New Rules of Sociological Method
A Positive Critique of Interpretative Sociologies
Anthony Giddens Second Edition
Polity Press
Introduction to the
First Edition
As we know them today, the social sciences were shaped by the spectacular advances of natural science and technology in the late eighteenth and the nineteenth centuries. I say this bluntly, iri awareness of the complexities which it conceals. It would certainly not be true to say that the successes of human beings in seemingly mastering nature intellectually in science, and materially in technology, were adopted uncritically as forming a model for social thought. Throughout the nineteenth century, idealism in social philosophy and romanticism in literature, in their various guises, maintained their distance from the intellectual standpoints fostered by the natural sciences, and normally expressed deep hostility to the spread of machine technology. But for the most part, authors within these traditions were as sceptical of the possibility of creating a science of society as they were distrustful of the claims of the sciences of nature, and their views served as no more than a critical foil to the much more influential writings of those who sought to create just such a science. Mentioning just one or two figures in isolation is risky, but I think it reasonable to regard Comte and Marx as the pre-eminent influences upon the subsequent development of the social sciences (I shall use this term primarily to refer to sociology and anthropology, but shall also on occasion make reference to economics and to history). Comte's influence is fundamental since, as projected through Durkheim's writings, his conception of sociological
Introduction to the First Edition    17
method can readily be traced through to some of the basic themes of 'academic sociology' and anthropology in the twentieth century. Following Marx's own scornful dismissal of Comte, Marxism set itself against those streams of social theory connected to the emphases of the former author. Comte's formulation of the idea of a natural science of society was actually a sophisticated one, as anyone can check for himself by glancing through no more than a few pages of the Philosophie Positive, even if it lacked the subtleties (and, it must be said, some of the logical difficulties) of Marx's work, informed as the latter was by a transposed Hegelian dialectic. Both Comte and Marx wrote in the shadow of the triumphs of natural science, and both regarded the extension of science to the study of human conduct in society as a direct outcome of the progressive march of human understanding towards humanity itself.
Comte sanctified this as a doctrine. The 'hierarchy of the sciences' expresses not only a logical order of relations but an historical one too. Human knowledge first of all dispels the shrouds of mysticism in those areas of nature furthest from human involvement and control, in which humanity appears to play no role as subject: first mathematics, and then astronomy. The development of science subsequently edges closer and closer to human life, moving through physics, chemistry and biology to the creation of sociology, the science of human conduct in society. It is easy to see how, even before Darwin, evolutionary theory in biology seemed to prepare the stage for the explication of human conduct according to principles of scientific reason, and to appreciate Marx's enthusiasm for the Origin of Species as offering a parallel to what he and Engels sought to accomplish in their work.
An end to mystery, and an end to mystification: this is what Comte and Marx alike anticipated and strove for. If nature could be revealed as a secular order, why should human social life remain enigmatic? For perhaps there is only a short step from scientific knowledge to technical mastery; with a precise scientific understanding of the conditions of their own social existence, why should not people be able rationally to shape their own destiny? The Marxian vision is ambiguous: and some versions of what Marx had to say, I believe, can be reconciled without
18    Introduction to the First Edition
difficulty, on the level of ontology at least, with this present study. I refer to those versions of Marx which regard Marxism, not as a natural science of society which happened to predict the demise of capitalism and its replacement by socialism, but as an informed investigation into the historical interconnections of subjectivity and objectivity in human social existence. But in so far as there were strongly naturalistic strains in Marx's writings, and most certainly there were, Marx can be categorized along with Comte as previsaging, and seeking to bring into being, a science of society which would reproduce, in the study of human social life, the same kind of sensational illumination and explanatory power already yielded up by the sciences of nature. By this token, social science must surely be reckoned a failure. Beside the seeming certainties, the system of precise laws attained in classical mechanics, that model for all aspiring sciences after Newton, which in the nineteenth century was unquestioningly assumed to be the goal to be emulated, the achievements of the social sciences do not look impressive.
This much is accepted, and necessarily so, by those in the social sciences today who cling to the same sort of ideal. The wish to establish a natural science of society, which would possess the same sort of logical structure and pursue the same achievements as the sciences of nature, remains prominent. Of course, many who accept it have relinquished the belief, for various reasons, that social science, in the near future, will be able to match the precision or the explanatory scope of even the less advanced natural sciences. However, a sort of yearning for the arrival of a social-scientific Newton remains common enough, even if today there are perhaps many more who are sceptical of such a possibility than still cherish such a hope. But those who still wait for a Newton are not only waiting for a train that will not arrive, they are in the wrong station altogether.
It is of the first importance, of course, to trace out the process whereby the certainties of natural science itself have been assaulted in the twentieth century. This has to a large extent come about through the internal transformation of physics and the setting aside of Newton by Einsteinian relativity, complementarity theory and the 'uncertainty principle'. But of equal significance, to this study at least, is the appearance of new forms
Introduction to the First Edition    19
of the philosophy of science. One might identify two intertwining yet ultimately opposed trends in the philosophy of science over the past forty or fifty years, in the wake of the perturbations experienced in classical physics. On the one side - and this is not at all paradoxical - there has been the attempt to sustain the claim that natural scientific knowledge, or a particular characterization of it, should be regarded as the exemplar of everything which can be regarded legitimately as 'knowledge'. If the famous 'verification principle' was itself rapidly shown to be incapable of verification, and the radical attempt to expunge metaphysics from human affairs was soon abandoned, the influence of logical positivism or logical empiricism remains strong, if not preponderant. In recent decades, this orthodoxy has been challenged with mounting success. In this challenge the works of Karl Popper played a pivotal, if not entirely unambiguous, role. Whatever Popper's original views may have been, his critique of inductive logic and his insistence that, though claims to knowledge in science have to begin somewhere, there is nowhere where they have to begin, were of decisive importance, not only for their own value, but as a springboard for many subsequent contributions.
Some such discussions in natural science have an immediate significance for epistemological problems in the social sciences. But in any case I want to assert that social science should move out of the shadow of the natural sciences, in whatever philosophical mantle the latter be clad. By this I do not mean to say that the logic and method of the study of human social conduct are wholly discrepant with those involved in the study of nature, which I certainly do not believe; nor do I propose to support the view expressed by those in the tradition of the Geisteswissenschaften, according to which any sort of generalizing social science is logically ruled out of court. But any approach to the social sciences which seeks to express their epistemology and ambitions as directly similar to those of the sciences of nature is condemned to failure in its own terms, and can only result in a limited understanding of human society.
The failure of social science, when thought of as a natural science of society, is manifest not only in the lack of an integrated corpus of abstract laws, whose circumstances of application
20    Introduction to the First Edition
are precisely known, and which meet with the acceptance of a 'professional community'; it is evident in the response of the lay public. Conceived as a project by Comte and Marx, social science was to be revelatory, to sweep away the opaque prejudices of earlier times and replace them with rational self-understanding. What appears as the 'resistance' of the lay public to the 'findings' of social science is often simply equated with the opposition that has sometimes been provoked by theories of the natural world: for example, a disinclination to accept that the world is spherical rather than flat. But that sort of resistance is aroused by scientific theories or discoveries which shake or disturb common sense (I do not want to touch here upon the opposition of vested interests to scientific ideas). The objection which lay members of society frequently have to the claims of sociology is just the opposite: that its 'findings' tell them nothing which they did not already know - or worse, dress up in technical language that which is perfectly familiar in everyday terminology. There is a disinclination among those involved in the social sciences to take this sort of protest seriously: after all, haven't the natural sciences often shown that beliefs which people took for granted, which they 'knew', were in fact mistaken? Why should we not merely say that it is the task of social science to check upon common sense, to see whether lay members of society do really know what they claim to know? I want to suggest, however, that we have to take the objection seriously, even if in the end it is not sustained: for, in some sense that is not at all easy to spell out, society is the outcome of the consciously applied skills of human agents.
The difference between society and nature is that nature is not a human product, is not created by human action. While not made by any single person, society is created and recreated afresh, if not ex nihilo, by the participants in every social encounter. The production of society is a skilled performance, sustained and 'made to happen' by human beings. It is indeed only made possible because every (competent) member of society is a practical social theorist; in sustaining any sort of encounter he or she draws upon social knowledge and theories, normally in an unforced and routine way, and the use of these practical
Introduction to the First Edition    21
resources is precisely the condition of the production of the encounter at all. Such resources (which I shall later call generi-cally 'mutual knowledge') as such are not corrigible in the light of the theories of social scientists, but are routinely drawn upon by them in the course of any researches they may prosecute. That is to say, a grasp of the resources used by members of society to generate social interaction is a condition of the social scientist's understanding of their conduct in just the same way as it is for those members themselves. While this is easily appreciated by an anthropologist who visits an alien culture, and who seeks to describe the conduct observed there, it is not as transparent to anyone studying conduct within a familiar cultural frame, who tends to take such mutual knowledge for granted.
Recent developments in sociology, drawing in large part upon not so recent developments in analytic philosophy and phenomenology, have been very much concerned with these matters. That such an interchange between the social sciences and philosophy should have occurred is not surprising, since what distinguishes some of the leading standpoints within these broad philosophical traditions - namely 'existential phenomenology', 'ordinary language philosophy' and the philosophy of the later Wittgenstein - is a resurgent interest in action, meaning and convention in the context of human social life. Now a concern with problems of action is certainly not alien to existing orthodoxies in the social sciences. The term 'action' itself, in the shape of the 'action frame of reference' occupies a prime place in the work of Talcott Parsons. In his earlier writings at least, Parsons specifically sought to incorporate a 'voluntaristic' frame within his approach. But Parsons (like J. S. Mill) went on to identify voluntarism with the 'internalization of values' in personality and hence with psychological motivation ('need-dispositions'). There is no action in Parsons's 'action frame of reference', only behaviour which is propelled by need-dispositions or role-expectations. The stage is set, but the actors only perform according to scripts which have already been written out for them. I shall try to trace out some further implications of this later on in this study. But is it any wonder that laypeople find it hard to recognize themselves in such theories? For although Parsons's writings are in these
22    Introduction to the First Edition
respects vastly more sophisticated than those of many others, we do not appear in them as skilled and knowledgeable agents, as at least to some extent masters of our own fate.
The first part of this study consists of a brief and critical Cook's tour through some prominent schools of social thought and social philosophy. There are striking, and not very widely acknowledged, points of connection between, on the more abstract level of the philosophy of being, Heidegger and the later Wittgenstein and, so far as the social sciences are concerned, the lesser figures of Schutz and Winch. There is one very substantial difference between the latter two: Schutz's philosophy remained wedded to the standpoint of the ego, and hence to the notion that we can never achieve more than a fragmentary and imperfect knowledge of the other, whose consciousness must forever remain closed to us^fwhile for Winch, following Wittgenstein, even our knowledge of ourselves is achieved through publicly accessible semantic categories. But both insist that, in formulating descriptions of social conduct, the observing social scientist does, and must, depend upon the typifications, in Schutz's term, used by members of society themselves to describe or account for their actions; and each, in his different way, underlines the significance of reflexivity or self-awareness in human conduct. Since what they have to say is in some respects not too dissimilar, it is not very surprising that their writings have much the same sort of limitations - limitations which I think are shared by many who have written about the 'philosophy of action', especially those, like Winch, influenced above all by the later Wittgenstein. 'Post-Wittgensteinian philosophy' plants us firmly in society, emphasizing both the multifold character of language and the way it is embedded in social practices. However, it also leaves us there. The rules governing a form of life are taken as a parameter, within and with reference to which modes of conduct may be 'deciphered' and described. But two things are left obscure: how one is to set about analysing the transformation of forms of life over time; and how the rules governing one form of life are to be connected to, or expressed in terms of, those governing other forms of life. As some of Winch's critics have pointed out (Gellner, Apel, Habermas), this easily terminates in a relativism which breaks off just where some of the basic issues
Introduction to the First Edition    23
which confront sociology begin: problems of institutional change and the mediation of different cultures.
It is remarkable how frequently conceptions which at least in certain important respects parallel that of 'forms of life' (language-games)  appear in schools  of philosophy or social theory which have little or no direct connection to Wittgenstein's Philosophical Investigations: 'multiple realities' (James, Schutz), 'alternate realities' (Castaneda)r'language structures' (Whorf), 'problematics'   (Bachelard,   Althusser),   'paradigms'   (Kuhn). There are, of course, very basic differences between the philosophical standpoints which these express, and the sorts of problems which their authors have developed to try to illuminate them. Each of them in some part signals a movement along a broad front in modern philosophy away from empiricism and logical atomism in the theory of meaning; but it is not difficult to see how the emp_hasis_ujjpn.discii^eJ.iiniyeises of meaning' can allow the principle of relativity of meaning and experience to become relativism caught in a vicious logical circle, and unable to deal with problems of meaning-variance. I shall try to show in the course of this study how it is possible, and important, to sustain a principle of relativity while rejecting relativism. This depends upon escaping from the tendency of some if not most of the authors just mentioned to treat universes of meaning as 'self-contained' or unmediated. Just as knowledge of the self is, from the earliest experience of the infant, acquired through knowledge of others (as G. H. Mead showed), so the learning of a language-game, the participation in a form of life, occurs in the context of learning about other forms of life that are specifically rejected or are to be distinguished from it. This is surely compatible with Wittgenstein, whatever some of his followers may have made of his ideas: a single 'culture' incorporates many types of language-game on levels of practical activity, ritual, play and art; and to become acquainted with that culture, as a growing infant or as an alien observer or visitor, is to come to grasp the mediations of these in moving between languages of representation, instrumentality, symbolism etc. In quite different contexts, Schutz talks of the 'shock' of moving between different 'realities', and Kuhn refers to the apprehension of a new 'paradigm' as a sudden 'Gestalt switch'. But although such sudden
24    introduction to the First Edition
transitions no doubt occur, the ordinary member of society quite routinely shifts between different orders of language and activity, as do scientists on the level of theoretical reflection.
Parsons argued that the most significant convergent idea in modern social thought concerns the 'internalization of values', as independently arrived at by Durkheim and Freud; I think a better case can be made for the notion of the social (and linguisticjíjíomidauon of reflexivity,* as independently arrived at, from widely varying perspectives, by Mead, Wittgenstein and Heidegger - and, following the latter, Gadamer. Self-consciousness has always been regarded, in positivistically inclined schools of social theory, as a nuisance to be minimized; these schools endeavour to substitute external observation for 'introspection'. The specific 'unreliability' of the 'interpretation of consciousness', indeed, whether by the self or by an observer, has always been the principal rationale for the rejection of Verstehen by such schools, i The intuitive or empathic grasp of consciousness is regarded by them merely as a possible source of hypotheses about human conduct (a view which is echoed even in Weber). In the tradition of the Geisteswissenschaften in the nineteenth and early twentieth centuries, Verstehen was regarded above all as a method, a means of studying human activity, and as such as depending upon the 'reliving' or 're-enactment' of the experiences of others. Such a view, as held by Dilthey and later in modified form by Weber, was certainly vulnerable to the strictures levelled against it by positivistic opponents, since both Dilthey and Weber, in their varying ways, wanted to claim that the 'method of understanding' yields material of an 'objective', and therefore intersubjectively verifiable, kind. But what these writers called 'understanding' is not merely a method for making sense of what others do, nor does it require an emphatic grasp of their consciousness in some mysterious or obscure fashion: /(is the very ontological condition of human life in society as such. This is the central insight of Wittgenstein and of certain versions of existentialist phenomenology; self-understanding is connected integrally to the understanding of others. Intentionality, in the phenomenological sense, is not thus to be treated as an expression of an ineffable inner world of private mental experiences, but as necessarily
Introduction to the First Edition    25
drawing upon the communicative categories of language, which in turn presuppose definite forms of life. Understanding what one does is only made possible by understanding, that is, being able to describe, what others do, and vice versa. It is a semantic matter, rather than a matter of empathy; an,d_ rfeflgxivity, as the distinctive property of the human species, k intimately and integrally dependent^upjonJhejsQciaLcharacter^language.
Language is first of all a symbolic or sign-system; but it is not simply, or even primarily, a structure of 'potential descriptions' - it is a medium of practical social activity. The organization of 'accountability', as has been made fully clear in existentialist phenomenology after Heidegger, is the fundamental condition of social life; the production of 'sense' in communicative acts is, like the production of society which it underpins, a skilled accomplishment of actors - an accomplishment that is taken for granted, yet is only achieved because it is never wholly taken for granted. Meaning in communicative...acts, as it is produced by lay actors, cannot be grasped simply in terms of a lexicon, any more than it can be transcribed within frameworks of formal logic that pay no attention to context-dependence. This is surely one of the ironies of some sorts of supposedly precise 'measures' employed in the social sciences, quite properly resented by the lay public since the categories often appear foreign and imposed.
In this study, I discuss several schools of thought in social theory and social philosophy, from the phenomenology of Schutz to recent developments in hermeneutic philosophy and critical theory. I shall try to make it clear what, if anything, I have borrowed from each of these schools, and shall attempt to indicate some of their shortcomings. This essay is not, however, intended to be a work of synthesis, and while I shall specifically draw attention to several parallel currents in social thought in the contemporary period, it is not my objective to seek to show an immanent process of convergence which will finally establish a secure logical framework for sociology. There are some standpoints in contemporary social thought which I have not analysed in a detailed way, even though much of what I have to say bears directly upon them. I have in mind functionalism, structuralism and symbolic interactionism - labels for an array of views which are diverse, to be sure, but each of which possesses
26    Introduction to the First Edition
certain central and distinctive themes of its own. I shall indicate only cursorily here why the arguments developed in this study diverge from those characteristic of such traditions of social theory.
There are four key respects in which I shall say that func-tionalism, as represented at least by Durkheim and Parsons, is essentially wanting. One I have already alluded to earlier: the reduction of human agency to the 'internalization of values'. Second: the concomitant failure to treat social life as actively constituted through the doings of its members. Third: the treatment of power as a secondary phenomenon, with norm or 'value' residing in solitary state as the most basic feature of social activity and consequently of social theory. Fourth: the failure to make conceptually central the negotiated character of norms, as open to divergent and conflicting 'interpretations' in relation to divergent and conflicting interests in society. The implications of these failures are so damaging, I think, that they undermine any attempt to remedy any rescue functionalism by reconciling it with other perspectives of a different sort.
Use of the term 'structure' has no particular connection with 'structuralism', any more than 'sign' has with semiology. 1 definitely want to maintain that 'structure' is a necessary concept in social theory and shall make use of it below. But I shall want to distinguish my version of the concept both from that characteristic of Anglo-American functionalism, where 'structure' appears as a 'descriptive' term, and from that of the French structuralists, who use it in a reductive way; both types of usage of the notion of structure, I shall say, lead to the conceptual blot-ting-out of the active subject.
Symbolic interactionism is the only one of these three schools of thought to accord primacy to the subject as a skilled and creative actor; in American social theory in particular it was for many decades the only major rival to functionalism. Mead's social philosophy, in an important sense, was built around reflexivity: the reciprocity of the T and the 'me'. But even in Mead's own writings, the constituting activity of the 'ľ is not stressed. Rather, it is the 'social self with which Mead was preoccupied; and this emphasis has become even more pronounced in the writings of most of his followers. Hence much of the possible
Introduction to the First Edition    27
impact of this theoretical style has been lost, since the 'social self can easily be reinterpreted as the 'socially determined self, and from  then on  the differences  between  symbolic interactionism and functionalism become much less marked. This explains why the two have been able to come together in American social theory, where the differentiation between symbolic interactionism - which from Mead to Goffman has lacked a theory of institutions and institutional change - and functionalism has become typically regarded as merely a division of labour between 'micro-' and 'macro-sociology'. I wish to emphasize in this study, however, that the problem of the relation between the constitution (or, as I shall often say, production and reproduction) of society by actors, and the constitution of those actors by the society of which they are members, has nothing to do with a differentiation between micro- and macro-sociology; it cuts across any such division.
162   The Form of Explanatory Accounts
nevertheless remain a slave. Yet it is fundamental to recognize that 'objective' causal conditions that influence human action can in principle be recognized and incorporated into that action in such a way as to transform it.
This observation concerns features of human activity that bear only a superficial resemblance to indeterminacy in physics. It is sometimes argued that self-fulfilling and self-negating predictions do not present a 'difficulty' unique to the social sciences, since in natural science also observations made about a series of events may influence the course of those events. However, in social science, 'indeterminacy' - a poor term in this connection - results from the incorporation of knowledge as a means to the securing of outcomes in purposeful conduct. Self-influencing observations or predictions represent one aspect of a much more far-reaching phenomenon in sociology than is true of natural science.
Conclusion: Some New Rules of Sociological Method
At this point I shall recapitulate some of the themes of this brief study and try to draw some of the threads together. The schools of 'interpretative sociology' which I discussed in chapter 1 have made some essential contributions to the clarification of the logic and method of the social sciences. In summary form, these are the following: the social world, unlike the world of nature, has to be grasped as a skilled accomplishment of active human subjects; the constitution of this world as 'meaningful', 'accountable' or 'intelligible' depends upon language, regarded, however, not simply as a system of signs or symbols but as a medium of practical activity; the social scientist of necessity draws upon the same sorts of skills as those whose conduct he or she seeks to analyse in order to describe it; generating descriptions of social conduct depends upon the hermeneutic task of penetrating the frames of meaning which lay actors themselves draw upon in constituting and reconstituting the social world.
These insights, however, derive from schools of thought which stand close to philosophical idealism and manifest the traditional shortcomings of that philosophy when transferred to the field of social analysis: a concern with 'meaning' to the exclusion of the practical involvements of human life in material activity (for while it is true that human beings do not produce the world of nature, they do none the less produce from it, and actively transform the conditions of their own existence by so
164   Conclusion
doing); a tendency to seek to explain all human conduct in terms of motivating ideals at the expense of the causal conditions of action; and a failure to examine social norms in relation to asymmetries of power and divisions of interest in society. These shortcomings cannot be rectified within the traditions of thought in which they originate, but nor can the positive contributions which they go along with be readily accommodated within rival theoretical schemes that have translated human agency into social determinism, and which have retained strong associations with positivism in philosophy. Three interlacing orders of problems have to be resolved in order to transcend the limitations of interpretative sociologies, concerning: the clarification of the concept of action and the correlate notions of intention, reason and motive; the connecting of the theory of action to the analysis of the properties of institutional structures; and the epistemological difficulties which confront any attempt to elucidate the logic of social-scientific method.
The failure of the Anglo-American philosophy of action to develop a concern with institutional analysis is reflected in an overconcentration upon purposive conduct. Thus many authors have been inclined to assimilate 'action' with 'intended action', and 'meaningful act' with 'intended outcome'; and they have not been much interested in analysing the origins of the purposes that actors endeavour to realize, which are assumed as given, or the unintended consequences that courses of purposive action serve to bring about. Freeing the concept of action as such, and the identification of the meaning of acts, from any necessary connection with intentions distances the hermeneutic tasks of social science from subjectivism, and makes possible a clarification both of the nature of the causal conditions of action and of the double hermeneutic with which the social sciences are inevitably involved.
'Intention', 'reason' and 'motive', I have argued, are all potentially misleading terms, since they already presuppose a conceptual 'cutting into' the continuity of action, and are aptly treated as expressing an ongoing reflexive monitoring of conduct that 'competent' actors are expected to maintain as a routine part of their day-to-day lives. The reflexive monitoring of conduct only becomes the statement of intentions, or the giving
Conclusion   165
of reasons, either when actors carry out retrospective enquiries into their own conduct or, more usually, when queries about their behaviour are made by others. The rationalization of action is closely bound up with the moral evaluations of 'responsibility' which actors make of each other's conduct, and hence with moral norms and the sanctions to which those who contravene them are subject; thus spheres of 'competence' are defined in law as what every citizen is 'expected to know about' and take account of in monitoring his or her action.
Orthodox functionalism, as represented most prominently by Durkheim and later by Parsons, does embody an attempt to connect intentional action and institutional analysis, via the theorem that the moral values upon which social solidarity rests also appear as motivating elements in personality. This view, I have tried to show, serves only to replace the notion of action with the thesis that the properties of social and personality systems have to be examined in conjunction with one another: the member of society does not figure here as a skilled, creative agent, capable of reflexively monitoring his or her behaviour (and in principle capable of doing so in the light of anything she or he may believe can be learned from Parsons's theories!).
I have therefore set out an alternative view, one capable of more detailed development, but whose outlines should be clear. The production of society is brought about by the active constituting skills of its members, but draws upon resources, and depends upon conditions, of which they are unaware or which they perceive only dimly. Three aspects of the production of interaction can be distinguished: the constitution of meaning, morality and relations of power. The means whereby these are brought into being can also be regarded as modalities of the reproduction of structure: the idea of the duality of structure is a central one here, since structure appears as both condition and consequence of the production of interaction. All organizations or collectivities 'consist of systems of interaction, and can be analysed in terms of their structural properties: but as systems, their existence depends upon modes of structuration whereby they are reproduced. The reproduction of modes of domination, one must emphasize, expresses asymmetries in the forms of meaning and morality that are made to 'count' in interaction,
166    Conclusion
thus tying them in to divisions of interest that serve to orient struggles over divergent interpretations of frames of meaning and moral norms.
The production of interaction as 'meaningful', I have proposed, can usefully be analysed as depending upon 'mutual knowledge' which is drawn upon by participants as interpretative schemes to make sense of what each other says and does. Mutual knowledge is not corrigible to the sociological observer, who must draw upon it just as lay actors do in order to generate descriptions of their conduct; in so far as such 'knowledge', however, can be represented as 'common sense', as a series of factual beliefs, it is in principle open to confirmation or otherwise in the light of social scientific analysis.
Some aspects of the philosophy of natural science, I have argued, are relevant to elucidating the logical status of claims to knowledge made in the social sciences. But their relevance is limited by features which have no immediate parallel in the natural sciences; and in any case such developments themselves have to be subjected to critical scrutiny. Kuhn's use of the term 'paradigm' shares important elements with other versions of the notion of what I have called 'frame of meaning', and as Kuhn applies it to analysing the history of science, also raises similar difficulties to these other versions. Thus Kuhn exaggerates the internal unity of 'paradigms', as Winch does 'forms of life', and consequently does not acknowledge that the problem of the mediation of different frames of meaning has to be treated as the starting-point of analysis. When conjoined to an insistence upon a distinction of sense and reference, this allows us to grasp the significance of the hermeneutic recognition of the authenticity of meaning-frames without slipping into a relativism which forecloses the possibility of any rational evaluation of them. The mediation of paradigms or widely discrepant theoretical schemes in science is a hermeneutic matter like that involved in the contacts between other types of meaning-frame.
Sociology, unlike natural science, deals with a pre-interpreted world, where the creation and reproduction of meaning-frames is a very condition of that which it seeks to analyse, namely human social conduct: this is, to repeat, why there is a double hermeneutic in the social sciences that poses as a specific
Conclusion   167
difficulty what Schutz, following Weber, calls the 'postulate of adequacy'. I have suggested that Schutz's formulation of this, based upon the thesis that the technical concepts of social science have to be in some way capable of being reduced to lay notions of everyday action, will not do. It has in fact to be reversed: rather than, in some sense, the concepts of sociology having to be open to rendition in terms of lay concepts, it is the case that the observing social scientist has to be able first to grasp those lay concepts, that is, penetrate hermeneutically the form of life whose features he or she wishes to analyse or explain.
The relation between technical vocabularies of social science and lay concepts is a shifting one: just as social scientists adopt everyday terms - 'meaning', 'motive', 'power', etc. - and use them in specialized senses, so lay actors tend to take over the concepts and theories of the social sciences and embody them as constitutive elements in the rationalization of their own conduct. The significance of this phenomenon is recognized only marginally in orthodox sociology, in the guise of 'self-fulfilling' or 'self-negating' prophecies, which are regarded simply as nuisances that inhibit accurate prediction. Yet although causal generalizations in the social sciences in some aspects may resemble natural scientific laws, they are in an essential way distinct from the latter because they depend upon reproduced alignments of unintended consequences; in so far as they are announced as generalizations, and are picked up as such by those to whose conduct they apply, their form is altered. This once more reunites us with the theme of reflexivity, central to this study. Social science stands in a relation of tension to its 'subject-matter' - as a potential instrument of the expansion of rational autonomy of action, but equally as a potential instrument of domination.
In conclusion, and in summary form, here are some new 'rules of sociological method'. The latter phrase is only intended ironically. I do not claim that the presuppositions that follow are 'rules' in the sense in which I have suggested that term is most appropriately used in the social sciences. Rather, they are a skeletal statement of some of the themes of the study as a whole, and are merely designed to exemplify its differences from the
168   Conclusion
famous sociological manifesto that Durkheim issued almost a century ago. This statement does not in and of itself constitute a 'programme' for sociological research, although I regard it as an integral part of such a programme. The sub-classification provided below works roughly as follows. Section A concerns the 'subject-matter of sociology': the production and reproduction of society; section B, the boundaries of agency, and the modes in which processes of production and reproduction may be examined; section C, the modes in which social life is 'observed' and characterizations of social activity established; section D, the formulation of concepts within the meaning-frames of social science as metalanguages.
A
1    Sociology is not concerned with a 'pre-given' universe of objects, but with one which is constituted or produced by the active doings of subjects. Human beings transform nature socially, and by 'humanizing' it they transform themselves; but they do not, of course, produce the natural world, which is constituted as an object-world independently of their existence. If in transforming that world they create history, and thence live in history, they do so because the production and reproduction of society is not 'biologically programmed', as it is among the lower animals. (Theories human beings develop may, through their technological applications, affect nature, but they cannot come to constitute features of the natural world as they do in the case of the social world.)
2    The production and reproduction of society thus has to be treated as a skilled performance on the part of its members, not as merely a mechanical series of processes. To emphasize this, however, is definitely not to say that actors are wholly aware of what these skills are, or just how they manage to exercise them; or that the forms of social life are adequately understood as the intended outcomes of action.
B
1 The realm of human agency is bounded. Human beings produce society, but they do so as historically located actors,
Conclusion    169
and not under conditions of their own choosing. There is an unstable margin, however, between conduct that can be analysed as intentional action, and behaviour that has to be analysed nomologically as a set of 'occurrences'. In respect of sociology, the crucial task of nomological analysis is to be found in the explanation of the structural properties of social systems.
2    Structure must not be conceptualized as simply placing constraints upon human agency, but as enabling. This is what I call the duality of structure. Structure can always in principle be examined in terms of its structuration. To enquire into the structuration of social practices is to seek to explain how it comes about that structure is constituted through action, and reciprocally how action is constituted structurally.
3    Processes of structuration involve an interplay of meanings, norms and power. These three concepts are analytically equivalent as the 'primitive' terms of social science, and are logically implicated in both the notion of intentional action and that of structure: every cognitive and moral order is at the same time a system of power, involving a 'horizon of legitimacy'.
C
1    The sociological observer cannot make social life available as a 'phenomenon' for observation independently of drawing upon her or his knowledge of it as a resource whereby it is constituted as a 'topic for investigation'. In this respect, the observer's position is no different from that of any other member of society; 'mutual knowledge' is not a series of corrigible items, but represents the interpretative schemes which both sociologists and lay actors use, and must use, to 'make sense' of social activity - that is, to generate 'recognizable' characterizations of it.
2   Immersion in a form of life is the necessary and only means whereby an observer is able to generate such characterizations. 'Immersion' here - say, in relation to an alien culture - does not, however, mean 'becoming a full member' of the community, and cannot mean this. To 'get to know' an alien form of life is to know how to find one's way about in it, to
170   Conclusion
be able to participate in it as an ensemble of practices. But for the sociological observer this is a mode of generating descriptions which have to be mediated, that is, transformed into categories of social-scientific discourse.
D
1   Sociological concepts thus obey a double hermeneutic:
(a) Any theoretical scheme in the natural or social sciences is in a certain sense a form of life in itself, the concepts of which have to be mastered as a mode of practical activity generating specific types of descriptions. That this is already a hermeneutic task is clearly demonstrated in the philosophy of science of Kuhn and others, (b) Sociology, however, deals with a universe which is already constituted within frames of meaning by social actors themselves, and reinterprets these within its own theoretical schemes, mediating ordinary and technical language. This double hermeneutic is of considerable complexity, since the connection is not merely a oneway one; there is a continual 'slippage' of the concepts constructed in sociology, whereby these are appropriated by those whose conduct they were originally coined to analyse, and hence tend to become integral features of that conduct (thereby in fact potentially compromising their original usage within the technical vocabulary of social science).
2    In sum, the primary tasks of sociological analysis are the following:
(a)     The hermeneutic explication and mediation of divergent forms of life within descriptive metalanguages of social science;
(b)     Explication of the production and reproduction of society as the accomplished outcome of human agency.
Notes
Introduction to the second edition
1    Giddens, Anthony, The Constitution of Society, Cambridge, 1984.
2    Mouzelis, Nicos, Back to Sociological Theory: The construction of social orders, London, 1991; Harbers, Hans, and de Vries, Gerard, 'Empirical consequences of the "double hermeneutic"', Social Epistemology, Vol. 6,1992.
3    Mouzelis, Back to Sociological Theory, pp. 27-8.
4    Ibid., p. 35.
5    cf. Giddens, The Consequences of Modernity, Cambridge, 1990.
6    This point is accepted by Mouzelis. Mouzelis, Back to Sociological Theory, pp. 32-4.
7    Giddens, The Consequences of Modernity.
8    Knorr-Cetina, Karen, 'Social and scientific method or what do we make of the distinction between the natural and social sciences?', Philosophy of the Social Sciences, Vol. 2,1981.
9    Harbers and de Vries, 'Empirical consequences of the "double hermeneutic"', p. 4.
10    Ibid., p. 11.
11    Lynch, William T., 'What does the double hermeneutic explain/ justify?', Social Epistemology, Vol. 6,1992.
12    Ibid., p. 16.
13    Ibid., p. 38.