Sociology of Knowledge and The Concept of Discourse

The field of sociology emerged in the first half of the 19th century with questions about the meaning of, and the transformations in, social relations of knowledge (“gesellschaftliche Wissensverhältnisse”).Footnote 1 Auguste Comte, who gave the discipline its name, agreed with other contemporary authors on the three historical stages of knowledge (the theocratic, the metaphysical, and the positive stages), which corresponded to particular social structures and institutions. This sequence of stages is conceptualized as a history of progression in which positive knowledge finally advances to become the dominating social form of knowledge. Not long afterwards, in 1845/1846, in The German Ideology, Karl Marx and Friedrich Engels outlined the origin of social knowledge as coming out of the active, socially mediated, human way of looking at nature, a line of thought that would be taken up, and formulated quite differently, decades later within US-American pragmatism. At the beginning of the 20th century, Emile Durkheim went against the Kantian postulate of a priori reason in his studies on the sociology of religion. Durkheim argued that the categories of time, space, and causality are also the result of a social history of knowledge, in which social structures and forms of knowledge as well as the content of knowledge are closely interlinked with one another. In the 1920s, Karl Mannheim suggested that sociology should study the social situatedness and perspective of all individual and collective knowledge. About the same time, Ludwig Fleck presented detailed sociological works on the genesis and discovery of ‘scientific facts’. Max Weber’s The Protestant Ethic, a work of discourse analysis avant la lettre, traced the importance of religious knowledge back to the dynamics of the development of capitalism. Sociology, as conceived by Weber, is from its onset a kind of “Kulturwissenschaft”, since social sense making or interpretations of the world are a central subject of analysis. Alfred Schütz, who was interested in the “methodology of understanding”, developed the concept of a collective social stock of knowledge from which the acting agents obtain blueprints for their actions and analyses of (and in) the world. In the USA, the Chicago School sociologists inquired into the importance of human “definitions of the situation”—the concept proposed by William I. Thomas and Dorothy Thomas—for sociologically observable actions and behaviours. The subsequent perspective of symbolic interactionism was then elaborated to conduct research on the interactive and collective consolidation and usage of such “definitions of the situation”.

In the 1960s, Peter L. Berger and Thomas Luckmann brought together many of the thoughts of the aforementioned sociology of knowledge traditions and neighbouring fields in their influential book, The Social Construction of Reality (Berger and Luckmann 1966). In their work, the authors differentiated between, on the one hand, society as an “objective reality” which is sedimented within institutions and stocks of knowledge, and on the other hand, how the acting subjects appropriate this reality in the diverse socialization processes. Above all, Berger and Luckmann emphasize the role of language and the daily “conversation machinery” for the construction of a shared social reality. They discussed how knowledge is typified and ‘realized’ through interactions and socially objectified in differing processes of institutionalization. It is also reified and becomes the foundation of social worlds differentiated by their symbolic horizons. Next, they talked about the legitimization of these knowledge/institutional complexes as well as about forms or steps of legitimization, which span from the simple usage of particular vocabularies through theoretical postulates and explicit legitimization theories to elaborate symbolic subuniverses. These legitimizations are supported by various forms of social organization. Along with the analysis regarding the structure of knowledge comes the question about the individuals, groups, actors, organizations, practices, artefacts and institutional structures which fix (or transform) such orders. The historically situated knowledge order within a society is internalized by the actors via socialization processes, and is then reproduced (and occasionally transformed) through the permanent use of language or other systems of signs and through nonverbal practices.

Shortly after the appearance of this influential book, and far removed from any sociology of knowledge context (although it did retain echoes of Emile Durkheim), Michel Foucault (Foucault 1972a [1969]) wrote his The Archaeology of Knowledge (see below).

In the following years, the sociology(ies) of knowledge field would continue to undergo incisive expansion. After a series of preliminary works in sociology of science, an empirical approach was developed at the beginning of the 1970s for Social Studies of Science which addressed the question of the concrete genesis of scientific knowledge. At the same time, researchers of symbolic interactionism established a comprehensive programme to analyse the collective battles of interpretation concerning contested social issues or ‘social problems’. In the 1980s, Niklas Luhmann integrated a particular sociology of knowledge approach into his systems theory around the question of the relationship between “social structure” and “semantics”. He no longer referred to a particular social location concerning the usage of knowledge, but rather to the connection between functional differentiation and the evolving hegemonic social semantics. In contrast, Pierre Bourdieu insisted, somehow more in both the Durkheimian and the Marxist tradition, on the connection between class struggles and classification struggles. In German-speaking regions, the hermeneutic sociology of knowledge, based on the works of Schütz, Berger and Luckmann, ultimately focused on research-oriented implementations and questions of methodology linked to the “social construction of reality”. Along with the focus on knowledge, the concept of communication also began to gain increasing importance. At the end of the 1980s, Thomas Luckmann and Hubert Knoblauch thus started to speak about the “communicative construction of reality”. They thereby highlight the actual forms and processes in which knowledge is socially objectified and then also once more subjectively adopted. The sociology of language department within the German Sociological Association [Deutsche Gesellschaft für Soziologie] was the first to deal with the questions posed by Knoblauch and Luckmann. The official re-naming of this department took place in 2000; hereafter it was labelled “sociology of knowledge” (see www.wissenssoziologie.de). This reflects the fact that questions regarding the meaning of knowledge and its communicative processes were given greater weight than the earlier influential questions about the connection between language use and its embedding within social structures, or those about the formal patterns found in communication genres.

Although Berger and Luckmann highlighted the role of “theoretical conceptions” (ideas, theories, and others) in social processes, they emphasized much more that their main interest (and therefore also that of the sociology of knowledge inspired by these two authors) applied to ‘common sense’ since in the end this seemed to them the most relevant level of social knowledge (Berger and Luckmann 1966: 14–15). With few exceptions, the hermeneutic sociology of knowledge which follows Berger and Luckmann in German-speaking regions has almost exclusively continued to focus on the sociology of knowledge questions discussed here, and therefore on common sense knowledge and individuals as the knowledge actors of daily life. In this respect, SKAD, although situated in this paradigm of knowledge research, is both an extension and a correction, elaborating on the “objective reality” side of Berger’s and Luckmann’s theory, that is on the (institutional) processes and structures in social relations of knowledge, in taking the discursive construction of “objective reality” into consideration.Footnote 2

Although it is precisely in the symbolic interactionist’s tradition that the concept of the “universe of discourse” and the analysis of public debates, public discourses, and the construction of social problems therein play an important role, this sociological paradigm as well as sociology of knowledge itself, regardless of their diversity, have not yet developed any genuine sociology of knowledge perspective on the analysis of social discourse. The corresponding initiative towards this goal came rather from Michel Foucault. On the one hand, as a philosopher working with an historical approach, he developed his discourse analysis about power/knowledge complexes quite removed from sociological positions. On the other hand, he can definitely be understood as a representative of the Durkheim tradition, which advances a genuine sociology of knowledge analysis of social “systems of thought”. Foucault’s fundamental achievement was to look at discourses as socio-historically situated “practices”, and not as the development of ideas or lines of argumentation, and to ‘liberate’ discourse analysis from the specific linguistic issues. In so doing, he laid important foundations for a sociological analysis of discourses. When he argued that his main concern was the “analysis of problematizations” (Foucault 1984), that is the appearance of central breaking or turning points in the history of social constitutions of subjectivities or particular orders of practice, he came rather close to the interests of the symbolic interactionist. Having discussed Foucault’s contributions in more detail in Keller (2005: 122–150; 2008), I want to focus here on the following points: in his seminal book The Archaeology of Knowledge, Foucault (Foucault 1972b [1969]), reflecting his own previous work (especially the Order of things, a historical study of the sciences, published in 1966), proposes a theoretical framework which takes ‘discourse’ as its central concept. Discourses are considered as historically situated ‘real’ social practices, not representing external objects but constituting them. This implies looking at concrete data—oral and written texts, articles, books, discussions, institutions, disciplines—in order to analyse ‘bottom up’ how discourses are structured and how they are structuring knowledge domains. Foucault speaks of “discursive formations” (p. 34–78) which can be reconstructed, for example the “formation of concepts” (which concepts are used and how they relate to each other) or the “formation of enunciative modalities” (as the ‘places for speakers’ and the established criteria to access them). His notion of “statement” (p. 79–117) refers to the typified core elements of discursive events and concrete utterances. Foucault, in his later works, never realized the kind of analysis he projected in the Archeology. But he returned to ‘discourse’ several times: L’ordre du discours [The order of discourse], presented as oral communication in 1970 and strangely translated as “The Discourse on Language” (included as an appendix in the American translation of Archeology, p. 215–238) in fact pursues the framework of discourse research by introducing more explicitly ideas of power and mechanisms of the ‘inner structuration’ of discourses (as the “commentary” which, for example, differentiates between important statements and the rest). In the Riviere case Foucault (Foucault 1982 [1973]) addresses discourses as battle fields, as power struggles around the legitimate definition of phenomena. This comes very close to symbolic interactionist traditions. In other texts, he introduced the notion of the “dispositif” (often translated as “apparatus”; see Foucault 1980: 194–228). In French everyday language, the word dispositif is commonly used to design a kind of infrastructure which is established by organizations or governments in order to fulfil a particular purpose. To give an example: waste incinerators, garbage cans, staff, administrative waste regulation, laws—all this is part of the dispositif established for solving some problems emerging because of today’s consumer society.

In later years, Foucault turned to other issues. But the concept of discourse spread, causing a whole field of approaches to discourse (Keller 2003), which, to be clear, do not all refer to Foucault. Today, corpus linguistics is analysing large corpuses of textual data in order to find statistical correlations between terms used. Discourse theory as established by Ernesto Laclau and Chantal Mouffe addresses discourses as political struggles for hegemony. Critical Discourse Analysis (CDA) represented by Norman Fairclough in the UK, Ruth Wodak in the UK and in Austria or Siegfried Jäger in Germany, originating from within sociolinguistics, establishes a critique of ideological functions of language use. And discourse analysis in the tradition of ethnomethodology and conversation analysis addresses issues of the ordering of communication in interactional settings, of “text and talk in action” (Teun van Dijk). Recent debates in the field have focused on mutual critiques between this and more theoretically orientated approaches like Laclau and Mouffe (e.g. Jørgensen and Philipps 2002), between the ‘all to micro’-orientation on the one hand, the ‘all to macro’ on the other. The main point I want to make against all these approaches to discourse refers back to Foucault’s interests in the discursive constitution of knowledges, and to Berger and Luckmann’s theory of the social construction of reality. The approaches mentioned above cannot (and maybe do not aim to) account for the sociohistorical processings of knowledge and symbolic orderings in larger institutional fields and social arenas.Footnote 3 This is particularly clear in research done by Critical Discourse Analysis, which implies that the researcher knows and unmasks the ideological and strategic use of language by ‘those in power’ in order to ‘manipulate the people’. The sociology of knowledge approach to discourse presented in chapter 2 directs its attention, with Berger, Luckmann and Foucault, towards the social construction of reality. It gives priority to Berger and Luckmann, because they establish a dialectical perspective on society both as “objective reality” and as “subjective reality”, becoming ‘real’ through all kinds of knowledge. It uses Foucault’s ideas and concepts—discourse as practice, discursive formation, statement, dispositif, discursive battles—in order to explore in more detail the processes of institutionalization and transformation of symbolic orderings neglected in the Berger and Luckmann tradition, and it refers to methodology and methods of qualitative (interpretative) inquiry, close to the perspective of the latter.Footnote 4

The Sociology of Knowledge Approach to Discourse

SKAD is not a method but rather a research programme embedded in the sociology of knowledge tradition in order to examine the discursive construction of symbolic orders which occurs in the form of conflicting social knowledge relationships and competing politics of knowledge.Footnote 5 Social relationships of knowledge are complex socio-historical constellations of production, stabilization, structuration, and transformation of knowledge within a variety of social arenas. In the context of social constructivism, the concept of knowledge not only refers to that which counts as socially recognized and confirmed positive knowledge. In fact, it constitutes the entirety of all social systems of signs, and in so doing, the symbolic orders and stocks of knowledge constituted by these systems which mediate between human beings and the world. Included among these are such things as religious doctrine, sociological theory, the interpretative knowledge about social situations, and the larger theorems of globalization, freedom, sustainability, and so on.

SKAD follows Foucault and examines discourses as performative statement practices which constitute reality orders and also produce power effects in a conflict-ridden network of social actors, institutional dispositifs, and knowledge systems. It is emphasized that discourse is concrete and material, it is not an abstract idea or free floating line of arguments. This means that discourse appears as speech, text, discussion, visual image, use of symbols, which have to be performed by actors following social instructions and therefore discourses are a real social practice. The sociology of knowledge analysis of discourse is concerned with reconstructing the processes which occur in social constructions, objectivization, communication, and the legitimization of meaning structures or, in other words, of interpretation and acting structures on the institutional, organizational or social actors’ level. It is also concerned with the analysis of the social effects of these processes. This includes various dimensions of reconstruction: sense making as well as subject formation, ways of acting, institutional/structural contexts, and social consequences; how, for example, they become apparent in the form of a dispositif (that means: an installed infrastructure designed to ‘solve a problem’, for instance, consisting of a law, administrative regulations, staff, things like cars, computers and so on) or in the adoption or rejection by social actors in their everyday life (e.g. actors refusing to ‘behave in an environmental-friendly way’). This perspective assumes the normality of symbolic battles, contested problematizations, and controversies, of competitive discourses, whose manifestations and effects can be traced back only in the rarest cases to the dominance and intentions of individual discourse actors (although one can perhaps not dismiss them upfront). Sociology of knowledge discourse research analyses social conventions and structuration of symbolic ordering as well as the symbolic structuring of social orders. It looks for fixed and fluid rules of interpretation practices and takes an interest in the participating actors’ part in conflicts on collective levels of “definition of the (collective’s) situations” (W. I. Thomas and D. Thomas). Last but not least, it targets the materialities of discourse, whether they appear as dispositifs (assemblages of actors, practices, things) performing discourse production or in the objectivizations and consequences of discursive claims manifesting themselves as artefacts, social practices, communication processes and subject positions. These levels can only be briefly outlined here.Footnote 6

The Constitution of Meaning in Consciousness

Berger and Luckmann’s theory of knowledge provides a theoretical foundation which considers both the institutional knowledge process and the life-world adoption and usage of stocks of knowledge. This explains its precedence over Foucault’s approach to discourse which only addresses institutional settings. Along with Alfred Schütz, it assumes that meaning is constituted in the human consciousness. Although Schütz, at least in his early writings (Schütz 1967 [1932]), does indeed have an excessively rational or cognitive bias (correctable through pragmatism) in his emphasis on the conceptual nature of actions (in other words, the assumption that we mentally anticipate actions along with their goals and processes and then put them into practice), as well as on the importance of “in-order-to” and “because-motives,” this does not detract from his analysis of the role of the consciousness in the transformation of sensual experience into conceptual experience. The meaning we ascribe to our actions and interactions, social situations and/or the world, is necessarily located in human consciousness. Without a process such as the layering of meaning, or the constitution of meaning, there is no separation between I and the world, no perception of space, time, the social, and so on. This capacity of the consciousness should not be understood as a genuine “production capacity”, as if consciousness creates the existence and the meaning of the world out of nothing in an act of solitary, productive creativity. Consciousnesses do indeed draw on social interpretation schemata in a fundamental typification process in order to perform their orientation capacity. This occurs by means of signs or knowledge schemata, which are taken from the socio-historically generated and established collective signs or stocks of knowledge, for the most part primarily within socialization processes. The specific, subjective stocks of knowledge of particular individuals are inconsistent, heterogeneous, complex sedimentations and actualizations of knowledge triggered from the outside, which always exist in a situational, pragmatically motivated relation between focalization and blurry horizons, actualized by ‘external’ stimulations.

George Herbert Mead and the tradition of symbolic interactionism considered in more depth how individual competence in the use of signs/knowledge or of significant symbols develops within socialization processes. Above all, Mead emphasized the primacy of communication and of the universe(s) of discourse that always historically ‘comes before’ the individual.Footnote 7 The existence of social-symbolic orders—never ultimately achieved but always being in the “process of ordering”—and the corresponding communication processes are a necessary prerequisite for the development of individual consciousnesses that are capable of intellectual reflection. Thought is therefore a form of communication turned inwards. Research into the social phenomenon of discourses is obsolete without such a theory of sign-processing consciousnesses (which does not mean that everything is already said here).Footnote 8 Signs as well as the ‘correct ways to use them’ are processed discursively, and the corresponding social rules are working as instructions in discursively embedded utterances. Typification stocks are nothing more than systems of difference made up of signs which emerge out of the practical usage of signs by social collectives and which, through their reciprocal relation or dissociation, both differ from one another and, at the same time, constitute each other. Historically, they make up the more or less solidly fixed pre-existing ‘supply’ to be used by particular individuals and consciousnesses. The language system of meaning is a pre-condition of the inevitable, necessary ‘desubjectification’ of the individual’s interpretation practice; in other words, the historical-social assignation of the possibilities for a ‘subjective’ orientation of individuals in the life-world. Its usage always presupposes the participating actors’ capacity for interpretation. Every long-term use of signs is a social practice regulated by social conventions. These kinds of convention form the basis of discourse practices as a set of instructing rules and are actualized in practical usage, thus simultaneously reproduced and altered, or changed, as needed. So the usage of typifications is socially regulated but not completely determined. There is therefore in principal a certain amount of freedom in interpretation and action in concrete situations as well as a surplus of forms of communication and models for the attribution of meaning. Societies differ from one another in the available spectrum and in their ways of producing such choices.

Discourses

I identify discourses, according to Foucault, as regulated, structured practices of sign usage. In this regard, Foucault’s chief merit is to have brought awareness to the materiality of social signs and idea production, that is, to their concrete manifestations in practices, institutional structures, objects, and textual documents. For example, a scientific discourse is manifest in texts, conferences, papers, talks, associations and so on which can all be studied as data. Furthermore, in his seminal works mentioned above he identified the ways in which they can be analysed as emergent discourse formations without recourse to the unmasking of ‘real’ or ‘covert’ reasons and intentions of particular social interest groups or actors. He then proposed corresponding dimensions of analysis of discursive formations which, when combined with historically situated institutionalization processes and the interwoven actions of social actors therein, can be a benefit for sociology. In discourses, the use of language or symbols by social actors constitutes the sociocultural facticity of physical and social realities. The meaning of signs, symbols, images, gestures, actions or things is more or less fixed in socially, spatially, and temporally or historically situated (and therefore transformable) orders of signs. It is affirmed, conserved or changed in the concrete usage of the signs. In this respect, every fixed meaning is a snapshot within a social process that is capable of generating an endless variety of possible readings and interpretations. Discourses can be understood as attempts to freeze meanings or, more generally speaking, to freeze more or less broad symbolic orders, that is, fix them in time and by so doing, institutionalize a binding context of meaning, values and actions/agency within social collectives. SKAD is concerned with this correlation between the sign usage as a social practice and the (re)-production/transformation of social orders of knowledge. It is called the sociology of knowledge approach to discourse (analysis) because the perspective towards discourses implied in SKAD can be situated in the sociology of knowledge tradition founded by Peter Berger and Thomas Luckmann. This is mainly due to SKAD’s research focus and because it benefits from its connection to this tradition. More specifically, this approach proposes a perspective on discourse that bridges the gap between either agency or structure orientated traditions in sociology of knowledge. By so doing, it is possible to overcome the unproductive opposition between approaches which focus on the emergence of collective knowledge orders, as Foucault did in The Archaeology of Knowledge, and others which emphasize the social actors’ definition battles, for example as in Foucault’s (1982) own writing about the Rivière Case or in the symbolic interactionism tradition by Joseph Gusfield (1981) with his analyses of “collective crusades against alcohol abuse”. Indeed, just as Berger and Luckmann addressed the manifestation of institutions out of processes of institutionalization, we can consider the processing of discourses through society as a dialectical interplay between actors producing statements, and the pre-given as well as emerging structurations and sociohistorical means they have to draw upon.

The speaker positions which are available within discursive battles and the corresponding discourse or issue arenas, as well as the social actors who are involved within them, are not ‘masters of the discourse universe’, but are rather (co-)constituted by the existing structuring of discursive orders or formations. Nevertheless, they in no way act as marionettes of discourses’ (or “cultural dopes,” as Garfinkel put it some time ago), but rather as lively, interested producers of statements, as articulators with more or less strong resource and creativity potentials. The symbolic orders that are produced and transformed in this process constitute the aggregated effects of their actions; unambiguous temporary forms of dominance or hegemony are probably rare, but they are non-standard configurations that should not be excluded from an empirical point of view.

Discourses are simultaneously both an expression and a constitutional prerequisite of the social; they become real through the actions of social actors, supply specific knowledge claims and contribute to the liquefaction and dissolution of the institutionalized interpretations and apparent unavailabilities. Discourses crystallize and constitute themes in a particular form as social interpretation and action issues. Discursive formations are discourse groupings which follow the same formation rules. Foucault himself proposed analysing the forming of objects, strategies, terminologies, and modalities of enunciation in discourses—primarily considering scientific knowledge formations or other ‘truth games’ (Foucault 1972b: 34–78). However, he offered very little further direction or analytical concepts for this exploration. SKAD therefore supplies additions to this which can only be briefly elucidated here. I describe discursive fields as being social arenas, constituting themselves around contested issues, controversies, problematizations, and truth claims in which discourses are in reciprocal competition with one another. The topics of sociology of knowledge discourse analysis are both public discourses as well as special discourses performed in close arenas for special publics. They are analysed with regard to their bearer, to matching or differing formation rules and content positionings, as well as to their effects. In the processing of discourses, specific discourse coalitions and statement bearers can ‘win out’ over others, by a wide range of means. As Thomas Kuhn demonstrated a long time ago for scientific revolutions: paradigm shifts do not have to emerge out of arguments; there are all kinds of other reasons. This holds true for discourses, too. However, the then occurring discursive formation cannot be understood as an intended and controlled effect of individual actors. What is at stake in these discourses is the fixing of collective symbolic orders through a more or less accurate repetition and stabilization of the same statements in singular utterances. Argumentative consensus building processes as projected in Habermas’s normative discourse ethics, where all participants are equal, and the best argument wins, may appear as a very particular and rather seldom occurring case in discourse processing. SKAD therefore addresses discourses as complexes of power/knowledge which are to be the object not of normative judgement but of empirical inquiry.

SKAD proposes some more terms in order to analyse utterances that are assumed to be part of the same discursive formation. The term discourse itself indicates a structuration context which is the basis of disseminated discursive events. The unity of the structuring context, that is, of the discourse in question, should be considered as a necessary construct for sociological observation, an essential research hypothesis. In the limited sequence of actual utterances (communications), social actors reproduce and transform discourse structures through the contingency of historically situated conditions and concrete action while they, with more or less excitement and competition, pursue their respective daily routine. Discursive orders are the results of a continuous communicative production within individual language and action events which are, however, not understood as spontaneous or chaotic, but rather as interwoven, structured practices which refer back to one another. Under this definition, discourses are defined as a real, manifest, observable, and describable social practice which finds its expression in various documents, in the use of oral and written language, images or more generally speaking, in the usage of signs. Discourses are realized through social actors’ communicative actions. A pamphlet, a newspaper article or a speech within the context of a demonstration, actualizes, for instance, an environmental policy discourse in differing concrete forms and with differing empirical scope. Discourses are subject to the conditions of institutional inertia: individual discursive events never actualize and reproduce a discourse’s structure in a completely identical way, but rather always in a more or less varied form. ‘Actualization’ can therefore be understood in two ways: as the transfer of a discourse structure into a real event and as the accompanying modification or adaption to the current conditions of a situational context. Qualitatively significant discourse transformations can rarely be related to such an individual event. Rather, they originate out of the sum of variations, in a kind of switch from the quantitative to the qualitative effect. The materiality of discourses (as discursive or non-discursive practices, ‘real speakers’, texts, speeches, discussions, things) simply means: the way discourses exist in societies, become ‘real’ in what could be used as ‘possible empirical data’. For example, this journal, Human Studies, is one bearer of the materiality of humanities and social science discourses, made of paper and papers, editorial board, processes of reviewing and so on. I suggest that discursive events, actors, practices, dispositifs, and knowledge structurings are the building blocks of this materiality of discourses. Therefore, they should be explained briefly here.

Discursive events (statement events) create the typifiable material form of utterances, in which a discourse appears. In Foucault’s terms (see Foucault 1972a: 79–117), an utterance is the concrete, in each case individual, singular, and unrepeatable discursive event. On the other hand, a statement is the core of that which is typical and can be identified as such in an utterance; the same statement can be made in very different utterances and situated forms, and can exist as text, image, graphic or audio-visual data. The relationship between discourse and discursive events corresponds to the relationship between structure, or structuring, and individual actions, i.e., in the words of Anthony Giddens, to the “duality of structure” (Giddens 1986: 24–26). Structures originate out of actions, and in turn, actions originate out of structures in the process of structuring. There is no discourse without statement events; without discourses, statement events could not be understood, typified and interpreted, and therefore could not constitute a collective reality. This kind of structure is both structured—i.e., is the result of previous structure-forming processes—and structuring in respect to the scopes of future discursive events. The real events are not a direct effect of the structural rules, but rather the result of how the social actors actively articulate, interpret, and deal with these orientation models. When they are actualized, it is through a (more or less) creative and performative achievement on the part of the social actors who draw on resources and use, interpret, and (co-)produce them for their practical needs, strategies, tactics and contexts in order to carry out their moves. Discourse structures are power structures; discursive conflicts are powerful struggles about the power of interpretation and action.

For producing/articulating interpretations, social actors use the rules and resources that are available as discourses in their discursive practice not as deterministic regulation, but as instruction, or they react to them as addressees. Only if discourse research accounts for this ‘agency of actors’ it can be understood how the more or less creative implementation of such practices happens. SKAD does not hastily mistake the discourse level as being a condition of possibilities or limitations of utterances with the factual interpretation and practices of social actors. Social actors are not only the empty addressees of knowledge supplies and the value assessments embedded therein, but are also socially configured incarnations of agency, according to the socio-historical and situational conditions, who more or less obstinately interpret social knowledge supplies as ‘offered rules’ in their everyday interpretation activities (Hitzler et al. 1999), standing in the crossfire of multiple and heterogeneous, maybe even contradicting discourses, trying to handle the situations they meet.

Social actors (in both the individual or collective form) are related to discourse in two ways: on the one hand, as the holders of the speaker position, or statement producers, who speak within a discourse; and on the other hand, as addressees of the statement practice. The differentiation between social actors who initially ‘exist’ independent of, or outside of, discourses, and their “discourse specific configuration,” which occurs in the form of taking on the available or ‘conquered’ speaker position, is helpful for sociological discourse research. Only then can it be taken into account that speakers don’t appear out of nothing in discourses. The sociological vocabulary of institutions, organizations, roles, and strategies of the individual or the collective—but always of social actors—can be used for a corresponding analysis of the structuration of speaker positions in discourses. Through their reflexive and practical interpretations of the structural conditions, they can also cause their transformation.

Social actors are ‘interpellated’ by discourses in some more ways: for example as problem initiators, holders of responsibility, objects of necessary interventions or potential consumers of specific services. So the contemporary discourse on environmental issues created two central subject positions in the beginning of the 1970s: that of the individual environmental offender on the one hand, and that of the environmentally responsible eco-friendly citizen on the other. We can switch between both interpellations daily, with regard to issues like environmentally friendly consumption or household waste separation (see Keller 1998). The way in which the addressed adopt the corresponding subject positions, and thus, how they ‘subjectify’ themselves according to their elements and rationalities, is not predetermined by this, but rather merits some analysis of its own. Dispositifs play a central role here, as institutional and organizational infrastructures that offer concrete situational settings for the corresponding programming efforts in the form of buildings, trainers, seminars, technologies of the self, codes of practice, laws, participants etc. To summarize, SKAD describes a three-fold relationship between discourses and actors:

  • Speaker positions depict positions of legitimate speech acts within discourses which can be taken on and interpreted by social actors under specific conditions (for instance, after the acquisition of specific qualifications) as role players.

  • Subject positions/Identity offerings depict positioning processes and ‘patterns of subjectivation’ which are generated in discourses and which refer to (fields of) addressees. Technologies of the self are understood as exemplary elaborated, applicable and available instructions for subjectivation.

  • Social actors are individuals or collectives which draw on the above-mentioned speaker or subject positions and, according to their more or less obstinate (role) interpretations and competences, accept, effect, translate, adopt, use or oppose them, and therefore ‘realize’ them in a versatile way which should be empirically investigated.

The term practice(s) depicts very generally conventionalized action patterns which are made available in collective stocks of knowledge as a repertoire for action, that is, in other words, a more or less explicitly known, often incorporated recipe or knowledge script about the ‘proper’ way of acting. This knowledge can originate, establish, and develop itself (further) in fields of social practice through experimenting and testing actions in relation to specific issues. SKAD differentiates between several forms of practice: Discursive practices are typical, realized communication patterns which are bound to a discourse context. They are not only interesting for discourse research as far as their formal process structure is concerned, as in genre theory and conversation analysis, but rather equally so in consideration of what was called by Foucault the (socio-historical emergence of) rules of formation, their adoption by social actors and their function in discourse production. Discursive practices are observable and describable, typical ways of acting out statement production whose implementation requires interpretative competence and active shaping by social actors. The social processing of discourses also takes place through non-discursive practices, in other words, through ways of acting which do not primarily use signs, but which are essential for the statements of a discourse (for example, the construction or assembly of measuring instruments in order to prove specific statements about environmental pollution).

SKAD differentiates between the latter and between model practices generated in discourses, that is, exemplary patterns (or templates) for actions which are constituted in discourses for their addressees. To continue with the above-mentioned example of environmental discourse, this includes recommendations for forms of eco-friendly behaviour (as for example: turning the shower off while you shampoo your hair, using your bike, preparing slow food). Similar to the subject positions discussed earlier, one should not think that the model practice will actually be implemented simply in the way it was imagined in discourse. Its ‘realization’, that is ‘becoming’ real, has to be considered in its own right.

The social actors who mobilize a discourse and who are mobilized by discourse establish a corresponding infrastructure of discourse production and problem solving which can be identified as a dispositif. Michel Foucault (1980: 194–228) introduced different notions of dispositifs. SKAD takes up that one which is most common in everyday French. ‘Dispositif’ then refers to what could be called an infrastructure established by social actors or collectivities in order to solve a particular ‘situation with its inherent problems of action’. Consider the state’s need to get some ‘money of its own’: Financial laws, administrative regulation, tax authorities, tax assessment, tax investigators all together, mixed up with texts, objects, actions and persons, constitute the dispositif in question—a ‘strategic’ ensemble of heterogeneous elements, drawn together, arranged in order to manage a situation, to respond to a kind of “urgency” (Michel Foucault). SKAD distinguishes between dispositifs of discourse production and dispositifs or infrastructures emerging out of a disourse (or out of several discourses) in order to deal with the real world phenomena addressed by the discourse in question. A dispositif is both: the institutional foundation, the total of all material, practical, personal, cognitive, and normative infrastructure of discourse production, and also the infrastructures of implementation emerging out of discursively configured problematizations of fields of practice. Consider the issue arena of ‘household waste’, recycling and so on, important issues of public debate and policy decisions in recent decades: with reference to the discourse (re)-production level, the discursive interventions of the various management, spokespersons, and press committees and also the research centres who diffuse and legitimize a specific construction of waste issues through their statements, brochures and so on, should be mentioned. With regard to implementation one could include among these, for example, the legal regulation of responsibilities, formalized proceedings, specific objects, technologies, sanctions, courses of studies, personal and other phenomena. For instance, waste separation systems are part of the dispositif and effects of discourses on waste. This includes the corresponding legal regulations, the waste removal company’s staff and, finally, also the waste separation and waste cleaning practices to which people submit (or refuse).Footnote 9 Dispositifs are the real means for the realization of the external ‘power-effects’ of a discourse, that is the changes it introduces or elicits in the addressed situations and fields of action, be they intentional or non-intended. Dispositifs mediate between discourses and fields of practice. SKAD is therefore not just textual analysis of signs in use, communication, text or image research. It is simultaneously case study, observation, and even a dense ethnographic description, which considers the link between statement events, practices, actors, organizational arrangements, and objects as more or less historical and far-reaching socio-spatial processes.

Analysis Dimensions

The sociology of knowledge analysis of discourse considers various analysis dimensions of social relations of knowledge and knowledge politics. In particular, this concerns questions not only about the structuring of the contents of a discourse (knowledge configurations), the material ‘foundation’ of the utterance production, the power-effects which emanate from them, but also about the levels of social macro, meso, and micro discourse contexts as well as the general processes of discourse transformation from a historical, spatial and social perspective (for example, the transnationalization of discourses, the ‘de-expertization,’ and the like). The first three of these dimensions—knowledge configuration, discourse production and power effects—will be discussed in more depth below.

Knowledge Configuration

Discourse research interested in knowledge configuration requires exploratory concepts to understand the typifiable statement content of a particular utterance. In order to analyse this level of knowledge structuring, I propose distinguishing between interpretative schemes or frames, classifications, phenomenal structures [Phänomenstrukturen], and narrative structures. Moreover, one can differentiate between (argumentative) legitimating elements (for example, scientific, moralistic, and voluntaristic patterns of legitimization), subject positions, and discourse-generated model practices as components of phenomenal structures. Together, these elements create the interpretative repertoire of a discourse.Footnote 10 I shall consider these concepts more closely in what follows.

The term interpretative scheme or frame (Deutungsmuster) depicts fundamental meaning and action-generating schemata, which are circulated through discourses and make it possible to understand what a phenomenon is all about. Discourses link different frames to specific interpretative frameworks. They draw on socially available stocks of knowledge; they are also capable, however, of generating new interpretative schemes and of positioning those within the social agenda—which is exactly what characterizes discourses. An example of this is the interpretative scheme of the “irreducible risk” of complex technologies which has found its way into social stocks of knowledge over the last few decades within, and because of, the various environmental discourses. A second element for the content-focused analysis of discourses is the exploration of the classifications (and therefore qualifications) of phenomena which are performed within them and by them. From the social constructionist sociology of knowledge perspective, classifications are a more or less elaborate, formalized, and institutionally fixed form of social typification processes. Like every form of sign use, language usage within discourses classifies the world, separates it into particular categories which are the foundation for its experience, interpretation and way of being dealt with. Competition for such classifications occurs, for example, between discourses about how (potential) technical catastrophes should be interpreted, which identity offerings can be considered legitimate, what the differences between correct and condemnable behaviour may look like, and if perpetrators are certifiably sane or not (one could consider Foucault’s Rivière Case here, see Foucault 1982). Classifications have specific impacts for action (see Bowker and Star 2000 or e.g. Douglas 1966, 1986).

Alongside interpretation schemes and classifications, the concept of phenomenal structure [Phänomenstruktur], referring somehow to Karl Mannheim’s classical notion of “Aspektstruktur,” offers a complementary third access to the levels of content-related structuring of discourse (see Table 1). For instance, constructing a theme as a problem on the public agenda, requires that the protagonists deal with the issue in several dimensions, and refer to argumentative, dramatizing, and evaluative statements; the determination of the kind of problem or theme of a statement unit, the definition of characteristics, causal relations (cause-effect), and their link to responsibilities, problem dimensions, value implications, moral and aesthetic judgments, consequences, possible courses of action, and others. The phenomena which are constituted by phenomenal structures do not necessarily in any way appear as a ‘problem’, even if they are always in a very general way about ‘interpretation and acting problems’ to a certain extent—but not in any way necessarily about ‘social’ problems. The existing state of discourse research provides insight into some important elements of such phenomenal structures. For example, the subject positions constituted by a discourse can be differentiated in a variety of ways. In this way, discourses carry out social actors’ positionings as heroes, rescuers, problem cases, sensibly, and responsibly acting individuals, villains and so on. However, this does not only occur with regard to the ‘agents’ of the narrative which is being offered, but rather also with regard to the various addressees of a discourse. This also includes discourse-generated model practices, which provide guidelines or templates for how one should act concerning issues about acting that have been defined by the discourse. The concept of phenomenal structure takes on these kinds of consideration and links them to the fact that discourses, in the constitution of their referential relation (and so, their ‘theme’), designate different elements or dimensions of their topic and link them to a specific form or to a specific phenomenal constellation. This does not describe any essential qualities of a discourse topic, but rather the corresponding discursive attributions.

Table 1 Phenomenal structure: Administrative discourse on waste issues, France (taken from Keller 1998: 232)
Full size table

A final element that is part of the content-related shaping of discourses should be discussed here. The structuring moments of statements and discourses, through which various interpretation schemes, classifications, and dimensions of the phenomenal structure (for example, actors, problem definitions) are placed in relation to one another in a specific way, can be described as narrative structures. Narrative structures are not simply techniques used to link linguistic elements together, but as “mise en intrigue” (emplotment; Paul Ricoeur), as a configurative act which links disparate signs and statements in the form of narratives, they are rather also a basic modus of the human ordering of the experience of the world (cf. Ricoeur 1984: 5). In the seriality of discursive events constituting a discourse, the above mentioned elements of knowledge configuration are tied together in a particular ‘narration,’ and are integrated via a common thread, a story line. Narrative structures include definable episodes, processes, the personal or the ‘actants’ and their specific positionings, the spatial and temporal structures as well as the dramaturgy (the plot) of a story line. In a synchronous perspective, they link the various interpretation elements of a discourse into a coherent, portrayable, and communicable form. From a diachronic perspective, the actualizations and transformations of the discourses are tied to one another over the course of time. They provide the acting schema for the narration with which the discourse can address an audience in the first place and with which it can construct its own coherence over the course of time.

Discourse Production

Alongside the analysis of knowledge configuration, research into the infrastructures of discourse production represents an important component of SKAD. In this case, the research is geared towards looking at the interaction between social actors and speaker positions, institutional and organizational arrangements, discursive and non-discursive practices as well as artefacts in the utterance processes. As has already been discussed, in discourse contexts social actors take on available speaker positions. They do this within the framework of more or less wide-reaching institutional/organizational forms; this includes, for example, media arenas, but also scientific courses of study and other settings of statement production. This encompasses artefacts (books, laboratory instruments, computers, etc.) but also a multitude of discursive and non-discursive practices.

Power-Effects

Foucault argued that power and knowledge are Siamese twins. Using the notion of ‘power-effects’, SKAD refers to different kinds of intended or non-intended consequences emerging out of a discursive field or discourse formation, that is the range of ‘changes in the world’ that are linked to the social processing of discourses. Discourses lead to inner-worldly consequences in two different respects. They (occasionally) create dispositifs or apparatuses of world intervention. This describes infrastructural interconnections between personnel (agents), institutional-organizational processes, artefacts, and discursive or non-discursive practices that are identified through research and which process the discursively constituted problematizations through time, space, and social collectivities and arenas although such devices are rather seldom generated quasi from nothing out of a discourse. “Creation” here is always entangled or has to cope with existing institutional-organizational infrastructures. Social fields of practice often function as mediating instances between multiple discursive impositions. This also especially applies to the question of the actual ways of acting which are a result of the discursive interpellations, without being controlled by them. Finally, one can assume intended as well as unintended aggregate effects, something which was analyzed perhaps most impressively by Max Weber in his Protestant Ethic (Weber 2002 [1904/1905]).

Interpretative Analytics: Doing SKAD Research

Hermeneutic sociology of knowledge, the current paradigm that emerged from the Berger and Luckmann tradition in Germany and which provides the context for SKAD, does not follow the “hermeneutics of suspicion” (as Ricoeur called it in the late 1960s, referring, for example, to Marx or Freud; see Ricoeur, 1970: 32–35) which even today informs Critical Discourse Analysis and Laclau and Mouffe orientated discourse research.Footnote 11 Nor does it look for the only true meaning of what speaking subjects “want to tell us”. Referring to ‘hermeneutics’ it only insists that any kind of data-oriented research that uses sign-based material requires reflected steps of interpretation. The discursive world-making can only be understood through reflexive interpretation. In order for sociological practices of interpretation to become scientifically sound, it is necessary to accompany and reflect upon interpretation processes by means of methodological provisions—although the main goal of the analysis is not ‘the ultimate truth,’ but rather the concept of conceivable ‘good reasons’ for socio-historically situated interpretative work. Discourse analyses imply interpretations even when they concentrate on formal structures, things or practices. Following the works of Hans-Georg Soeffner, the newer sociological hermeneutics is occupied with the possibilities and strategies for the methodical control of interpretation processes (cf. Soeffner 1989; Hitzler and Honer 1997). The concept of hermeneutics as “attitude and action” (Hitzler and Soeffner 1994) here refers to the ‘understanding of understanding’, and therefore a (qualitative) research methodology which, on the one hand, reflects upon the position of the researcher, and, on the other hand, develops strategies of data interpretation that focus on the comprehensibility and social objectification of the steps of interpretation. Ronald Hitzler and Anne Honer have succinctly formulated this position:

The basic problem for the sociological researcher when he is reflecting upon his work, is making it transparent for himself and for others how he understands that which he believes to understand, and how he knows what he thinks he knows. (…) Their claim entails absolutely stripping the basic operations in sociological research and theory construction of their epistemological naïveté, to reconstruct them and elucidate them (Hitzler and Honer 1997: 23f).

If sociology seeks to be an empirical science, that is, a specifically reasonable form of reality-related analysis versus being a writer’s novel or journalist’s reportage, then the formulated claim of general disclosure and transparency of the steps of interpretation must be maintained. This requires a systematic procedure of analysis and applies independently of whether or not subjective or collective stocks of knowledge (or the forms of externalization/articulation which document them or are indicative of such: books, speeches, newspaper articles, films) are analysed. Like Grounded Theory, Conversation Analysis and other qualitative approaches, SKAD therefore favours sequential analysis of textual data directed towards its own research questions, to give an account of discursive claims and statements beyond the single utterance or discursive event: line by line, step by step development, debate and choice of interpretations, in order to build up a socially accountable analysis of frames (Deutungsmuster), phenomenal structure, classifications and so on. The open coding procedure elaborated by Grounded Theory indicates this way of ‘methods’. In referring to such traditions SKAD takes care of the fragile relation between ‘questions towards’ and ‘answers given’ by empirical data. For example, in my own research on waste issues, a ‘risk’ frame was elaborated out of newspaper data. This interpretation scheme entered German discourses on waste in the early 1980s, and appeared in many different ways: as textual utterance, as newspaper title illustration, as scientific analysis of waste incineration; in French discourse on waste the main organizing frame to be found was the importance of the French engineers mastering of all kinds of technological procedure, including types of waste disposal.

I speak of interpretative analytics in order to emphasize that discourse research places various data types and interpretation steps in relation to one another, for example, more classical sociological strategies of individual case analysis or case studies combined with detailed close analyses of textual data. I also speak of interpretative analytics, because, in contrast to other qualitative approaches in sociology, SKAD is not per se interested in the ‘consistency of meaning’ inherent to one particular document of discourse, but rather assumes that such data is articulating some (not all) elements’ of discourse or maybe appear as crossing point of several discourses (as lots of books or newspaper articles). So discourse research has to break up the material surface unity of utterances. The mosaic of the analysed discourse or discourses develops incrementally out of this process—this is certainly one of the most important modifications of ‘traditional qualitative approaches’ in the social sciences (cf. Keller 2003), which very often take one interview, for example, as a ‘coherent’ and ‘sufficient’ case of its own.

SKAD is characterized, like all discourse-focused approaches, by a relation of self-reflexivity. It is no more and no less than a discourse about discourses which follows its own discourse production rules, ways of enabling and disciplining. Statements about individual data as well as generalizing hypotheses formulations and conclusions must be argued and explained. As the discussion up to this point has illustrated, the questions about self-reflexivity and about the constructivism of sociology of knowledge discourse analysis are closely linked to one another. Constructivism does not indicate any kind of escape from reality and its occasionally painful materiality. Discourses are to begin with positive and materially occurring linguistic actions and communication processes which process (disputable) statements and bodies of knowledge. The concrete existence of discourses and dispositifs is therefore assumed and in no way disputed. Constructivism, as the basic approach of a discourse-theoretical and analytical program, means focusing the analysis on the socially produced ‘order of things’ in the medium of discursive knowledge politics, and so to make the contingency of the symbolic order the basis for the questions about those processes which it transforms into temporarily fixed crystallizations and structural contexts. In this context, neither the resistant character of reality nor the existence of physical phenomena and processes that are independent from assignment of meaning are denied. Therefore, not everything can be ‘successfully’ said and practically ‘done’ in all kinds of ways about everything. However, the criteria for the evaluation of evidence and inconsistencies themselves are a part of discourses, and in this way there is no escape from the net of meanings. It cannot be ignored that the SKAD reconstruction work is also irreducibly construction work. The interpretation can be called reconstructive because it refers to data and its goal is to reveal something about the data’s interrelation and peculiarities (for instance, contained interpretation schemes, meaning structurings, and so on). In this general sense, all discourse research necessarily proceeds in a reconstructive way. Such analyses proceed constructively (including those in discourse research) because they generate interpretations, conceptual schemata, and so on out of the data, and in so doing they generate types of statements that were not in the actual data as such and could not have been. Since the construction process is determined first of all by the relevancies—the questions, analysis concepts and strategies—of sociological discourse research, these are geared towards giving the “field’s own relevancies” a chance.

One final point should be addressed here. SKAD uses the rich tradition of qualitative data analysis to proceed in discourse research. It refers to case studies and fieldwork methods as developed in symbolic interactionism and interpretative sociology in order to grasp the materialities and dispositifs of discourse. As mentioned above, it uses sequential analysis in order to analyse frames, classifications, phenomenon structures and the like, on the utterances level, whether they appear as texts or audio-visual data. It draws upon the strategies of data collection and data analysis proposed by grounded theory, first and foremost its ideas of theory-oriented sampling and concepts of minimal and maximal contrast in structuring data work (see Strauss 1987: 22–40; Strauss and Corbin 1998: 201–216). Theoretical sampling means the step-by-step building up of data corpora, in starting analysis early and in following argued criteria for continuing data collection. Minimal and maximal contrasting is a systematic strategy to cross the field of inquiry in order to establish the range of important findings and to achieve detailed accounts of particular elements of analysis. Those strategies, together with sequential analysis and coding-like development of concepts (cf. Strauss 1987: 55–81), are established in SKAD in order to account for the ‘doing’ of research. That mainly means: to achieve a reasoned analysis where others could agree by argument with what is said about one single piece of data or the whole data corpus. To be clear: SKAD, unlike Grounded Theory, does not aim to explore particular ‘situations and (inter)actions’, but ‘discourses’. So it does not take up the whole of grounded theory research methodology (cf. Keller 2003: 93–108). And it is well aware that the idea of ‘complete transparency of research’ works as a guiding horizon that will never be definitely reached in this world.

Conclusion

SKAD proposes to take seriously Foucault’s interests in discourse as practice of power/knowledge. It therefore claims to be more than text or language-in-use analysis: it considers the knowledge side and the ‘power effects of discourses’, the infrastructures of discourse production as well as the institutional effects and ‘external’ impacts on practice emerging out of discourses meeting fields of practices. Neither linguistic discourse research nor Critical Discourse Analysis (and analysis of discursive hegemonies only to a small extend) are dealing with this Foucauldian program. In some way, discourse research in symbolic interactionism always has been much closer to Foucault’s arguments than it appeared to be in the disciplinary fights which emerged over recent decades between poststructuralism and sociology. Discourses do not speak for themselves, but are rather first ‘brought to life’ in historically situated processes of interaction and institution building by social actors, and their communication (inter-)acts within pre-existing social fields of practice and institutional structures. These are always intersecting orderings of practice or the corresponding ordering processes and efforts, whose actual scope probably rarely matches the discursively projected models, and which—as Max Weber already argued a hundred years ago—are all more or less transitory (Kendall and Wickham 2001; Law 1994). By taking the ideas of power/knowledge and the ‘existence’ of discourses as practices and dispositifs seriously, SKAD is not only engaged in researching communication, texts and images, but rather also in dispositif analysis, and thus, in case studies, observations, focused ethnography which considers the practical accomplishment of utterances or other elements of discourse in their socio-historical settings.