Analysing discourses and dispositives

A Foucauldian approach to theory and methodology

By Siegfried Jäger and Florentine Maier

Noch nicht begutachtetes Draft: Ruth Wodak/Michael Meyer: Methods of Critical Discourse Analysis, 3. Auflage, Sage Publications Ltd, London 2014

1 Introduction

This article aims to give an introduction to the methodology of analysing discourses and dispositives, building on the theoretical insights of Michel Foucault. The article aims towards novices to this approach. Critical discourse and dispositive analysis based on Michel Foucault’s discourse theory centres on the following questions:

• What is valid knowledge at a certain place and time?
• How does this knowledge arise and how is it passed on?
• What functions does it have for constituting subjects?
• What consequences does it have for the shaping of society?

By knowledge we understand all elements of thinking and feeling in human minds, or in other words, all contents that make up human consciousness.

People derive this knowledge from the discursive surroundings into which they are born and in which they are enmeshed throughout their lives. Knowledge is therefore conditional, i.e. its validity depends on people’s location in history, geography, class relations etc.

Critical discourse analysis and dispositive analysis aim to identify the knowledges contained in discourses and dispositives, and how these knowledges are connected to power relations in power/knowledge complexes. All sorts of knowledge can be analysed, for example: common knowledge transmitted through everyday communication, scientific knowledge, knowledge transmitted by the media, by schools, etc.

To provide an overview of the methodology for critically analysing discourses and dispositives, we first need to establish the foundations of Foucauldian discourse theory on which critical discourse and dispositive analysis build. An outline of this theoretical background is given in the ensuing section. We will show that the concept of dispositive can be understood as – in the end – identical with the concept of discourse. In the third section, practical methodological guidelines for analysing discourses and dispositives are provided. We conclude with general recommendations how to use and further develop the proposed methodology.

2 Theoretical foundations of discourse and dispositive analysis

In this chapter, we show the way from discourse analysis to dispositive analysis. We connect both methods, by showing that both aim to illuminate the knowledge that people need to speak, act, and create things.

2.1  The concept of discourse

By discourse we understand an ‚an institutionalized way of talking [and, we may add: non-linguistically performed acting] that regulates and reinforces action and thereby exerts power‘ (Link, 1983:60, authors’ own translation).

With this understanding we build on the work of literary and cultural scientist Jürgen Link and his team at the University of Dortmund, who have developed an approach, based on Foucauldian theory, to analyse discourses and their power effects, to uncover the linguistic and iconographic means by which discourses work, and to analyse how discourse legitimize and secure hegemony in bourgeois-capitalist modern society (see Link, 1982). We extend their definition by adding non-linguistically performed action, including the creation of material objects, which Foucault refers to in his later works. These are also manifestations of the knowledge located in people’s minds.

This definition can be illustrated by the image of discourses and dispositives as flows of knowledge throughout time and space. Different discourses and dispositives are intimately entangled with each other and together form the giant milling mass of overall societal discourse and dispositives. This milling mass is growing constantly and exuberantly.

2.2 Discourses and reality

Unlike disciplines such as the natural sciences that view material reality as an objective given, discourse and dispositive analysis examine how reality is brought into being by human beings assigning meanings. Only by being assigned a meaning does reality come into existence for people.

Discourses thus do not merely reflect reality. Rather, discourses shape and enable (social) reality. Without discourses, there would be no (social) reality. Discourses can thus be understood as material reality sui generis. They are not a second-class material reality, not ‚less material‘ than ‚real‘ reality, not passive media into which reality is imprinted. Discourses are fully valid material realities among others (Link, 1992).Therefore, discourse cannot be reduced to a notion of ‚false consciousness‘ or ‚distorted view of reality‘, as in some orthodox Marxist approaches to ‚ideology critique‘. Discourse is a material reality of its own. It is neither ‚much ado about nothing‘, nor a distortion, nor a lie about reality. This characterization of discourse as material reality implies that discourse theory is a materialist theory. Contrary to a common misconception, probably based on the fact that discourse analysis deals with language, discourse theory is not an idealist theory. In other words, discourse theory deals with material realities, not with ‚mere‘ ideas. Discourses may be conceptualized as societal means of production. Discourses are not ‚mere ideology‘; they produce subjects and reality.

We can therefore say: Discourses determine reality, though of course always via intervening active subjects in their social contexts as co-producers and co-agents of discourses. The subjects are able to do this because they are entangled into discourse and therefore have knowledge at their disposal. Discourse analysis is therefore not only about the retrospective analysis of allocations of meaning, but also about the analysis of the on-going production of reality through discourse, conveyed by active subjects. We have thus expounded the fundamentals of knowledge, discourse and reality. This leads us to the question how they are related to materializations and non-linguistically performed action. To answer this question, we introduce the concept of dispositive.

2.3 Dispositives

By dispositive (building on Foucault, 1980a:194), we mean a constantly evolving synthesis of knowledge that is built into linguistically performed practices (i.e. thinking, speaking, writing), non-linguistically performed practices (vulgo “doing things”), and materializations (i.e., natural and produced things).

The dispositive as a whole comprises the net that is spun between these linguistically and non-linguistically performed elements. It’s linguistically and non-linguistically elements are interrelated and unable to exist on their own. Together, they constitute reality. Figure 1, attempts to give a simplified illustration of such a system.

Figure 1: A simple example of a dispositive

Discourse analysis and dispositive analysis have become known as separate approaches not because of sharp differences in theoretical content or methodological approach, but because they have emerged from different receptions of Foucauldian theory. This is because Foucault has laid a foundation for analysing dispositives in terms of their power/knowledge implications, but has not spelled out a complete and explicit theory on this matter. The obstacle that prevented him from doing so was that in most of his work he saw only linguistically performed practices (and talking and writing more so than thinking) as discursive. Even though on some occasions he crossed that line (for example writing about „the latent discourse of the painter […] the murmur of his intentions, which are not transcribed into words, but into lines, surfaces, and colours“, Foucault, 1972:193), most of the time he did not see non-linguistically performed practices and artefacts as discursive. In that, ironically confirming his own theory, he showed himself as a product of his times and origins, where the bourgeoisie valued intellectual activity and saw physical activity as something separate and unintellectual. We shall therefore, for the purpose of this article, leave aside those lines of Foucault’s thought that turned out as dead end streets (interested readers can find out more about them for example in Jäger & Maier, 2009; Waldenfels, 1991:291; Bublitz, 1999:82-115). Instead, we shall develop the fruitful ones and propose an approach to dispositive analysis that rests on the notion of knowledge as the connecting force. We do so in the spirit of bricolage, picking up Foucault’s toolbox of theoretical and practical instruments, and developing his ideas further.

The linguistically and non-linguistically performed practices and materializations are connected by knowledge or, to put it more precisely, by a common power/knowledge complex. By knowledge, as stated above, we understand all kinds of elements of thinking and feeling in human minds. Everything in human consciousness is discursive, i.e. constituted by knowledge. People derive this knowledge from the discursive surroundings into which they are born: What other people tell them, what they read or hear via the media, what treatments they experience, what material objects they are in contact with, etc. Thereby they learn the conventions of assigned meanings, which helps them to interpret reality in the way it has previously been interpreted by others.

People’s knowledge is the basis for their thinking, speaking, silent acting and production of things. Thereby knowledge, in human consciousness and transformed into physical action, creates and becomes reality. Foucault explicates this mechanism in the “Archaeology of Knowledge”, where he shows how discourses systematically form the objects (i.e., not just material objects but also topics of thought and speech) of which they speak (Foucault, 2002:54). He also describes how non-linguistically performed practices and materializations (‘discursive relations’) offer objects of which discourses can speak, and determine the relations that discourse must establish in order to speak of this or that object (ibid.:50f.). In doing so, his aim is to define objects ‚by relating them to the body of rules that enable them to form as objects of a discourse and thus constitute the conditions of their historical appearance‘ (ibid.: 53.). This is also the aim of the approach to critical dispositive analysis that we present here.

For a more precise understanding of knowledge, it is helpful to distinguish between explicit and tacit knowledge. Some of the knowledge that people have is explicit, for example expressed in words, mathematical formulas or diagrams. (Such as a sign that states: “Children must be accompanied on lift at all times.”) Much knowledge, however, is implicit. It is hardly verbalized (e.g.: Do not stare at people when you are inside the lift with them.) or even cannot be adequately verbalized (e.g.: What exactly is the borderline between staring at somebody and just normally looking at them?). The tacit knowledge of a particular culture is passed on in non-linguistic practices and materializations. In other words, the knowledge is in the practices and materializations, and people learn it by watching others and trying out themselves. Researchers can reconstruct and explicate tacit knowledge by closely analysing non-linguistic practices and materializations, and – as in case of some ethnographies – also by participating in the practices themselves.

As we have explained above, discourse and dispositive analysis view reality as something that is created by human beings assigning meanings. In dispositive analysis, this process of reality creation is made salient with respect to non-linguistically performed practices and materializations. Of course, this does not mean that human beings are the creators of the raw matter of material reality. But people shape and use these raw materials. The assignment of meanings includes tangible physical acts. For example, when the Three Gorges Dam was built at the Yangtze River, this happened as part of a dispositive of “development”. The river was given a new meaning as a major source of hydroelectric power. For the rice farmers who were resettled to make place for the dam, the river lost the meaning of being a source of irrigation for their now flooded paddies. The river lost one meaning and gained another. As the knowledge assigned to a particular part of reality changed, this part of reality turned into a different thing.

An object that is not assigned any meaning is not an object. It is totally nondescript, invisible, even non-existent. I don’t see it because I overlook it. We can trace this phenomenon by looking at cases where discourses have withdrawn from the reality that used to be built on them. In such cases, those parts of reality become meaningless in the truest sense of the word. They return to a blank state. For example: After the Second World War, firewood in Vienna was so scarce that people foraged parks for down wood. For them, a broken branch was a valuable piece of firewood. Today, when we walk through the park and see a broken branch on the ground, we most likely do not take particular notice of it.

The point is that all meaningful reality exists for us because we make it meaningful for us, or because our ancestors and neighbours assigned meaning to it, and this meaning is still valid for us. It is like King Midas with his gold: everything he touched turned into gold. Similarly, everything that human beings assign meaning to becomes a particular kind of reality, according to the meaning it was assigned. Ernesto Laclau expressed this connection elegantly:

By ‚discursive‘ I do not mean that which refers to ‚text‘ narrowly defined, but to the ensemble of the phenomena in and through which social production of meaning takes place, an ensemble which constitutes a society as such. The discursive is not, therefore, being conceived as a level nor even as a dimension of the social, but rather as being co-extensive with the social as such. This means that the discursive does not constitute a superstructure (since it is the very condition of all social practice) or, more precisely, that all social practice constitutes itself as such insofar as it produces meaning. Because there is nothing specifically social which is constituted outside the discursive, it is clear that the non-discursive is not opposed to the discursive as if it were a matter of two separate levels. History and society are an infinite text. (Laclau, 1980:87)

We may therefore say that reality is meaningful, that reality exists in the way it does, only insofar as it is assigned meaning by people, who are themselves entangled into and constituted by discourses. If people no longer assign the same meaning to an object, the object changes or loses its meaning. This meaning may then at most be reconstructed as a former meaning that has mixed with other meanings or has ceased to be valid. Even if we just watch the night sky and see constellations of stars there, we see them as a result of a discourse. We see the constellations because we have learnt to see them. To assign meaning is not a noncommittal, ‚merely symbolic‘ act. To assign meaning is to animate whatever one comes across, to re-shape and change. For example, from the collective symbolism used with regard to immigrants, it is apparent that many people have learnt to assign negative meanings to immigrants, and now actually perceive them as floods that need to be held back, or even as lice or pigs that should be crushed or slaughtered.

The elements of a dispositive are connected not just by a certain kind of knowledge, but also by a common purpose they serve, namely the purpose of dealing with an urgent need. In fact, Foucault saw this as the major inner bond that holds a dispositive together. He defined a dispositive as

[…] a sort of— shall we say — formation which has as its major function at a given historical moment that of responding to an urgent need [urgence]. The [dispositive] thus has a dominant strategic function. (Foucault, 1980b:195)

If a society – or rather: its hegemonic forces – are confronted with an urgent need that has arisen due to some shift in power/knowledge relations (for example a state has become insolvent, or an airplane has crashed into the World Trade Center), they will gather all the elements they can get hold of to deal with it. These may be speeches, full-body scanners and biometric passports, new laws, a new committee such as the Troika, and so on. By these means, they mend the ‚leak‘, the urgent need that has arisen (see Balke, 1998; Deleuze, 1988). The dispositive then stays in place until a new shift occurs.

Discourses and dispositives are therefore specific to particular times and places. Different cultures have different discursive objects, which can make it difficult to translate texts from one language into another, or to understand texts and other symbolic practices from other cultures and historical periods. Changes of discourse do not happen out of the blue. Often discursive changes take place in long processes that impalpably but thoroughly change everything. But also abrupt changes can happen, when a discourse that previously seemed like a quiet little brook turns into a torrent (for example when the self-incineration of a Tunisian street vendor sparked off the Arab spring). In other words, the contents of discourses and dispositives are malleable to change.

2.4 Discourses and power

As we have noted above, discourses are not only expressions of social practice but also serve particular ends, namely the exercise of power. It is therefore necessary to discuss the Foucauldian concept of power. The question how the production of discourses and the temporarily and spatially contingent knowledge they contain are connected to mechanisms and institutions of power has been central for Foucault (as he states in the introduction to the German version of The History of Sexuality — see Foucault, 1983:8).

Power, in a Foucauldian sense, refers, to ‘a whole series of particular mechanisms, definable and defined, that seem capable of inducing behaviours or discourses‘ (Foucault, 1996:394).

Discourses exercise power because they institutionalize and regulate ways of talking, thinking and acting. Based on the above outline of the connection between discourses and reality, we can distinguish two kinds of connections between discourses and power: On the one hand, there is the power of discourse. On the other hand, there is also something like the power over discourse.

The power of discourse lies in the fact that discourses delineate a range of elements of knowledge, which are sayable, makeable and seeable. This means that they simultaneously inhibit other elements of knowledge that are not sayable, not makeable and not seeable (cf. Jäger, 2012; Link & Link-Heer, 1990). As flows of knowledge through time and space, discourses determine the way in which a society interprets reality and organizes further linguistically and non-linguistically performed discursive practices (i.e. further talking, thinking, acting and seeing). To put it more precisely, we can distinguish two effects of discourse: Firstly, discourses form individual and mass consciousness and thereby constitute individual and collective subjects. Secondly, since consciousness determines action, discourses determine action, and action creates materializations. Discourses thus guide the individual and collective creation of reality.

From a discourse-theoretical point of view, it is thus not the subject who makes the discourses, but the discourses that make the subject (which may be irritating for those attached to the idea of the uniqueness of the individual). The subject is of interest not as an actor, but as a product of discourses. As Foucault argues:

One has to dispense with the constituent subject, to get rid of the subject itself, that’s to say, to arrive at an analysis which can account for the constitution of the subject within a historical framework. And this is what I would call genealogy, that is, a form of history which can account for the constitution of knowledges, discourses, domains of objects etc. without having to make reference to a subject which is either transcendental in relation to the field of events or runs in its empty sameness throughout the course of history. (Foucault, 1980b:117)

Foucauldian discourse theory, though often wrongly accused of so doing, does not deny the subject. It aims to analyse the constitution of the subject in its historical and social context from a diachronic (i.e. longitudinal) and synchronic (i.e. cross-sectional) perspective: who was conceived of as a subject at a particular point in time? How, and how come? For example, unlike in the past, women and children in Western society usually have subject status today. In modern management, in contrast to traditional bureaucracy, employees are depicted as subjects. They are ‚empowered‘, with all the responsibilities that involves. The subject status of foetuses and apes is hotly debated. Asylum seekers and criminals are denied subject status when they are portrayed as maniacs, dogs or viruses (these are collective symbols – see the section below on collective symbols). In a nutshell, Foucauldian discourse theory contests the existence of an autonomous subject, but that does not mean that it is against the subject. The active individual is fully involved when it comes to realizing power relations in practice. The individual thinks, plans, constructs, interacts and fabricates. Individuals also face the problem of having to prevail, to assert themselves, to find their place in society.

When analysing the power effects of discourse, it is important to distinguish between the effects of a text and the effects of a discourse. A single text has minimal effects that are hardly noticeable and almost impossible to prove. In contrast, a discourse, with its recurring contents, symbols and strategies, leads to the emergence and solidification of knowledge and therefore has sustained effects. What is important is not the single text, the single film, the single photograph, etc. but the constant repetition of statements. The philologist Victor Klemperer recognized this mechanism as early as the 1930s, when he observed the language of the Nazis. In his analysis of the language of the Third Reich (Klemperer, 2001, 2006), he contends that fascist language works like the continuous administration of small doses of arsenic, which unfold their effect only over the long term. An analogous observation can be applied to today’s proliferation of neoliberal economic discourse (e.g. Balanoff, 2013).

With regard to power over discourse, different individuals and groups have different chances of influence. However, none of them can simply defy hegemonic discourse, and none of them alone has full control over discourse. Discourses are supra-individual. Everybody is co-producing discourse, but no single individual or group controls discourse or has precisely intended its final result. Discourses take on a life of their own as they evolve. They transport more knowledge than the single subject is aware of. In Foucault’s words, ‚[p]eople know what they do; they frequently know why they do what they do; but what they don’t know is what they do does‘ (sic, personal communication, quoted in Dreyfus & Rabinow, 1982:187). The power effects of discourses should therefore not necessarily be interpreted as the conscious and manipulative intent of some individual or group. There may be a difference between a speaker’s reasons for using a particular discourse, and the social consequences of doing so (Burr, 2003:61). Yet, in the long run actors can accomplish changes in discourse. This – by definition – qualifies them as powerful actors. They may be members of political and economic elites who have greater financial resources or privileged access to the media. For example, the Basic Constitutional Law that governs the right of asylum in Germany was tightened after over ten years of intensive political and media lobbying. Exclusions inherent to the structure of discourse can thus be amplified by institutional conditions. But also less elite actors are able to shift discourses, if they are skilful communicators (which to no small degree requires education). For example the Occupy movement, with its slogan “We are the 99%”, has put the issue of economic inequality back into daily political discourse in the Global North, and Arab Spring uprisings have brought the issue of democracy to the table in countries that previously seemed like unshakable authoritarian regimes. It is probably no coincidence that both of those movements were driven by groups of rather well-educated people.

2.5 Critique and the aims of critical discourse analysis

After having shed some light on the issue of power, we can further clarify the notion of critique and the aims that underlie critical discourse analysis. As stated in the introduction, critical discourse and dispositive analysis aim to identify the knowledges contained in discourses and dispositives, and how these knowledges are connected to power relations in power/knowledge complexes. This comprises two aspects: Firstly, it entails disentangling the giant milling mass of discourse, charting what is said in a given society, in a particular time and place, with regard to its qualitative spectrum (What is said? How is it said?), and uncovering the techniques through which discursive limits are extended or narrowed down. Secondly, it entails subjecting these workings of power/knowledge to critique.

Critique does not mean to bluntly evaluate whether a particular discourse is good or bad. It means to expose the evaluations that are inherent in a discourse, to reveal the contradictions within and between discourses, the limits of what can be said, done and shown, and the means by which a discourse makes particular statements, actions and things seem rational and beyond all doubt, even though they are only valid at a certain time and place.

In doing so, critical analysts orient themselves towards ethical principles (for example that all human beings are of equal value and that their physical and mental integrity should not be violated). Obviously, since all knowledges are valid only in particular times and places, these ethical principles also cannot claim the status of objective truths. They have emerged out of a long discursive process of debate and struggle, and their validity is ever precarious. Nonetheless – or rather: as a result of this process – the analyst has come to understand them as valid.

Critical discourse analysts need to be clear about the fact that their critique is not situated outside discourse. They participate in discourse as parrhesiasts: They make clear that they are expressing their own opinion which they understand to be true. They express it in the most direct words they can find. And they do so even at personal risk, for example when they say something inconvenient to the audience and contradict mainstream opinions (Foucault, 2001). This kind of critique is not ideological, because unlike ideology it does not make claims to absolute truth. Critical discourse analysts aim for a democratic discourse, where people exchange ideas on equal footing, try to understand each other, and are open to modifying their position based on sound arguments.

The following section provides some ideas on how discourses and dispositives can be analysed systematically. For this purpose, we first introduce concepts to analyse the structure of discourses and dispositives. After that, we discuss under what conditions discourse analyses can be considered as complete. We proceed by going through a “little toolbox” of discourse analytic methods, giving step-by-step guidelines for matters ranging from the choice of a subject, to analysis proper. We close with thoughts on the peculiarities of dispositive analysis.

3 Methods for analysing discourses and dispositives

The theoretical considerations introduced in the previous section of this chapter are the basis for the development of concepts and methods that facilitate analysis. A range of these concepts and methods is outlined in this section. The scope of this chapter does not allow for detailed methodological justifications of each of these concepts and methods, but such explanations can be found in the book on methods of critical discourse analysis by Jäger (2004, 2012).

Methods of critical discourse analysis also include linguistic concepts (e.g. figurativeness, vocabulary, pronominal structure, argumentation) which can be used to examine the more subtle workings of texts. However, these linguistic instruments are not described in detail here, as explanations can be found in works on style and grammar. Linguistic concepts fill only one slot in the ‚toolbox‘ of discourse analysis. Depending on the research question and subject matter, various other tools can be added to the toolbox. In any case, certain methods are part of a standard repertoire. These are described in the remainder of this section, putting the emphasis on the analysis of linguistically performed discursive practices – what is usually known as discourse analysis – and shining a more cursory light on the analysis of non-linguistically performed discursive practices and materializations in an extended form of discourse and dispositive analysis.

3.1 The structure of discourses and disposives

The following suggestions on terminology, in a first step, aim to provide some help in making the structure of linguistically performed discourses more transparent and amenable to analysis.

Special discourses and interdiscourse

A basic distinction can be drawn between special discourses and interdiscourse. Special discourses are discourses in the sciences, while interdiscourse refers to all non-scientific discourses. Elements of special discourses continuously feed into interdiscourse and vice versa.

Discourse strands

In general societal discourse, a great variety of topics arise. Flows of discourse that centre on a common topic are called discourse strands. Each discourse strand comprises several sub-topics, which can be summarized into groups of sub-topics.

The concept of discourse strands is similar to the one of discourses. The difference is that discourse is the more abstract concept, located at the level of statements (enoncés, i.e. all the kernels of meaning that constitute the “atoms” of a particular discourse). Discourse strands, in contrast, are conceived of at the level of concrete utterances (énonciations) or performances located on the surface of texts (cf. Foucault, 2002).

Every discourse strand has a diachronic and a synchronic dimension. A synchronic analysis of a discourse strand examines the finite spectrum of what is said and sayable at a particular point in time and place. A diachronic analysis cuts through a discourse strand at various points in time and place, for example at particular discursive events. By comparing these synchronic cuts, it provides insights into the changes and continuities of discourse strands over time.

In a way, a synchronic cut through a discourse strand is always also a diachronic one. This is because each topic has a genesis, a historical a priori. When analysing a topic, the analyst has to keep an eye on its history. To identify the knowledge of a society on a topic, the analyst has to reconstruct the genesis of this topic. Foucault has undertaken several attempts to do so, not only with regard to the sciences, but also with regard to everyday life and institutions (e.g. the hospital and prison in France).

Discursive limits and techniques for extending them or narrowing them down

Each discourse delineates a range of statements that are sayable and thereby inhibits a range of other statements, which are not sayable (cf. Link & Link-Heer, 1990).The borders to what is not sayable are called discursive limits.

Through the use of certain rhetorical strategies, discursive limits can be extended or narrowed down. Such strategies for example include direct prescriptions, relativizations, defamations, allusions and implicatures. Discourse analysis examines these strategies in their own right, and also uses them as analytic clues to identify discursive limits: if ‚tricks‘ are used, this is an indicator that certain statements cannot be said directly without risking negative sanctions. For example in modern day racism, statements are often introduced with the clause “I am not racist, but…”, which extends the limit of what can be said without being accused as racist. When politicians say that “there is no alternative” to a particular course of action, this narrows down discursive limits, because it suggests that there is no possibility to call this action into question and publicly debate it.

Discourse fragments

Each discourse strand consists of a multitude of elements that are traditionally called texts. We prefer the term ‚discourse fragment‘, because one text may touch on various topics and thus contain various discourse fragments. A discourse fragment therefore refers to a text or part of a text that deals with a particular topic. For example, if we are interested in the discourse of immigration, we may find relevant discourse fragments in immigration laws and news articles that focus on immigration, but also in news articles that focus on other topics and mention immigration only in passing. For example in an article with the title “Man cuts his ex-girlfriend’s throat: 20 years of prison” we may read: “’I could not live without her’, said the accused, a Romanian who has been living in Austria since 2006.” Or in a shop we may find a Halloween costume, consisting of an orange prison jumpsuit and a space alien mask, sold as “Illegal Alien” costume (which is a splendid example of a dehumanizing collective symbol). All such discourse fragments on the same topic form a discourse strand.

Entanglements of discourse strands

A text usually refers to various topics and therefore to various discourse strands. In other words, it usually contains fragments from various discourse strands. These discourse strands are usually entangled with each other. An entanglement of discourse strands can take the form of one text addressing various topics to equal degrees, or of one text addressing mainly one topic and referring to other topics only in passing.

A statement where several discourses are entangled is called a discursive knot. For example, in the statement “integrating immigrants into our society costs a lot of money”, the discourse strand of immigration is entangled with the discourse strand of the economy. In the statement “in [insert any Islamic country here], they still live in a patriarchal society”, the discourse strand of immigration is entangled with the discourse strand of gender.

Two discourse strands can be entangled more or less intensively. For example, in everyday discourse in Germany, the discourse strand of immigration is intensively entangled with the discourse strand of gender, as sexist attitudes and behaviours are attributed to immigrants (see Jäger, 1996).

Collective symbols

An important means of linking up discourse strands is the use of collective symbols. Collective symbols are ‚cultural stereotypes‘, also called ‚topoi‘, which are handed down and used collectively (Drews, Gerhard, & Link, 1985: 265). They are known to all members of a society. They provide the repertoire of images from which we construct a picture of reality for ourselves. Through collective symbols we interpret reality, and have reality interpreted for us, especially by the media.

An important technique for connecting collective symbols is catachreses (also called image fractures). Catachreses establish connections between statements, link up spheres of experience, bridge contradictions and increase plausibility. Thereby, catachreses amplify the power of discourse. An example of a catachresis is the statement “the locomotive of progress can be slowed down by floods of immigrants”. Here, the symbols of the locomotive (meaning progress) and floods (meaning a threat from the outside) are derived from different sources of images. The first one is taken from traffic and the second from nature. With a catachresis, the images are connected.

As a special form of collective symbols, pragma-symbols deserve mention. These are terms that refer to material objects while at the same time pointing to a meaning beyond that. For example: “In this civil war, it is tanks against stones.” This refers not just to the concrete fighting situation, but also to the unequal strength of conflict parties.

Discourse planes and sectors

Different discourse strands operate on different discourse planes, such as the sciences, politics, the mass media, education, everyday life, business, administration, etc. These discourse planes can be characterized as social locations from which speaking takes place.

Discourse planes influence each other and relate to each other. For example, on the mass media plane, discourse fragments from scientific specialist discourse or political discourse are taken up. The mass media also take up everyday discourse, bundle it, bring it to the point, or – especially in the case of the yellow press – spice it up with sensational and populist claims. In this way, mass media regulate everyday thinking and exert a considerable influence on what is and what can be done in politics and everyday life. For example, the larger-than-life image of the now defunct Austrian populist Jörg Haider would hardly have come about without the help of media reports that normalized right-wing populism.

A discourse plane consists of various sectors. For example, TV, newspapers and the internet are different sectors of mass media. When analysing discourse planes, it is important to consider the relative importance of various sectors for the research question at hand. For example, social media and Web 2.0 are gaining importance as forms of mass media hat open up unprecedented opportunities for two-way communication and – at the same time – for surveillance.

A discourse plane is tightly interwoven in itself. For example, on the discourse plane of the mass media, a TV broadcast may repeat and build on contents that have been brought up in social media, and vice versa. It is therefore all the more justified to talk about the mass media discourse plane, which – especially with regard to traditional mainstream media in a society – can be considered as integrated in its major aspects.

Discursive events and discursive context

All events are rooted in discourse. However, an event only counts as a discursive event if it appears on the discourse planes of politics and the mass media intensively, extensively and for a prolonged period of time.

Discursive events are important because they influence the future development of discourse. For example, the Three Mile Island nuclear accident near Harrisburg was comparable to the one in Chernobyl. But while the Three Mile Island accident was covered up for years, the Chernobyl accident was a major media event and influenced global politics. The nuclear accident of Fukushima again changed the discourse of nuclear power worldwide. Whether an event, such as a nuclear accident, becomes a discursive event or not depends on the power constellations at work in politics and the media. Discourse analysis can examine whether an event becomes a discursive event or not. If it becomes a discursive event, it influences the further development of discourse.

Another example for a discursive event is the Eurozone crisis. It is a combined crisis, or series of crises, in several dimensions as such sovereign debt, banking, unemployment, etc. People roughly agree that it started in 2008 or 2009, but there is no consensus about its duration and precise nature. Repeatedly, politicians, economists and journalists have declared the end of the crisis. We see here a discursive struggle whether or not a discursive event is still going on. If the crisis is on, further measures need to be taken, probably more far-reaching ones than the ones taken so far. If the crisis is over, there is no such need. The attempts at crisis management can be understood as attempts to get the urgencies of various interest groups (nations, different groups of the population, corporations,…) under control.

Another reason why the identification of discursive events is important is that they outline the discursive context that a discourse strand relates to. For example, a synchronic (i.e. cross-sectional) analysis of a discourse strand can be enriched with diachronic (i.e. longitudinal) elements by adding a chronicle of the discursive events belonging to it. Such historic references can be helpful for synchronic analyses of discourse strands (as, for example, demonstrated by Caborn, 1999).

Discourse positions

A discourse position describes the ideological position from which subjects, including individuals, groups and institutions, participate in and evaluate discourse. Also, the media take up discourse positions, which become evident in their reporting. (As noted above, subject status is nothing natural and obvious, but something that in itself needs to be established through discourse.)

Subjects develop a discourse position because they are enmeshed in various discourses. They are exposed to discourses and work them into a specific ideological position or worldview in the course of their life. This relationship also works the other way around. Discursive positions contribute to and reproduce the discursive enmeshments of subjects (Jäger, 1996:47).

Discourse positions can be identified through discourse analysis. But a rough outline of discourse positions is also part of people’s everyday knowledge. People know roughly which politicians and newspapers tend towards the left, the right or the centre. Everyday self-descriptions of one’s discourse position, however, should be taken with a grain of salt. For example, newspapers often describe themselves as “independent” and “impartial”, which from a discourse-theoretical perspective is impossible.

Subjects can take up widely diverging positions. For example with regard to the discourse strand of the economy, some subjects take up a neo-liberal discourse position and favour privatization, free trade, low taxes, fiscal policy discipline, and so on. Others, in contrast, reject neo-liberalism and take up a neo-Keynesian position or something even more unorthodox.

Discourse positions are homogeneous only in their core and become diffuse with regard to less central issues. For example, subjects who embrace the discourse position of neo-liberalism agree that it is in principle right and important to reduce the nation’s budget deficit. They do not question the current economic system. However, they may have differing views on the best way to reduce the budget deficit.

Within a hegemonic discourse, discourse positions are fairly homogeneous, which itself is already an effect of hegemonic discourse. Dissenting discourse positions often belong to remote counter-discourses (e.g. a fundamental questioning of the current economic system may not arise from economic discourse, but from ecology or ethics). However, these counter-discourses can pick up arguments from hegemonic discourse and subvert their meaning (for example an Occupy Wall Street protester who defied the rain, holding up a sign saying “Do you feel it trickle down?”).

Overall societal discourse and global discourse

All the entangled discourse strands in a society together form the overall societal discourse. A society is never totally homogeneous but consists of different subcultures. In Germany, since reunification in 1989, overall societal discourse has become ideologically more homogeneous, and it seems unlikely that this is going to change easily (Teubert, 1999). The overall societal discourse of a society, in turn, is part of global discourse. Even though global discourse is very heterogeneous, there is evidence of homogenizing tendencies (see, for example, Meyer, 2009).

Overall societal discourse is a complex network. Discourse analysis aims to disentangle this net. The usual procedure is to first identify single discourse strands on single discourse planes or sectors (for example, the discourse strand of immigration on discourse sectors of traditional mass media). Subsequently, analyses of this discourse strand on further discourse planes, such as politics or everyday communication, can be added. At the end of such analyses, the question is usually how the different discourse planes and sectors of a discourse strand relate to each other. For example, one may examine whether and how the political discourse plane is linked to discourse sectors of traditional mass media or to the discourse plane of everyday communication, whether and how traditional mass media influence everyday communication, and so on.

The history, present, and future of discourse strands

Discourse strands have a history, a present and a future. In order to identify the changes, ruptures, ebbings and recurrences of a discourse strand, it is necessary to analyse longer periods of time. To put it into Foucault’s words, an ‘archaeology of knowledge’ or a ‚genealogy‘ is needed. On the basis of such an analysis, even prognoses about discourse can be undertaken. These can take the form of scenarios based on different future discursive events. Discourses may change, but normally they do not vanish totally and suddenly. Therefore discourse analysis allows prognoses.

Of course, an analysis of the history, present and future of overall societal or even global discourse is an enormous endeavour and can only be tackled in the form of many single projects. Such single projects create reliable knowledge about certain subzones of overall societal discourse. This scientific knowledge can be the basis for a change of everyday, political and media knowledge, and can change behaviours and policies. Work on the discursive plane of science can thus influence the further development of a particular discourse strand.

3.2 On the completeness of discourse analyses

A discourse analysis fully captures the qualitative range of what can be said and how it is said in one or more discourse strands. It is complete if further analysis leads to no further new findings. Social scientists who mainly work with large amounts of quantitative data will be surprised to learn that in discourse analysis, a relatively small amount of qualitative data suffices to reach this point. The arguments and contents that can be read or heard about a particular topic (e.g. immigration) at a particular time in a particular social location are amazingly limited (often in both senses of the word). With regard to methodology, this means that analysts continue to analyse new materials until they notice that arguments begin to repeat themselves. If this is the case, completeness (in the sense of theoretical saturation) has been achieved.

While qualitative analysis is the bedrock of discourse analysis, quantitative analyses can also be interesting. The analyst can examine with what frequency particular statements occur. In this way, focal issues in discourse strands, or statements that have the character of slogans and are therefore accompanied by a bulk of judgments and prejudices, can be identified. If a statement occurs frequently, it has sustained effects and strongly solidifies a particular knowledge. In diachronic analysis, frequencies can be used to identify trends (e.g. Meyer, Buber, & Aghamanoukjan, 2013). However, for the explanatory power of a discourse analysis, the qualitative aspect is of greater importance than the quantitative.

3.3 A little toolbox for discourse analysis

In this section, a brief summary of our toolbox for discourse analysis is presented. As noted above, within the scope of this chapter, we cannot provide detailed methodological justifications for each of the tools, but these can be found in the volumes by Jäger (2004, 2012). In our own research projects, we use short hand-outs like the following as memory aids or checklists when first dealing with materials.

Choosing a subject matter

The first step in a discourse analysis project is usually to choose a subject matter. In the project report (usually in the introduction), a rationale for the project and its subject matter has to be given.It needs to be kept in mind that the relationship between a phenomenon of interest and particular discourse strands is often not straightforward because a phenomenon may permeate many discourse strands. For example, in a research project that aims to examine how racism permeates the media, the researcher has to decide which discourse strand(s) to focus on. To make the choice, the researcher has to have an initial concept of racism in mind. This concept may be developed further in the course of the analysis. Theoretical concepts are always debatable, and the researcher needs to clarify and justify which concept he is working with. (( For example, a definition of racism that is generally justifiable and well-accepted in the sciences encompasses the following three elements: (1) One or several people are for biological or cultural reasons constructed as an ethnic group or even a race. (2) This group is evaluated (negatively or positively, e.g. when blacks are assumed to be superior jazz musicians). (3) The construction and evaluation takes place from a position of power (which in discourse analysis is obvious, since discourse is per se “powerful”).)) Equipped with this concept, the researcher can think about promising discourse strands where racism may be found. In the case of racism, it is, for example, the discourse strand of immigration, refugees and asylum-seeking. Of course, the discourse strand of immigration could also be interpreted in the light of other research interests. To choose a subject matter means to choose a phenomenon of interest and a discourse strand which will be examined. This discourse strand delineates the scope of materials for analysis.

Choosing a discourse plane and a sector and characterizing them

Typically, it will be necessary, at least initially, to confine the analysis to one discourse plane (for example mass media) and one place (for instance Cairo, or international airports). When examining a discourse plane, analysis may cover one or several sectors of this plane (for example, the sector of newspapers) and some important place. The choice of sector needs to be justified. For example, a sector may be exemplary for how an issue is dealt with in the mass media, or a sector may not have previously been examined in any research project. In the latter case, of course, a review of previous research should summarize the findings from an analysis of other sectors.

In some cases, it may be possible to examine several discourse planes at once. The analysis of interactions of several discourse planes in the regulation of mass consciousness is extremely interesting, but also time-consuming. To achieve this task, it is necessary to base the analysis on well-justified examples of sectors of these discourse planes and instances of their interaction. The task becomes even more complex if entanglements of discourse strands are also considered.

Accessing and preparing the materials

As a next step, the concrete corpus for analysis needs to be delineated. For example, when analysing newspapers, the particular newspapers and the time periods under consideration need to be selected. Often it will be advisable to select several important newspapers from a particular country, or from several countries, covering an extended period of time. In contrast, a project that examines the portrayal of women in pop songs could probably rely on a few exemplary songs (though this would have to be demonstrated in the particular project). In any case, the selections have to be justified.

As a preparation for analysis, a general characterization of data sources needs to be provided. For example in case of newspapers: What is their political orientation, who are their readers, what is their circulation, and so on?

Analysis

Analysis typically takes place in three steps: structural analysis of the discourse strand, detailed analysis of discourse fragments, and synoptic analysis. Usually, these steps have to be gone through several times. Their sequence can be modified.In the cycles of analysis, connections between different levels of analysis are discovered, interpretations are developed and weak arguments are discarded.

A structural analysis of the discourse strand

The first typical step is the structural analysis of discourse strands. Its detailed steps are as follows:

1. A list of all articles of relevance for the discourse strand is compiled. This list should include bibliographical information, notes about topics covered in the article, the literary genre, any special characteristics and the section in which the article appears.
2. Structural analysis should roughly capture the characteristics of articles on particular aspects of interest, such as any illustrations, the layout, the use of collective symbols, the argumentation, the vocabulary and so on, and identify which forms are typical for the newspaper. This outline will be needed later to identify typical articles for the detailed analysis of typical discourse fragments.
3. A discourse strand encompasses various subtopics. These are first identified and then summarized into groups. For example, in the case of the discourse strand of stem cell research, sub-topics may be summarized into groups such as the ‚legal implications of stem cell research‘, the ‚benefits of stem cell research‘, the ‚technical procedures of stem cell research‘, the ‚ethical problems of stem cell research‘, the ‚costs of stem cell research‘ and so on. The development of groups of sub-topics is an iterative process, which should lead to a good compromise between parsimony and discriminatory power.
4. The next step is to examine with what frequency particular groups of subtopics appear. Which ones are focused on and which ones are neglected? Are there any subtopics that are conspicuous by their absence?
5. If the analysis is diachronic, it will also examine how subtopics are distributed over the course of time. Are some subtopics particularly frequent at particular times or places? How does this relate to discursive events?
6. Discursive entanglements are then identified. For example, the discourse strand of stem cell research is entangled with the discourse strands of ethics, business and medicine.

The findings from these steps of analysis are combined and interpreted together. Thereby, a characterization of the newspaper’s discourse position begins to emerge. For example, does the newspaper perceive stem cell research positively or negatively?

The structural analysis of a discourse strand can and should already yield ideas for the ensuing detailed analysis of typical discourse fragments (see the next subsection) and for the final synoptic analysis (see the subsection on synoptic analysis). These ideas should be written down immediately and marked accordingly.

Detailed analysis of typical discourse fragments

To identify the fine detail within the newspaper’s discourse position and to assess the effects of this discourse on readers, certain discourse fragments are subjected to detailed analysis. Discourse fragments that are typical of the particular newspaper are selected for this purpose. Criteria for typicality are, for example, typical illustrations, typical use of collective symbols, typical argumentation, typical vocabulary and so on. The typical forms of these aspects have been identified in the preceding structural analysis.

To select typical discourse fragments, the researcher can proceed in several steps and rate the articles according to defined criteria. To ensure that the selection is intersubjectively plausible, several researchers can engage in this rating. The articles that score highest on typicality are then subjected to detailed analysis. If time restraints require it, or if one article exhibits all typical characteristics of the discourse strand, detailed analysis may be confined to one article only. If structural analysis has shown that the discourse strand is very heterogeneous, and if no single homogeneous discourse position can be discerned, the researcher can address several typicalities, i.e. several kinds of ‚typical‘ articles.

The procedures for selecting typical articles should be systematic and transparent, but not mechanical. What is an appropriate procedure depends on the concrete research project and the discourse strand in question. The detailed analysis of typical discourse fragments should cover the following aspects:

1. Context

• Why was this article selected? Why is this article typical?
• What is the general topic of this article?
• Who is the author? What is her position and status within the newspaper? What are her special areas of coverage, and so on?
• What was the occasion for the article?
• In which section of the newspaper does the article appear?

2. Surface of the text

• What is the layout like? What kinds of pictures or graphs accompany the text?
• What are the headings and subheadings?
• How is the article structured into units of meaning?
• What topics are touched upon in the article? (In other words, what discourse strands is the article a fragment of?)
• How do these topics relate to each other and overlap (entanglements of discourse strands)?

3. Rhetorical means

• What kind and form of argumentation does the article follow? What argumentation strategy is used?
• What logic underlies the composition of the article?
• What implications and allusions does the article contain?
• What collective symbolism is used (linguistic and graphic, involving, for example, statistics, photographs, pictures, caricatures, etc.)?
• What idioms, sayings and clichés are used?
• What are the vocabulary and style?
• What actors are mentioned, and how are they portrayed (persons, pronouns used)?
• What references are made (e.g. references to science, information about the sources of knowledge used)?

4. Content and ideological statements

• What concept of humankind does the article presuppose and convey?
• What concept of society does the article presuppose and convey?
• What concept of (for example) technology does the article presuppose and convey?
• What perspective regarding the future does the article give?

5. Other peculiarities of the article
6. Discourse position and overall message of the article.

In analysing each of these aspects, the researcher has to ask herself what this peculiarity of the article means, what it implies. For example, what does it mean that a particular image accompanies this text? What effect does this image create? Each of these interpretations remains open to revisions. At the end of the detailed analysis, the interpretations of single aspects are combined into a total interpretation of the article. Usually, the interpretations of the single aspects fit together like the pieces of a jigsaw puzzle and form a unitary picture. If one aspect stands out, it is often due to special circumstances, such as when a photo or a headline has not been provided by the author but by the editor, who had other purposes in mind, such as spicing up the article. Such discrepancies also provide important insights into the newspaper’s discourse position. Together with the findings from structural analysis, the findings from detailed analysis form the basis for synoptic analysis.

Synoptic analysis

In synoptic analysis, a final assessment of the newspaper’s discourse position is made. For this purpose, the findings from structural analysis and detailed analysis are interpreted in relation and comparison to each other.

3.4 Some thoughts on analysing dispositives

Although a dispositive has certain durability, it is subject to historical changes constant influence by other dispositives. A synchronic analysis serves to identify the current state of a dispositive. A particular materialization, linguistically or non-linguistically performed discursive practices can be relevant for various dispositives. For example, the dispositive of ‚traffic‘ encompasses streets, cars, traffic jams, drivers, traffic signs and so on. But ‚traffic‘ is also an economic problem that creates costs and affects business. ‚Traffic‘ is therefore embedded into the economic dispositive. The economic dispositive, in turn, is embedded into the political dispositive. In a society, dispositives overlap and are entangled with each other. These entanglements may be what unifies a society.

A dispositive analysis has to include the following steps:

1. Reconstructing the knowledge that is built into linguistically performed practices (through discourse analysis, as described above). This analysis is the basis for the further steps in a dispositive analysis. It already creates an awareness of important aspects of the dispositive, such as uncharted territories in discourse, significant materializations and so on.
2. Reconstructing the knowledge that is built into non-linguistically performed discursive practices.
3. Identifying the knowledge that is built into materialization.

Reconstructing this knowledge usually results in texts. A dispositive analysis thus translates knowledge about non-linguistically performed practices and materializations into linguistically expressed knowledge.

A dispositive analysis should consider the form in which the examined knowledge occurs: is the knowledge manifestly apparent? Or is it implicit, for example hidden in implicatures? Into what arguments is the knowledge packed? And so on. It should be noted again that the concept of knowledge is here a broad one, including not only cognitions but also emotions.

Since the analysis of discursive elements of a dispositive has already been discussed extensively above, the remainder of this section will focus on reconstructing the knowledge that is built into non-lingualistically performed discursive practices.

Knowledge about non-linguistically performed practices

Non-linguistically performed discursive practices, streichen: i.e actions can be observed and described. The task is to identify the knowledge that enables and accompanies these practices.

For example, the analyst can observe a man who crosses a street and walks into a bakery, where he buys a loaf of bread. The analyst’s task is now to find out what this man knows and wants. The man knows that he has to go to a certain place to be able to buy bread. He knows that for this purpose he has to dress in a certain way (e.g. put on shoes and a coat). He knows that when crossing a road, he has to pay attention to the traffic and observe the traffic rules. Moreover, he knows that the bakery is located in a particular place, or how to look for a bakery. He knows that in a bakery he can buy bread, and that he needs money for that. The simple act of buying bread is thus already based on a considerable amount of knowledge, and this analysis only gives a small hint of it.

The following is a more complex example. Suppose the analyst observes a man who has dug a hole at the side of a road, and is now standing in this hole and manipulating a large pipe. To reconstruct the knowledge built into this practice, the analyst has to share in this knowledge and understand what the man is doing. Suppose that to a large extent she is lacking this knowledge. There are basically three things she can do to gain an understanding of what the man is doing.

Firstly, the researcher can draw on existing texts. For example, she can consult previous research, but also more mundane documents such as practitioners‘ literature, instruction sheets, or field manuals.

Secondly, the analyst can ask the man what he is doing. In ethnographic methodology, this is called an ethnographic interview (see, for example, Spradley, 1979) ((Ethnographic interviews take place in the course of participant observation, i.e., while research subjects engage in the action of interest. Such an interview can be long, or short, when the researcher asks just a quick question and then continues with silent observation and note-taking.)). When the researcher asks the man what he is doing, the man may answer: ‚I am repairing a burst pipe.‘ With this information, she already understands better what he is doing. Next, she may ask him: ‚Why are you doing that?‘ He may answer something like: ‚Because the pipe has burst‘, or ‚That’s my job‘, or ‚I need to earn money somehow‘, and so on. The knowledge built into his activity is thus fairly complex and can be followed up to the economic practice of dependent wage labour.

A large part of knowledge is only available to people in their practices (tacit knowledge), and people cannot easily explicate it in talk. In other words, people will know more than they can tell. As a third option, the researcher can therefore rely on participant observation (see, for example, Agar, 2002; Emerson, Fretz, & Shaw, 1995; Hammersley & Atkinson, 2007; Spradley, 1980) to learn about this implicit knowledge and make it explicit in her research. In the extreme case, the researcher may herself learn to dig holes and mend burst pipes. A fascinating example of such work is Wacquant’s (2004) study of boxing.

Knowledge with regard to materializations

Knowledge in general is not resting within actions or things. It is assembled in the minds of people only. When an analyst observes an object, such as a house, a church or a bicycle, obviously he cannot ask this object about its meaning. However, there are indirect ways of reconstructing knowledge about materializations. Methodological guidelines for doing so can, for example, be found in multimodal discourse analysis (van Leeuwen, 2005) and artefact analysis (Froschauer, 2002; Lueger, 2004; for an example of empirical work that combines multimodal discourse analysis and artefact analysis, see Maier, 2009).

To analyse materializations, the researcher has to rely on his own and his fellow researchers‘ background knowledge. In addition, he should extend this knowledge by drawing on the pertinent literature, and by asking users, producers and other persons who are experts on the activities and materialization in question.

Artefact analysis, as developed by Lueger (2004) and Froschauer (2002), suggests that one of the first steps in analysing a materialization is to deconstruct the materialization by dividing it into its constituent parts and transcribing it into a field protocol. The material object is thus transformed into a text. Here, another problem arises, which incidentally also applies to the field notes and observation protocols produced in the participant observation of non-linguistic practices. The field protocols written by the researcher are not neutral. Like any text, they pursue particular interests, and in the ideal case, this interest should be to answer the research question.

In some cases, the researcher may even be able to draw on previous research that has already discursified the materialization in question as, for example, Caborn (1999, 2006) has done with regard to state architecture in Berlin after the reunification.

It should be emphasized again that the meaning of materializations is not fixed. The knowledge one imputes into a materialization today may be different from the knowledge that it conveyed in the past. ‚Legends‘ might have formed around it, and meanings may have changed. Moreover, a materialization may have different meanings for members of different cultures (as well as cultures understood in the broad sense of the term) and inhabitants of different places. A good case in point is the Aztec crown of feathers, which is exhibited in the Anthropological Museum in Vienna. In pre-Columbian times, it was a ritual headdress, worn by priests or even by the Aztec emperor Montezuma. For the Spanish conquistador Hernán Cortes, it was a treasure which he stole. The Habsburg emperors bought it as an exotic curiosity. In today’s Anthropological Museum, it is an exhibit of scientific value. For today’s descendants of the Aztecs, the crown is a symbol of their cultural identity and stands for the blossoming and ensuing destruction of their culture. They argue that the crown was stolen from them, and that the Museum should return it. For Austrian politicians and diplomats, the crown has thus become a cause for political dissonance with Mexico (for an in-depth analysis of the transformations of exhibits in anthropological museums, see Döring & Hirschauer, 1997).

As the example of the feather crown shows, each of the meanings assigned to a materialization is tightly related to power relations (e.g. whether the crown is a ritual item, a scientific exhibit, or a symbol of collective identity). The object itself does not change, but its meanings change, as people apply to it new kinds of knowledge. In the materialization as such, these power relations are invisible, and the task of the analyst is to bring them out into the open. The analyst can only do this if he considers historical contexts.

As these initial thoughts on methods for analysing dispositives indicate, the task of a dispositive analysis is complex. It encompasses the analysis of knowledge about linguistically and non-linguistically performed discourses. An example of such an analysis is Michel Foucault’s (1979) book Discipline and Punish. Also, Victor Klemperer’s (2001) Diary of the Nazi Years can be read as a dispositive analysis. Both authors provided little explicit information on their methodology. They apply their methodology implicitly, or as Foucault calls it, in the form of bricolage. They analyse discourse, assembled knowledge, consult statistics, deconstruct them, draw conclusions, add their own opinions, and so on.

The thoughts on dispositive analysis presented in this chapter do not provide a recipe or schema. However, they do give some ideas on how to approach dispositive analysis. A central part of dispositive analysis is the discourse analysis of texts. Moreover, dispositive analysis comprises the analysis of non-linguistic practices, for which methods developed in ethnography, such as ethnographic interviews and participant observation, provide important means. A final component is the analysis of materializations, which can draw on methods such as multimodal discourse analysis and artefact analysis. An explicit methodology for combining these approaches has yet to be developed. Such an endeavour can only be achieved by means of concrete research projects that devote space and time to explicit reasoning about methodology. This promotes the development of dispositive analysis, and contributes to bridging the gaps between discourse analysis and other methods of empirical social research.

4 Conclusion

The methodology we have presented here has been continuously developed since the mid-1980s, and has been applied in a wide range of studies ((See for example the project summaries in Margarete Jäger and Siegfried Jäger (2007).)). It is not a rigid formula that can be followed mechanically. It is a flexible approach and a systematic incitement for researchers to develop their own analytic strategies, depending on the research question and type of materials at hand. An article like this one can provide initial insights, but – since every study needs a customized approach – it cannot anticipate the full range of possibilities. This is in accordance with Foucault’s understanding of methodology:

If you want an image, think of a network of scaffolding that functions as a point of relay between a project being concluded and a new one. Thus I don’t construct a general method of definitive value for myself or for others. What I write does not prescribe anything, neither to myself nor to others. At most, its character is instrumental and visionary or dream-like. (Foucault, 1991:29)

However, Foucault by no means implies that „anything goes“. As he emphasizes, he is very much interested in discovering truths, albeit truths that are valid in a certain time and place:

In the course of my works, I utilize methods that are part of the classic repertory: demonstration, proof by means of historical documentation, quoting other texts, referral to authoritative comments, the relationship between ideas and facts, the proposal of explanatory patterns, etc. There’s nothing original in that. From this point of view, whatever I assert in my writing can be verified or refuted as in any other history book. (Foucault, 1991:32f.)

In this spirit, we encourage researchers to develop a thorough theoretical understanding that underlies their methodology and – on this basis – to innovate, adapt, mix and match the methods as it fits their research purpose. The best way to learn critical discourse analysis is to do it.

5 Suggestions for Further Reading

Dreyfus, H. L., & Rabinow, P. (1982). Michel Foucault: Beyond Structuralism and Hermeneutics. Sussex: The Harvester Press.
A classical exposition of Foucault’s work as a whole, judged accurate by Foucault himself.

Foucault, M. (1990). The History of Sexuality.Volume 1: An Introduction, vol. New York: Vintage; Reissue edition.
One of the thinner and more accessible works by Foucault, this gripping little book can be recommended as a starting point for reading Foucault in the original. It is actually more about the general workings of power/knowledge than about the specific issue of sexuality.

Jäger, S. (2012). Kritische Diskursanalse. Eine Einführung (6th, revised edition). Münster: Unrast-Verlag.
A thorough and comprehensive outline of discourse theory and the method of critical discourse analysis developed by Siegfried Jäger streichen: und kollegen (in German language).

Jäger, S., & Zimmermann, J. (Eds.). (2010). Lexikon kritische Diskursanalyse: eine Werkzeugkiste: Unrast-Verlag.
A dictionary of key concepts of critical discourse analysis (in German language).

Wetherell, M., Taylor, S., & Yates, S. J. (2001). Discourse theory and practice: A reader: Sage.
This reader is a good starting point for reading primary literature on discourse theory and discourse analysis. It covers major authors, various strands of theorizing, and key epistemological and methodological issues.

6 References

Agar, M. (2002). The Professional Stranger (3rd unrevised edition ed.). San Diego, CA: Academic Press.

Balanoff, E. K. (2013). A Special, Set-Apart Place No Longer? Deconstructing the Discourse of Meaning and Mission in Nonprofit Newsletters. Administrative Theory & Practice, 35(1), 11-27.

Balke, F. (1998). Was zu denken zwingt. Gilles Deleuze, Felxi Guattari und das Außen der Philosophie. In J. Jurt (Ed.), Zeitgenössische Französische Denker: Eine Bilanz (pp. 187-210). Freiburg i.B.: Rombach Litterae.

Bublitz, H. (1999). Foucaults Archäologie des kulturellen Unbewußten: Zum Wissensarchiv und Wissensbegehren moderner Gesellschaften. Frankfurt am Main: Campus.

Burr, V. (2003). Social Constructionism (2nd edition ed.). Hove: Routledge.

Caborn, J. (1999). Die Presse und die „Hauptstadtdebatte“: Konstrukte der deutschen Einheit. In U. Kreft, H. Uske & S. Jäger (Eds.), Kassensturz: Politische Hypotheken der Berliner Republik (pp. 61-84). Duisurg: DISS.

Caborn, J. (2006). Schleichende Wende: Diskurse von Nation und Erinnerung bei der Konstituierung der Berliner Republik. Münster: Unrast.

Deleuze, G. (1988). Foucault. Minneapolis: University of Minnesota Press.

Döring, H., & Hirschauer, S. (1997). Die Biographie der Dinge: Eine Ethnographie musealer Representation. In S. Hirschauer & K. Amann (Eds.), Die Befremdung der eigenen Kultur (pp. 267-297). Frankfurt am Main: Suhrkamp.

Drews, A., Gerhard, U., & Link, J. (1985). Moderne Kollektivsymbolik: Eine diskurstheoretisch orientierte Einführung mit Auswahlbiographie. Internationales Archiv für Sozialgeschichte der deutschen Literatur (IASL), 1. Sonderheft Forschungsreferate, Tübingen, 256-375.

Dreyfus, H. L., & Rabinow, P. (1982). Michel Foucault: Beyond Structuralism and Hermeneutics. Sussex: The Harvester Press.

Emerson, R. M., Fretz, R. I., & Shaw, L. L. (1995). Writing Ethnographic Fieldnotes. Chicago, ILL et al.: The University of Chicago Press.

Foucault, M. (1972). The Archaeology of Knowledge. New York: Pantheon Books.

Foucault, M. (1979). Discipline and Punish: The Birth of the Prison. Harmondsworth: Penguin Books.

Foucault, M. (1980a). The Confession of the Flesh. In C. Gordon (Ed.), Power/Knowledge: Selected Interviews and Other Writings 1972-1977 by Michel Foucault (pp. 194-228). New York: Pantheon Books.

Foucault, M. (1980b). Truth and Power. In C. Gordon (Ed.), Power/Knowledge: Selected Interviews and Other Writings 1972-1977 by Michel Foucault (pp. 107-133). New York: Pantheon Books.

Foucault, M. (1983). Der Wille zum Wissen: Sexualität und Wahrheit 1. Frankfurt am Main: Suhrkamp.

Foucault, M. (1991). Remarks on Marx: Conversations with Duccio Trombadori. New York: Semiotext(e).

Foucault, M. (1996). What is Critique? In J. Schmidt (Ed.), What is Enlightenment? Eighteenth-Century Answers and Twentieth-Century Questions (pp. 382-398). Berkeley et al.: University of California Press.

Foucault, M. (2001). Fearless Speech. Los Angeles, CA: Semiotext(e).

Foucault, M. (2002). The Archaeology of Knowledge. London: Routledge.

Froschauer, U. (2002). Artefaktanalyse. In S. Kühl & P. Strodtholz (Eds.), Methoden der Organisationsforschung (pp. 361-395). Reinbek: rororo.

Hammersley, M., & Atkinson, P. (2007). Ethnography: Principles in Practice (3rd edition ed.). London et al.: Routledge.

Jäger, M. (1996). Fatale Effekte: Die Kritik am Patriarchat im Einwanderungsdiskurs. Duisburg: Unrast-Verlag.

Jäger, M., & Jäger, S. (2007). Deutungskämpfe: Theorie und Praxis Kritischer Diskursanalyse. Wiesbaden: VS Verlag.

Jäger, S. (2004). Kritische Diskursanalse. Eine Einführung (4th unrevised edition ed.). Münster: UNRAST-Verlag.

Jäger, S. (2012). Kritische Diskursanalse. Eine Einführung (6th, revised edition ed.). Münster: Unrast-Verlag.

Jäger, S., & Maier, F. (2009). Theoretical and methodological aspects of Foucauldian critical discourse analysis and dispositive analysis. Methods of critical discourse analysis, 2, 34-61. ????

Klemperer, V. (2001). I will Bear Witness: A Diary of the Nazi Years. New York: Modern Library.

Klemperer, V. (2006). The Language of the Third Reich: LTI – Lingua Tertii Imperii: A Philologist’s Notebook. London et al.: Continuum.

Laclau, E. (1980). Populist rupture and discourse. Screen Education, 34, 87-93.

Link, J. (1982). Kollektivsymbolik und Mediendiskurse. kulturRRevolution, 1, 6-21.

Link, J. (1983). Was ist und was bringt Diskurstaktik. kulturRRevolution, 2, 60-66.

Link, J. (1992). Die Analyse der symbolischen Komponente realer Ereignisse: Ein Beitrag der Diskurstheorie zur Analyse neorassistischer Äußerungen. In S. Jäger & F. Januschek (Eds.), Der Diskurs des Rassismus. Oldenburg: Osnabrücker Beiträge zur Sprachtheorie 46.

Link, J., & Link-Heer, U. (1990). Diskurs/Interdiskurs und Literaturanalyse. Zeitschrift für Linguistik und Literaturwissenschaft (LiLi), 77, 88-99.

Lueger, M. (2004). Grundlagen qualitativer Feldforschung (2nd edition ed.). Wien: WUV.

Maier, F. (2009). Doing Hair – Doing Age: Perspectives of Emancipated Ageing. In M. Beisheim, B. Gusenbauer, L. Kreil & F. Maier (Eds.), Perspectives of Women’s Age at the Work Place. Seiten? Bern et al.: Peter Lang Europäischer Verlag der Wissenschaften.

Meyer, J. W. (2009). World Society: The Writings of John W. Meyer. Oxford: Oxford University Press.

Meyer, M., Buber, R., & Aghamanoukjan, A. (2013). In Search of Legitimacy: Managerialism and Legitimation in Civil Society Organizations. Voluntas: International Journal of Voluntary and Nonprofit Organizations, 24(1), 167-193.

Spradley, J. P. (1979). The Ethnographic Interview. New York: Holt, Rinehart and Winston.

Spradley, J. P. (1980). Participant Observation. Fort Worth et al.: Harcourt Brace Jovanovich College Publishers.

Teubert, W. (1999). zum Verlust von Pluralität im politisch-gesellschaftlichen Diskurs: Das Beispiel Besitzstände. In U. Kreft, H. Uske & S. Jäger (Eds.), Kassensturz: Politische Hypotheken der Berliner Republik (pp. 29-48). Duisburg: DISS.

van Leeuwen, T. (2005). Introducing Social Semiotics. London and New York: Routledge.

Wacquant, L. (2004). Body & Soul: Notebooks of an Apprentice Boxer. Oxford et al.: Oxford University Press.

Waldenfels, B. (1991). Ordnung in Diskursen. In F. Ewald & B. Waldenfels (Eds.), Spiele der Wahrheit: Michel Foucaults Denken (pp. 277-297). Frankfurt am Main: Suhrkamp.