It is with these three core sets of issues, and their mutual connections, that the theory of structuration, as I represent it in this book, is concerned. 'Structuration' is an unlovely term at best, although it is less inelegant in the Gallic context from which it came. I have not been able to think of a more engaging word for the views I want to convey. In elaborating the concepts of structuration theory, I do not intend to put forward a potentially new orthodoxy to replace the old one. But structuration theory is sensitive to the shortcomings of the orthodox consensus and to the significance of the convergent developments noted above.[1]

In case there is any doubt about terminology here, let me emphasize that I use the term 'social theory' to encompass issues that I hold to be the concern of all the social sciences. These issues are to do with the nature of human action and the acting self; with how interaction should be conceptualized and its relation to institutions; and with grasping the practical connotations of social analysis. I understand 'sociology', by contrast, to be not a generic discipline to do with the study of human societies as a whole, but that branch of social science which focuses particularly upon the 'advanced' or modern societies. . . . This book is written with a definite sociological bias, in the sense that I tend to concentrate upon material particularly relevant to modern societies. But as an introduction to structuration theory it is also intended in substantial degree as a formulation of the tasks of social theory in general and is 'theory' in the same sense. That is to say, the focus is upon the understanding of human agency and of social institutions.[2]

'Social theory' is not a term which has any precision, but it is a very useful one for all that. As I represent it, 'social theory' involves the analysis of issues which spill over into philosophy, but it is not primarily a philosophical endeavour. The social sciences are lost if they are not directly related to philosophical problems by those who practise them. To demand that social scientists be alive to philosophical issues is not the same as driving social science into the arms of those who might claim that it is inherently speculative rather than empirical. Social theory has the task of providing conceptions of the nature of human social activity and of the human agent which can be placed in the service of empirical work. The main concern of social theory is the same as that of the social sciences in general: the illumination of concrete processes of social life. To hold that philosophical debates can contribute to this concern is not to suppose that such debates need to be resolved conclusively before worthwhile social research can be initiated. On the contrary, the prosecution of social research can in principle cast light on philosophical controversies just as much as the reverse. In particular, I think it wrong to slant social theory too unequivocally towards abstract and highly generalized questions of epistemology, as if any significant developments in social science had to await a clear-cut solution to these.

A few remarks are necessary about the 'theory' in social theory. There are certain senses often attributed to 'theory' in the social sciences from which I want to maintain some considerable distance. One conception used to be popular among some of those associated with the orthodox consensus, although it is no longer widely held today. This is the view -- influenced by certain versions of the logical empiricist philosophy of natural science -- that the only form of 'theory' worthy of the name is that expressible as a set of deductively related laws or generalizations. This sort of notion has turned out to be of quite limited application even within the natural sciences. If it can be sustained at all, it is only in respect of certain areas of natural science. Anyone who would seek to apply it to social science must recognize that (as yet) there is no theory at all; its construction is an aspiration deferred to a remote future, a goal to be striven for rather than an actual part of the current pursuits of the social sciences.

Although this view does have some adherents even now, it is far removed from anything to which I would hold that social theory could or should aspire -- for reasons which will emerge clearly enough in the body of the book which follows. But there is a weaker version of it which still commands a very large following and which invites rather longer discussion even in this prefatory context. This is the idea that the 'theory' in social theory must consist essentially of generalizations if it is to have explanatory content. According to such a standpoint, much of what passes for 'social theory' consists of conceptual schemes rather than (as should be the case) 'explanatory propositions' of a generalizing type. {pp. xvi-xviii}

Most of the controversies stimulated by the so-called 'linguistic turn' in social theory, and by the emergence of post-empiricist philosophies of science, have been strongly epistemological in character. They have been concerned, in other words, with questions of relativism, problems of verification and falsification and so on. Significant as these may be, concentration upon epistemological issues draws attention away from the more 'ontological' concerns of social theory, and it is these upon which structuration theory primarily concentrates. Rather than becoming preoccupied with epistemological disputes and with the question of whether or not anything like 'epistemology' in its time-honoured sense can be formulated at all, those working in social theory, I suggest, should be concerned first and foremost with reworking conceptions of human being and human doing, social reproduction and social transformation. Of prime importance in this respect is a dualism that is deeply entrenched in social theory, a division between objectivism and subjectivism. Objectivism was a third -ism characterizing the orthodox consensus, together with naturalism and functionalism. In spite of Parsons' terminology of 'the action frame of reference', there is no doubt that in his theoretical scheme the object (society) predominates over the subject (the knowledgeable human agent). Others whose views could be associated with that consensus were very much less sophisticated in this respect than was Parsons. By attacking objectivism -- and structural sociology -- those influenced by hermeneutics or by phenomenology were able to lay bare major shortcomings of those views. But they in turn veered sharply towards subjectivism. The conceptual divide between subject and social object yawned as widely as ever.[3]

Structuration theory is based on the premise that this dualism has to be reconceptualized as a duality -- the duality of structure. Although recognizing the significance of the 'linguistic turn', it is not a version of hermeneutics or interpretive sociology. While acknowledging that society is not the creation of individual subjects, it is distant from any conception of structural sociology. The attempt to formulate a coherent account of human agency and of structure demands, however, a very considerable conceptual effort. An exposition of these views is offered in the opening chapter and is further developed throughout the book. It leads on directly to other main themes, especially that of the study of space-time relations. The structural properties of social systems exist only in so far as forms of social conduct are reproduced chronically across time and space. The structuration of institutions can be understood in terms of how it comes about that social activities become 'stretched' across wide spans of time-space. Incorporating time-space in the heart of social theory means thinking again about some of the disciplinary divisions which separate sociology from history and from geography. The concept and analysis of history is particularly problematic. This book, indeed, might be accurately described as an extended reflection upon a celebrated and oft-quoted phrase to be found in Marx. Marx comments that 'Men [let us immediately say human beings] make history, but not in circumstances of their own choosing'. Well, so they do. But what a diversity of complex problems of social analysis this apparently innocuous pronouncement turns out to disclose! {pp. xx-xxi}

Thus, for example, I acknowledge the call for a decentring of the subject and regard this as basic to structuration theory. But I do not accept that this implies the evaporation of subjectivity into an empty universe of signs. Rather, social practices, biting into space and time, are considered to be at the root of the constitution of both subject and social object. I admit the central significance of the 'linguistic turn', introduced especially by hermeneutical phenomenology and ordinary language philosophy. At the same time, however, I hold this term to be in some part a misleading one. The most important developments as regards social theory concern not so much a turn towards language as an altered view of the intersection between saying (or signifying) and doing, offering a novel conception of praxis. The radical transmutation of hermeneutics and phenomenology initiated by Heidegger, together with the innovations of the later Wittgenstein, are the two main signal markers of the new path. But to pursue the path further means precisely to shake off any temptation to become a full-blown disciple of either of these thinkers.[4]

Let me offer here a short summary of the organization of the book. Having given in the first chapter an outline of the chief concepts involved in structuration theory, in the second I begin the more substantive part of the volume with a discussion of consciousness, the unconscious and the constitution of day-to-day life. Human agents or actors -- I used these terms interchangeably -- have, as an inherent aspect of what they do, the capacity to understand what they do while they do it. The reflexive capacities of the human actor are characteristically involved in a continuous manner with the flow of day-to-day conduct in the contexts of social activity. But reflexivity operates only partly on a discursive level. What agents know about what they do, and why they do it -- their knowledgeability as agents -- is largely carried in practical consciousness. Practical consciousness consists of all the things which actors know tacitly about how to 'go on' in the contexts of social life without being able to give them direct discursive expression. The significance of practical consciousness is a leading theme of the book, and it has to be distinguished from both consciousness (discursive consciousness) and the unconscious. While accepting the importance of unconscious aspects of cognition and motivation, I do not think we can be content with some of the more conventionally established views of these. I adopt a modified version of ego psychology but endeavour to relate this directly to what, I suggest, is a fundamental concept of structuration theory -- that of routinization.

The routine (whatever is done habitually) is a basic element of day-to-day social activity. I use the phrase 'day-to-day social activity' in a very literal sense, not in the more complex, and I think more ambiguous, way which has become familiar through phenomenology. The term 'day-to-day' encapsulates exactly the routinized character which social life has as it stretches across time-space. The repetitiveness of activities which are undertaken in like manner day after day is the material grounding of what I call the recursive nature of social life. (By its recursive nature I mean that the structured properties of social activity -- via the duality of structure -- are constantly recreated out of the very resources which constitute them.) Routinization is vital to the psychological mechanisms whereby a sense of trust or ontological security is sustained in the daily activities of social life. Carried primarily in practical consciousness, routine drives a wedge between the potentially explosive content of the unconscious and the reflexive monitoring of action which agents display. Why did Garfinkel's 'experiments with trust' stimulate such a very strong reaction of anxiety on the part of those involved, seemingly out of all proportion to the trivial nature of the circumstances of their origin? Because, I think, the apparently minor conventions of daily social life are of essential significance in curbing the sources of unconscious tension that would otherwise preoccupy most of our waking lives.[5]

The situated character of action in time-space, the routinization of activity and the repetitive nature of day-to-day life -- these are phenomena which connect discussion of the unconscious with Goffman's analyses of co-presence.[6] {pp. xxii-xxiv}

Goffman's emphasis on trust and tact strikingly echoes themes found in ego psychology and generates an analytically powerful understanding of the reflexive monitoring of the flux of encounters involved in daily life.

Fundamental to social life is the positioning of the body in social encounters. 'Positioning' here is a rich term. The body is positioned in the immediate circumstances of co-presence in relation to others: Goffman provides an extraordinarily subtle but telling set of observations about face work, about gesture and reflexive control of bodily movement as inherent in the continuity of social life. Positioning is, however, also to be understood in relation to the seriality of encounters across time-space. Every individual is at once positioned in the flow of day-to-day life; in the life-span which is the duration of his or her existence; and in the duration of 'institutional time', the 'supra-individual' structuration of social institutions. Finally, each person is positioned, in a 'multiple' way, within social relations conferred by specific social identities; this is the main sphere of application of the concept of social role. The modalities of co-presence, mediated directly by the sensory properties of the body, are clearly different from social ties and forms of social interaction established with others absent in time or in space.[7] {pp. xxiv-xxv}

Locales are not just places but settings of interaction; as Garfinkel has demonstrated particularly persuasively, settings are used chronically -- and largely in a tacit way -- by social actors to sustain meaning in communicative acts. But settings are also regionalized in ways that heavily influence, and are influenced by, the serial character of encounters. Time-space 'fixity' also normally means social fixity; the substantially 'given' character of the physical milieux of day-to-day life interlaces with routine and is deeply influenced in the contours of institutional reproduction. Regionalization also has strong psychological and social resonance in respect of the 'enclosure' from view of some types of activities and some types of people and the 'disclosure' of others.[8] {pp. xxv-xxvi}

The points of connection of structuration theory with empirical research are to do with working out the logical implications of studying a 'subject matter' of which the researcher is already a part and with elucidating the substantive connotations of the core notions of action and structure. Some of the points I have made on the abstract level of theory apply directly on the level of research. A good deal of social theory, especially that associated with structural sociology, has treated agents as much less knowledgeable than they really are. The results of this can be very easily discerned in empirical work, in respect of a failure to gain information that allows access to the full range of agents' knowledgeability in at least two ways. What actors are able to say about the conditions of their action and that of others is foreshortened if researchers do not recognize the possible significance of a range of discursive phenomena to which, as social actors themselves, they would certainly pay close attention but which in social research are often simply discounted. These are aspects of discourse which in form are refractory to being rendered as statements of propositional belief or which, like humour or irony, derive their meaning not so much from the content of what is said but from the style, mode of expression or context of utterance. But to this we must add a second factor of greater importance: the need to acknowledge the significance of practical consciousness. Where what agents know about what they do is restricted to what they can say about it, in whatever discursive style, a very wide area of knowledgeability is simply occluded from view. The study of practical consciousness must be incorporated into research work. It would be an error to suppose that non- discursive components of consciousness are necessarily more difficult to study empirically than the discursive, even though agents themselves, by definition, cannot comment directly on them. The unconscious, on the other hand, poses altogether a different order of problem, certainly demanding techniques of interrogation distinct from those involved in descriptive social research.[9] {pp. xxx-xxxi}

In structuration theory 'structure' is regarded as rules and resources recursively implicated in social reproduction; institutionalized features of social systems have structural properties in the sense that relationships are stabilized across time and space. 'Structure' can be conceptualized abstractly as two aspects of rules -- normative elements and codes of signification. Resources are also of two kinds: authoritative resources, which derive from the co-ordination of the activity of human agents, and allocative resources, which stem from control of material products or of aspects of the material world. {p. xxxi}

If the social sciences are understood as they were during the period of dominance of the orthodox consensus, their attainments do not look impressive, and the relevance of social research to practical issues seems fairly slight. . . . Both cognitively and practically, the social sciences seem distinctly inferior to the natural sciences. But if we accept that social science should no longer be some sort of replica of natural science and is in some respects a quite divergent enterprise, a very different view of their relative achievements and influence can be defended. There are no universal laws in the social sciences, and there will not be any -- not, first and foremost, because methods of empirical testing and validation are somehow inadequate but because, as I have pointed out, the causal conditions involved in generalizations about human social conduct are inherently unstable in respect of the very knowledge (or beliefs) that actors have about the circumstances of their own action. The so-called 'self-fulfilling prophecy', of which Merton and others have written, is a special case of a much more generic phenomenon in the social sciences. This is a mutual interpretive interplay between social science and those whose activities compose its subject matter -- a 'double hermeneutic'. The theories and findings of the social sciences cannot be kept wholly separate from the universe of meaning and action which they are about. But, for their part, lay actors are social theorists, whose theories help to constitute the activities and institutions that are the object of study of specialized social observers or social scientists. There is no clear dividing line between informed sociological reflection carried on by lay actors and similar endeavours on the part of specialists. I do not want to deny that there are dividing lines, but they are inevitably fuzzy, and social scientists have no absolute monopoly either upon innovative theories or upon empirical investigations of what they study. {pp. xxxii-xxxiii}

The point is that reflection on social processes (theories, and observations about them) continually enter into, become disentangled with, and re-enter the universe of events that they describe. {p. xxxiii}

The best and most interesting ideas in the social sciences (a) participate in fostering the climate of opinion and the social processes which give rise to them, (b) are in greater or lesser degree entwined with theories-in-use which help to constitute those processes and (c) are thus unlikely to be clearly distinct from considered reflection which lay actors may bring to bear in so far as they discursively articulate, or improve upon, theories-in-use. {p. xxxiv}

Theories in natural science are original, innovative and so on to the degree to which they place in question what either lay actors or professional scientists previously believed about the objects or events to which they refer. But theories in the social sciences have to be in some part based upon ideas which (although not necessarily discursively formulated by them) are already held by the agents to whom they refer. Once reincorporated within action, their original quality may become lost; they may become all too familiar. {p. xxxiv}

If they are correct, these ruminations lead on in a direct way to a consideration of social science as critique -- as involved in a practical fashion with social life. . . . The formulation of critical theory is not an option; theories and findings in the social sciences are likely to have practical (and political) consequences regardless of whether or not the sociological observer or policy-maker decides that they can be 'applied' to a given practical issue. {p. xxxv}

It might be useful at this point to recapitulate some of the basic ideas contained in the preceding chapters. I shall summarize these as a number of points: taken together, they represent the aspects of structuration theory which impinge most generally upon problems of empirical research in the social sciences.

(1) All human beings are knowledgeable agents. That is to say, all social actors know a great deal about the conditions and consequences of what they do in their day-to-day lives. Such knowledge is not wholly propositional in character, nor is it incidental to their activities. Knowledgeability embedded in practical consciousness exhibits an extraordinary complexity -- a complexity that often remains completely unexplored in orthodox sociological approaches, especially those associated with objectivsm. Actors are also ordinarily able discursively to describe what they do and their reasons for doing it. However, for the most part these faculties are geared to the flow of day-to-day conduct. The rationalization of conduct becomes the discursive offering of reasons only if individuals are asked by others why they acted as they did. Such questions are normally posed, of course, only if the activity concerned is in some way puzzling -- if it appears either to flout convention or to depart from the habitual modes of conduct of a particular person.

(2) The knowledgeability of human actors is always bounded on the one hand by the unconscious and on the other by unacknowledged conditions/unintended consequences of action. Some of the most important tasks of social science are to be found in the investigation of these boundaries, the significance of unintended consequences for system reproduction and the ideological connotations which such boundaries have.

(3) The study of day-to-day life is integral to analysis of the reproduction of institutionalized practices. Day-to-day life is bound up with the repetitive character of reversable time -- with paths traced through time-space and associated with the constraining and enabling features of the body. However, day-to-day life should not be treated as the 'foundation' upon which the more ramified connections should be understood in terms of an interpretation of social and system integration.

(4) Routine, psychologically linked to the minimizing of unconscious sources of anxiety, is the predominant form of day-to-day social activity. Most daily practices are not directly motivated. Routinized practices are the prime expression of the duality of structure in respect of the continuity of social life. In the enactment of routines agents sustain a sense of ontological security.

(5) The study of context, or of the contextualities of interaction, is inherent in the investigation of social reproduction. 'Context' involves the following: (a) the time-space boundaries (usually having symbolic or physical markers) around interaction strips; (b) the co-presence of actors, making possible the visibility of a diversity of facial expressions, bodily gestures, linguistic and other media of communication; (c) awareness and use of these phenomena reflexively to influence or control the flow of interaction.

(6) Social identities, and the position-practice relations associated with them, are 'markers' in the virtual time-space of structure. They are associated with normative rights, obligations and sanctions which, within specific collectivities, form roles. The use of standardized markers, especially to do with the attributes of age and gender, is fundamental in all societies, notwithstanding large cross-cultural variations which can be noted.

(7) No unitary meaning can be given to 'constraint' in social analysis. Constraints associated with the structural properties of social systems are only one type among several others characteristic of human life.

(8) Among the structural properties of social systems, structural principles are particularly important, since they specify overall types of society. It is one of the main emphases of structuration theory that the degree of closure of societal totalities -- and of social systems in general -- is widely variable. There are degrees of 'systemness' in societal totalities, as in other less or more inclusive forms of social system. It is essential to avoid the assumption that what a 'society' is can be easily defined, a notion which comes from an era dominated by nation-states with clear-cut boundaries that usually conform in a very close way to the administrative purview of centralized governments. Even in nation-states, of course, there are a variety of social forms which cross-cut societal boundaries.[10]

(9) The study of power cannot be regarded as a second-order consideration in the social sciences. Power cannot be tacked on, as it were, after the more basic concepts of social science have been formulated. There is no more elemental concept than that of power. However, this does not mean that the concept of power is more essential than any other, as is supposed in those versions of social science which have come under a Nietzschean influence. Power is one of several primary concepts of social science, all clustered around the relations of action and structure. Power is the means of getting things done and, as such, directly implied in human action. It is a mistake to treat power as inherently divisive, but there is no doubt that some of the most bitter conflicts in social life are accurately seen as 'power struggles'. Such struggles can be regarded as to do with efforts to subdivide resources which yield modalities of control in social systems. By 'control' I mean the capability that some actors, groups or types of actors have of influencing the circumstances of action of others. In power struggles the dialectic of control always operates, although what use agents in subordinate positions can make of the resources open to them differs very substantially between different social contexts.[11]

(10) There is no mechanism of social organization or social reproduction identified by social analysts which lay actors cannot also get to know about and actively incorporate into what they do. In very many instances the 'findings' of sociologists are such only to those not in the contexts of activity of the actors studied. Since actors do what they do for reasons, they are naturally likely to be disconcerted if told by sociological observers that what they do derives from factors that somehow act externally to them. Lay objections to such 'findings' may thus have a very sound basis. Reification is by no means purely characteristic of lay thought.

These points suggest a number of guidelines for the overall orientation of social research. First, all social research has a necessarily cultural, ethnographic or 'anthropological' aspect to it. This is an expression of what I call the double hermeneutic which characterizes social science. The sociologist has as a field of study phenomena which are already constituted as meaningful. The condition of 'entry' to this field is getting to know what actors already know, and have to know, to 'go on' in the daily activities of social life. The concepts that sociological observers invent are 'second-order' concepts in so far as they presume certain conceptual capabilities on the part of the actors to whose conduct they refer. But it is in the nature of social science that these can become 'first-order' concepts by being appropriated within social life itself.[12] What is 'hermeneutic' about the double hermeneutic? The appropriateness of the term derives from the double process of translation or interpretation which is involved. Sociological descriptions are interpretive categories which also demand an effort of translation in and out of the frames of meaning involved in sociological theories. Various considerations concerning social analysis are connected with this:

(1) Literary style is not irrelevant to the accuracy of social descriptions. This is more or less important according to how far a particular piece of social research is ethnographic -- that is, is written with the aim of describing a given cultural milieu to others who are unfamiliar with it.

(2) The social scientist is a communicator, introducing frames of meaning associated with certain contexts of social life to those in others. Thus the social sciences draw upon the same sources of description (mutual knowledge) as novelists or others who write fictional accounts of social life. Goffman is able quite easily to intersperse fictional illustrations with descriptions taken from social science research because he seeks very often to 'display' the tacit forms of mutual knowledge whereby practical activities are ordered, rather than trying to chart the actual distribution of those activities.

(3) 'Thick description' will be called for in some types of research (especially that of a more ethnographic kind) but not in others. It is usually unnecessary where the activities studied have generalized characteristics familiar to those to whom the 'findings' are made available, and where the main concern of the research is with institutional analysis, in which actors are treated in large aggregates or as 'typical' in certain respects defined as such for the purposes of the study.

Second, it is important in social research to be sensitive to the complex skills which actors have in co-ordinating the contexts of their day-to-day behaviour. In institutional analyses these skills may be more or less bracketed out, but it is essential to remember that such bracketing is wholly methodological. Those who take institutional analysis to comprise the field of sociology in toto mistake a methodological procedure for an ontological reality. Social life may very often be predictable in its course, as such authors are prone to emphasize. But its predictability is in many of its aspects 'made to happen' by social actors; it does not happen in spite of the reasons they have for conduct. If the study of unintended consequences and unacknowledged conditions of action is a major part of social research, we should none the less stress that such consequences and conditions are always to be interpreted within the flow of intentional conduct. We have to include here the relation between reflexively monitored and unintended aspects of the reproduction of social systems, and the 'longitudinal' aspect of unintended consequences of contingent acts in historically significant circumstances of one kind or another.

Third, the social analyst must also be sensitive to the time-space constitution of social life. In part this is a plea for a disciplinary identity, which, if it is not an exclusive concern with structural constraint, is bound up with a conceptual focus upon 'society'. Historians and geographers, for their part, have been willing enough to connive at this disciplinary dissection of social science. The practitioners of a discipline, apparently, do not feel secure unless they can point to a sharp conceptual delimitation between their concerns and those of others. Thus 'history' may be seen as about sequences of events set out chronologically in time or perhaps, even more ambiguously, about 'the past'. Geography, many of its representatives like to claim, finds its distinctive character in the study of spatial forms. But if, as I have emphasized, time-space relations cannot be 'pulled out' of social analysis without undermining the whole enterprise, such disciplinary divisions actively inhibit the tackling of questions of social theory significant for the social sciences as a whole. Analyzing the time-space coordination of social activities means studying the contextual features of locales through which actors move in their daily paths and the regionalization of locales stretching away across time-space. As I have accentuated frequently, such analysis in inherent in the explanation of time-space distanciation and hence in the examination of the heterogeneous and complex nature assumed by larger societal totalities and intersocietal systems in general. {pp. 281-86}

All social interaction is expressed at some point in and through the contextualities of bodily presence. In moving from the analysis of strategic conduct to a recognition of the duality of structure, we have to begin to 'thread outwards' in time and space. That is to say, we have to try to see how the practices followed in a given range of contexts are embedded in wider reaches of time and space -- in brief, we have to attempt to discover their relation to institutionalized practices. {pp. 297-98}

I said earlier that I do not propose to analyse the relevance that structuration theory may or may not have for evaluating specific types of research methods -- participant observation, survey research, and so on. It is, however, both possible and worth while to look more generically at the tasks of social research informed by structuration theory and at the consequences of the foregoing discussion of research work for the traditional debate between 'qualitative' and 'quantitative' methods in social research.

The methodological 'insertion' of the research investigator into whatever material is the object of study can be made at any of the four levels indicated above. All social research presumes a hermeneutical moment, but the presumption may remain latent where research draws upon mutual knowledge that is unexplicated because researcher and research inhabit a common cultural milieu. The more vociferous advocates of quantitative research repress the essential significance of (1) in two ways. They either take (1) to be purely descriptive rather than explanatory, or else they fail to see that it enters into the formulation of their research work at all. But research concerned with (1) may be both explanatory and generalizing. It has to do with answering why-questions that stem from the mutual unintelligibility of divergent frames of meaning. Naturally, such questions arise across the varying contexts of single societies as well as between societies. Research which is geared primarily to hermeneutic problems may be of generalized importance in so far as it serves to elucidate the nature of agents' knowledgeability and thereby their reasons for action, across a wide range of action-contexts. Pieces of ethnographic research . . . like, say, the traditional small-scale community research of fieldwork anthropology -- are not in themselves generalizing studies. But they can easily become such if carried out in some numbers, so that judgements of their typicality can justifiably be made.

Hermeneutic aspects of social research are not necessarily illuminating to those who are the subjects of that research, since their main outcome is the elucidation of settings of action considered as 'alien milieux'. Such is not the case with the investigation of practical consciousness. Studying practical consciousness means investigating what agents already know, but by definition it is normally illuminating to them if this is expressed discursively, in the metalanguage of social science. Only for ethnomethodology is the analysis of practical consciousness a circumscribed 'field' of study. For all other types of research the interpretation of practical consciousness is a necessary element, implicitly understood or explicitly stated, of broader features of social conduct.

As I have consistently stressed, identifying the bounds of agents' knowledgeability in the shifting contexts of time and space is fundamental to social science. The investigation of (3), however, presumes some considerable knowledge of levels (1), (2), and (4). Without them we are back with an untutored form of structural sociology. The study of the unintended consequences and unacknowledged conditions of action . . . can and should be carried on without using functionalist terminology. What is 'unintended' and 'unacknowledged', in any context or range of contexts of action, is usually by no means a simple matter to discover. No study of the structural properties of social systems can be successfully carried on, or its results interpreted, without reference to the knowledgeability of the relevant agents -- although many proponents of structural sociology imagine that this is exactly what defines the province of 'sociological method'.[13]

Level (4), the specifying of institutional orders, involves analysing the conditions of social and system integration via identification of the main institutional components of social systems. Those institutional forms are most important which, in terms of designated structural properties, can be specified as overall 'societies'. Once more, however, I have been at some pains to stress that it is only with many reservations that the main unit of analysis in social science can be said to be a 'society'. Institutional orders frequently cross-cut whatever decisions can be recognized between overall societies.

It is in the relation between (1) and (2) on the one hand and between (3) and (4) on the other that a division between 'qualitative' and 'quantitative' methods is often located. A fondness for quantitative methods has, of course, long been a trait of those attracted to objectivism and structural sociology. According to this type of standpoint, analysing conditions of social life that stretch well beyond any immediate contexts of interaction is the prime objective of social science, and grasping the 'hardened' nature of the institutional components of social life can best be accomplished through classification, measurement and statistical methods. Obviously the idea that the overriding concern of the social sciences is with uncovering law-like generalizations about social conduct is closely related to this proclivity. There is a strong, and often deliberate, echoing of the 'macro'/'micro' division here. Those who favour quantitative methods as the main basis of what makes social science 'science' are prone to emphasize the primacy of so-called macrosociological analysis. Those who advocate qualitative methods as the foundation of empirical research in the social sciences, on the other hand, emphasize (1) and (2) in order to point up the necessarily situated and meaningful character of social interaction. They tend often to be directly hostile to the use of quantitative methods in social science, on the grounds that quantification and the use of statistical method impose a fixity on social life that it does not in fact have. It is not difficult to see in the conflict between these positions a methodological residue of the dualism of structure and action, and showing such a dualism to be spurious will allow us to tease out further some of the empirical implications of the duality of structure. {pp. 327-30}

Once the point of this is fully understood, the idea that there is either a clear-cut division or a necessary opposition between qualitative and quantitative methods disappears. Quantitative techniques are likely to be demanded when a large number of 'cases' of a phenomenon are to be investigated, in respect of a restricted variety of designated characteristics. But both the collection and interpretation of quantitative materials depends upon procedures methodologically identical to the gathering of data of a more intensive, 'qualitative' sort. {p. 333}

(1) and (2) are thus as essential for understanding (3) and (4) as vice versa, and qualitative and quantitative methods should be seen as complementary rather than antagonistic aspects of social research. Each is necessary to the other if the substantive nature of the duality of structure is to be 'charted' in terms of the forms of institutional articulation whereby contexts of interaction are co-ordinated within more embracing social systems. {p. 334}

Empirical research self-evidently has no rationale if it does not somehow generate new knowledge which was not available before. Since all social actors exist in situated contexts within larger spans of time-space, what is novel to some such actors is not to others -- including, among those others, social scientists. It is, of course, in these 'information gaps' that ethnographic research has its specific importance. In a broad sense of the term this sort of research is explanatory, since it serves to clarify puzzles presented when those from one cultural setting encounter individuals from another which is in some respects quite different. The query 'Why do they act (think) as they do?' is an invitation to enter the culturally alien milieu and to make sense of it. To those already within that milieu, as Winch and many others have pointed out, such an enterprise may be inherently unenlightening. However, much social research, in terms of both the empirical material it generates and the theoretical interpretations which may be linked to it, has critical connotations for beliefs which agents hold. To investigate what such connotations might be we have to consider the question of exactly in what sense the social sciences reveal new knowledge and how such knowledge might connect with the critique of false belief. These matters are complex, and I shall not attempt to deal here with more than certain aspects of them.

The critical endeavours of the social sciences, like those of natural science, are bound up with the logical and empirical adequacy of reported observations and theories associated with them. As Schutz and many others have quite rightly emphasized, the critical character of social science in this respect normally departs quite sharply from the beliefs and theories-in-use incorporated within the conduct of day-to-day social life. All social actors, it can properly be said, are social theorists, who alter their theories in the light of their experiences and are receptive to incoming information which they may acquire in doing so. Social theory is by no means the special and insulated province of academic thinkers. However, lay actors are generally concerned above all with the practical utility of the 'knowledge' that they apply in their daily activities, and there may be basic features of the institutional organization of society (including, but not limited to, ideology) which confine or distort what they take to be knowledge.

It is surely plain that the 'revelatory model' of natural science cannot be directly transferred to the social sciences. Common-sense beliefs about the natural world are corrigible in the light of the findings of the natural sciences. There are no particular logical difficulties in understanding what is going on in such circumstances, even though there may be social barriers to the reception of scientific ideas. That is to say, lay beliefs are open to correction, in so far as this is necessary, by the input of novel scientific theories and observations. The natural sciences can in principle demonstrate that some of the things that the lay member of society believes about the object world are false, while others are valid. It is more complicated, for better or for worse, in the social sciences. The 'findings' of the social sciences, as I have emphasized, are not necessarily news to those whom those findings are about.

The issues involved here have become very murky indeed as a result of the push and pull between objectivist and interpretive formulations of social sciences. The former have tended to apply the revelatory model in an uninhibited way to the social sciences. That is to say, they have regarded common-sense beliefs involved in social life to be unproblematically corrigible in terms of the enlightenment which the social sciences can deliver. Those influenced by hermeneutics and ordinary-language philosophy, however, have established powerful objections to this naive standpoint. Common-sense beliefs, as incorporated in day-to-day language use and action, cannot be treated as mere impediments to a valid or veridical characterization of social life. For we cannot describe social activity at all without knowing what its constituent actors know, tacitly as well as discursively. Empiricism and objectivism simply suppress the whole issue of the generation of social descriptions via the mutual knowledge which sociological observers and lay members of society hold in common. The trouble is, having reached this conclusion, those advocating interpretive forms for social science find it difficult or impossible to maintain that critical edge which the opposite type of tradition has rightly insisted upon in juxtaposing social science and common sense. The tasks of social science then seem precisely limited to ethnography -- to the hermeneutic endeavour of the 'fusion of horizons'. Such a paralysis of the critical will is as logically unsatisfactory as the untutored use of the revelatory model.[14]

A way out of this impasse can be found by distinguishing mutual knowledge from 'common sense'. The first refers to the necessary respect which the social analyst must have for the authenticity of belief or the hermeneutic entree into the description of social life. 'Necessary' in this statement has logical force to it. The reason why it characteristically makes more sense to speak about 'knowledge' rather than 'belief' when speaking of how actors find their way around in the contexts of social life is that the generation of descriptions demands the bracketing of scepticism. Beliefs, tacit and discursive, have to be treated as 'knowledge' when the observer is operating on the methodological plane of characterizing action. Mutual knowledge, regarded as the necessary mode of gaining access to the 'subject matter' of social science, is not corrigible in the light of its findings; on the contrary, it is the condition of being able to come up with 'findings' at all.

It is because mutual knowledge is largely tacit -- carried on the level of practical consciousness -- that it is not obvious that respect for the authenticity of belief is a necessary part of all ethnographic work in the social sciences. The attacks led by those influenced by phenomenology and ethnomethodology upon more orthodox conceptions of social science have undoubtedly been of major importance in elucidating the nature of mutual knowledge. But in speaking of 'common sense' I mean to reserve the latter concept to refer to the propositional beliefs implicated in the conduct of day-to-day activities. The distinction is largely an analytical one; that is to say, common sense is mutual knowledge treated not as knowledge but as fallible belief. However, not all mutual knowledge can be expressed as propositional beliefs -- beliefs that some states of affairs or others are the case. Moreover, not all such beliefs are capable of being formulated discursively by those who hold them.

Distinguishing between mutual knowledge and common sense does not imply that these are always easily separable phases of study in actual social research. For one thing, the descriptive language used by sociological observers is always more or less different from that used by lay actors. The introduction of social scientific terminology may (but does not necessarily) call in question discursively formulated beliefs (or, where connected in an ensemble, 'theories-in-use') which actors hold. Where contested descriptions are already employed by the agents studied, any description given by observers, even using actors' categories, is directly critical of other available terminologies that could have been used. What is a 'liberation movement' from one perspective might be a 'terrorist organization' from another. The choice of one term rather than the other, of course, implies a definite stance on the part of the observer. It is less immediately apparent that the choice of a more 'neutral' term does as well; its use, however, also indicates a critical distance which the observer takes from the concepts applied by the actors directly involved.

In any research situation there may be beliefs accepted by participants which so grate upon those held by the observer that the observer expresses critical distance from them, even in what is otherwise a purely ethnographic study. An anthropologist may feel no qualms about asserting, 'The X grow their crops by planting seeds every autumn', since it is mutually held as knowledge between him or her and the members of culture X that the planting of seeds at an appropriate time of the year eventuates in a particular crop. But that anthropologist is likely to say, 'The X believe their ceremonial dance will bring rain' indicating a gap between what he or she and those in culture X believe to be the case about the conditions under which rainfall occurs.

The examples mentioned in the above paragraph indicate that even purely ethnographic social research -- that is, research which follows the confined goal of descriptive reportage -- tends to have a critical moment. While this does not comprise the logical distinction between mutual knowledge and common sense, it does mean specifying more directly what is involved in that moment of critique, which in other types of research is usually more directly developed.

I have to emphasize at this point the modest dimensions of the discussion which follows. Analysing logically what is involved in the garnering of mutual knowledge, as well as what is involved in the critique of common-sense belief, raises questions of epistemology which it would be out of the question to discuss exhaustively here. The ideas I shall develop in what follows are intended to supply no more than an outline format, which presumes a definite epistemological view without supporting it in detail. There are two senses, I want to claim, in which social science is relevant to the critique of lay beliefs construed as common sense (which includes, but does not give any special priority to, the critique of ideology). The critical activities in which social scientists engage as the core of what they do have direct implications for the beliefs which agents hold, in so far as those beliefs can be shown to be invalid or inadequately grounded. But such implications are especially important where the beliefs in question are incorporated into the reasons actors have for what they do. Only some of the beliefs which actors hold or profess form part of the reasons they have for their conduct. When these are subjected to critique in the light of claims or findings of social science, the social observer is seeking to demonstrate that those reasons are not good reasons.

The identification of agents' reasons is normally intimately bound up with the hermeneutic problems posed by the generating of mutual knowledge. Given that this is so, we should distinguish what I shall call 'credibility criteria' from the 'validity criteria' relevant to the critique of reasons as good reasons. Credibility criteria refer to criteria, hermeneutic in character, used to indicate how the grasping of actors' reasons illuminates what exactly they are doing in light of those reasons. Validity criteria concern criteria of factual evidence and theoretical understanding employed by the social sciences in the assessment of reasons as good reasons. Consider the famous case of the red macaws, much discussed in the anthropological literature. The Bororo of Central Brazil say, 'We are red macaws.' Debated by Von den Steinem, Durkheim and Mauss, among others, the statement has seemed to many to be either nonsensical or hermeneutically impenetrable. The issue was, however, recently taken up by an anthropologist who had the chance to reinvestigate the matter at source, among the Bororo. He found that the statement is made only by men; that Bororo women tend to own red macaws as pets; that in various ways in Bororo society men are peculiarly dependent upon women; and that contact with the spirits is made by men and red macaws independently of women. It seems plausible to infer that 'We are red macaws' is a statement in which men ironically comment upon their indebtedness to women and at the same time assert their own spiritual superiority to them. Investigation of why the statement is made helps to clarify the nature of the statement. The investigation of credibility criteria, in respect of discursively formulated beliefs at any rate, usually depends upon making clear the following items: who expresses them, in what circumstances, in what discursive style (literal description, metaphor, irony, etc.) and with what motives.

Assessment of validity criteria is governed solely by the conjunction of 'internal' and 'external' critique generated by social science. That is to say, validity criteria are the criteria of internal critique which I hold to be substantially constitutive of what social science is. The main role of the social sciences in respect of the critique of common sense is the assessment of reasons as good reasons in terms of knowledge either simply unavailable to lay agents or construed by them in a fashion different from that formulated in the metalanguages of social theory. I see no basis for doubting that the standards of internal critique in the social sciences carry over directly to external critique in this respect. This statement is a strong one, and it is particularly at this juncture that a specific epistemological standpoint is presupposed. It presumes, and I presume, that it is possible to demonstrate that some belief claims are false, while others are true, although what 'demonstrate' means here would need to be examined as closely as would 'false' and 'true'. It presumes, and I presume, that internal critique -- the critical examinations to which social scientists submit their ideas and claimed findings -- is inherent in what social science is as a collective endeavour. I intend to risk the disfavour of the philosophically sophisticated by asserting, without further ado, that I hold these things to be the case. In a different context, however, it would clearly be necessary to defend such contentions at some considerable length.[15]

It can be shown, I think, that there is a non-contingent relation between demonstrating a social belief to be false, and practical implications for the transformation of action linked to that belief. Criticizing a belief means (logically) criticizing whatever activity or practice is carried on in terms of that belief, and has compelling force (motivationally) in so far as it is a reason for action. Where the belief in question informs a segment or aspect of conduct in relation to the natural world, showing it to be false will (ceteris paribus) cause the agent to change his or her behaviour in whatever respects are relevant. If this does not happen, the presumption is that other considerations are overriding in the agent's mind, that the implications of the falsity of the belief are misunderstood, or that the actor does not in fact accept that its falsity has been convincingly shown. Now social beliefs, unlike those to do with nature, are constitutive elements of what it is they are about. From this it follows that criticism of false belief (ceteris paribus) is a practical intervention in society, a political phenomenon in a broad sense of that term.[16] {pp. 334-40}

Analysis of strategic conduct -- Social analysis which places in suspension institutions as socially reproduced, concentrating upon how actors reflexively monitor what they do; how actors draw upon rules and resources in the constitution of interaction.

Contextuality -- The situated character of interaction in time-space, involving the setting of interaction,actors co-present, and communication between them.

Credibility criteria -- The criteria used by agents to provide reasons for what they do, grasped in such a way as to help to describe validly what it is that they do.

Discursive consciousness -- What actors are able to say, or to give verbal expression to, about social conditions, including expecially the conditions of their own action; awareness which has a discursive form.

Double hermeneutic -- The intersection of two frames of meaning as a logically necessary part of social science, the meaningful social world as constituted by lay actors and the metalanguages invented by social scientists; there is a constant 'slippage' from one to the other involved in the practice of the social sciences.

Duality of structure -- Structure as the medium and outcome of the conduct it recursively organizes; the structural properties of social systems do not exist outside of action but are chronically implicated in its production and reproduction.

External critique -- Critique of lay agents' beliefs and practices, derived from the theories and findings of the social sciences.

Homeostatic loops -- Causal factors which have a feedback effect in system reproduction, where that feedback is largely the outcome of unintended consequences.

Institutional analysis -- Social analysis which places in suspension the skills and awareness of actors, treating institutions as chronically reproduced rules and resources.

Internal critique -- The critical apparatus of social science, whereby theories and findings are subjected to evaluation in the light of logical argument and the provision of evidence.

Knowledgeability -- Everything which actors know (believe) about the circumstances of their action and that of others, drawn upon in the production and reproduction of that action, including tacit as well as discursively available knowledge.

Mutual knowledge -- Knowledge of 'how to go on' in forms of life, shared by lay actors and sociological observers; the necessary condition of gaining access to vaild descriptions of social activity.

Ontological security -- Confidence or trust that the natural and social worlds are as they appear to be, including the basic existential parameters of self and social identity.

Practical consciousness -- What actors know (believe) about social conditions, including especially the conditions of their own action, but cannot express discursively; no bar of repression, however, protects practical consciousness as is the case with the unconscious.

Rationalization of action -- The capability competent actors have of 'keeping in touch' with the grounds of what they do, as they do it, such that if asked by others, they can supply reasons for their activities.

Reflexive Monitoring of action -- The purposive, or intentional, character of human behaviour, considered within the flow of activity of the agent; action is not a string of discrete acts, involving an aggregate of intentions, but a continuous process.[17]

Reflexive Self-regulation -- Causal loops which have a feedback effect in system reproduction, where that feedback is substantially influenced by knowledge which agents have of the mechanisms of system reproduction and employ to control it.

Reproduction circuit -- An institutionalized series of reproduction relations, governed either by homeostatic causal loops or by reflexive self-regulation.

Routinization -- The habitual, taken-for-granted character of the vast bulk of the activities of day-to-day social life; the prevalence of familiar styles and forms of conduct, both supporting and supported by a sense of ontological security.

Stratification model -- An interpretation of the human agent, stressing three 'layers' of cognition/motivation: discursive consciousness, practical consciousness, and the unconscious.

Structuration -- The structuring of social relations across time and space, in virtue of the duality of structure.

Structural properties -- Structured features of social systems, especially institutionalized features, stretching across time and space.

Structure -- Rules and resources, recursively implicated in the institutional articulation of social systems. To study structures, including structural principles, is to study major aspects of the transformation/mediation relations which influence social and system integration.

System -- The patterning of social relations across time-space, understood as reproduced practices. Social systems should be regarded as widely variable in terms of the degree of 'systemness' they display and rarely have the sort of internal unity which may be found in physical and biological systems.

System integration -- Reciprocity between actors or collectivities across extended time-space, outside conditions of co-presence.

[2] In his lectures on The Consequences of Modernity (Stanford, 1990), Giddens extends his analysis of modern society, warns of the dangers of the inherently "juggernaut" character of modernity, and advocates "a critical theory without guarantees" that will foster "models of utopian realism" into order to make modernity once again an "empancipatory" project. Similarly Promethean or prophetic concerns long have been one strand of social theory. An interesting predecessor of Giddens in the previous generation was Ernest Becker. Although near the end of his life Becker's writing seems to have become increasingly pessimistic, culminating in the incomplete, posthumously published Escape from Evil (Free Press, 1975), in his "middle period" Becker put forward an expansive vision in his trilogy The Birth and Death of Meaning: A Perspective in Psychiatry and Anthropology (1st edn., Free Press, 1962), The Revolution in Psychiatry: The New Understanding of Man (Free Press, 1964), and The Structure of Evil: An Essay on the Unification of the Science of Man (Braziller, 1968). Both Giddens and Becker represent an important aspect of Enlightenment humanism. The "middle" Becker, who was perhaps more expansive and integrative than Giddens, ended The Structure of Evil with an epilogue titled "The Merger of Science, Philosophy, and Religion."

[3] Talcott Parsons was a very influential teacher in the Department of Social Relations at Harvard. His former students include Clifford Geertz in anthropology at Princeton, Robert Bellah in sociology at Berkeley, Edward Shils in social thought at Chicago, and Joseph Elder in Indian Studies at Wisconsin. The entry of the hermeneutical and "linguistic turn" into what approximates an "orthodox consensus" in more recent social theory was marked by Paul Rabinow and William M. Sullivan, eds., Interpretive Social Science: A Reader (1st edn., California, 1979).

[4] Most of the range of theories and anti-theories that may be lumped under the label 'post-modern' or 'postmodern' take for granted "the evaporation of subjectivity into an empty universe of signs." A newly available guide is Post-Modernism and the Social Sciences (Princeton, 1992) by Pauline Marie Rosenau. Also relevant, entertaining, and troubling is Walker Percy's semiotic modest proposal Lost in the Cosmos (FS&G, 1983), which can be brought down to earth by reading Walker Percy and the Old Modern World (LSU, 1985) by Patricia Poteat.

[5] Harold Garfinkel was a founder of ethnomethodology. Two existential-psychological thrillers, that propose to disclose what may happen when routine fails, are Ernest Becker's The Denial of Death (Free Press, 1973) and Robert Jay Lifton's The Broken Connection (Simon & Schuster, 1979). A more positive version of what may happen when one experiences a "rupture" of mundane reality is offered in two phenomenological-theological studies by Peter Berger, The Heretical Imperative (Doubleday, 1979) and A Rumor of Angels (expanded edn.; Doubleday, 1990).

[6] Like Peter Berger, the late Erving Goffman was one of this century's best-selling sociological writers. Arguably his most important book is Frame Analysis: An Essay on the Organization of Experience (Harvard, 1974). Among his others are Asylums, Stigma, Interaction Ritual, and The Presentation of Self in Everyday Life.

[7] Among the many works on the theme of the body, two that have some affinities with structuration theory are Bryan S. Turner, The Body and Society: Explorations in Social Theory (Blackwell, 1984) and John O'Neill, Five Bodies: The Human Shape of Society (Cornell, 1985).

[8] See, e.g., E. V. Walter, Placeways: A Theory of the Human Environment (North Carolina, 1988).

[9] An analysis that complements the one offered by Giddens, and runs circles around it stylistically, is The Social Construction of Reality by Peter Berger and Thomas Luckmann (Doubleday, 1966). It is a dangerous and deeply subversive book, replete with dialectical reversals, paradoxes, and high ironies. Berger followed a similar interpretive pattern, which he applied to religion and termed "methodological atheism," in The Sacred Canopy (Doubleday, 1967).

[10] However, Mary Douglas makes some interesting generalizations about 'systemness' in her book Natural Symbols (1st edn., Pantheon, 1970). Later editions have been revised and read more clearly.

[11] For a "classical" example of the use of the category of power in the interpretation of religion, see G. van der Leeuw, Religion in Essence and Manifestation (2nd edn., Harper, 1964).

[12] The distinction between the three intersubjectivities that Gananath Obeyesekere makes in his The Work of Culture (Chicago, 1990) further enriches the model of the circulation of knowledge. Obeyesekere seeks to deal with the unconscious, too.

[13] These observations may help to account for (a) the possible relevance of An Interpretation of Religion (Yale, 1989) by John Hick to professionals engaged in comparative study of religion and (b) resistance to or criticism of his interpretation from the point of view of the day-to-day "practical consciousness" of some religious persons.

[14] An example of an ethnography with hermeneutical intent that may or may not represent "a paralysis of the critical will," is G. R. Thursby, "Siddha Yoga: Swami Muktananda and the Seat of Power," in T. Miller, ed., When Prophets Die (SUNY, 1991). A similar kind of study, but with prominent references to 'second order' scholarly debates while maintaining its commitment to the hermeneutical-phenomenological level of discourse, is G. R. Thursby, The Sikhs (Brill, 1992). Compare James L. Peacock and Ruel W. Tyson, Jr., Pilgrims of Paradox: Calvinism and Experience among the Primitive Baptists of the Blue Ridge (Smithsonian, 1989).

[15] Telling exemplifications of these critiques, posed in terms of the natural sciences, social sciences, and the humanities, are found in recent feminist literature: Sandra Harding, The Science Question in Feminism (Cornell, 1986); Sandra Harding, ed., Feminism and Methodology: Social Science Issues (Indiana, 1987); Christie Farnham, ed., The Impact of Feminist Research in the Academy (Indiana, 1987); Mary McCanney Gergen, ed., Feminist Thought and the Structure of Knowledge (New York University, 1988); and Louise Levesque-Lopman, Claiming Reality: Phenomenology and Women's Experience (Rowman & Littlefield, 1988).

[16] This characterization as 'political' would apply to much of the literature cited in Elizabeth Castelli & James McBride, "Beyond the Language and Memory of the Fathers: Feminist Perspectives in Religious Studies," in P. Frese & J. Coggeshall, eds., Transcending Boundaries: Multi-disciplinary Approaches to the Study of Gender (Bergin & Garvey, 1991). A like interest in the political implications of religious studies is evidenced in Ninian Smart, Religion and the Western Mind (SUNY, 1987).