessary or sufficient conditions (see Stich, 1992: 249). It is customary in APE to analyse the nature of human cognition and knowledge through its products, which are invariably taken to be rule-based symbol systems, from natural languages and informal sign systems through to the artificial systems of logic, mathematics, and computing. Certainly, a belief in rules underlying performance would lend credence to the otherwise quite brittle and implausible demands of analysis.

However, when it comes to explaining the human learning of these competencies, it is important nowadays to attend to theories of brain functioning (see Grossberg, 1982). But from a neutral perspective, categorisation and concept formation (and, more controversially, reasoning) are better seen as products of pattern recognition and processing – processes characteristically driven by the application of soft constraints – rather than as being rule-based with precise conditions of application (see Bechtel and Abrahamson, 1991: 220–54; Evers, 1990.)

Notice that analytic methodology needs to challenge more than the truth value of these psychological reflections. In maintaining a conceptual/empirical distinction, it needs to challenge their relevance. The task of meshing a theory of conceptual analysis with a theory of concept acquisition was attempted in the early APE literature – and most clearly seen in the writings of Peters (1974: 119– 50) – by a shift in favour of Piagetian stage theories (shorn of their empirical excrescences). The last major criticism raises the question of whose concepts are being analysed. It is important, because various answers strike at the assumption of epistemic privilege analysis allots to ordinary usage.

For example, theories which see ordinary language as an ideology functioning to reproduce capitalist economic social relations of production, will count almost all the APE claims to conceptual truth as the most central falsehoods of a systematically false theory of education. (The best early critiques along these lines are Adelstein, 1972, and Harris, 1977.) More generally, the concepts of those for whom language is actually a structured misrepresentation of reality will not bear the epistemic weight required to justify educational claims. So gender theorists, for example, who see familiar discursive practices as gendered and part of the male symbolic order, would place little value in ambitious versions of analysis (see Davies, 1989).

A more drastic depature from analysis is to deny the possibility of a preferred knowledge of the semantics of language, and standards of rationality, sufficient to give sentence tokens the stable representational structure required for what are in any case debatable demands of educational justification. Rorty’s (1979: 357– 94) account of philosophy without mirrors (and some of the literature of post-modernism) raises the option. In what follows, I shall canvass benign rather than drastic post-analytic consequences for philosophy of education.
Read More : Criticisms of analytic of Education methodology