CHAPTER ONE INTERPRETATION AND THE SCIENCES OF MAN~:' I Is there a sense in which interpretation is essential to explanation in the sciences of man? The view that it is, that there is an unavoidably 'hermeneutical' component in the sciences of man, goes back to Dilthey. But recently the question has come again to the fore, for instance, in the work of Gadamer,l in Ricoeur's interpretation of Freud,2 and in the writings of Habermas.3 Interpretation, in the sense relevant to hermeneutics, is an attempt to make clear, to make sense of, an object of study. This object must, therefore, be a text, or a text-analogue, which in some way is confused, incomplete, cloudy, seemingly contradictory - in one way or another, unclear. The interpretation aims to bring to light an underlying coherence or sense. This means that any science which can be called 'hermeneutical', even in an extended sense, must be dealing with one or another of the confusingly interrelated forms of meaning. Let us try to see a little more clearly what this involves. We need, first, an object or field of objects, about which we can speak in terms of coherence or its absence, of making sense or nonsense. Second, we need to be able to make a distinction, even if only a relative one, between the sense or coherence made, and its embodiment in a particular field of carriers or signifiers. For otherwise the task of making clear what is fragmentary or confused would be radically impossible. No * I have greatly benefited in preparing this paper from discussions held under the auspices of the Study Group for the Unity of Knowledge, whose meetings were supported by the Ford Foundation. 1 E.g., H. G. Gadamer, Wahrheit und Methode (Tiibingen, 1960). Paul Ricoeur, De ['interpretation (Paris, 1965). E.g., J. Habermas, Erkenntnis und Interesse (Frankfurt, 1968). 16 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 17 sense could be given to this idea. We have to be able to make for our interpretations claims of the order: the meaning confusedly present in this text or text-analogue is clearly expressed here. The meaning, in other words, is one which admits of more than one expression, and, in this sense, a distinction must be possible between meaning and expression. The point of the above qualification, that this distinction may be only relative, is that there are cases where no clear, unambiguous, non-arbitrary line can be drawn between what is said and its expression. It can be plausibly argued (I think convincingly, although there is no space to go into it here) that this is the normal and fundamental condition of meaningful expression, that exact synonymy, or equivalence of meaning, is a rare and localized achievement of specialized languages or uses of civilization. But this, if true (and I think it is), does not do away with the distinction between meaning and expression. Even if there is an important sense in which a meaning re-expressed in a new medium cannot be declared identical, this by no means entails that we can give no sense to the project of expressing a meaning in a new way. It does of course raise an interesting and difficult question about what can be meant by expressing it in a clearer way: what is the 'it' which is clarified if equivalence is denied? I hope to return to this in examining interpretation in the sciences of man. Hence the object of a science of interpretation must be describable in terms of sense and nonsense, coherence and its absence; and must admit of a distinction between meaning and its expression. There is also a third condition it must meet. We can speak of sense or coherence, and of their different embodiments, in connection with such phenomena as gestalts, or patterns in rock formations, or snow crystals, where the notion of expression has no real warrant. What is lacking here is the notion of a subject for whom these meanings are. Without such a subject, the choice of criteria of sameness and difference, the choice among the different forms of coherence which can be identified in a given pattern, among the different conceptual fields in which it can be seen, is arbitrary. In a text or text-analogue, on the other hand, we are trying to make explicit the meaning expressed, and this means expressed by or for a subject or subjects. The notion of expression refers us to that of a subject. The identification of the subject is by no means necessarily unproblematical, as we shall see further on; it may be one of the most difficult problems, an area in which prevailing epistemological prejudice may blind us to the nature of our object of study. I think this has been the case, as I will show below. And moreover, the identification of a subject does not assure us of a clear and absolute distinction between meaning and expression, as we saw above. But any such distinction, even a relative o~e, is without any anchor at all, is totally arbitrary, without appeal to a subject. The object of a science of interpretation must thus have: sense, distinguishable from its expression, which is for or by a subject. 2 Before going on to see in what way, if any, these conditions are realized in the sciences of man, I think it would be useful to set out more clearly what rides on this question, why it matters whether or not we think of the sciences of man as hermeneutical, what the issue is at stake here. The issue here is at root an epistemological one. But it is inextricable from an ontological one, and, hence, cannot but be relevant to our notions of science and of the proper conduct of enquiry. We might say that it is an ontological issue which has been argued ever since the seventeenth century in terms of epistemological considerations which have appeared to some to be unanswerable. The case could be put in these terms: what are the criteria of judgement in a hermeneutical science? A successful interpretation is one which makes clear the meaning originally present in a confused, fragmentary, cloudy form. But how does one know that this interpretation is correct? Presumably because it makes sense of the original text: what is strange, mystifying, puzzling, contradictory is no longer so, is accounted for. The interpretation appeals throughout to our understanding of the 'language' of expression, which understanding allows us to see that this expression is puzzling, that it is in contradiction to that other, and so on, and that these difficulties are cleared up when the meaning is expressed in a new way. But this appeal to our understanding seems to be crucially inadequate. What if someone does not 'see' the adequacy of our interpretation, does not accept our reading? We try to show him how it makes sense of the original nonsense or partial sense. But for him to follow us he must read the original language as we do, he must recognize these expressions as puzzling in a certain way, and hence be looking for a solution to our problem. If he does not, what can we do? The answer, it would seem, can,only be more of the same. We have to show him through the reading of other expressions why this expression must be read in the way we propose. But success here requires that he follow us in these other readings, and so on, it would seem, potentially forever. We cannot escape an ultimate appeal to a common understanding of the expressions, of the 'language' involved. 18 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 19 This is one way of trying to express what has been called the 'hermeneutical circle'. What we are trying to establish is a certain reading of text or expressions, and what we appeal to as our grounds for this reading can only be other readings. The circle can also be put in terms of partwhole relations: we are trying to establish a reading for the whole text, and for this we appeal to readings of its partial expressions; and yet because we are dealing with meaning, with making sense, where expressions only make sense or not in relation to others, the readings of partial expressions depend on those of others, and ultimately of the whole. Put in forensic terms, as we started to do above, we can only convince an interlocutor if at some point he shares our understanding of the language concerned. If he does not, there is no further step to take in rational argument; we can try to awaken these intuitions in him, or we can simply give up; argument will advance us no further. But of course the forensic predicament can be transferred into my own judging: if I am this illequipped to convince a stubborn interlocutor, how can I convince myself? How can I be sure? Maybe my intuitions are wrong or distorted, maybe I am locked into a circle of illusion. Now one, and perhaps the only, sane response to this would be to say that such uncertainty is an ineradicable part of our epistemological predicament; that even to characterize it as 'uncertainty' is to adopt an absurdly severe criterion of 'certainty', which deprives the concept of any sensible use. But this has not been the only or even the main response of our philosophical tradition. And it is another response which has had an important and far-reaching effect on the sciences of man. The demand has been for a level of certainty which can only be attained by breaking beyond the circle. There are two ways in which this break-out has been envisaged. The first might be called the 'rationalist' one and could be thought to reach a culmination in Hegel. It does not involve a negation of intuition, or of our understanding of meaning, but rather aspires to attainment of an understanding of such clarity that it would carry with it the certainty of the undeniable. In Hegel's case, for instance, our full understanding of the whole in 'thought' carries with it a grasp of its inner necessity, such that we see how it could not be otherwise. No higher grade of certainty is conceivable. For this aspiration the word 'break-out' is badly chosen; the aim is rather to bring understanding to an inner clarity which is absolute. The other way, which we can call 'empiricist', is a genuine attempt to go beyond the circle of our own interpretations, to get beyond subjectivity. The attempt is to reconstruct knowledge in such a way that there is no need to make final appeal to readings or judgements which cannot be checked further. That is why the basic building block of knowledge on this view is the impression, or sense-datum; a unit of information which is not the deliverance of a Judgement, which has by definition no element in it of reading or interpretation, which is a brute datum. The highest ambition would be to build our knowledge from such building blocks by judgements which could be anchored in a certainty beyond subjective intuition. This is what underlies the attraction of the notion of the association of ideas, or if the same procedure is viewed as a method, induction. If the original acquisition of the units of information is not the fruit of judgement or interpretation, then the constatation that two such elements occur together need not be the fruit of interpretation either, of a reading or intuition which cannot be checked. For if the occurrence of a single element is a brute datum, then so is the co-occurrence of two such elements. The path to true knowledge would then repose crucially on the correct recording of such co-occurrences. This is what lies behind an ideal of verification which is central to an important tradition in the philosophy of science, whose main contemporary protagonists are the logical empiricists. Verification must be grounded ultimately in the acquisition of brute data. By 'brute d~ta' I mean here and throughout data whose validity cannot be questIOned by offering another interpretation or reading, data whose credibility cannot be founded or undetermined by further reasoning. 4 If such a difference of interpretation can arise over given data, then it must be possible to structure the argument so as to distinguish the basic, brute data from the inferences made on the basis of them. The inferences themselve;s, of course, to be valid, must similarly be beyond the challenge of a rival interpretation. Here the logical empiricists added to the armoury of traditional empiricism, which set great store by the method of induction, the whole domain of logical and mathematical inference which had been central to the rationalist 4 The notion of brute data here has some relation to, but is not at all the same as, the 'brute facts' discussed by Elizabeth Anscombe, 'On brute facts', Analysis, 18 (1957-5 8), pp. 69-72., and John Searle, Speech Acts (Cambridge, 1969), pp. 50-3· For Anscombe and Searle brute facts are contrasted to what may be called 'institutional facts', to use Searle's term, i.e., facts which presuppose the existence of certain institutions. Voting would be an example. But, as we shall see below in section II, some institutional facts, such as X's having voted Liberal, can be verified as brute data in the senseused here, and thus find a place in the category of political behaviour. What cannot as eaSily be descnbed In terms of brute data are the institutions themselves. Cf. the discussion below in section II. 20 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 21 positIOn (with Leibniz at least, although not with Hegel), and which offered another brand of unquestionable certainty. Of course, mathematical inference and empirical verification were combined in such a way that two theories or more could be verified of the same domain of facts. But this was a consequence to which logical empiricism was willing to accommodate itself. As for the surplus meaning in a theory which could not be rigorously co-ordinated with brute data, it was considered to be quite outside the logic of verification. As a theory of perception, this epistemology gave rise to all sorts of problems, not least of which was the perpetual threat of scepticism and solipsism inseparable from a conception of the basic data of knowledge as brute data, beyond investigation. As a theory of perception, however, it seems largely a thing of the past, in spite of a surprising recrudescence in the Anglo-Saxon world in the 193os and 1940s. But there is no doubt that it goes marching on, among other places, as a theory of how the human mind and human knowledge actually function. In a sense, the contemporary period has seen a better, more rigorous statement of what this epistemology is about in the form of computerinfluenced theories of intelligence. These try to model intelligence as consisting of operations on machine-recognizable input which could themselves be matched by programs which could be run on machines. The machine criterion provides us with our assurance against an appeal to intuition or interpretations which cannot be understood by fully explicit procedures operating on brute data - the input.s The progress of natural science has lent great credibility to this epistemology, since it can be plausibly reconstructed on this model, as for instance has been done by the logical empiricists. And, of course, the temptation has been overwhelming to reconstruct the sciences of man on the same model; or rather to launch them in lines of enquiry that fit this paradigm, since they are constantly said to be in their 'infancy'. Psychology, where an earlier vogue of behaviourism is being replaced by a boom of computer-based models, is far from the only case. The form this epistemological bias - one might say obsession - takes is different for different sciences. Later I would like to look at a particular case, the study of politics, where the issue can be followed out. But in general, the empiricist orientation must be hostile to a conduct of enquiry Cf. discussion in M. Minsky, Computation (Englewood Cliffs, N], 1967), pp. 104-7, where Minsky explicitly argues that an effective procedure, which no longer requires intUItIOn or Interpretation, is one which can be realized by a machine. which is based on interpretation, and which encounters the hermeneutical circle as this was characterized above. This cannot meet the requirements of intersubjective, non-arbitrary verification which it considers essential to science. And along with the epistemological stance goes the ontological belief that reality must be susceptible to understanding and explanation by science so understood. From this follows a certain set of notions of what the sciences of man must be. On the other hand, many, including myself, would like to argue that these notions about the sciences of man are sterile, that we cannot come to understand important dimensions of human life within the bounds set by this epistemological orientation. This dispute is of course familiar to all in at least some of its ramifications. What I want to claim is that the issue can be fruitfully posed in terms of the notion of interpretation as I began to outline it above. I think this way of putting the question is useful because it allows us at once to bring to the surface the powerful epistemological beliefs which -underlie the orthodox view of the sciences of man in our academy, and to make explicit the notion of our epistemological predicament implicit in the opposing thesis. This is in fact rather more way-out and shocking to the tradition of scientific thought than is often admitted or realized by the opponents of narrow scientism. It may not strengthen the case of the opposition to bring out fully what is involved in a hermeneutical science as far as convincing waverers is concerned, but a gain in clarity is surely worth a thinning of the ranks - at least in philosophy. 3 Before going on to look at the case of political science, it might be worth asking another question: why should we even pose the question whether the sciences of man are hermeneutical? What gives us the idea in the first place that men and their actions constitute an object or a series of objects which meet the conditions outlined above? The answer is that on the phenomenological level or that of ordinary speech (and the two converge for the purposes of this argument) a certain notion of meaning has an essential place in the characterization of human behaviour. This is the sense in which we speak of a situation, an action, a demand, a prospect having a certain meaning for a person. Now it is frequently thought that 'meaning' is used here in a sense which is a kind of illegitimate extension from the notion of linguistic meaning. Whether it can be considered an extension or not is another 22 PHHOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN matter; it certainly differs from linguistic meaning. But it would be very hard to argue that it is an illegitimate use of the term. When we speak of the 'meaning' of a given predicament, we are using a concept which has the following articulation. (1) Meaning is for a subject: it is not the meaning of the situation in vacuo, but its meaning for a subject, a specific subject, a group of subjects, or perhaps what its meaning is for the human subject as such (even though particular humans might be reproached with not admitting or realizing this). (2) Meaning is of something; that is, we can distinguish between a given element - situation, action, or whatever - and its meaning. But this is not to say that they are physically separable. Rather we are dealing with two descriptions of the element, in one of which it is characterized in terms of its meaning for the subject. But the relations between the two descriptions are not symmetrical. For, on the one hand, the description in terms of meaning cannot be, unless descriptions of the other kind apply as well; or put differently, there can be no meaning without a substrate. But on the other hand, it may be that the same meaning may be borne by another substrate - for instance, a situation with the same meaning may be realized in different physical conditions. There is a necessary role for a potentially substitutable substrate; or all meanings are of something. (3) Things only have meaning in a field, that is, in relation to the meanings of other things. This means that there is no such thing as a single, unrelated meaningful element; and it means that changes in the other meanings in the field can involve changes in the given element. Meanings can not be identified except in relation to others, and in this way resemble words. The meaning of a word depends, for instance, on those words with which it contrasts, on those which define its place in the language (e.g., those defining 'determinable' dimensions, like colour, shape), on those which define the activity or 'language game' it figures in (describing, invoking, establishing communion), and so on. The relations between meanings in this sense are like those between concepts in a semantic field. Just as our colour concepts are given their meaning by the field of contrast they set up together, so that the introduction of new concepts will alter the boundaries of others, so the various meanings that a subordinate's demeanour can have for us, as deferential, respectful, cringing, mildly mocking, ironical, insolent, provoking, downright rude, are established by a field of contrast; and as with finer discrimination on our part, or a more sophisticated culture, new possibilities are born, so other terms of this range are altered. And as the meaning of our terms 'red', 'blue', 'green' is fixed by the definition of a field of contrast through the determinable term 'colour', so all these alternative demeanours are only available in a society which has, among other types, hierarchical relations of power and command. And corresponding to the underlying language game of designating coloured objects is the set of social practices which sustain these hierarchical structures and are fulfilled in them. Meaning in this sense -let us call it experiential meaning - thus is for a subject, of something, in a field. This distinguishes it from linguistic meaning which has a four- and not a three-dimensional structure. Linguistic meaning is for subjects and in a field, but it is the meaning of signifiers and it is about a world of referents. Once we are clear about the likenesses and differences, there should be little doubt that the term 'meaning' is not a misnomer, the product of an illegitimate extension into this context of experience and behaviour. There is thus a quite legitimate notion of meaning which we use when we speak of the meaning of a situation for an agent. And that this concept has a place is integral to our ordinary consciousness and hence speech about our actions. Our actions are ordinarily characterized by the purpose sought and explained by desires, feelings, emotions. But the language by which we describe our goals, feelings, desires is also a definition of the meaning things have for us. The vocabulary defining meaning words like 'terrifying', 'attractive' - is linked with that describing feeling _ 'fear', 'desire' - and that describing goals - 'safety', 'possession'. Moreover, our understanding of these terms moves inescapably in a hermeneutical circle. An emotion term like 'shame', for instance, essentially refers us to a certain kind of situation, the 'shameful', or 'humiliating', and a certain mode of response, that of hiding oneself, of covering up, or else 'wiping out' the blot. That is, it is essential to this feeling's being identified as shame that it be related to this situation and give rise to this type of disposition. But this situation in its turn can only be identified in relation to the feelings which it provokes; and the disposition is to a goal which can similarly not be understood without reference to the feelings experienced: the 'hiding' in question is one which will cover up my shame; it is not the same as hiding from an armed pursuer; we can only understand what is meant by 'hiding' here if we understand what kind of feeling and situation is being talked about. We have to be within the circle. An emotion term like 'shame' can only be explained by reference to other concepts which in turn cannot be understood without reference to 24 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN shame. To understand these concepts we have to be in on a certain experience, we have to understand a certain language, not just of words, but also a certain language of mutual action and communication, by which we blame, exhort, admire, esteem each other. In the end we are in on this because we grow up in the ambit of certain common meanings. But we can often experience what it is like to be on the outside when we encounter the feeling, action, and experiential meaning language of another civilization. Here there is no translation, no way of explaining in other, more accessible concepts. We can only catch on by getting somehow into their way of life, if only in imagination. Thus if we look at human behaviour as action done out of a background of desire, feeling, emotion, then we are looking at a reality which must be characterized in terms of meaning. But does this mean that it can be the object of a hermeneutical science as this was outlined above? There are, to remind ourselves, three characteristics that the object of a science of interpretation has: it must have sense or coherence; this must be distinguishable from its expression, and this sense must be for a subject. Now, in so far as we are talking about behaviour as action, hence in terms of meaning, the category of sense or coherence must apply to it. This is not to say that all behaviour must 'make sense', if we mean by this be rational, avoid contradiction, confusion of purpose, and the like. Plainly a great deal of our action falls short of this goal. But in another sense, even contradictory, irrational action is 'made sense of' when we understand why it was engaged in. We make sense of action when there is a coherence between the actions of the agent and the meaning of his situation for him. We find his action puzzling until we find such a coherence. It may not be bad to repeat that this coherence in no way implies that the action is rational: the meaning of a situation for an agent may be full of confusion and contradiction; but the adequate depiction of this contradiction makes sense of it. Thus we necessarily have a hermeneutical circle. Our conviction that the account makes sense is contingent on our reading of action and situation. But these readings cannot be explained or justified except by reference to other such readings, and their relation to the whole. If an interlocutor does not understand this kind of reading, or will not accept it as valid, there is nowhere else the argument can go. Ultimately, a good explanation is one which makes sense of the behaviour; but then to appreciate a good explanation, one has to agree on what makes good sense; what makes good sense is a function of one's readings; and these in turn are based on the kind of sense one understands. But how about the second characteristic, that sense should be distinguishable from its embodiment? This is necessary for a science of interpretation because interpretation lays a claim to make a confused meaning clearer; hence there must be some sense in which the 'same' meaning is expressed, but differently. This immediately raises a difficulty. In talking of experiential meaning above, I mentioned that we can distinguish between a given element and its meaning, between meaning and substrate. This carried the claim that a given meaning may be realized in another substrate. But does this mean that we can always embody the same meaning in another situation? Perhaps there are some situations, standing before death, for instance, which have a meaning which cannot be embodied otherwise. But fortunately this difficult question is irrelevant for our purposes. For here we have a case in which the analogy between text and behaviour implicit in the notion of a hermeneutical science of man only applies with important modifications. The text is replaced in the interpretation by another text, one which is clearer. The text-analogue of behaviour is not replaced by another such text-analogue. When this happens we have revolutionary theatre, or terroristic acts designed to make propaganda of the deed, in which the hidden relations of a society are supposedly shown up in a dramatic confrontation. But this is not scientific understanding, even though it may perhaps be based on such understanding, or claim to be. But in science the text-analogue is replaced by a text, an account. Which might prompt the question how we can even begin to talk of interpretation here, of expressing the same meaning more clearly, when we have two such utterly different terms of comparison, a text and a tract of behaviour? Is the whole thing not just a bad pun? This question leads us to open up another aspect of experiential meaning which we abstracted from earlier. Experiential meanings are defined in fields of contrast, as words are in semantic fields. But what was not mentioned above is that these two kinds of definition are not independent of each other. The range of human desires, feelings, emotions, and hence meanings is bound up with the level and type of culture, which in turn is inseparable from the distinctions and categories marked by the language people speak. The field of meanings in which a given situation can find its place is bound up with the semantic field of the terms characterizing these meanings and the related feelings, desires, predicaments. But the relationship involved here is not a simple one. There are two 26 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN simple types of models of relation which could be offered here, but both are inadequate. We could think of the feeling vocabulary as simply describing pre-existing feelings, as marking distinctions which would be there without them. But this is not adequate because we often experience in ourselves or others how achieving, say, a more sophisticated vocabulary of the emotions makes our emotional life, not just our descriptions of it, more sophisticated. Reading a good, powerful novel may give me the picture of an emotion which I had not previously been aware of. But we cannot draw a neat line between an increased ability to identify and an altered ability to feel emotions which this enables. The other simple inadequate model of the relationship is to jump from the above to the conclusion that thinking makes it so. But this clearly will not do either, since not just any new definition can be forced on us, nor can we force it on ourselves; and some which we do gladly take up can be judged inauthentic, or in bad faith, or just wrong-headed by others. These judgements may be wrong, but they are not in principle illicit. Rather we make an effort to be lucid about ourselves and our feelings, and admire a man who achieves this. Thus, neither the simple correspondence view is correct, nor the view that thinking makes it so. But both have prima facie warrant. There is such a thing as self-lucidity, which points us to a correspondence view; but the achievement of such lucidity means moral change, that is, it changes the object known. At the same time, error about oneself is not just an absence of correspondence; it is also in some form inauthenticity, bad faith, self-delusion, repression of one's human feelings, or something of the kind; it is a matter of the quality of what is felt just as much as what is known about this, just as self-knowledge is. H this is so, then we have to think of man as a self-interpreting animal. He is necessarily so, for there is no such thing as the structure of meanings for him independently of his interpretation of them; for one is woven into the other. But then the text of our interpretation is not that heterogeneous from what is interpreted; for what is interpreted is itself an interpretation; a self-interpretation which is embedded in a stream of action. It is an interpretation of experiential meaning which contributes to the constitution of this meaning. Or to put it in another way: that of which we are trying to find the coherence is itself partly constituted by self- interpretation. Our aim is to replace this confused, incomplete, partly erroneous selfinterpretation by a correct one. And in doing this we look not only to the self-interpretation but to the stream of behaviour in which it is set; just as in interpreting a historical document we have to place it in the stream of events which it relates to. But of course the analogy is not exact, for here we are interpreting the interpretation and the stream of behaviour in which it is set together, and not just one or the other. There is thus no utter heterogeneity of interpretation to what it is about; rather there is a slide in the notion of interpretation. Already to be a living agent is to experience one's situation in terms of certain meanings; and this in a sense can be thought of as a sort of proto-'interpretation'. This is in turn interpreted and shaped by the language in which the agent lives these meanings. This whole is then at a third level interpreted by the explanation we proffer of his actions. In this way the second condition of a hermeneutical science is met. But this account poses in a new light the question mentioned at the beginning: whether the interpretation can ever express the same meaning as the interpreted. And in this case, there is clearly a way in which the two will not be congruent. For if the explanation is really clearer than the lived interpretation then it will be such that it would alter in some way the behaviour if it came to be internalized by the agent as his self-interpretation. In this way a hermeneutical science which achieves its goal, that is, attains greater clarity than the immediate understanding of agent or observer, must offer us an interpretation which is in this way crucially out of phase with the explicandum. Thus, human behaviour seen as action of agents who desire and are moved, who have goals and aspirations, necessarily offers a purchase for descriptions in terms of meaning - what I have called 'experiential meaning'. The norm of explanation which it posits is one which 'makes sense' of the behaviour, which shows a coherence of meaning. This 'making sense of' is the proferring of an interpretation; and we have seen that what is interpreted meets the conditions of a science of interpretation: first, that we can speak of its sense or coherence; and second, that this sense can be expressed in another form, so that we can speak of the interpretation as giving clearer expression to what is only implicit in the explicandum. The third condition, that this sense be for a subject, is obviously met in this case, although who this subject is is by no means an unproblematical question, as we shall see later on. This should be enough to show that there is a good prima facie case to the effect that men and their actions are amenable to explanation of a hermeneutical kind. There is, therefore, some reason to raise the issue and challenge the epistemological orientation which would rule interpretation out of the sciences of man. A great deal more must be said to bring PHILOSOPHY AND SOCIAL SCIENCE out what is involved in the hermeneutical sciences of man. But before getting on to this, it might help to clarify the issue with a couple of examples drawn from a specific field, that of politics. II I In politics, too, the goal of a verifiable science has led to the concentration on features which can supposedly be identified in abstraction from our understanding or not understanding experiential meaning. These - let us call them brute data identifications - are what supposedly enable us to break out from the hermeneutical circle and found our science four square on a verification procedure which meets the requirements of the empiricist tradition. But in politics the search for such brute data has not gone to the lengths which it has in psychology, where the object of science has been thought of by many as behaviour qua 'colourless movement', or as machine-recognizable properties. The tendency in politics has been to stop with something less basic, but - so it is thought - the identification of which cannot be challenged by the offering of another interpretation or reading of the data concerned. This is what is referred to as 'behaviour' in the rhetoric of political scientists, but it has not the rock bottom quality of its psychological homonym. Political behaviour includes what we would ordinarily call actions, but ones that are supposedly brute data identifiable. How can this be so? Well, actions are usually described by the purpose or end-state realized. But the purposes of some actions can be specified in what might be thought to be brute data terms; some actions, for instance, have physical end-states, like getting the car in the garage or climbing the mountain. Others have end-states which are closely tied by institutional rules to some unmistakable physical movement; thus, when I raise my hand in the meeting at the appropriate time, I am voting for the motion. The only questions we can raise about the corresponding actions, given such movements or the realization of such end-states, are whether the agent was aware of what he was doing, was acting as against simply emitting reflex behaviour, knew the institutional significance of his movement, and so on. Any worries on this score generally turn out to be pretty artificial in the contexts political scientists are concerned with; and where they do arise they can be checked by relatively simple devices, for example asking the subject: did you mean to vote for the motion? INTERPRETATION AND THE SCIENCES OF MAN Hence, it would appear that there are actions which can be identified beyond fear of interpretative dispute; and this is what gives the foundation for the category of 'political behaviour'. There are some acts of obvious political relevance which can be specified thus in physical terms, such as killing, sending tanks into the streets, seizing people and confining them to cells; and there is an immense range of others which can be specified from physical acts by institutional rules, such as voting. These can be the object of a science of politics which can hope to meet the stringent requirements of verification. The latter class particularly has provided matter for study in recent decades - most notably in the case of voting studies. But of course a science of politics confined to such acts would be much too narrow. For on another level these actions also have meaning for the agents which is not exhausted in the brute data descriptions, and which is often crucial to understanding why they were done. Thus, in voting for the motion I am also saving the honour of my party, or defending the value of free speech, or vindicating public morality, or saving civilization from breakdown. It is in such terms that the agents talk about the motivation of much of their political action, and it is difficult to conceive a science of politics which does not come to grips with it. Behavioural political science comes to grips with it by taking the meanings involved in action as facts about the agent, his beliefs, his affective reactions, his 'values', as the term is frequently used. For it can be thought verifiable in the brute data sense that men will agree to subscribe or not to a certain form of words (expressing a belief, say); or express a positive or negative reaction to certain events, or symbols; or agree or not with the proposition that some act is right or wrong. We can thus get at meanings as just another form of brute data by the techniques of the opinion survey and content analysis. An immediate objection springs to mind. If we are trying to deal with the meanings which inform political action, then surely interpretive acumen is unavoidable. Let us say we are trying to understand the goals and values of a certain group, or grasp their vision of the polity; we might try to probe this by a questionnaire asking them whether they assent or not to a number of propositions, which are meant to express different goals, evaluations, beliefs. But how did we design the questionnaire? How did we pick these propositions? Here we relied on our understanding of the goals, values, vision involved. But then this understanding can be challenged, and hence the significance of our results questioned. Perhaps the finding of our study, the compiling of proportions of assent and 3° PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN dissent to these propositions, is irrelevant, is without significance for understanding the agents or the polity concerned. This kind of attack is frequently made by critics of mainstream political science, or for that matter social science in general. To this the proponents of this mainstream reply with a standard move of logical empiricism: distinguishing the process of discovery from the logic of verification. Of course, it is our understanding of these meanings which enables us to draw up the questionnaire which will test people's attitudes in respect to them. And, of course, interpretive dispute about these meanings is potentially endless; there are no brute data at this level, every affirmation can be challenged by a rival interpretation. But this has nothing to do with verifiable science. What is firmly verified is the set of correlations between, say, the assent to certain propositions and certain behaviour. We discover, for instance, that people who are active politically (defined by participation in a certain set of institutions) are more likely to consent to certain sets of propositions supposedly expressing the values underlying the system.6 This finding is a firmly verified correlation no matter what one thinks of the reasoning, or simple hunches, that went into designing the research which established it. Political science as a body of knowledge is made up of such correlations; it does not give a truth value to the background reasoning or hunch. A good interpretive nose may be useful in hitting on the right correlations to test, but science is never called on to arbitrate the disputes between interpretations. .Thus, in addition to those overt acts which can be defined physically or institutionally, the category of political behaviour can include assent or dissent to verbal formulae, or the occurrence or not of verbal formulae in speech, or expressions of approval or rejection of certain events or measures as observed in institutionally defined behaviour (for instance, turning out for a demonstration). Now there are a number of objections which can be made to this notion of political behaviour; one might question in all sorts of ways how interpretation-free it is in fact. But I would like to question it from another angle. One of the basic characteristics of this kind of social science is that it reconstructs reality in line with certain categorial principles. These allow for an inter-subjective social reality which is made up of brute data, identifiable acts and structures, certain institutions, procedures, actions. It allows for beliefs, affective reactions, evaluations as the psychological 6 See H. McClosky, 'Consensus and ideology in American politics', American Political Science Review, 5: 58 (r964), pp. 361-82. properties of individuals. And it allows for correlations between these twO orders or reality: for example, that certain beliefs go along with certain acts, certain values with certain institutions, and so on. To put it another way, what is objectively (inter-subjectively) real is brute data identifiable. This is what social reality is. Social reality described in terms of its meaning for the actors, such that disputes could arise about interpretation which could not be settled by brute data (e.g., are people rioting to get a hearing, or are they rioting to redress humiliation, or out of blind anger, or because they recover a sense of dignity in insurrection?), is given subjective reality, that is, there are certain beliefs, affective reactions, evaluations which individuals make or have about or in relation to social reality. These beliefs or reactions can have an effect on this reality; and the fact that such a belief is held is a fact of objective social reality. But the social reality which is the object of these attitudes, beliefs, reactions can only be made up of brute data. Thus any description of reality in terms of meanings which is open to interpretive question is only allowed into this scientific discourse if it is placed, as it were, in quotes and attributed to individuals as their opinion, belief, attitude. That this opinion, belief, etc. is held is thought of as a brute datum, since it is redefined as the respondent's giving a certain answer to the questionnaire. This aspect of social reality which concerns its meanings for the agents has been taken up in a number of ways, but recently it has been spoken of in terms of political culture. Now the way this is defined and studied illustrates clearly the categorial principles above. For instance, political culture is referred to by Almond and Powell as the 'psychological dimension of the political system'.? Further on they state: 'Political culture is the pattern of individual attitudes and orientations towards politics among the members of a political system. It is the subjective realm which underlies and gives meaning to political actions'.8 The authors then go on to distinguish three different kinds of orientations, cognitive (knowledge and beliefs), affective (feelings), and evaluative (judgements and opinions). From the point of view of empiricist epistemology, this set of categorial principles leaves nothing out. Both reality and the meanings it has for actors are coped with. But what it in fact cannot allow for are intersubjective meanings, that is, it cannot allow for the validity of descriptions of ? Gabriel A. Almond and G. Bingham Powell, Comparative Politics: a Developmental Approach (Boston and Toronto, 1966), p. 23. 8 Ibid., p. 50. 32 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 33 social reality in terms of meanings, hence not as brute data, which are not in quotation marks and attributed as opinion, attitude, etc. to individual(s). It is this exclusion that I would like to challenge in the name of another set of categorial principles, inspired by a quite other epistemology. 2 We spoke earlier about the brute data identification of acts by means of institutional rules. Thus, putting a cross beside someone's name on a slip of paper and putting this in a box counts in the right context as voting for that person; leaving the room, saying or writing a certain form of words, counts as breaking off the negotiations; writing one's name on a piece of paper counts as signing the petition, and so on. But what is worth looking at is what underlies this set of identifications. These identifications are the application of a language of social life, a language which marks distinctions among different possible social acts, relations, structures. But what underlies this language? Let us take the example of breaking off negotiations above. The language of our society recognizes states or actions like the following: entering into negotiation, breaking off negotiations, offering to negotiate, negotiating in good (bad) faith, concluding negotiations, making a new offer, and so on. In other more jargon-infested language, the semantic 'space' of this range of social activity is carved up in a certain way, by a certain set of distinctions which our vocabulary marks; and the shape and nature of these distinctions is the nature of our language in this area. These distinctions are applied in our society with more or less formalism in different contexts. But of course this is not true of every society. Our whole notion of negotiation is bound up for instance with the distinct identity and autonomy of the parties, with the willed nature of their relations; it is a very contractual notion. But other societies have no such conception. It is reported about the traditional Japanese village that the foundation of its social life was a powerful form of consensus, which put a high premium on unanimous decision.9 Such a consensus would be considered shattered if two clearly articulated parties were to separate out, pursuing opposed aims and attempting either to vote down the opposition or push it into a 9 See. Thomas C. Smith, The Agrarian Origins of Modern Japan (Stanford, 1959), chap. 5. ThiS type of consensus is also found in other traditional societies. See, for instance, the desa system of the Indonesian village. settlement on the most favourable possible terms for themselves. Discussion there must be, and some kind of adjustment of differences. But our idea of bargaining, with the assumption of distinct autonomous parties in willed relationship, has no placei:here; nor does a series of distinctions, like entering into and leaving negotiation, or bargaining in good faith (sc. with the genuine intention of seeking agreement). Now difference between our society and one of the kind just described could not be well expressed if we said we have a vocabulary to describe negotiation which they lack. We might say, for instance, that we have a vocabulary to describe the heavens that they lack, viz., that of Newtonian mechanics; for here we assume that they live under the same heavens as we do, only understand it differently. But it is not true that they have the same kind of bargaining as we do. The word, or whatever word of their language we translate as 'bargaining', must have an entirely different gloss, which is marked by the distinctions their vocabulary allows in contrast to those marked by ours. But this different gloss is not just a difference of vocabulary, but also one of social reality. But this still may be misleading as a way of putting the difference. For it might imply that there is a social reality which can be discovered in each society and which might exist quite independently of the vocabulary of that society, or indeed of any vocabulary, as the heavens would exist whether men theorized about them or not. And this is not the case; the realities here are practices; and these cannot be identified in abstraction from the language we use to describe them, or invoke them, or carry them out. That the practice of negotiation allows us to distinguish bargaining in good or bad faith, or entering into or breaking off negotiations, presupposes that our acts and situation have a certain description for US, for example, that we are distinct parties entering into willed relations. But they cannot have these descriptions for us unless this is somehow expressed in our vocabulary of this practice; if not in our descriptions of the practices (for we may as yet be unconscious of some of the important distinctions), in the appropriate language for carrying them on. (Thus, the language marking a distinction between public and private acts or contexts may exist even where these terms or their equivalents are not part of this language; for the distinction will be marked by the different language which is appropriate in one context and the other, be it perhaps a difference of style, or dialect, even though the distinction is not designated by specific descriptive expressions.) The situation we have here is one in which the vocabulary of a given social dimension is grounded in the shape of social practice .in this 34 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 35 dimension; that is, the vocabulary would not make sense, could not be applied sensibly, where this range of practices did not prevail. And yet this range of practices could not exist without the prevalence of this or some related vocabulary. There is no simple one-way dependence here. We can speak of mutual dependence if we like, but really what this points up is the artificiality of the distinction between social reality and the language of description of that social reality. The language is constitutive of the reality, is essential to its being the kind of reality it is. To separate the two and distinguish them as we quite rightly distinguish the heavens from our theories about them is forever to miss the point. This type of relation has been recently explored, for instance by John Searle, with his concept of a constitutive rule. As Searle points out,IO we are normally induced to think of rules as applying to behaviour which could be available to us whether or not the rule existed. Some rules are like this, they are regulative like commandments: do not take the goods of another. But there are other rules, for example, those governing the Queen's move in chess, which are not so separable. If one suspends these rules, or imagines a state in which they have not yet been introduced, then the whole range of behaviour in question, in this case chess playing, would not be. There would still, of course, be the activity of pushing a wooden piece around on a board made of eight squares by eight; but this is not chess any longer. Rules of this kind are constitutive rules. By contrast again, there are other rules of chess, such as that one say 'J'adoube' when one touches a piece without intending to play it, which are clearly regulative. II I am suggesting that this notion of the constitutive be extended beyond the domain of rule-governed behaviour. That is why 1suggest the vaguer word 'practice'. Even in an area where there are no clearly defined rules, there are distinctions between different sorts of behaviour such that one sort is considered the appropriate form for one action or context, the other for another action or context; for example doing or saying certain things amounts to breaking off negotiations, doing or saying other things amounts to making a new offer. But just as there are constitutive rules, that is rules such that the behaviour they govern could not exist without them, and which are in this sense inseparable from that behaviour, so I am suggesting that there are constitutive distinctions, constitutive ranges of 10 Searle, Speech Acts, pp. 33-42 • 11 See the discussion in Stanley Cavell, Must We Mean What We Say? (New York, 1969), pp. 21-3 I. language which are similarly inseparable, in that certain practices are not without them. We can reverse this relationship and say that all the institutions and practices by which we live are constituted by certain distinctions and hence a certain language which is thus essential to them. We can take voting, a practice which is central to large numbers of institutions in a democratic society. What is essential to the practice of voting is that some decision or verdict be delivered (a man elected, a measure passed), through some criterion of preponderance (simple majority, two-thirds majority, or whatever) out of a set of micro-choices (the votes of the citizens, MPs, delegates). If there is not some such significance attached to our behaviour, no amount of marking and counting pieces of paper, raising hands, or walking out into lobbies amounts to voting. From this it follows that the institution of voting must be such that certain distinctions have application: for example, that between someone being elected, or a measure passed, and their failing of election, or passage; that between a valid vote and an invalid one which in turn requires a distinction between a real choice and one which is forced or counterfeited. For no matter how far we move from the Rousseauian notion that each man decide in full autonomy, the very institution of the vote requires that in some sense the enfranchised choose. For there to be voting in a sense recognizably like ours, there must be a distinction in men's self-interpretations between autonomy and forced choice. This is to say that an activity of marking and counting papers has to bear intentional descriptions which fall within a certain range before we can agree to call it voting, just as the intercourse of two men or teams has to bear descriptions of a certain range before we will call it negotiation. Or in other words, that some practice is voting or negotiation has to do in part with the vocabulary established in a society as appropriate for engaging in it or describing it. Hence implicit in these practices is a certain vision of the agent and his relation to others and to society. We saw in connection with negotiation in our society that it requires a picture of the parties as in some sense autonomous, and as entering into willed relations. And this picture carries with it certain implicit norms, such as that of good faith mentioned above, or a norm of rationality, that agreement correspond to one's goals as far as attainable, or the norm of continued freedom of action as far as attainable. These practices require that one's actions and relations be seen in the light of this picture and the accompanying norms, good faith, autonomy, and rationality. But men do not see themselves in PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATiON AND THE SCIENCES OF MAN 37 this way in all societies, nor do they understand these norms in all societies. The experience of autonomy as we know it, the sense of rational action and the satisfactions thereof, are unavailable to them. The meaning of these terms is opaque to them because they have a different structure of experiential meaning open to them. We can think of the difference between our society and the simplified version of the traditional Japanese village as consisting in this, that the range of meaning open to the members of the two societies is very different. But what we are dealing with here is not subjective meaning which can fit into the categorial grid of behavioural political science, but rather intersubjective meanings. It is not just that the people in our society all or mostly have a given set of ideas in their heads and subscribe to a given set of goals. The meanings and norms implicit in these practices are not just inthe minds of the actors but are out there in the practices themselves, practices which cannot be conceived as a set of individual actions, but which a/ie essentially modes of social relation, of mutual action. . The actors may have all sorts of beliefs and attitudes whith may be nghtly thought of as their individual beliefs and attitudes, even if others share them; they may subscribe to certain policy goals or certain forms of th~oryabout the polity, or feel resentment at certain things, and so on. They bnng these with them into their negotiations, and strive to satisfy them. But what they do not bring into the negotiations is the set of ideas and norms constitutive of negotiation themselves. These must be the common ~ropertyo~ th.e society before there can be any question of anyone entering mto negotiatIOn or not. Hence they are not subjective meanings, the pr~pertyof one or some individuals, but rather inter-subjective meanings, whICh are constitutive of the social matrix in which individuals find themselves and act. The inter-subjective meanings which are the background to social action ar~ ~ften treated by political scientists under the heading 'consensus'. By thiS IS meant convergence of beliefs on certain basic matters, or of attitude. But the two are not the same. Whether there is consensus or not the condition of there being either one or the other is a certain set of comm~nterms of reference. A society in which this was lacking would not be a society in the normal sense of the term, but several. Perhaps some multi-racial or multitribal states approach this limit. Some multi-national states are bedevilled by consistent cross-purposes, e.g., my own country, Canada. But consensus as a convergence of beliefs or values is not the opposite of this kind of fundamental diversity. Rather the opposite of diversity is a high degree of inter-subjective meanings. And this can go along with profound cleavage. Indeed, inter-subjective meanings are a condition of a certain kind of very profound cleavage, such as was visible in the Reformation, or the American Civil War, or splits in left wing parties, where the dispute is at fever pitch just because each side can fully understand the other. In other words, convergence of belief or attitude or its absence presupposes a common language in which these beliefs can be formulated, and in which these formulations can be opposed. Much of this common language in any society is rooted in its institutions and practices; it is constitutive of these institutions and practices. It is part of the inter-subjective meanings. To put the point another way, apart from the question of how much people's beliefs converge is the question of how much they have a common language of social and political reality in which these beliefs are expressed. This second question cannot be reduced to the first; inter-subjective meaning is not a matter of converging beliefs or values. When we speak of consensus we speak of beliefs and values which could be the property of a single person, or many, or all; but inter-subjective meanings could not be the property of a single person because they are rooted in social practice. We can perhaps see this if we envisage the situation in which the ideas and norms underlying a practice are the property of single individuals. This is what happens when single individuals from one society interiorize the notions and values of another, for example children in missionary schools. Here we have a totally different situation. We are really talking now about subjective beliefs and attitudes. The ideas are abstract, they are mere social 'ideals'. Whereas in the original society, these ideas and norms are rooted in their social relations, and they can formulate opinions and ideals on the basis of them. We can see this in connection with the example we have been using all along, that of negotiations. The vision of a society based on negotiation is coming in for heavy attack by a growing segment of modern youth, as are the attendant norms of rationality and the definition of autonomy. This is a dramatic failure of 'consensus'. But this cleavage takes place in the ambit of this inter-subjective meaning, the social practice of negotiation as it is lived in our society. The rejection would not have the bitter quality it has if what is rejected were not understood in common, because it is part of a social practice which we find hard to avoid, so pervasive is it in our society. At the same time there is a reaching out for other forms which stilI have the 'abstract' quality of ideals which are subjective in this sense, that is, not rooted in practice; which is what makes the rebellion look so 'unreal' to outsiders, and so irrational. PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 39 3 Inter-subjective meanings, ways of experiencing action in society which are expressed in the language and descriptions constitutive of institutions and practices, do not fit into the categorial grid of mainstream political science. This allows only for an inter-subjective reality which is brute data identifiable. But social practices and institutions which are partly constituted by certain ways of talking about them are not so identifiable. We have to understand the language, the underlying meanings, which constitute them. We can allow, once we accept a certain set of institutions or practices as our starting point and not as objects of further questioning, that we can easily take as brute data that certain acts are judged to take place or certain states judged to hold within the semantic field of these practices; for instance, that someone has voted Liberal, or signed the petition. We can then go on to correlate certain subjective meanings - beliefs, attitudes, etc. - with this behaviour or its lack. But this means that we give up trying to define further just what these practices and institutions are, what the meanings are which they require and hence sustain. For these meanings do not fit into the grid; they are not subjective beliefs or values, but are constitutive of social reality. In order to get at them we have to drop the basic premise that social reality is made up of brute data alone. For any characterization of the meanings underlying these practices is open to question by someone offering an alternative interpretation. The negation of this is what was meant as brute data. We have to admit that inter-subjective social reality has to be partly defined in terms of meanings; that meanings as subjective are not just in causal interaction with a social reality made up of brute data, but that as inter-subjective they are constitutive of this reality. We have been talking here of inter-subjective Ill,eaning. Earlier I was contrasting the question of inter-subjective meaning with that of consensus as convergence of opinions. But there is another kind of non-subjective meaning which is also often inadequately discussed under the head of 'consensus'. In a society with a strong web of inter-subjective meanings, there can be a more or less powerful set of common meanings. By these I mean notions of what is significant, which are not just shared in the sense that everyone has them, but are also common in the sense of being in the common reference world. Thus, almost everyone in our society may share a susceptibility to a certain kind of feminine beauty, but this may not be a common meaning. It may be known to no one, except perhaps market researchers, who play on it in their advertisements. But the survival of a national identity as francophones is a common meaning of Quebecois; for it is not just shared, and not just known to be shared, but its being a common aspiration is one of the common reference points of all debate, communication, and all public life in the society. We can speak of a shared belief, aspiration, etc. when there is convergence between the subjective beliefs, aspirations, of many individuals. But it is part of the meaning of a common aspiration, belief, celebration, etc. that it be not just shared but part of the common reference world. Or to put it another way, its being shared is a collective act, it is a consciousness which is communally sustained, whereas sharing is something we do each on his own, as it were, even if each of us is influenced by the others. Common meanings are the basis of community. Inter-subjective meaning gives a people a common language to talk about social reality and a common understanding of certain norms, but only with common meanings does this common reference world contain significant common actions, celebrations, and feelings. These are objects in the world that everybody shares. This is what makes community. Once again, we cannot really understand this phenomenon through the usual definition of consensus as convergence of opinion and value. For what is meant here is something more than convergence. Convergence is what happens when our values are shared. But what is required for common meanings is that this shared value be part of the common world, that this sharing be shared. But we could also say that common meanings are quite other than consensus, for they can subsist with a high degree of cleavage; this is what happens when a common meaning comes to be lived and understood differently by different groups in a society. It remains a common meaning, because there is the reference point which is the common purpose, aspiration, celebration. Such is for example the American Way, or freedom as understood in the USA. But this common meaning is differently articulated by different groups. This is the basis of the bitterest fights in a society, and this we are also seeing in the USA today. Perhaps one might say that a common meaning is very often the cause of the most bitter lack of consensus. It thus must not be confused with convergence of opinion, value, attitude. Of course, common meanings and inter-subjective meanings are closely interwoven. There must be a powerful net of inter-subjective meanings for there to be common meanings; and the result of powerful common meanings is the development of a greater web of inter-subjective meanings as people live in community. On the other hand, when common meanings wither, which they can do 4° PHI-LOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN through the kind of deep dissensus we described earlier, the groups tend to grow apart and develop different languages of social reality, hence to share fewer inter-subjective meanings. To take our above example again, there has been a powerful common meaning in our civilization around a certain vision of the free society in which bargaining has a central place. This has helped to entrench the social practice of negotiation which makes us participate in this intersubjective meaning. But there is a severe challenge to this common meaning today, as we have seen. Should those who object to it really succeed in building up an alternative society, there would develop a gap between those who remain in the present type of society and those who had founded the new one. Common meanings, as well as inter-subjective ones, fall through the net of mainstream social science. They can find no place in its categories. For they are not simply a converging set of subjective reactions, but part of the common world. What the ontology of mainstream social science lacks is the notion of meaning as not simply for an individual subject; of a subject who can be a 'we' as well as an '1'. The exclusion of this possibility, of the communal, comes once again from the baleful influence of the epistemological tradition for which all knowledge has to be reconstructed from the impressions imprinted on the individual subject. But if we free ourselves from the hold of these prejudices, this seems a wildly implausible view about the development of human consciousness; we are aware of the world through a 'we' before we are through an '1'. Hence we need the distinction between what is just shared in the sense that each of us has it in our individual worlds and, that which is in the common world. But the very idea of something which is in the common world in contradistinction to what is in all the individual worlds is totally opaque to empiricist epistemology. Hence it finds no place in mainstream social science. What this results in must now be seen. III I Thus, to sum up the last pages: a social science which wishes to fulfil the requirements of the empiricist tradition naturally tries to reconstruct social reality as consisting of brute data alone. These data are the acts of people (behaviour) as identified supposedly beyond interpretation either by physical descriptions or by descriptions clearly defined by institutions and practices; and secondly, they include th~ subjective reality of individuals' beliefs, attitudes, values, as attested by their responses to certain forms of words, or in some cases their overt non-verbal behaviour. What this excludes is a conside~ation of social reality as characterized by inter-subjective and common meanings. It excludes, for instance, an attempt to understand our civilization, in which negotiation plays such a central part both in fact and in justificatory theory, by probing the selfdefinitions of agent, other and social relatedness which it embodies. Such definitions which deal with the meaning for agents of their own and others' action, and of the social relations in which they stand, do not in any sense record brute data, in the sense that this term is being used in this argument; that is, they are in no sense beyond challenge by those who would quarrel with our interpretations of these meanings. Thus, I tried to adumbrate above the vision implicit in the practice of negotiation by reference to certain notions of autonomy and rationality. But this reading will undoubtedly be challenged by those who have differe~t fundamental conceptions of man, human motivation, the human condition; or even by those who judge other features of our present predicament to have greater importance. If we wish to avoid these disputes, and have a science grounded in verification as this is understood by the logical empiricists, then we have to avoid this level of study altogether and hope to make do with a correlation of behaviour which is brute data identifiable. A similar point goes for the distinction between common meanings and shared subjective meanings. We can hope to identify the subjective meanings of individuals if we take these in the sense in which there are adequate criteria for them in people's dissent or assent to verbal formulae or their brute data indentifiable behaviour. But once we allow the distinction between such subjective meanings which are widely shared and genuine common meanings, then we can no longer make do with brute data identification. We are in a domain where our definitions can be challenged by those with another reading. The profound option of mainstream social scientists for the empiricist conception of knowledge and science makes it inevitable that they should accept the verification model of political science and the categorial principles that this entails. This means in turn that a study of our civilization in terms of its inter-subjective and common meanings is ruled out. Rather this whole level of study is made invisible. On the mainstream view, therefore, the different practices and institutions of different societies are not seen as related to different clusters of PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 43 inter-subjective or common meanings; rather, we should be able to differentiate them by different clusters of 'behaviour' and/or subjective meaning. The comparison between societies requires on this view that we elaborate a universal vocabulary of behaviour which will allow us to present the different forms and practices of different societies in the same conceptual web. Now present-day political science is contemptuous of the older attempt at comparative politics via a comparison of institutions. An influential school of our day has therefore shifted comparison to certain practices, or very general classes of practices, and proposes to compare societies according to the different ways in which these practices are carried on. Such are the 'functions' of the influential 'developmental approach'.u But it is epistemologically crucial that such functions be identified independently of those inter-subjective meanings which are different in different societies; for otherwise, they will not be genuinely universal; or will be universal only in the loose and unilluminating sense that the function-name can be given application in every society but with varying, and often widely varying, meaning - the same being 'glossed' very differently by different sets of practices and inter-subjective meanings. The danger that such universality might not hold is not even suspected by mainstream political scientists since they are unaware that there is such a level of description as that which defines intersubjective meanings, and are convinced that functions and the various structures which perform them can be identified in terms of brute data behaviour. But the result of ignoring the difference in inter-subjective meanings can be disastrous to a science of comparative politics, viz., that we interpret all other societies in the categories of our own. Ironically, this is what seems to have happened to American political science. Having strongly criticized the old institution-focussed comparative politics for its ethnocentricity (or Western bias), it proposes to understand the politics of all society in terms of such functions, for instance, as 'interest articulation' and 'interest aggregation' whose definition is strongly influenced by the bargaining culture of our civilization, but which is far from being guaranteed appropriateness elsewhere. The not surprising result is a theory of political development which places the Atlantic-type polity at the summit of human political achievement. Much can be said in this area of comparative politics (interestingly 12 See Almond and Powell, Comparative Politics. explored by Alasdair MacIntyre).J3 But I would like to illustrate the significance of these two rival approaches in connection with another common problem area of politics: This is the question of what is called 'legitimacy'.14 2 It is an obvious fact, with which politics has been concerned since at least Plato, that some societies enjoy an easier, more spontaneous cohesion which relies less on the use of force than others. It has been an important question of political theory to understand what underlies this differen.ce. Among others, Aristotle, Machiavelli, Montesquieu, and Tocqueville have dealt with it. Contemporary mainstream political scientists approach this question with the concept 'legitimacy'. The use of the word here can be easily understood. Those societies which are more spontaneously cohesive can be thought to enjoy a greater sense of legitimacy among their members. But th~ application of the term has been shifted. 'Legitimacy' is a term in which we discuss the authority of the state or polity, its right to our allegiance. However we conceive of this legitimacy, it can only be attributed to a polity in the light of a number of surrounding conceptions _ for example, that it provides men freedom, that it emanates from their will, that it secures them order, the rule of law, or that it is founded on tradition, or commands obedience by its superior qualities. These conceptions are all such that they rely on definitions of what is significant for men in general or in some particular society or circumstances, definitions of paradigmatic meaning which cannot be identifiable as brute data. Even where some of these terms might be given an 'operational definition' in terms of brute data - a term like 'freedom', for instance, can be defined in terms of the absence of legal restriction, afa Hobbes - this definition would not carry the full force of the term, and in particular that whereby it could be considered significant for men. According to the empiricist paradigm, this latter aspect of the meaning of such a term is labelled 'evaluative' and is thought to be utterly heterogeneous from the 'descriptive' aspect. But this analysis is far from firmly established; no more so in fact than the empiricist paradigm of knowledge itself with which it is closely bound up. A challenge to this 13 'How is a comparative science of politics possible?', in Alasdair Mcintyre, Against the Self-Images ofthe Age (London, 1971). 14 Macintyre's article also contains an interesting discussion of 'legitimacy' from a different, although I think related, angle. 44 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 45 paradigm in the name of a hermeneutical science is also a challenge to the distinction between 'descriptive' and 'evaluative' and the entire conception of Wertfrieheit which goes with it. In any case, whether because it is 'evaluative' or can only be applied in connection with definitions of meaning, 'legitimate' is not a word which can be used in the description of social reality according to the conceptions of mainstream social science. It can only be used as a description of subjective meaning. What enters into scientific consideration is thus not the legitimacy of a polity but the opinions or feelings of its member individuals concerning its legitimacy. The differences between different societies in their manner of spontaneous cohesion and sense of community are to be understood by correlations between the beliefs and feelings of their members towards them on one hand and the prevalence of certain brute data identifiable indices of stability in them on the other. Thus Robert Dahl in Modern Political Analysisl5 speaks of the different ways in which leaders gain 'compliance' for their policies. The more citizens comply because of 'internal rewards and deprivations', the less leaders need to use 'external rewards and deprivations'. But if citizens believe a government is legitimate, then their conscience will bind them to obey it; they will be internally punished if they disobey; hence government will have to use less external resources, including force. Less crude is the discussion of Seymour Lipset in Political Man. 16 But it is founded on the same basic ideas, viz. that legitimacy defined as subjective meaning is correlated with stability. 'Legitimacy involves the capacity of the system to engender and maintain the belief that the existing political institutions are the most appropriate ones for the society.'17 Lipset is engaged in a discussion of the determinants of stability in modern polities. He singles out two important ones in this chapter, effectiveness and legitimacy. 'Effectiveness means actual performance, the extent to which the system satisfies the basic functions of government as most of the population and such powerful groups within it as big business or the armed forces see them.'18 Thus we have one factor which has to do with objective reality, what the government has actually done; and the other which has to do with subjective beliefs and 'values'. 'While effectiveness is primarily instrumental, legitimacy is evaluative.'19 Hence 15 Foundation of Modern Political Science Series (Englewood Cliffs, 1963), pp. 31-2. 16 New York, 1963, chap. 3. 17 Ibid., p. 64. 18 Ibid. 19 Ibid. from the beginning the stage is set by a distInction between social reality and what men think and feel about it. Lipset sees two types of crisis of legitimacy that modern societies have affronted more or less well. One concerns the status of major conservative institutions which may be under threat from the development of ~odern industrial democracies. The second concerns the degree to which all political groups have access to the political process: Thus, under th~ first head, some traditional groups, such as landed arIstocracy or clerIcals, have been roughly handled in a society like France, and have remained alienated from the democratic system for decades afterwards; whereas in England the traditional classes were more gently handled, themselves were willing to compromise and have been slowly integrated and transformed into the new order. Under the second head, some societies managed to integrate the working class or bourgeoisie into the politi~al pr.ocess at an early stage, whereas in others they have been kept out till qUite recently, and consequently have developed a deep sense of alienation from the sy~tem, have tended to adopt extremist ideologies, and have generally contributed to instability. One of the determinants of a society's performance on these two heads is whether or not it is forced to affront the different conflicts of democratic development all at once or one at a time. Another important determinant of legitimacy is effectiveness. This approach, which sees stability as partly the result of legitimacy beliefs and these in turn as resulting partly from the way the status, welfare, and access to political life of different groups fare, seems at first blush eminently sensible and well designed to help us understand the history of the last century or two. But this approach has no place for a study of the inter-subjective and common meanings which are constitutive of modern civilization. And we may doubt whether we can understand the cohesion of modern societies or their present crisis if we leave these out of account. Let us take the winning of the allegiance of the working class to the new industrial regimes in the nineteenth and twentieth century. This is far from being a matter simply or even perhaps most significantly of the speed with which this class was integrated into the political process and the effectiveness of the regime. Rather the consideration of the granting of access to the political process as an independent variable may be misleading. . It is not just that we often find ourselves invited by historians to account for class cohesion in particular countries in terms of other factors, such as the impact of Methodism in early nineteenth century PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 47 England (Elie Halevy)20 or the draw of Germany's newly successful nationalism. These factors could be assimilated to the social scientist's grid by being classed as 'ideologies' or widely held 'value systems' or some other such concatenations of subjective meaning. But perhaps the most important such 'ideology' in accounting for the cohesion of industrial democratic societies has been that of the society of work, the vision of society as a large-scale enterprise of production in which widely different functions are integrated into interdependence; a vision of society in which economic relations are considered as primary, as it is not only in Marxism (and in a sense not really with Marxism) but above all with the tradition of Classical Utilitarianism. In line with this vision there is a fundamental solidarity between all members of society who labour (to use Arendt's language),21 for they are all engaged in producing what is indispensable to life and happiness in far-reaching interdependence. This is the 'ideology' which has frequently presided over the integration of the working class into industrial democracies, at first directed polemically against the 'unproductive' classes, for example in England with the anti-Corn Law League, and later with the campaigns of Joseph Chamberlain ('when Adam delved and Eve span, who was then the gentleman?'), but later as a support for social cohesion and solidarity. But, of course, the reason for putting 'ideology' in quotes above is that this definition of things, which has been well integrated with the conception of social life as based on negotiation, cannot be understood in the terms of mainstream social science, as beliefs and 'values' held by a large number of individuals. For the great interdependent matrix of labour is not just a set of ideas in people's heads but is an important aspect of the reality which we live in modern society. And at the same time, these ideas are embedded in this matrix in that they are constitutive of it; that is, we would not be able to live in this type of society unless we were imbued with these ideas or some others which could call forth the discipline and voluntary coordination needed to operate this kind of economy. All industrial civilizations have required a huge wrench from the traditional peasant populations on which they have been imposed; for they require an entirely unprecedented level of disciplined sustained, monotonous effort, long hours unpunctuated by any meaningful rhythm, such as that of seasons or festivals. In the end this way of life can only be accepted 20 Histoire du peuple anglais au XIXe siecle (Paris, 1913). 21 The Human Condition (New York, 1959). when the idea of making a living is endowed with more significance than that of just avoiding starvation; and this it is in the civilization of labour. Now this civilization of work is only one aspect of modern societies, along with the society based on negotiation and willed relations (in Anglo-Saxon countries), and other common and inter-subjective meanings which have different importance in different countries. My point is that it is certainly not implausible to say that it has some importance in explaining the integration of the working class in modern industrial democratic society. But it can only be called a cluster of inter-subjective meaning. As such it cannot come into the purview of mainstream political science; and an author like Lipset cannot take it into consideration when discussing this very problem. But, of course, such a massive fact does not escape notice. What happens rather is that it is re-interpreted. And what has generally happened is that the interdependent productive and negotiating society has been recognized by political science, but not as one structure of inter-subjective meaning among others, rather as the inescapable background of social action as such. In this guise it no longer need be an object of study. Rather it retreats to the middle distance, where its general outline takes the role of universal framework, within which (it is hoped) actions and structures will be brute data identifiable, and this for any society at any time. The view is then that the political actions of men in all societies can be understood as variants of the processing of 'demands' which is an important part of our political life. The inability to recognize the specificity of our inter-subjective meanings is thus inseparably linked with the belief in the universality of North Atlantic behaviour types or 'functions' which vitiates so much of contemporary comparative politics. The notion is that what politics is about perennially is the adjustment of differences, or the production of symbolic and effective 'outputs' on the basis of demand and support 'inputs'. The rise of the inter-subjective meaning of the civilization of work is seen as the increase of correct perception of the political process at the expense of 'ideology'. Thus Almond and Powell introduce the concept of 'political secularization' to describe 'the emergence of a pragmatic, empirical orientation' to politics.22 A secular political culture is opposed not only to a traditional one, but also to an 'ideological' culture, which is characterized by 'an inflexible image of political life, closed to conflicting information' and 'fails to develop the open, bargaining attitudes associated with full 22 Comparative Politics, p. 58. 24 Thus Lewis Feuer in The Conflict ofGenerations (New York, 1969), attempts to account for the 'misperception of social reality" in the Berkeley student uprising of 1968 In terms of a generational conflict (pp. 466-70), which in ~urn isrooted In the psychology of adolescence and attaining adulthood. Yet Feuer himself m hiS first chapter notes the comparative recency of self-defining political generations, a phenomenon which dates from the post-Napoleonic era (p. 33). But an adequate attempt to explam thiS hlstoncal shift which after all underlay the Berkeley rising and many others, would I beheve have to take us beyond the ambit of individual psychology to psycho-h.istory, to a ~tudy of t~e intrication of psychological conflict and inter-subjective meanmgs. A vanant of thiS form of study has been adumbrated in the work of Erik Erikson. . ty cannot be interpreted as bargaining "gambits in the name of anySOCle . demands, old or new. These can only be interpreted withm the accepted f ework of our social science as a return to ideology, and hence as .ra~ional. Now in the case of some of the more bizarre and bloody forms lrra b . d d' . 1 of protest, there will be little disagree~e~t; they will e JU ge lff.atiOn~ b all but their protagonists. But wlthm the accepted categones thiS i7rationality can only be understood in terms of individual psychology; it is the public eruption of private pathology; it ~ann.ot be un~er~tood as a malady of society itself, a malaise which afflicts Its constitutive mean~~M . No one can claim to begin to have an adequate explanatiOn for these major changes which our civilization is undergoing. But in contrast t.o the incapacity of a science which remains within the accepted cate.gones, a hermeneutical science of man which has a place for a study of l~tersub. ctive meaning can at least begin to explore fruitful avenues. Plamly the ~isciplinewhich was integral to the civilization of work and bargaining is beginning to fail. The structures of this civilization, interdepend~nt.work, bargaining, mutual adjustment of individual ends, are begmnmg to change their meaning for many, and are beginning to be felt not as normal and best suited to man, but as hateful or empty. And yet we are all caught in these inter-subjective meanings in so far as we live in this society, and in a sense more and more all-pervasively as it progresses. Hence the virulence and tension of the critique of our society which is always in some real sense a self-rejection (in a way that the old socialist opposition never was). Why has this set of meanings gone sour? Plainly, we have to accept t.hat they are not to be understood at their face value. The free, productive, bargaining culture claimed to be sufficient for man. If it was not, then we have to assume that while it did hold our allegiance, it also had other meanings for us which commanded this allegiance and which have now gone. PHILOSOPHY AND SOCIAL SCIENCE secularization'.13 The clear understanding here is that a secularized culture is one which essentially depends less on illusion, which sees things as they are, which is not infected with the 'false consciousness' of traditional or ideological culture (to use a term which is not in the mainstream vocabulary). 3 This way of looking at the civilization of work, as resulting from the retreat of illusion before the correct perception of what politics perennially and really is, is thus closely bound up with the epistemological premises of mainstream political science and its resultant inability to recognize the historical specificity of this civilization's inter-subjective meanings. But the weakness of this approach, already visible in the attempts to explain the rise of this civilization and its relation to others becomes even more painful when we try to account for its present malaise: even crisis. The strains in contemporary society, the breakdown of civility, the rise of deep alienation, which is translated into even more destructive action tend to shake the basic categories of our social science. It is not just tha; such a development was quite unpredicted by this science, which saw in the rise of affluence the cause rather of a further entrenching of the bargaining culture, a reduction of irrational cleavage, an increase of tolerance, in short 'the end of ideology'. For prediction, as we shall see below, cannot be a goal of social science as it is of natural science. It is rather that this mainstream science hasn't the categories to explain this breakdown. It is forced to look on extremism either as a bargaining gambit of the desperate, deliberately raising the ante in order to force a hearing. Or, alternatively, it can recognize the novelty of the rebellion by accepting the hypothesis that heightened demands are being made on the system owing to a revolution of 'expectations', or else to the eruption of new desires or aspirations which hitherto had no place in the bargaining process. But these new desires or aspirations must be in the domain of individual psychology, that is, they must be such that their arousal and satisfaction is to be understood in terms of states of individuals rather than in terms of the inter-subjective meanings in which they live. For these latter have no place in the categories of the mainstream, which thus cannot accommodate a genuine historical psychology. But some of the more extreme protests and acts of rebellion in our 2.l Ibid.,p.6r. INTERPREtATION AND THE SCIENCES OF MAN 49 5° PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN This is the starting point of a set of hypotheses which attempt to redefine our past in order to make our present and future intelligible. We might think that the productive, bargaining culture offered in the past common meanings (even though there was no place for them in its philosophy), and hence a basis for community, which were essentially linked with its being in the process of building. It linked men who could see themselves as breaking with the past to build a new happiness in America, for instance. But in all essentials that future is built; the notion of a horizon to be attained by future greater production (as against social transformation) verges on the absurd in contemporary America. Suddenly the horizon which was essential to the sense of meaningful purpose has collapsed, which would show that like so many other Enlightenment-based dreams the free, productive, bargaining society can only sustain man as a goal, not as a reality. Or we can look at this development in terms of identity. A sense of building their future through the civilization of work can sustain men as long as they see themselves as having broken with a millennial past ofinjustice and hardship in order to create qualitatively different conditions for their children. All the requirements ofa humanly acceptable identity can be met by this predicament, a relation to the past (one soars above it but preserves it in folkloric memory), to the social world (the inter-dependent world of free, productive men), to the earth (the raw material which awaits shaping), to the future and one's own death (the everlasting monument in the lives of prosperous children), to the absolute (the absolute values of freedom, integrity, dignity). But at some point the children will be unable to sustain this forward thrust into the future. This effort has placed them in a private haven of security, within which they are unable to reach and recover touch with the great realities: their parents have only a negated past, lives which have been oriented wholly to the future; the social world is distant and without shape; rather one can only insert oneself into it by taking one's place in the future-oriented productive juggernaut. But this now seems without any sense; the relation to the earth as raw material is therefore experienced as empty and alienating, but the recovery of a valid relation to the earth is the hardest thing once lost; and there is no relation to the absolute where we are caught in the web of meanings which have gone dead for us. Hence past, future, earth, world, and absolute are in some way or another occluded; and what must arise is an identity crisis of frightening proportions. These two hypotheses are mainly focussed on the crisis in US civilization and they would perhaps help account for the fact that the USA is in som~ sense going first through this crisis of all Atlantic nations; not, that is, only because it is the most affluent, but more because it has ,been more ~ully based on the civilization of work than European countnes who retamed omething of more traditional common meanings. s But they might also help us to understand why alienation is ~o~t sever,e among groups which have been but marginal in ~fflu~nt ~~r~al~mg SO~Ieties. These have had the greatest strain in living m thiS cIVIlization ,while their identity was in some ways antithetical to it. Such are blacks m the USA, and the community of French-speaking Canadians, each in different ays. For many immigrant groups the strain was also great, but they ~rced themselves to surmount the obstacles, and the new identity is sealed in the blood of the old, as it were. But for those who would not or could not succeed in thus transforming themselves, but always lived a life of strain on the defensive, the breakdown of the central, powerful identity is the trigger to a deep turnover. It can be thought of as a liberation but at the same time it is deeply unsettling, because the basic parameters of former life are being changed and there are not yet the new images and definitions to live a new fully acceptable identity. In a sense we are in a condition where a new social compact (rather the first social compact) has to be made between these groups and those they live with, and no one knows where to start. In the last pages, I have presented some hypotheses which may appear very speculative; and they may indeed turn out to be without foundation, even without much interest. But their aim was mainly illustrative. My principal claim is that we can only come to grips with this phenomenon of breakdown by trying to understand more clearly and profoundly the common and inter-subjective meanings of the society in which we have been living. For it is these which no longer hold us, and to understand this change we have to have an adequate grasp of these meanings. But this we cannot do as long as we remain within the ambit of mainstream social science, for it will not recognize inter-subjective meaning, and is forced to look at the central meanings of our society as though they were the inescapable background of all political action. Breakdown is thus inexplicable in political terms; it is an outbreak of irrationality which must ultimately be explained by some form of psychological illness. Mainstream science may thus venture into the area explored by the above hypotheses, but after its own fashion, by forcing the psychohistorical facts of identity into the grid of an individual psychology, in short, by re-interpreting all meanings as subjective. The result might be a psychological theory of emotional maladjustment, perhaps traced to certain features of family background, analogous to the theories of the PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 53 authoritarian personality and the California F-scale. But this would no longer be a political or social theory. We would be giving up the attempt to understand the change in social reality at the level of its constitutive inter-subjective meanings. IV It can be argued, then, that mainstream social science is kept within certain limits by its categorial principles which are rooted in the traditional epistemology of empiricism; and secondly that these restrictions are a severe handicap and prevent us from coming to grips with important problems of our day which should be the object of political science. We need to go beyond the bounds of a science based on verification to one which would study the inter-subjective and common meanings embedded in social reality. But this science would be hermeneutical in the sense that has been developed in this paper. It would not be founded on brute data; its most primitive data would be readings of meanings, and its object would have the three properties mentioned above: the meanings are for a subject in a field or fields; they are moreover meanings which are partially constituted by self-definitions, which are in this sense already interpretations, and which can thus be re-expressed or made explicit by a science of politics. In our case, the subject may be a society or community; but the inter-subjective meanings, as we saw, embody a certain self-definition, a vision of the agent and his society, which is that of the society or community. But then the difficulties which the proponents of the verification model foresee will arise. Ifwe have a science which has no brute data, which relies on readings, then it cannot but move in a hermeneutical circle. A given reading of the inter-subjective meanings of a society, or of given institutions or practices, may seem well founded, because it makes sense of these practices or the development of that society. But the conviction that it does make sense of this history itself is founded on further related readings. Thus, what I said above on the identity crisis which is generated by our society makes sense and holds together only if one accepts this reading of the inter-subjective meanings of our society, and if one accepts this reading of the rebellion against our society by many young people (Sc. the reading in terms of identity crisis). These two readings make sense together, so that in a sense the explanation as a whole reposes on the readings, and the readings in their turn are strengthened by the explanation as a whole. But if these readings seem implausible, or even more, if they are not nderstood by our interlocutor, there is no verification procedure which :e can fall back on. We can only continue to offer interpretations; we are in an interpretative circle.· . But the ideal of a science of verification is to find an appeal beyond differences of interpretation. Insight will always be useful in discovery, but should not have to play any part in establishing the truth of its findings. This ideal can be said to have been met by our natural sciences. But a hermeneutic science cannot but rely on insight. It requires that one have the sensibility and understanding necessary to be able to make and comprehend the readings by which we can explain the reality concerned. In physics we might argue that if someone does not accept a true theory, then either he has not been shown enough (brute data) evidence (perhaps not enough is yet available), or he cannot understand and apply some formalized language. But in the sciences of man conceived as hermeneutical, the non-acceptance of a true or illuminating theory may come from neither of these, indeed is unlikely to be due to either of these, but rather from a failure to grasp the meaning field in question, an inability to make and understand readings of this field. In other words, in a hermeneutical science, a certain measure of insight is indispensable, and this insight cannot be communicated by the gathering of brute data, or initiation in modes of formal reasoning or some combination of these. It is unformalizable. But this is a scandalous result according to the authoritative conception of science in our tradition, which is shared even by many of those who are highly critical of the approach of mainstream psychology, or sociology, or political science. For it means that this is not a study in which anyone can engage, regardless of their level of insight; that some claims of the form: 'if you don't understand, then your intuitions are at fault, are blind or inadequate' will be justified; that some differences will be non-arbitrable by further evidence, but that each side can only make appeal to deeper insight on the part of the other. The superiority of one position over another will thus consist in this, that from the more adequate position one can understand one's own stand and that of one's opponent, but not the other way around. It goes without saying that this argument can only have weight for those in the superior position. Thus a hermeneutical science encounters a gap in intuitions, which is, the other side, as it were, of the hermeneutical circle. But the situation is graver than this; for this gap is bound up with our divergent options in politics and life. 54 PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 55 We speak of a gap when some cannot understand the kind of selfdefinition which others are proposing as underlying a certain society or set of institutions. Thus some positivistically minded thinkers will find the language of identity-theory quite opaque; and some thinkers will not recognize any theory which does not fit with the categorial presuppositions of empiricism. But self-definitions are not only important to us as scientists who are trying to understand some, perhaps distant, social reality. As men we are self-defining beings, and we are partly what we are in virtue of the self-definitions which we have accepted, however we have come by them. What self-definitions we understand and what ones we do not, is closely linked with the self-definitions which help to constitute what we are. If it is too simple to say that one only understands an 'ideology' which one subscribes to, it is nevertheless hard to deny that we have great difficulty grasping definitions whose terms structure the world in ways which are utterly different from or incompatible with our own. Hence the gap in intuitions doesn't just divide different theoretical positions, it also tends to divide different fundamental options in life. The practical and the theoretical are inextricably joined here. It may not just be that to understand a certain explanation one has to sharpen one's intuitions, it may be that one has to change one's orientation - if not in adopting another orientation, at least in living one's own in a way which allows for greater comprehension of others. Thus, in the sciences of man in so far as they are hermeneutical there can be a valid response to 'I don't understand' which takes the form, not only 'develop your intuitions', but more radically 'change yourself'. This puts an end to any aspiration to a value-free or 'ideology-free' science of man. A study of the science of man is inseparable from an examination of the options between which men must choose. This means that we can speak here not only of error, but of illusion. We speak of 'illusion' when we are dealing with something of greater substance than error, error which in a sense builds a counterfeit reality of its own. But errors of interpretation of meaning, which are also selfdefinitions of those who interpret and hence inform their lives, are more than errors in this sense: they are sustained by certain practices of which they are constitutive. It is not implausible to single out as examples two rampant illusions in our present society. One is that of the proponents of the bargaining society who can recognize nothing but either bargaining gambits or madness in those who rebel against this society. Here the error is sustained by the practices of the bargaining culture, and given a semblance of reality by the refusal to treat any protests on other terms; it hence acquires the more substantive reality of illusion. The second example is provided by much 'revolutionary' activity in our society which in desperate search for an alternative mode of life purports to see its situation in that of an Andean guerilla or Chinese peasants. Lived out, this passes from the stage of laughable error to tragic illusion. One illusion cannot recognize the possibility of human variation, the other cannot see any limits to man's ability'to transform itself. Both make a valid science of man impossible. Inface of all this, we might be so scandalized by the prospect of such a hermeneutical science, that we will want to go back to the verification model. Why can we not take our understanding of meaning as part of the logic of discovery, as the logical empiricists suggest for our unformalizable insights, and still found our science on the exactness of our predictions? Our insightful understanding of the inter-subjective meanings of our society will then serve to elaborate fruitful hypotheses, but the proof of these puddings will remain in the degree they enable us to predict. The answer is that if the epistemological views underlying the science of interpretation are right, such exact prediction is radically impossible, for three reasons of ascending order of fundamentalness. The first is the well-known 'open system' predicament, one shared by human life and meteorology, that we cannot shield a certain domain of human events, the psychological, economic, political, from external interference; it is impossible to delineate a closed system. The second, more fundamental, is that if we are to understand men by a science of interpretation, we cannot achieve the degree of fine exactitude of ascience based on brutedata. The data of natural science admit of measurement to virtually any degree of exactitude. But different interpretations cannot be judged in this way. At the same time different nuances of interpretation may lead to different predictions in some circumstances, and these different outcomes may eventually create widely varying futures. Hence it is more than easy to be wide of the mark. But the third and most fundamental reason for the impossibility of hard prediction is that man is a self-defining animal. With changes in his selfdefinition go changes in what man is, such that he has to be understood in different terms. But the conceptual mutations in human history can and frequently do produce conceptual webs which are incommensurable, that is, where the terms cannot be defined in relation to a common stratum of expressions. The entirely different notions of bargaining in our society and in some primitive ones provide an example. Each will be glossed in terms of practices, institutions, ideas in each society which have nothing corresponding to them in the other. PHILOSOPHY AND SOCIAL SCIENCE INTERPRETATION AND THE SCIENCES OF MAN 57 i The success of prediction in the natural sciences is bound up with the f fact that all states of the system, past and future, can be described in the t same range of concepts, as values, say, of the same variables. Hence all future states of the solar system can be characterized, as past ones are, in the language of Newtonian mechanics. This is far from being a sufficient condition of exact prediction, but it is a necessary one in this sense, that only if past and future are brought under the same conceptual net can one understand the states of the latter as some function of the states of the former, and hence predict. This conceptual unity is vitiated in the sciences of man by the fact of conceptual innovation which in turn alters human reality. The very terms in which the future will have to be characterized if we are to understand it properly are not all available to us at present. Hence we have such radically unpredictable events as the culture of youth today, the Puritan rebellion of the sixteenth and seventeenth centuries, the development of Soviet society, and so on. Thus it is much easier to understand after the fact than it is to predict. Human science is largely ex post understanding. Or often one has the sense of impending change, of some big reorganization, but is powerless to make clear what it will consist in: one lacks the vocabulary. But there is a clear asymmetry here, which there is not (or not supposed to be) in natural science, where events are said to be predicted from the theory with exactly the same ease with which one explains past events and by exactly the same process. In human science this will never be the case. Of course, we strive ex post to understand the changes, and to do this we try to develop a language in which we can situate the incommensurable webs of concepts. We see the rise of Puritanism, for instance, as a shift in man's stance to the sacred; and thus, we have a language in which we can express both stances - the earlier medieval Catholic one and the Puritan rebellion - as 'glosses' on this fundamental term. We thus have a language in which to talk of the transition. But think how we acquired it. This general category of the sacred is acquired not only from our experience of the shift which came in the Reformation, but from the study of human religion in general, including primitive religion, and with the detachment which came with secularization. It would be conceivable, but unthinkable, that a medieval Catholic could have this conception - or for that matter a Puritan. These two protagonists only had a language of condemnation for each other: 'heretic', 'idolator'. The place for such a concept was pre-empted by a certain way of living the sacred. After a big change has happened, and the trauma has been resorbed, it is possible to to understand it, because one now has available the new language, the try dO 0 b f 0 k transformed meaning world. But hard pre IctlOn e ore Just ma es one a laughing stock. Really to be able to predict the future would be to have licited so clearly the human condition that one would already have exp f 0 ThO 0 h dl 0 pre-empted all cultural innovation and trans ormatIOn. IS IS ar y III the bounds of the possible. Sometimes men show amazing prescience: the myth of Faust, for instance, which is treated several times at the beginning of the modern a There is a kind of prophesy here, a premonition. But what characer . terizes these bursts of foresight is that they see through a glass darkly, for they see in terms of the old language: Faust sells his soul to the devil. Theoy are in no sense hard predictions. Human science looks backward. It IS inescapably historical. 0 There are thus good grounds both in epistemological arguments and III their greater fruitfulness for opting for hermeneutical sciences of man. But we cannot hide from ourselves how greatly this opinion breaks with certain'commonly held notions about our scientific tradition. We cannot measure such sciences against the requirements of a science of verification: we cannot judge them by their predictive capacity. We have to accept that they are founded on intuitions which all do not share, and what is worse that these intuitions are closely bound up with our fundamental options. These sciences cannot be wertfrei; they are moral sciences in a more radical sense than the eighteenth century understood. Finally, their successful prosecution requires a high degree of self-knowledge, a freedom from illusion, in the sense of error which is rooted and expressed in one's way of life; for our incapacity to understand is rooted in our own self-definitions, hence in what we are. To say this is not to say anything new: Aristotle makes a similar point in Book I of the Ethics. But it is still radically shocking and unassimilable to the mainstream of modern science.