2 Chapter 1

This contemplative account, as it can be called, unites a number of notions. It describes knowledge as the product of isolated individuals. And it assumes that the individuals intrude minimally between reality and its representation: they apprehend reality passively, and, as it were, let it speak for itself; their perception is independent of their interests, their expectations or their previous experience. Hence the knowledge they produce is essentially only a function of reality itself. It can be tested by any individual who is able to compare it with reality, since its property of correspondence with reality is entirely independent of the situation wherein it was produced. These various notions tend to be associated because they are all indicated by a simple, memorable, concrete model: learning and knowledge generation are thought of in terms of visual apprehension, and verbal knowledge by analogy with pictorial representation. Indeed, it is probably our intuitive sense of correspondence between a picture and the appearance of something real, which sustains much of the credibility of the contemplative account, at least at the everyday level.

Certainly, our everyday epistemological notions appear to be thoroughly permeated with this conception, and the analogy between learning and passive visual apprehension. We talk of understanding as 'seeing', or 'seeing clearly'; we are happy to talk of valid descriptions giving us a 'true picture'. Similarly, we are able to characterise inadequate knowledge as 'coloured', 'distorted', 'blind to relevant facts', and so on. The overall visual metaphor is a resource with which we produce accounts of the generation and character of truth and error. And in many ways these accounts serve us well. Nonetheless, in sociology, the contemplative account has always co-existed with a sharply contrasted alternative, and at the present time it is the latter toward which the general trend of thought is moving. Increasingly, knowledge is being treated as essentially social, as a part of the culture which is transmitted from generation to generation, and as something which is actively developed and modified in response to practical contingencies.

Such a conception stands in polar opposition to most of the elements of the contemplative account. Knowledge is not produced by passively perceiving individuals, but by interacting social groups engaged in particular activities. And it is evaluated communally and not by isolated individual judgments. Its generation cannot be understood in terms of psychology, but must be accounted for by reference to the social and cultural context in which it arises. Its maintenance is not just a matter of how it relates to reality, but also of how it relates to the objectives and interests a society possesses by virtue of its historical development. An appropriate concrete model which integrates these various themes can be provided by considering a society's knowledge as analogous to its techniques or its conventional forms of artistic expression, both of which are readily understood as culturally transmitted, and as capable of modification and development to suit particular requirements.

The relationship of these two opposed conceptions has always been an uneasy one within the context of sociology, with the

3 Chapter 1

tension between the two always apparent, but with individual writers rarely situating themselves consistently and unambiguously on one side or the other. Thus, Karl Mannheim's 'Ideology and Utopia' (1936) opens with a clear indication of its commitment to the second, active conception:

Strictly speaking it is incorrect to say that the single individual thinks. 
Rather it is more correct to insist that he
participates in thinking further what other men have thought
before him. He finds himself in an inherited situation with
patterns of thought which are appropriate to this situation and
attempts to elaborate further the inherited modes of response
or to substitute others for them in order to deal more adequately
with the new challenges which have arisen out of the shifts and
changes in his situation. (Chapter 1.1)

But, although these points are reasserted a number of times throughout the work, a great part of its argument and much of its concrete discussion is, in fact, predicated upon the contemplative model. Natural science and mathematics, Mannheim tells us, are forms of knowledge which bear no mark of the context of their production and which can properly be assessed entirely in terms of their correspondence with reality. Moreover, precisely because they are the products of disinterested contemplation, they are preferable to other kinds of knowledge, to sociology or history or political thought.

In his treatment of these latter kinds of knowledge Mannheim continues to be inconsistent. Sometimes he insists that this knowledge can in no way be assessed in context-independent, contemplative terms. Then he develops an argument which implies the opposite. He states that such knowledge, knowledge of social reality, is always in practice related to social standpoints and interests, and thus context-dependent. This makes the knowledge inadequate or, at best, of restricted validity. However, under ideal but realisable conditions, context-independent knowledge, corresponding to social reality, could be produced. A class of disinterested intellectuals, able to take a properly contemplative approach could produce it.

It is true that some of this inconsistency is the product of Mannheim's combining essays written at different times. Chapters 1 and 5 are those most inclined to an active, contextual and social treatment of knowledge, and they were the last written. But these are also the least concrete chapters of the book. And even in these chapters, the contemplative account and its associated metaphors remain important components, without which the results of Mannheim's thinking would be bereft of all plausibility and coherence. It is clear that in spite of himself, Mannheim produced a work largely based upon the contemplative account. Although he explicitly rejected it, he apparently could not help but think in terms of it. Even the most original and insightful points in 'Ideology and Utopia' are conceptualised in terms of contemplation, and the associated visual metaphors. (1)

Thus, Mannheim's work reveals just how difficult it can be to move away from a contemplative position. The associated pictorial metaphor for knowledge is so pervasive, intuitively attractive and, indeed, valuable as an explanatory resource, that it can be

4 Chapter 1

difficult in practice to structure one's thought independently of it. Mannheim knew, and advanced, many good arguments against the contemplative account, and in favour of the alternative he explicitly advocated, but this did not suffice to reorient his practical approach. Hence, given that a form of the active, social conception of knowledge is to be put forward as a working orientation here, it seems appropriate to present it in a way which is designed to counteract the appeal of the pictorial metaphors incorporated in the contemplative account. Such a presentation cannot hope to count in any way as a justification; it merely offers a model for consideration, and for use in following the subsequent discussion. But there is in any case no space in which to develop a detailed discussion of the problems involved.

It might be thought that the best procedure for moving away from the contemplative account would be to break the equivalence of pictorial and verbal representations and emphasise the differences between passive-visual apprehension and understanding generally. If verbal statements cannot be matched against reality like pictures are, then the need for an alternative metaphor to characterise the nature of verbal knowledge is indicated. In fact, the opposite strategy is the more expedient. We should emphasise the equivalence of all representations, pictorial or verbal, and accept observing as a typical kind of learning. It is the treatment of visualisation and depiction as passive processes which marrs contemplative conceptions in the sociology of knowledge, and makes their visual and pictorial metaphors unsatisfactory. Our strategy should be to reveal pictorial representation, the most favourable case for the contemplative conception, as essentially an active and a socially mediated process, and in this respect typical of representation and knowledge generation generally.

In fact, this is something that has been done for us already by those academic fields directly concerned with the study of pictorial representations and their creation. Work in fields as different as the psychology of perception and the history of art could be used to make the points we need. Let us take the latter field, where the close relationship, if not the complete equivalence, of pictorial and verbal representation is more or less taken for granted, and references to the 'language' or 'vocabulary' of an artist or illustrator are commonplace. A particularly relevant work is that of Ivins (1953) on the history of prints and engravings. Here language is looked to as a model on the basis of which to understand pictures. Ivins devotes his extremely concrete and well-illustrated book to showing how the 'syntax' of 'pictorial statements' has changed from the Renaissance to the present day. And he makes it clear that the 'pictorial statements' he considers simply cannot be treated as passive reflections of real appearances; rather they render scenes and objects in terms of conventions.

Gombrich's important study 'Art and Illusion' (1959) makes similar points. It reveals the difficulties which arise in talking of the extent to which a representation can correspond to reality or the direct appearance of reality. And it makes clear that, at least for intuitively straightforward conceptions of correspondence, representations not only do not correspond with appearances but they



5 Chapter 1

cannot (not even if they are photographs). Representations may, when viewed under particular conditions in particular contexts, achieve a 'trompe l'oeil', but such deceptions are generally produced by conventions of representation which involve obvious distortions of what the painter or illustrator sees. The capacity to produce 'realistic' representations tends to depend upon the study of existing paintings which use appropriate conventions, rather than upon an open observant attitude to what is depicted.

In Gombrich, Ivins and similar work, we find an account of the construction of pictorial representations which serves admirably as an informal working model for the construction of knowledge. Pictorial representations are actively constructed from conventions available as the resources of some culture or sub-culture. The successful realisation of paintings, for example, depends upon familiarity with existing paintings and illustrations and the conventions implicit in them. Such conventions are meaningful as words are meaningful, and are actively manipulated and organised in the light of particular aims or interests.

Extending this account, when a representation conveys knowledge or information about, say, an object, it is by classifying it, by making it an instance of one or more kinds of entity recognised by the culture whose resources are drawn upon. In this way the representation makes it possible for existing knowledge to be applied to its referent, and it makes the referent a source of meaningful information, a potential check upon existing knowledge. Knowledge and object are connected by the representation. (We can, admittedly, deploy knowledge directly as we act, but this is because our perception organises and pre-classifies what we perceive; we read the world, rather as we read handwriting, as an assemblage of symbols.) (2)

All representations are indeed then, as Ivins says, kinds of statements. They must be distinguished both from the objects they represent and from the appearances of those objects. Any representation is one of numerous possibilities which the resources of a culture make available. And the resources of a culture are themselves reasonably treated as a particular selection from an endless number of possibilities. In both cases, we are entitled to seek an explanation of why some possibilities rather than others are actually encountered.

It may well be that particular individuals frequently notice resemblances between aspects of their environment in a random, undirected way, and build up particular beliefs and representations in a fashion which cannot be explained systematically. But public knowledge typically evolves much more coherently, and the people who contribute representations to it operate in what is cumulatively a much more orderly way. Typically they are concerned, directly or indirectly, in the performance of some institutionalised activity, designed to further particular aims or ends. This means that the knowledge they produce is designed from the start to facilitate certain kinds of prediction, or function in the performance of particular kinds of competence. And its evaluation is pre-structured to an extent by these design requirements; to anticipate Habermas's term, discussed in the next section, it is pre-structured by a situated technical 'interest in prediction and control'.

6 Chapter 1

Representations are not assessed with any particular stress on their rendering of appearances, but instrumentally, in conjunction with whatever the activities are with which they function. Hence, the growth of knowledge should not be thought of as the result of random learning about reality, but as the correlate of the historical development of procedures, relevant in various degrees to the competences and techniques and ends or objectives of cultures or sub-cultures. Of course, many such competences and associated representations find such wide instrumental applicability that once introduced into practically any culture they are almost
guaranteed an enduring position therein.

Representations are actively manufactured renderings of their referents, produced from available cultural resources. The particular forms of construction adopted reflect the predictive or other technical cognitive functions the representation is required to perform when procedures are carried out, competences executed, or techniques applied. Why such functions are initially required of the representation is generally intelligible, directly or indirectly, in terms of the objectives of some social group.

This very informal conception should suffice as a basis for the following discussion, although for many purposes it would be altogether inadequate as it stood. It would need considerable qualification, for example, if activities like scientific research were the central foci of discussion, with their basic orientation to the creative extension of knowledge. It is often pointed out that theories and representations employed for creative scientific work are often not those which have proved the most instrumentally adequate. Scientists often impute instrumental adequacy to one set of representations (say those of classical mechanics or geometrical optics) but regard others, those they use in their work, as having greater ontological adequacy. This is often taken to indicate that knowledge must be, and is, evaluated as a direct rendering of reality and not simply as an aid to activity. Unfortunately yesterday's ontologies have a depressing tendency to become tomorrow's instrumentally adequate representations, and on that basis, and other grounds which cannot be gone into here, the general outlines of the present account can be adequately defended. Nonetheless, the actors' distinction between instrumentally applicable theories and those suitable as guides to research is of great relevance and interest, and would merit extended discussion in other contexts. (3)

Let us however concentrate on our informal conception as it stands, and try to make it a little more concrete by reference to some examples. In order to continue to erode the appeal of a contemplative conception of knowledge, pictorial representations will be used. And so that the representations will be generally accepted as embodying knowledge, the illustrations chosen will be of a kind which have utility in the context of natural science. They will be considered in order, from those which are easily reconciled with the above account, to those which may not immediately appear to be so. Hopefully, the sequence will act as a 'bridge' to the most problematic cases, and indicate the fully general scope of the account.

7 Chapter 1

Imagine then that some students in a physics laboratory are requested to draw some apparatus set out before them, and that the result is Figure 1.1; such a result is not empirically unreasonable. Presumably, there is no problem in arguing that the figure is a pictorial statement constructed from existing cultural resources; it is assembled from signs meaningful as concepts in physical theories of electricity, and is obviously reminiscent



Figure 1.1 (I will show the figures in class--use your imaginations).

of a verbal statement. Perhaps the commonest immediate concern of students who construct diagrams like Figure 1.1 is to conform to expectations, but basically such representations are sustained in our culture as adjuncts to competences. In this case, it scarcely makes sense to ask whether the referent of Figure 1.1 is truly what the figure indicates it to be: the referent could be a battery and a resistance box wired together, a length of metal, a nerve fibre, a building or indeed practically anything at all. The appropriateness of the figure cannot be assessed in isolation, by examination of its referent. All that can be assessed is the use of the figure, how it is actively employed.

The real problem with Figure 1.1 is likely to lie in establishing that it is a typical representation. In particular, there is no vestige of resemblance between its appearance and that of what it is used to represent. Let us move then to Figure 1.2. Maps frequently show an intuitive resemblance to the appearance of reality itself, as, for example, when it is seen from the air; sometimes they are deliberately designed to resemble appearances. But they remain compatible with the above account. They are constructed entirely in terms of conventions. Their particular form depends upon what procedures they are designed to facilitate. Their value is assessed functionally and not by reference to appearance. Maps indeed afford one of the clearest and most accessible contexts in which to examine the connection between the structure of representations and their function. (If ever physics needs to be supplemented as a paradigm of knowledge, there is much to be said for turning to cartography.)

Figure 1.3 is taken from an anatomy textbook, and depicts some muscles of 'the arm'. It is designed to facilitate recognition and naming in the context of an esoteric activity. Therefore, it is not a rendering of a particular arm. Despite being apparently realistic it is intentionally a schemata. It cannot be taken as an attempt passively to imitate reality. Indeed its effect is to modify perception so that students can perceive arms in terms of its scheme of representation. As an aid to seeing and naming,
9 Chapter 1

8 Chapter 1

[Again, I will show the figures in class].

Figure 1.2

its schematic character is accentuated at the expense of its possibilities as a rendition of appearances.


There is no particular arm to which it relates as a representation; it is a typification constructed from available symbols. (That it is indeed constructed from symbols can only escape our notice if we forget that symbols are involved in perception as well as representation.)

Like all scientific representations, Figure 1.3 is reliably applicable only to aid particular kinds of procedure. In this case the procedures, together with directly associated instrumental interests, are embodied in the role of the anatomist and his student audience. Those who make practical use of such representations are generally well aware that their reliability and applicability is restricted; this awareness is automatically generated in learning to use the representations. Other instrumental interests and other activities, located in other scientific roles, engender other kinds of representation. But this limitation upon the scope of anatomical representations is not normally taken as grounds for scepticism about their validity; they are accepted and accorded credibility co-extensively with the acceptance of anatomists' competences and techniques.

When representations are photographs and not diagrammatic figures, the same interpretation applies. Such photographs remain constructs for use in activity. Admittedly, the photographer cannot simply assemble conventional symbols when he takes a photograph of some real object. But he can work his material so that his finished product can be seen in terms of, or as, such an assemblage. (4) Examination of the photographs sometimes used in the teaching of human anatomy illustrates this point convincingly. Their manufacture does, of course, involve the use of particular human bodies, but their representational adequacy is again evaluated in use, and not by comparing them with the particular bodies from which they were manufactured.

The argument is then that all representations, pictorial or verbal, realistic or abstract, are actively constructed assemblages of conventions or meaningful cultural resources, to be understood and assessed in terms of their role in activity. (5) Essentially this amounts to making representations analogous to techniques, artistic conventions or other typical forms of culture, rather than considering them in terms of the contemplative conception. Sociologists have often found it appropriate to adopt this treatment in dealing with everyday knowledge. But, like Mannheim, they have sometimes felt that a special kind of knowledge exists in the natural sciences and mathematics, intelligible only in contemplative terms. Scientific knowledge, however, is always assessed in conjunction with the institutionalised technical procedures of its specialties and is entirely typical


10 Chapter 1

of knowledge in general: science is in many ways a constellation of craft skills. (6) As for mathematical knowledge, we have here a developed set of generally utilisable procedures and representations to which no reality can even be said to correspond (7) It is precisely their extraordinary versatility in furthering a vast range of objectives, which results in their widespread use and sustains their credibility as knowledge.

However, in thus rejecting a contemplative conception of knowledge and adopting a view which emphasises its social dimension, it is important not to lose sight of the connection which does exist between knowledge and the real world. This is properly stressed in the contemplative account, albeit usually in terms of an unclear notion of truth as correspondence. Knowledge is not related to activity rather than reality; it is related to activity which consists precisely in men attempting to manipulate, predict and control the real world in which they exist. Hence knowledge is found useful precisely because the world is as it is; and it is to that extent a function of what is real, and not the pure product of thought and imagination. Knowledge arises out of our
encounters with reality and is continually subject to feedback-correction from these encounters, as failures of prediction manipulation and control occur. We seek to eliminate such failures, but so far reality has sustained its capacity to surprise us and dash our expectations. Indeed, our liability to be surprised in this way, to be confounded in our expectations, constitutes an important argument against a purely idealist theory of knowledge. (8)

2 KNOWLEDGE AND INTERESTS

Many important insights into how our knowledge of reality is mediated by interests and activity are to be found in Marx's work. They have been taken up by European writers on what might loosely be called the idealist wing of Marxism, and incorporated into general theories of knowledge which offer a striking contrast with the predominant conceptions in our own academic culture. These conceptions, a strict separation of fact and value, a stress on the objects of knowledge almost to the extent of excluding the role of the knowing subject, a view of that subject as an isolated contemplative individual without social dimensions or historical situation, and an atomistic concept of validation which sets isolated bits of knowledge in comparison with individual fragments of reality, are all condemned under the somewhat confusing umbrella label of 'positivism'. They must be swept aside, it is claimed, and replaced by a more down-to-earth account which treats knowledge as the actual product of men as they live and work in society.

Clearly, work which develops such important themes as these is of great importance and must be examined. Paradoxically, however, one of its most characteristic features is its lack of contact with actual instances of knowledge in its social context. That intimate involvement with the specifics of concrete historical situations, so laudable in Marx's work, and, one would

Chapter I

have thought, an implied necessity in terms of their own theories, is generally absent from the writings of the idealist Marxists, where, with few exceptions, one finds only large-scale speculations. There is no doubt that the work to be discussed here has suffered as a consequence. The theories of Lukacs and Habermas both involve weaknesses which attention to concrete examples would have exposed and helped to eliminate.

Lukacs set out his views in his famous polemic 'History and Class-Consciousness' (1923). Here he contemptuously rejected contemplative positions and asserted that consciousness and thus knowledge, of all kinds, in all contexts, was necessarily related to human interests; it was always the product of the activity of particular groups of men, rationally generating it in the course of furthering their interests. (9) Indeed, for Lukacs, men's rationality was manifest not in their thinking alone, but in their thought and activity considered as one phenomenon, that is, in their practice. Unlike Mannheim, whose work he directly inspired, Lukacs consistently stressed the need to consider practice, rather than thought alone, whenever knowledge and consciousness were under sociological investigation. In this he was entirely justified. What is open to question is whether even Lukacs's work takes sufficient account of the man-created character of knowledge.

Lukacs believed that under ideal conditions reality is fully accessible to the rational appraisal of men; the totality of what is real can potentially be understood. Men generate knowledge in the course of practice, to further their particular interests. Were practice unconstrained, the totality of these interests would amount to the universal, fully general, interests of mankind as a whole, and would generate the fullest possible understanding of reality. But in existing societies, practice is never unconstrained. That of oppressed classes is restricted by coercion and by ideological control. And that of dominant social classes is ultimately restricted by their own particular, restricted social interests, which limit the possible scope of their rationality. Awareness of some aspects of reality is irrelevant, or in some cases positively discomforting to them, and consequently is not developed; thus they never attain more than a partial understanding of reality. Moreover, there is a sense in which this partial understanding is a total misunderstanding. We can only properly understand an aspect of reality by considering it in context, in relation to everything else; hence to understand anything fully and correctly we must understand everything. It follows that the particular restricted interests of a class set limits upon the whole of its thinking, and logically determine the most that it can hope to produce in the way of knowledge. To every class there corresponds an ideal class consciousness. (10)

Without in any way addressing the problematic question of what Lukacs's own views were on the subject, it is worth noting that his account is readily intelligible if we assume that the knowledge which men generate in the course of their practice is in some sense a copy, reflection or picture of an aspect of reality. This assumption justifies the notion of a full and final under-

12 Chapter

standing of reality as a whole. It allows us to conceive of 'partial consciousness' in terms of actors having access only to parts of a complete picture of something. And it suggests that the missing pieces of the picture will be aspects of the whole irrelevant or disquieting to a particular class, and hence not reflected in its consciousness. If we do not make this assumption then it is difficult to see how to integrate and justify the various themes of Lukacs's argument, or how to interpret a number of other points in his work. (11) Perhaps pictorial metaphors did, at some level, help to structure even Lukacs's thought.

In any case, among its many weaknesses, Lukacs's position overlooks or ignores the fact that men's thinking is always an extension of earlier thought, that the production of new knowledge involves the use of existing knowledge and existing cultural resources, and that consciousness is to this extent always the product of history. It assumes instead an unproblematic interaction between men and reality, with a third variable, interest, effectively doing no more than accounting for the restricted scope of that interaction. Once it is realised that, as Lukacs might have said, new knowledge is dialectically generated from old, then the entire structure of his account falls to pieces. Knowledge has to be understood naturalistically in terms of its cultural antecedents and its present causes, not teleologically in terms of a future state it is or is not moving towards. An ideal 'complete' state of knowledge, a complete understanding of reality, can no longer be assumed; it is indeed no longer clear what the meaning of such a conception can be. And, accordingly, interest can no longer determine consciousness by restricting it to involvement with some section of the whole of reality.

Like Mannheim, Lukacs made promising programmatic statements about the general character of consciousness, and failed to develop them into a satisfactory framework for the sociology of knowledge. His failure is however differently rooted to that of Mannheim. Lukacs seems to have lacked any real curiosity about knowledge and consciousness, and to have written largely to legitimate projected courses of action. He disdained to consider concrete instances, and thus was incapable of learning; his thinking was cut off from dialectical interaction with experienced reality.

The work of Jurgen Habermas has many of the weaknesses of Lukacs's, and for similar reasons. Nonetheless, his 'Knowledge and Human Interests' (1972) is a significant text, which, although only fully intelligible in terms of an intellectual tradition entirely alien to that in which this book must reside, can still be exploited as a source of particular insights. (12) Let us start by outlining his conception of modern scientific knowledge, which he takes as the predominant current form of technical, instrumentally oriented knowledge, and typical of such knowledge.

Habermas sees scientific knowledge as the product of communities of interacting men who operate upon and perceive reality, not idly and contemplatively, but in terms of particular instrumental, manipulative and predictive interests. Such interests are constituted into the process of knowledge generation and evaluation. What scientists mean by the validity of their knowledge is predetermined by these interests. What scientists take to be facts

13 Chapter 1

or data is determined by the pragmatic pre-organisation of experience implied by existing systems of instrumental activity. (13) Scientific knowledge has a function only in the context of systems of instrumental activity.

Since particular areas of scientific knowledge are bound up with systems of instrumental action, they relate to each other just as our instrumental activities relate to each other: they constitute an overall body of knowledge available in the execution of goal-oriented, instrumental action, a general interpretation of reality with a view to all possible forms of technical control and prediction. According to Habermas, scientific knowledge, and technical knowledge generally, is oriented, with transcendental necessity, by a knowledge constitutive interest (KCI) in prediction and control; this interest is the natural basis for scientific knowledge. And far from this reflecting adversely upon the worth of science, it is, for Habermas, the source of its justification: modern science is to be valued in that it is the most developed form of instrumentally oriented knowledge.

Essentially, the tendency to treat interest as an adverse influence upon knowledge, and to represent science as the product of disinterested contemplation, stems from a justified distrust of the effects of particular narrow individual and social interests, which generate rationalisations and ideology. All scientific disciplines correctly guard themselves against such interests. But this has led to an incorrect understanding of the general relationship of knowledge and interest:

Because science must secure the objectivity of its statements
against the pressure and seduction of particular interests, it
deludes itself about the fundamental interests to which it owes
not only its impetus but the conditions of possible objectivity
themselves. (p. 311).

Whether or not this diagnosis is correct, Habermas's assertion of the necessary connection of scientific knowledge and technical interests, and his consequent instrumentalist account of science, are probably, in general terms, justified. (14) It is true that they involve difficulties and obscurities, and that his discussion of 'transcendental' KCIs is particularly tentative and unsatisfactory. (15) But most of these problems are more pertinent to grand speculative philosophy than to the vulgar naturalistic concerns of the present volume, where we can simply take the point that technical and scientific knowledge is generated and evaluated out of an active interest in prediction and control.

Unfortunately, the remaining themes in Habermas's discussion of knowledge are worth attention primarily because Of the plausible but disastrous misconceptions which they involve. Science, we are told, is the best we can achieve in the way of instrumental knowledge; but men have other interests besides instrumental ones, and knowledge can be constituted in relation to these interests also. Moreover, such knowledge can exist in institutionalised forms, with their own agreed standards of validity, fully on a par with the institutionalised forms of scientific knowledge. Habermas is not calling our attention to such things as our personal memories with their emotional or aesthetic meanings. Nor is he reminding us that our knowledge may be modulated, at the
14 Chapter

public level, by the desire for self-consolation or the aim of deceiving others; as we have seen, he accepts that particular interests of this kind should be prevented from influencing the processes of knowledge generation. He is suggesting that further ideal conceptions of knowledge must be accepted, every bit as important as the scientific-instrumental ideal, but relating to other KCIs.

Habermas offers us two further such ideals, related to two further distinct KCIs (pp. 308- 11): historical-hermeneutic knowledge relates to a practical KCI in meaningful communication and the achievement of consensus; (16) knowledge involving self reflection (as in philosophy and 'critical' sciences) arises out of an emancipatory KCI in autonomy and responsibility. (17) Consideration of the former ideal will suffice to illustrate the problems he thereby creates. Habermas claims that the evaluation of, for example, historical knowledge is structured by - interests different to those operative in the case of science, and that the knowledge itself is intrinsically different in character from scientific knowledge. He fails to substantiate his first claim due to inadequate consideration of the nature of historical knowledge (or 'hermeneutic' knowledge generally); he errs in his second claim due to his inadequate familiarity with scientific knowledge. He is correct in maintaining that men possess diverse interests, and that their consciousness is not entirely dominated by the instrumental aims of prediction and control. But his specific equation of different kinds of interest with different kinds of knowledge does not stand up to detailed consideration.

According to Habermas, whereas science is evaluated by the extent to which it facilitates instrumental operations with things, historical knowledge is evaluated by success in the 'preservation and expansion of the intersubjectivity of possible action- orienting mutual understanding' (p. 310). It is the kind of knowledge in terms of which people achieve identity and self- integration, and in terms of which they interact with others to achieve an 'unconstrained consensus'. Unfortunately, we are provided with no satisfactory concrete exemplification of this abstract statement, nor any other relevant indication of how precisely interests in interaction and consensus structure the evaluation of historical knowledge. Nor is it easy to imagine what Habermas has in mind.

It is inconceivable that history should be treated as purely expedient myth or fable, constructed solely with a view to what an audience wants to hear, or what would best serve the cause of social solidarity. Conceptions of what actually happened and what factors were relevant to men's actions, are obviously of great importance in history, together with scholarly methods of inferring such things from sources and records. Indeed, who would argue that such concerns should not take priority over any other considerations, when historical accounts are being evaluated? Certainly Habermas gives no indication of such scant regard for scholarly historical standards, and never anywhere suggests that they should be set aside out of expediency.

Yet, if the framework of evaluation of historical knowledge is primarily defined by these considerations, rather than by

15 Chapter 1

expedient considerations, it is very difficult to see why that framework should primarily be related to a KCI in dialogue and consensus. It would seem rather to be analogous to the evaluative framework of science. Doubtless, there are ways of countering this objection. It might be said, for example, that we should consider the evaluation of concepts in science and history, not of statements, or actual knowledge claims: whereas the prior evaluation of scientific concepts is instrumentally informed, the prior evaluation of historical concepts and their meanings is informed by their potential in the maintenance of dialogue and the attainment of consensus. (18) This is a possibility that deserves exploration, but, again, it is impossible to discern what it might involve from consideration of Habermas's work. How we rationally evaluate concepts and meanings in a way which reflects a prior interest in dialogue and consensus is unclear. It is no less problematic a notion than that of the evaluation of actual knowledge claims in terms of the same interest. Habermas simply does not satisfactorily justify and illustrate his point of view here.

Admittedly, the character of historical knowledge and how it compares with the knowledge of science is an extraordinarily difficult question. But let us for the sake of a clear discussion move right to the opposite pole to Habermas, and hold that historical knowledge is instrumental in just the same way as is scientific knowledge. To the extent that historians prefer the evidence of their sources to the requirements of their community or their audience, they are surely operating in terms of an interest in prediction and control rather than in consensus. Their findings are properly thought of as predictions of subsequent archaeological or paleological discoveries; their reconstructions of the past may constitute virtual experiments on the basis of which to learn how to predict, or even influence, the course of social change.

Historians, we might suggest, typically and properly evaluate their knowledge (and their concepts) in a framework pre-structured by interests in prediction or control, even if often with a view to using it to serve a variety of further interests. It may then assist individuals to orient themselves within their communities, or it may facilitate predictions of social or even individual behaviour. On this view, history differs from science not by virtue of the general interests which are constituted into the process of its production and evaluation, but by virtue of the interests it typically serves and the subject matter out of which it arises. Thus, there is a strong case for treating historical knowledge as primarily instrumental in character; it is not perhaps entirely analogous to physics or mechanical engineering knowledge, but is fully comparable with say palaeontology, or other sciences where immediate manipulative interests are not relevant, but the characteristic general cognitive operations involved in prediction and control are nonetheless manifest.

Along the same lines, it could be argued that all knowledge, 'scientific', 'hermeneutic' or otherwise, is primarily produced and evaluated in terms of an interest in prediction and control. We further our interest in communication and mutual understanding
16 Chapter 1

on the basis of any body of shared knowledge, which we make the basis of interaction by utilising communicative competences. Consensus in a community is not achieved by the application of a particular kind of public knowledge, but by the exercise of communicative skills and proclivities against the background of whatever it is in the way of knowledge that members of the community generally possess. Such skills and proclivities are currently the subject of study by some 'ethnomethodologists' and 'cognitive sociologists'. By their use we achieve an everyday consensus on the basis of what is accounted the common sense of our culture; historians achieve a consensus on the basis of their esoteric findings; and scientists achieve a consensus in their esoteric interactions on the basis of their theories and models. According to this view, knowledge has the character of a resource, communally exploited in the achievement of whatever interests actors decide. And precisely because of this, knowledge is always primarily linked, in its generation and initial evaluation, to an interest in prediction and control. Natural science, history, sociology, are (or potentially are) bodies of knowledge which serve as resources to facilitate prediction and control in different contexts. They do not differ in their essential relationship to KCIs. They all arise out of an active instrumental interest; they may all serve diverse particular interests; they may all be made the basis of interaction and unconstrained consensus.

With all of this, however, Habermas would disagree; he would characterise it as rampant scientism, an illegitimate and dangerous extension of an instrumental conception of knowledge, a misconceived analysis of interaction as based upon morally indifferent knowledge and arbitrary, irrational evaluation. Unfortunately, we are provided with no arguments to justify such disagreement, and no positive alternative account. Habermas does not show how knowledge can legitimately develop in a context of evaluation which is not primarily shaped by predictive, instrumental interests. Basically he just doesn't like the idea that history and the human sciences are bodies of instrumentally oriented knowledge, or that, in interactions between people, attempts to predict and modify the actions of the other occur literally from second to second. Habermas cannot accept the application of instrumental knowledge to people as normal and appropriate in interaction: to him it is equivalent to treating people as objects; it is a form of reification. The difference between people and things should be evident in the forms of knowledge that apply to them; Habermas does not consider that it may reside instead in the different procedures and forms of activity which we deploy in orienting ourselves to people. A properly morally aware orientation to another person is, for Habermas, manifest not just in attitudes or behaviour, but in the intrinsic character of the knowledge which it involves. The truth of knowledge should be explicitly assessed in terms of its relationship to 'the intention of a good life'. Knowledge is more than a resource for consciousness, it is a strong determinant of consciousness. But this is asserted and not shown.

Habermas's insistence upon linking science and history to

17 Chapter 1

distinct general interests is based upon only one substantive factor: this is his firm conviction that science and history are intrinsically different as bodies of knowledge. Historical know- and 'hermeneutic' knowledge generally, arises from treating thought and activity as meaningful, as intelligible only in terms of some hypothesised general coherent system of meaning. Hence the 'facts' it deals with are in a sense the products of its own hypotheses, and do not offer a fully independent check upon in the way that facts about real objects could do: acts of legislation, rights of inheritance, religious observances and the like are not available as independent existences which can serve to test historical accounts of legal systems or religions, since their perceived nature, or even that they are perceived, is a consequence of the accounts themselves. It follows that the theoretical speculations of 'hermeneutic' sciences are capable of evaluation only to the extent that they produce a consistent, coherent and plausible overall interpretation of activity as meaningful and intentional. Moreover, since history seeks to make the past intelligible as the meaningful product of men's though; and activity, it must reflect all the inconsistency, fluidity of meaning and adjustment to context characteristic of that thought and activity itself. Men communicate in ordinary language, which permits the perpetual renegotiation of meaning, and its adjustment to context in the course of dialogue. Hence the language of 'hermeneutic' science must also be imprecise and its meaning context-dependent. If we accept this description of history and 'hermeneutic' knowledge (as, in its essentials, we should do), it is clear that any consensus it achieves cannot be explained in terms of rational appraisal of an independent reality. This raises the fascinating and difficult question of how such a consensus is achieved, if indeed it is ever genuinely 'unconstrained' and more than a consequence of the application of power. For Habermas, consensus in the hermeneutic sciences is achieved like consensus in the everyday world: people enter into dialogue with a view to the achievement of consensus and evaluate the knowledge they produce with a view to its relevance to that achievement. Thus, it is part of the character of hermeneutic knowledge that it is capable of sustaining a moral community: [In hermeneutic inquiry] the understanding of meaning is directed in its very structure toward the attainment of possible consensus among actors in the framework of a self-understanding derived from tradition. (p. 310). Natural science, and instrumental knowledge generally, is taken by Habermas to be diametrically opposed in its characteristics to everything set out above. Its referents are held to be fully independent of its theories and capable of providing separate, external tests of their instrumental validity; its concepts and formulations are thought to attain exact definitions and stable meanings altogether independent of context; its status as 'pure instrumentality' is held to preclude its operation as a basis for communal consensus, and it is conceived instead as embedded within the practice of a scientific community which is sustained by ordinary language interaction involving hermeneutic knowledge (cf. pp 1 138-9)

8 Chapter 1

It is clear that Habermas bases these views on his study of 'positivist' philosophies of science, and is not at all familiar with its concrete practice. Thus, for all that he provides an interesting alternative conception of science to that of the 'positivism' he so detests, Habermas's thought, like Lukacs's, remains profoundly influenced and misdirected by that very 'positivism'. Openness to fully independent testing, independence of context, objectively demonstrable rather than communally assigned validity, are all features which have been imputed to knowledge by orthodox 'positivist' philosophers of to maximise its credibility. Given that all these are incorrect, it is ironical that Habermas should speak, as it were, out of the mouths of his enemies. More
still, orthodox 'positivist' philosophy and history has, in the meantime, itself become aware of the inadequacies of its earlier formulations and, profiting from its detailed particular studies of science, has been able to show the essential similarity of scientific and 'hermeneutic' knowledge.

There is no need here to illustrate the context-dependent, inherently fluid and imprecise character of scientific discourse, the lack of a clear fact-theory distinction within it, and the extent to which its status is communally sustained. These points have been well documented by Mary Hesse (1972) in an excellent review of Habermas's work, and they are firmly substantiated by extensive historical study and concretely based argument. (19) That they are justified is in no way incongruent with the instrumental character of scientific knowledge. All knowledge is made by men from existing cultural resources; old knowledge is part of the raw material involved in the manufacture of new; hence, whatever the interests which guide knowledge generation, socially sustained consensus and a modification of existing meanings will always be involved in the process. Habermas, like Lukacs, ignores this essential connection of scientific knowledge with its cultural antecedents, and this constitutes the crucial formal inadequacy in his account, the central misconception to which all else can be related. This is why Habermas does not realise that in describing 'hermeneutic' knowledge, he is merely pointing out certain universal features of all knowledge. (20)

Let us return now to our earlier hypothesis, that all knowledge is primarily instrumental. Clearly, nothing said by Habermas counts against the merits of this view. He does not show knowledge being evaluated without primary regard for predictive or instrumental standards. He establishes no effective distinction between instrumental knowledge and other kinds. He offers no specific criticisms of, for example, an instrumental account of historical knowledge. He does intend to demonstrate some of these things when he describes 'hermeneutic' knowledge and the way it is socially sustained. But what has really to be remembered here is that al l knowledge is socially sustained, a set of agreed conventions, as well l as being instrumental in character. In what follows, it will be assumed that knowledge generally is primarily instrumental, in the sense that it is generated and evaluated in a way that is pre- organised by an interest in prediction and control, and normative, in the sense that it is

19 Chapter 1

sustained by a communal consensus which is decided, and not a rational necessity. (21)

This is not to imply that there are no problems with such an interpretation. On the contrary, the problems are many, and too involved to be dealt with here, where we must simply take what seems the most plausible position and proceed. Perhaps the most difficult problem is how to deal with evaluatively oriented knowledge, on an instrumental view. Everyday discourse is commonly both explicitly evaluative and descriptive. The news media tell us of murders, terrorist attacks, miscarriages of justice and so on. If the above account is to be sustained, we must imagine that the knowledge we acquire from such discourse is to some extent the product of decoding. Terms like 'murder' have a place in networks of concepts theoretically organised to predict and infer what is physically the case. But they also convey information about evaluative orientations. Everyday discourse using such terms can be treated as conveying two kinds of information, rather as an electrical signal can carry both sound and picture information for a television receiver. In both cases, the superimposition of two kinds of information can be achieved with scarcely any interference between the two. We might suggest that everyday discourse typically carries two messages in one signal in this way, although whether crossover distortion is typically negligible is another matter. (22)

Such a treatment of the moral component in knowledge is highly contentious. Many serious objections can be brought against it; several philosophical accounts of the essence of moral discourse are difficult to reconcile with it. But alternative empirically oriented accounts are hard to find, as are concretely conceived alternatives to the general instrumental perspective presented here. Habermas would probably condemn the whole account for its insistent separation of the descriptive and the evaluative. But people have always drawn a distinction between the real and the ideal, that with which they are confronted, and that which they would wish to bring about. Hence, to suggest that they can and do decode discourse to obtain information about actualities and information about ideals is as plausible, in concrete, empirical terms as the rest of the general, instrumental view. Such a view surely deserves consideration as as promising an account of the character of knowledge as any we possess. Certainly, it should not be set aside on the basis of abstract principles which are themselves even more in need of justification

3 CULTURE AND HISTORY

All knowledge is actively produced by men with particular technical interests in particular contexts; its significance and its scope can never be generalised to the extent that no account is taken of those contexts and interests. Mannheim made this point in the abstract, but never successfully incorporated it into his concrete work. Lukacs and Habermas also stressed it, but solely as a basis for large-scale speculation; they both overlooked the character of scientific knowledge as the product of a historical

20 Chapter 1

development. Invins and Gombrich are the only authors so far cited whose understanding of culture, change provides a sufficient basis for a general conception of knowledge. Only they appreciated the way that representations are always built out of preexisting cultural resources, and hence have always to be explained as developments within an ongoing cultural tradition. (23) Only they gave detailed examples of how cultural forms actually have developed and changed over time. (24)

It is tempting to suggest that concrete, specific investigation is essential to an adequate general understanding of the character of knowledge and the way that it grows and changes. Familiarity with specific instances would seem a necessary, although certainly not a sufficient, condition of such an understanding. But there is probably a further reason why Gombrich, in writing a detailed commentary on particular pictorial representations, produced a work of greater general theoretical insight than Habermas's cosmological speculations. Gombrich's essay had no need to address the problem of validity; the paintings and other representations he considered were not for the most part thought of as knowledge at all. Habermas, on the other hand, wrote as an epistemologist: validity was his central problem; to pronounce upon the merit and the scope of possible forms of knowledge was his explicit intention.

If genuine knowledge is uniquely determined by the actual, presently existing relationship between the knower and the known, the subjects) and the objects(s)) of knowledge, such problems can be approached with confidence. Only one corpus of genuine knowledge can emerge from the rational perception of reality (as 'positivists' would have it), or the rational investigation of reality in terms predetermined by interests. Such a corpus can be used as a criterion in detecting and criticising error and ideology, and as an end-point for an hypothesised progressive movement in the growth of knowledge. The characteristic epistemological activity of passing judgment upon the knowledge claims of others is thus automatically justified. The most that men can actually hope to achieve in the way of knowledge is conceivable as a final, finished corpus. But if knowledge must also be the product of given cultural resources, if rational men must generate knowledge on the basis of what is already thought and believed, then the evaluation of knowledge becomes altogether more problematic.

If old knowledge is indeed a material cause in the generation of new, then man's rationality alone no longer suffices to guarantee him access to a single permanent corpus of genuine knowledge; what he can achieve will depend upon what cognitive resources are available to him, and in what ways he is capable of exploiting such resources. To begin to understand the latter involves abandoning simplistic theories of learning, and undertaking a detailed examination of knowledge generation. To discover the former involves examining knowledge generation in its social context, as part of the history of a particular society and its culture; rational men in different cultures may represent reality in different, even contradictory ways Hence, the evaluation of knowledge claims is shot through with difficulties; in particular

21 Chapter

the existing knowledge on the basis of which new knowledge is generated, the culturally given component, can never be independently checked; its origins and justifications in the past are largely inaccessible, nor is there an Archimedean point without the domain of culture, from which to make an assessment of it. To many, this raises the daunting spectre of relativism; for they rightly perceive that standards formulated to judge knowledge must by themselves manufactured from existing resources and historically contingent, if the above account is correct. Small wonder that epistemological writings rarely get directly to grips with these themes. (25)

The problem of relativism should not be of direct concern to a sociological study, and the issues involved cannot in any case be properly considered in the present context. It should suffice us simply to adopt the instrumental ideal of knowledge we have arrived at, and proceed. However, there is a good deal of sociological interest in the problem of relativism, and its discussion does raise some points of naturalistic interest, so a very brief digression on the issue is in order.

For those who wish to avoid relativism, the trouble with the above account is that it offers no naturalistic basis for the objective evaluation of competing knowledge claims, and for the view that knowledge is progressive. Let us then consider whether its essentials can be retained, but its relativistic implications eliminated. Two attempts to do this will be examined; both prove to be unsatisfactory but it is interesting to see why this is so. (26) The first attempt involves postulating that the rational processes by which men learn suffice to produce a convergence in the knowledge of different cultures. Although men have to use their existing knowledge and concepts to make the world intelligible and hence to learn about it, in learning they modify their knowledge in the direction of an ideal final form. They have indeed to start somewhere, but that starting point does not affect where they will eventually end up. A sculptor has to start with a given block of marble when he makes a figure, and the initial shape of the block may continue to influence his work as he proceeds, but we credit him with the ability eventually to realise his figure, whatever initial block he chooses.

This interesting possibility has been very thoroughly investigated by philosophers of science in the inductivist tradition, who would have welcomed its confirmation. So far, their work has produced no grounds for assuming a tendency to such convergence, and the general indication is that no such grounds can be expected to emerge. We must take it, as a provisional, revisable answer to this empirical question, that the cognitive processes which routinely are involved in learning do not suffice to shake off the effect of the given, culturally variable, starting point from which they proceed (cf. Hesse, 1974).

A second possibility is to concentrate upon the cultural resources out of which new knowledge is produced, and question whether these given resources are merely conventionally meaningful and consensually sustained. New knowledge, it is agreed, is actively produced from existing knowledge, without necessarily any regard for appearances, or the random flow of phenomena as they
22 Chapter I

are experienced generally. But this is because men are seeking to capture the constantly operating underlying agencies which generate appearances, the real continuing mechanisms at work in the world. To do this they imagine, or create out of existing knowledge, theories about the world - putative mechanisms and agencies, held to exist, and to explain why things are as they are. And then they actively intervene in the course of events to check their theories. Since many mechanisms and forces are thought to exist, they prevent the operation of some, and calculate the effect of others, so that the effect of the mechanism they wish to check becomes apparent. Given that this is what men do, and find profitable, the world must surely be made up of continuing mechanisms and agencies as men imagine. And given that existing knowledge, which postulates particular agencies, is predictively successful, these agencies must surely bear some resemblance to those which really exist.

When scientists attempt to further our understanding of the human body they exploit existing accounts of muscular and skeletal organisation, theories of organic function, and so on. When they investigate chemical compounds and their structures they utilise taken-for- granted knowledge of stable electronic configurations and orbitals. When members analyse their own society they deploy given notions like the 'power' of unions or political groupings, or the 'interests' of classes or occupational groups. In all these cases, knowledge may be developed and extended from a taken-for-granted base. But the base is not arbitrary and merely conventional; to have gained acceptance as existing knowledge, it must have come close to describing real existing mechanisms and powers underlying appearances, and presumably it must therefore be capable of describing them more closely still if it is further articulated in the course of active investigation. This gives us a kind of modified correspondence theory of truth: knowledge is not made up of facts which correspond with appearances; it is always a set of given theories, which are evaluated to the extent that they correspond with the powers and mechanisms constantly operative in the world and thus basically constitutive of reality. Our concepts are thus putative real universals which may eventually be modified and developed until they are indeed real universals. They are not just any set of signs and conventions.

There is much to be said for such a position. It is correct to say that the very structure of the knowledge which men produce presumes that reality is constituted in terms of enduring agencies and mechanisms; this is how knowledge gains its essential coherence, and why its verbal component is viably a finite system of symbols. It is also correct to insist that existing knowledge, the material cause of new knowledge, will always embody already the results of learning, and to this extent be more than arbitrary. But neither of these points suffices to discriminate and evaluate different conflicting bodies of knowledge.

Clearly, any group of men believing in some set of real universals can take these universals as the best available rendition of reality, and use them to evaluate different beliefs. We can and do evaluate in this way, but so do those in other cultures, and so did our intellectual ancestors. If we are to regard our


23 Chapter 1

evaluations as special, we must be able to show that cur favoured real explanatory mechanisms and agencies are inherently superior to or better grounded than anybody elses, that they really are closest to the real state of things. It is the evident lack of' any way of doing this which deprives our own beliefs about the basic character of reality of any value as justified independent standards for the evaluation of knowledge claims generally. (27)

Men in different cultures and societies have understood reality in a wide variety of ways, invoking diverse causative agencies and powers allegedly at work in the universe. In simple, tribal, societies, quasi human agencies - spirits or personified forces - have often been invoked to explain natural events and human fortunes But despite assiduous investigation on the part of social anthropologists, we have no firm evidence that such beliefs are inherently unstable, nor is it clear that men who rationally endeavour to predict and control reality within such anthropomorphic cultures must eventually transcend their received perspective and recognise that their scheme of things is erroneous. (28)

It might be thought, nonetheless, that the anthropological record is not sufficiently powerful evidence in this context. Tribal beliefs are sometimes alleged not to be related to attempts to predict and control reality at all, but to be primarily related to other interests (Douglas, 1966). Hence, it is appropriate to reinforce the argument by reference to the culture of the natural sciences, the primarily instrumental interests of which can scarcely be doubted.

It is well known that as scientific knowledge has developed numerous mechanisms and theories have been postulated and successively set aside. This is, indeed, why so many philosophers of science have struggled to maintain a fact/theory distinction, and to base their justificatory rhetoric on the accumulation of facts. But there has also been a good deal of informal faith placed in the progressive quality of this sequence of theories and mechanisms. Recent historical studies, however, in particular those of. T. S. Kuhn (1970), effectively undermine this faith; they demonstrate that fundamental theoretical transitions in science are not rational responses to increased knowledge of reality, predictable in terms of context-independent standards of inference and evaluation. Such transitions make very good sense as responses to perceived practical problems, or as correlates of technical and procedural reorganisation within particular scientific communities. They are intelligible enough when referred to actual situations where new findings or new instrumentations are emerging. To this extent, they certainly are not manifestations of scientific irrationality, or mysterious emotional reorientations. But they do not possess the kind of general features which would be required by the progressive realism we are considering: it cannot be said that there is less of reality left to explain after such a transition, or that any part of the world is finally explained, or even necessarily that scientists perceive themselves as having fewer problems afterwards. Nor are we ever in a position to say that scientists could not properly have done other than they did. We simply do not find, when actual instances are studied, that the case for a particular
24 Chapter

theoretical change can be established in context-independent terms. It is never unambiguously clear that existing theories could not have reasonably been maintained, or that yet other theories night not have been produced with just as much to recommend them.

Progressive realism is one of the ideal accounts of scientific knowledge which has it moving toward something, in this case a description of the real existing mechanisms in the world. There are now several independent strands of work which imply that such theories are misconceived, and that all knowledge generation and cultural growth should be regarded as endlessly dynamic and susceptible to alteration just as is human activity itself, with every actual change or advance a matter of agreement and not necessity. Even the long-standing Popperian tradition provides an adequate feel for these points; it provides many examples of the dialectical character of science, and the way it feeds upon an ever expanding number of self-generated problems producing more work for itself with every accomplishment (rather than less, through disposing of 'part of reality'). Imre Lakatos's brilliant study of the history of Euler's Theorem (1963) is an outstanding illustration of how much there is to be learned from this tradition. But two recent general approaches to semantic change, which cannot be discussed here, convey even more clearly the merits of such a view. One is the interaction view of metaphor, and the fully general account of meaning and meaning change it involves. The other is the ethnomethodological treatment of the indexical and reflexive properties of verbal utterances. Although apparently distinct independent academic traditions are involved, there are interesting parallels between them, which derive from their common reliance on the late work of Ludwig Wittgenstein. (29)

The upshot of all this is that our current scientific models and mechanisms are likely to be seen at some future time as part of what is an endlessly unfolding chain of such mechanisms, constructed and eventually abandoned (or stripped of their ontological standing) as the activity of knowledge generation proceeds. Clearly then our present theories should stand symmetrically with earlier scientific theories, and for that matter with any other instrumentally oriented knowledge, in all sociologically relevant respects. The diverse real universals postulated at different times and in different cultures and contexts, should be regarded alike as inventions of the mind, sustained to the extent that they are instrumentally valuable in the settings where they are found. There is no means of going further and ranking or evaluating them in a way which does not simply assume the priority of one or other of them.

Knowledge cannot be understood as more than the product of men operating in terms of an interest in prediction and control shaped and particularised by the specifics of their situations. }It is not the unique possession of any particular culture or type of culture. Wherever men deploy their cultural resources to authentic tasks of explanation and investigation indicated by their interests, what they produce deserves the name of knowledge. (30) It deserves sociological study (and naturalistic or scientific study generally) as a typical example of knowledge. There is no more strictly defined conception which would discriminate say

25 Chapter 1

between 'scientific' knowledge, and other kinds, and justify different forms of sociological investigation in the two cases. We can study the process of knowledge generation, and fill out our general understanding of how it unfolds, by observing any culture wherein change is occurring under the impetus of an interest in prediction and control. (31)

What then of the problem of relativism? The first thing to be said of this is that whatever conclusions are reached on the matter should not count against the preceding discussion. If one is interested in exploring and extending the possibilities of naturalistic thought and investigation, one does not turn back because its consequences prove unpleasant. If we cannot find any naturalistic basis for differentially evaluating the knowledge of different cultures, then that is that. If epistemologists and ontologists face problems as a consequence, they must simply be accepted. What matters is that we recognise the sociological equivalence of different knowledge claims. We will doubtless continue to evaluate beliefs differentially ourselves, but such evaluations must be recognised as having no relevance to the task of sociological explanation; as a methodological principle we must not allow our evaluation of beliefs to determine what form of sociological account we put forward to explain them.

It is sometimes felt that such arguments must be rejected simply because they represent a concession to relativism. Relativism is often opposed in sociology as a matter of passion and commitment, even by those who recognise the lack of any good arguments for their case. It is felt that to do otherwise is to provide a licence for any kind of nonsensical thought, and to display a lack of interest in what the world is really like.

Although there is no need to offer concessions to such an unsatisfactory position, it should be emphasised that the merits of relativism as a philosophical position are not argued for here Nobody is enjoined to value all knowledge equally, or to choose which they will employ with a coin or a die. The prejudice of the argument is rather thoroughly naturalistic; it is naturalism which is being employed and advocated. The naturalistic equivalence of the knowledge of different cultures is merely a finding, something r which happens to be the case. To be sure, it implies the conventional status of naturalism itself, but this is no disaster. It does not imply the abandonment of naturalism in favour of a frantic search for necessity elsewhere. One can choose to continue with the relevant activities.

Naturalism, moreover, implies the most intensely serious concern with what is real, and a particular, concretely relevant conception of it is actually advocated here. Everything of naturalistic significance would indicate that there is indeed one world, one reality, 'out there', the source of our perceptions if not their total determinant, the cause of our expectations being fulfilled or disappointed, of our endeavours succeeding or being frustrated. But this reality should not be identified with any linguistic account of it, or, needless to say, with any way of perceiving it, or pictorial representation of it. Reality is the source of primitive causes, which, having been pre-processed by our perceptual apparatus, produce changes in our knowledge and

22, Chapter 1

the verbal representations of it which we possess. All cultures relate symmetrically to this reality. Men in all cultures are capable of making reasonable responses to the causal inputs they receive from reality - that is, are capable of learning. (32) That the structure of our verbal knowledge does not thereby necessarily converge upon a single form, isomorphous with what is real, should not surprise us. Why ever should we expect this to be a property of our linguistic and cognitive capabilities?

NOTES

CHAPTER 1 THE PROBLEM OF KNOWLEDGE

1 It is these metaphors which Mannheim uses to make some of his most radical criticisms of existing conceptions of knowledge. Rather than attacking the predominantly individualistic and contemplative epistemologies of his time in terms of his explicit alternative conception, we find him instead resorting to a more sophisticated usage of the same contemplative standpoint. Take, for example, his treatment of the diverse, apparently mutually incompatible, views of a society, characteristically associated with its different social classes or sub-cultures. The usual way of accounting for these in contemplative terms was to hold that since there was only one reality with which verbal accounts could correspond, only one such account, at most, could be correct. The other incorrect, accounts would probably be ideologies generated in response to social interests. Mannheim applies contemplative conceptions in a more sophisticated way, employing a pictorial metaphor to advantage. We are told to look at a physical object and consider what we see. It will be a partial view of the object, a particular perspective depending upon our particular position with respect to the object. If we found other observers, working from different standpoints, in possession of completely different conceptions or perspectives, we should not assume that those perspectives were erroneous simply because they differed from our own. Why then should we not treat different conceptions of society as the product of different standpoints and acknowledge that all of them may have value, or as Mannheim would say, limited validity? And why should we not recognise that we can learn from all the different perspectives, just as we can learn more and more about a physical object by observing it from different standpoints (cf. Mannheim (1936), ch.. 5, pt 4)?

2 A parallel discussion of perception should, ideally, accompany this discussion of representation. We do indeed learn to see the world in terms of meaningful symbols - as assemblages of culturally meaningful components. The raw material gathered by our senses is actively processed and schematised before it becomes perceived sensation. Perception is selective: we see in terms of the interests which affect us directly, or indirectly through their effect on our socialisation.

3 Every representation can be put to two kinds of use. It can be routinely applied, in conjunction with the procedures into which it fits, to any of the ends for which the procedure is routinely appropriate. And it can be taken itself as a cultural resource in the generation of new knowledge, just as it was made from previous cultural resources (cf. Kuhn, Postscript, 1970, for a discussion). Normally, the former is the institutionalised usage and the latter an occasional variant. But in science the reverse can sometimes be the case, and this is what accounts for the particular problems alluded to here.

4 Ivins (1953) notes that when the camera first came into use there was much interest in 'photographic distortion'. This interest only declined as we allowed the camera to define our idea of accurate representation. Photographs appear infrequently in scientific textbooks. They play a significant role in some astronomical and geological works and occasionally in biology. But, given their cheapness in production, what Ivins calls their 'exact repeatability', and their elimination of much personal idiosyncracy from the illustra-
tive process, it is perhaps initially surprising that they are not used more. The probable reason for this is that other illustrative techniques offer much greater possibilities for intervention in the process of manufacture. The esoteric theories and abstractions of science necessitate esoteric ways of seeing and highly schematic illustration. It is the diagram which reigns supreme in physics and chemistry texts, and plays a major role elsewhere.

5 It is interesting to note how people intuitively translate their awareness of the conventions of a pictorial representation into a statement of how realistic it is. Thus, Daniel Gasman (1971) reproduces some drawings of marine organisms by the German scientist-philosopher Ernst Haeckel and comments:
they are not quite objectively rendered and the information
they are supposed to convey is hardly neutral. Their orna
mental lay-out and hypertrophied patterning, and the fantastic
and bizarre look of the unfamiliar flora and fauna, transform
them in the direction of disquieting, even nightmarish repre
sentations that seem to be related to the type of naturalistic
mysticism which can be observed in late nineteenth-century Art
Nouveau and symbolist artists like Obvist and Redon. (pp. 73-4)

It is clear that Gasman could not have checked with the organisms themselves - the 'reality' in question. His opinion derives from the conventions of the drawings he reproduces in his book. Their sinuous lines, heavy contrast, directional lighting, stressed planes of symmetry and implied upward motion are far from our
presently accepted conventions of scientific representation, and have typically been used to convey, in the art of Western cultures, profundity and emotional intensity. It remains, none the less, an intriguing question as to why Haeckel chose to employ these conventions in a 'scientific' work; Gasman was right to find the
drawings interesting.

6 For science as a craft skill cf. Polanyi (1958), Ravetz (1971), For a discussion of the detailed relationship of diagrams, and competences or 'techniques of inference', in science cf. Toulmin (1953).

7 Cf. Bloor (1973, 1976). It is interesting that some people, impressed with the power of mathematical knowledge and puzzled as to its origin, have claimed that there exists a world of mathematical objects, accessible to thought. The knowledge is thus provided with a reality to which it can correspond.

8 This very important point is either explicitly accepted by the writers discussed in the next section, or is compatible with their work. Habermas has perspicaciously discussed Marx's position on this matter, showing how he retained a belief in the primacy of 'external nature' even as he held that man made his world in the process of work (1972, pt 1.2).

9 Two general points are worth bearing in mind about Lukacs's account. First, wherever possible, Lukacs talks of consciousness and its determinants rather than of knowledge. He prefers to think in terms of active mental processes rather than in terms of the nature of knowledge; in this he is at the opposite extreme to Habermas. And indeed the question of the relationship of mental processes and the hypothesised knowledge on the basis of which they proceed is a topic of great complexity and fascination. But it would not be correct to imagine that Lukacs, like some modern ethnomethodologists, considered knowledge as a term which merely reifies consciousness, and misleads however it is used. It is perfectly in order to talk of Lukacs's account of the character of knowledge.

Second, Lukacs was primarily concerned with men's understanding of social reality, and rarely considered natural knowledge. He did, none the less, include natural science within the ambit of his views, and purported to have exposed the inadequacies of its method. Lukacs was, however, monumentally ignorant of scientific practice and, against his own precepts, equated natural science with the abstractions of positivist philosophy; hence we have the absurdities of his vituperations against science. 10 Lukacs's account was eschatological in character. A total understanding of reality would eventually be achieved by the proletariat. This element of Lukacs's thought, with its obvious and well-documented inadequacies, will not be discussed here, although doubtless his desire to demonstrate and justify the historical role of the proletariat influenced his work at many levels. 11 Consider, for example, another important theme in Lukacs's 'History and Class-Consciousness' (1923) - that, since our knowledge is of a changing, man-influenced reality, it must itself perpetually change. Clearly there is some point to this claim: if we were to change the rules of chess, our knowledge of how to play the game well would change: if a laissez-faire economy were slowly to metamorphose into monopoly capitalism, many associated economic rules and theories would doubtless be modified: if the last remaining giant pandas were to be exterminated, doubtless our catalogues of existing fauna would be adjusted accordingly. But what justifies giving this relationship general significance? After all, for over two millennia the central techniques and descriptive categories of Euclidean geometry and Archimedean mechanics have been found applicable and acceptable in diverse societies. Do not people seek to embody in their knowledge principles which they find in a sense invariant over a wide range of contexts, applicable to systems even during change? Is not much knowledge precisely knowledge of what is involved in change? A changing reality only implies the transient character of knowledge generally on the assumption that knowledge copies or reflects
reality at a superficial level. Yet Lukacs made the implication and made much of it. Ironically, it was one of the components in Lukacs's attack upon science which he conceived of as a reified body of knowledge seeking to pass off contingent facts of present reality as manifestations of eternal laws. The practice of natural
science is in fact thoroughly dialectical.

12 Habermas's discussions of Marx's and Pierce's epistemologies
are especially rewarding.

13 Habermas does not interrupt the flow of his text with examples. And, indeed, to illustrate his case convincingly, as it can be, would be a lengthy task in the context of most scientific fields. There are, however, some fields, like cartography, which are sufficiently accessible to 'outsiders' and provide immediate and intuitively satisfying support for Habermas's account. If one can see how an atlas is the product of various instrumental interests, rather than of undirected contemplation, and how it is communally sustained as a repository of knowledge, one has an excellent concrete model of how Habermas's account can treat of
knowledge generally.

14 Had the interest of this book been primarily epistemological, it would have been necessary to emphasise the differences between Habermas's and traditional instrumentalist epistemologies (cf. Habermas. 1973). Habermas's own interest is itself primarily epistemological; like Lukacs his primary goal is the refutation
of 'positivism' and the construction of an alternative form of 'self-understanding' for the sciences. Unfortunately, in doing so, he retains far too much of the 'positivism' he criticises.

15 Habermas is clear that KCIs emerged in the course of man's natural history, during the self-constituting evolutionary process which is both our past, and our present condition. But he makes little further progress in characterising this emergence. KCIs cannot be considered entirely as the products of cultural evolution. for culturally defined rules, problems and standards appear as such only within frames of reference defined by KCIs
themselves. On the other hand, Habermas is most anxious that KCIs should not be considered entirely by thinking of 'reason as an organ of adaptation' or knowledge as an instrument of adaptation to a changing environment. He ends by asserting that KCIs derive both from nature (biologically evolved cognitive capacities?) and from the cultural break with nature (cf. Habermas, (1972), pp. 312, 196-7).

16 For Habermas 'practical' implies a contrast with 'technical': it is taken in its Germanic sense to imply 'moral' or 'ethical', usually with reference to the political context.

17 Habermas's ideal of self-reflective knowledge will not be considered further here. A detailed critique would follow the same lines as, and emerge as isomorphous with, the discussion of the hermeneutic ideal which follows in the main text.

18 A formulation, or reformulation, of this kind seems to be
implied in Habermas (1973).

19 Among the work illustrating the consensual character of science, that of T. S. Kuhn (1970) is most noteworthy. For the indexical character of scientific knowledge and its context-dependence cf. Barnes and Law (1976). A general discussion of the social character of natural scientific knowledge with further references
to concrete studies is Barnes (1974).

20 To say this is not to propose an alternative explanation to that which relates Habermas's views to expedient social interests. It is frequently and plausibly suggested that Habermas is a humanist intellectual, responding to the threat of redundancy as the scope of the technique of natural science is extended. Realising that to attack the validity of natural science itself is an unrealistic strategy in the modern world, he seeks to limit its scope and assert its subservience, in the last analysis, to the field of learning he himself represents. Hence he attacks scientistic philosophy and what he considers to be the extension of science into the realm of human affairs. Since Habermas's views are strongly criticised here, it is perhaps worth noting that I share
his suspicion of the growth of certain quantitative methods in the social sciences: systems analysis, econometrics, cybernetics and the rest. Much of this material (although certainly not all) is worse than useless. But it does not represent the extension of natural scientific techniques. Its techniques, like its
interests, spring from a different source. They are, however, usually legitimated by reference to some extreme form of positivist philosophy of science, and, as elsewhere, Habermas fails to make the essential distinction between philosophy of science and natural science itself.
21 It follows from this view that reference to the 'disinterested evaluation' of knowledge is in most contexts a harmless enough formulation, which can be taken as practically equivalent to 'evaluation in terms of an authentic interest in prediction and control'.
22 In some spheres, notably natural science, an attempt is made to enforce the transmission of one message only, presumably to minimise the effects of crossover distortion. There is no 'a priori' reason why such an attempt should not succeed and produce an entirely non-evaluative information flow. If this occurred
in a natural science it could be legitimately said to be non evaluative in a certain restricted sense. Its discourse and knowledge would, of course, still be sustained socially and would remain normative in that sense. The electrical metaphor is useful again, in distinguishing these two senses in which science can be said to be normative. To talk of the normative component, in one sense, is like talking of the visual information in a TV signal - information which could be entirely eliminated to leave the sound only. To talk of the normative component in the second sense is to talk of the conventions in the code which is used to convey any information by the TV signal: there must always be such conventions, but to an extent they are a matter of choice and agreement.


92 Notes to Chapter 1

23 Much the same view of knowledge generation is found in Bhaskar (1975). Where I emphasise the instrumental features of the account by talking of cultural resources, Bhaskar, who is a realist, uses Aristotelean terminology and talks of material causes or transitive objects of knowledge. But the terms are substantially equivalent.
24 For a discussion of pictorial representation in science which deals with both cultural resources and instrumental interests cf. Rudwick (1976).
25 Thus, Habermas, whose perception is dependent upon the tradition of epistemological writing, argues that there is only one possible natural science; it is not, as Marcuse would have it, a historical project which could be different. In doing so he remains unaware of the material causes of scientific knowledge, and talks vaguely of science as 'pure instrumentality'. Popper's early epistemology (1934) is, in contrast, engagingly
direct upon these matters. In the terms it specifies, rational men can indeed hold to diverse bodies of knowledge; there is little restriction on the nature of what can be rationally believed. Popper's epistemology does not identify the best knowledge but the most rational men. Moreover, it identifies the most rational men
in conventional terms, not in absolute ones; Popper is clear that his epistemological standards have only the standing of conventions. Hence, Popper provides us with no naturalistic basis upon which to differentiate and evaluate knowledge claims. His position is just as relativistic in its implications as that which follows
here. The truly remarkable thing is how rarely this is noted (cf. Barnes, 1976).
26 Another possibility, not discussed here, is to assert the progressive nature of knowledge, and the possibility of differentially evaluating different knowledge claims, on purely instrumental grounds. The best knowledge is that which enables its possessors to do the most, to achieve their ends the most successfully.
Unfortunately, there seems no easy way of applying this criterion, since different cultures possess different competences and seek different ends. It would appear necessary to set prior evaluations upon different aims and activities before an instrumental assessment of knowledge claims could be carried out (just as later in
the text we find that prior evaluations of real universals are essential before a realist assessment of knowledge claims can be carried out). It remains an empirical possibility that men in all cultures would, on acquaintance, admit the superior instrumental efficacy of western science. But this does not effect the thesis
that different knowledge claims should stand symmetrically for sociological purposes. 
27 Bhaskar, whose (1975) was the model for the foregoing argument, recognises this conclusion by acknowledging that there is no way of avoiding epistemological relativism.
28 Cf. Polanyi (1958), Wilson (1971), Horton and Finnegan (1973), Barnes (1974).
29 The relevant ethnomethodological literature is too well known to need citing here, but references to the interaction view, which is less familiar to sociologists, can be found in Barnes (1974). The interaction view is applied to science in Hesse (1974), and an indication of the applicability of the notion of indexicality to scientific expressions is given in Barnes and Law (1976). 30 I believe it also to be the case that knowledge everywhere is based upon the same range of shared cognitive propensities. If there is indeed such a psychic unity among men, it reinforces the case for treating all institutionalised beliefs symmetrically as knowledge. 31 The general cognitive processes involved in knowledge generation cannot be considered here. A stimulating attempt to formulate them and exemplify their operation is Hesse (1974). 32 Analogously, moral and evaluative beliefs are doubtless modifiable by primitive causal inputs with a real basis. Evaluations and ethical views are no more immune to change with changing experience than are descriptive views. Nor are they any more varied and diverse. Curiously, moral relativism is much easier to accept today than descriptive relativism (Lukes, 1974). But the arguments for and against are identical in both cases. On the one hand, alternative beliefs are rationally possible, and actually found, in both cases. On the other hand, in neither case is belief so arbitrary and uninfluenced by real primitive causes that we can choose to believe whatever we like. Actions we can choose, beliefs, strangely, we cannot. We cannot simply decide to believe that bullets are harmless, nor that child killing is every man's duty. We could say as much, but our actions would betray us in both cases. (Needless to say, both the above are believable, and have been believed, in other contexts, but however one develops the argument the essential symmetry
between the two cases remains.)