Researching Discourse and Behaviour as Elements of Law in Action Bettina Lange


Researching Discourse and Behaviour as Elements of Law in Action


This chapter critically examines and advocates the combination of a discourse analytic and a qualitative approach to socio-legal research.1 It aims to contribute to the literature on socio-legal research methods by telling the behind-the-scenes story of how a research method emerged through the process of engaging in real-life, messy and mundane research activities. It suggests that practical research design can not be separated from an engagement with fundamental social science questions about the relationship between ideas and practices, and postmodernist and modernist perspectives on these. The chapter also highlights the importance of methodology for substantive socio-legal debates. It suggests that combining qualitative and discourse analysis may generate new insights into European Union (EU) law in action.

The first section of this chapter briefly introduces the research project, while the second section discusses reasons for the combination of a discourse analytical and qualitative approach in this specific study. Section three further explains how some approaches to discourse analysis and qualitative research can be considered as distinct. The fourth and main section of this chapter shows—also on the basis of a data extract and introductory analysis—how in practice a discourse analytical and a qualitative approach were combined.


What can be understood as ‘law’ and how normativity is constructed have been key concerns for socio-legal researchers2. These questions need further exploration in the context of EU law. Theories of EU integration have emphasised the role of law in the integration process, but have often relied on formal, autonomous, quasi-state law conceptions of normativity.3 For instance, formal EU law has been considered as moving integration forward when political initiatives among Member States have been waning.4 Hence, the research project explores the meaning of EU law in action and its contribution to integration processes. The study analyses the key legal obligation under the EU Directive on Integrated Pollution Prevention and Control.5 Under this Directive Member State regulatory authorities have powers to require operators of mainly industrial installations6 to employ the ‘best available techniques’ (BAT), in order to prevent and reduce their emissions to all three environmental media, air, water and land.7 Article 2. Nr. 11 of the IPPC Directive provides a rudimentary definition of ‘best available techniques’:

BAT shall mean the most effective and advanced stage in the development of activities and their methods of operation which indicate the practical suitability of particular techniques for providing in principle the basis for emission limit values designed to prevent and, where that is not practicable, generally to reduce emissions and the impact on the environment as a whole.8

Annex IV to the Directive spells out further criteria for the determination of BAT. Some of these address environmental considerations, such as waste minimisation, energy efficiency of regulated processes and the principles of prevention of damage to the environment and precaution. Others refer to the ‘costs and benefits of a measure’.9 Finally, Member State regulatory authorities must take into account, but are not bound by, BAT Reference Documents (BREFs) when determining BAT for a specific plant.10 These BREFs are the result of an EU-wide information exchange between representatives of industry, Member State regulatory authorities and—upon invitation of the EU Commission—environmental non-governmental organisations.11 Implementing the Directive requires determining in more detail what constitutes BAT. At the EU level this occurs mainly during the BREF writing process. At the national and local level this happens during the drafting of national implementing legislation and when licences for specific plants are issued. I refer to the various BAT options which different social actors advance as ‘BAT law in action’ and to the final legally authoritative choice of one particular BAT definition in implementing legislation or licences as ‘state law BAT’.12 But how can ‘BAT law in action’ and its interaction with state law be researched?


Three criteria informed the choice of methodology for the IPPC project. The methods had to be able to answer the specific research question asked and they had to fit the characteristics of the BAT determination process and help to manage restrictions on access to data. The combination of a qualitative and discourse analytical approach fulfilled all three criteria. I use the term discourse analysis here to refer to an examination of discourse in both its linguistic and Foucauldian dimension. A linguistic notion of discourse refers to ‘informal and formal, including institutionalized, spoken interaction and written texts’.13 Foucault’s concept of discourse, in contrast, is more comprehensive and complex. First, it comprises speech acts and any system of signs, not just language. These are taken to represent knowledge about ‘a topic at a particular historical moment’.14 Second, Foucault’s concept of discourse covers not just statements but also the regulated practices which account for statements.15 Hence, groups of statements on a particular topic, such as the discourse on what constitutes ‘best available techniques’ is captured, as well as the—sometimes hidden—rules and structures which produce this discourse.16 Discourses can generate exclusion when certain statements are kept in circulation while others are marginalised. Hence, discourse is an essential aspect of relations of power. In fact, discourse can be an instrument and effect of power as well as a starting point for resistance.17 According to Foucault, expert status is an important resource for the production of discourse, since its successful circulation depends on whether statements will be judged as ‘true’ rather than ‘false’. Discourse relies on the idea that there will be limitations on who will be considered to speak authoritatively.18 Hence, discourse analysis pays attention to the discursive resources which social actors use, such as ‘category systems, narrative characters and interpretative repertoires’.19 It also examines the distribution, exchange and control of discourse.20

In contrast to this, qualitative research analyses a whole range of social interactions, including non-verbal behaviour. Participant observation and unstructured interviewing are often considered as its key data collection techniques.21 Qualitative research assumes that actors create social life through a range of interpretative practices. Definitions of situations are key to how people act and can thus produce real consequences.22 Qualitative researchers attempt to understand these through entering the social actors’ behavioural world and becoming familiar with its perspectives.23

Combining a discourse analytical and a qualitative approach allows us to shift the emphasis from behavioural to discursive aspects of the ‘law in action’ and thus to depart from the classical sociology of law literature and contemporary studies influenced by it. For instance, Ehrlich defines the living law through reference to patterns of behaviour from which a rule can be deduced.24 These can be empirically identified through observations of further behaviour, such as the experience of informal social sanctions by those who do not comply with the living law.25 Similarly, Pound’s concept of the law in action, though different from Ehrlich’s, also considers social actors’, in particular law makers’ and law enforcers’, behaviour, as key.26 It is an evaluation of their activities—in the light of the normative benchmark of state law—that helps to identify the law in action that they generate. Hence, some contemporary studies of the law in action have focused on behaviour by asking how legal actors make discretionary decisions, how they receive state law in regulated and regulatory organisations, how they avoid law and what alternative social norms they create.27

A shift to discursive aspects of normativity is not new. Conversations, for instance, have been perceived as helping to solve problems raised by regulating through legal rules.28 Conversations, however, are seen here as separate from rule formation and are analysed after formal legal rules have been defined and established.29 In contrast to this, the discourse analytical approach in the IPPC project aims to analyse how law in action feeds at an earlier stage into the creation of state law. It works with a broad definition of discourse, including text, not just conversations. By combining a discourse analysis and qualitative approach the project aims at a full explanation of BAT discourse. This should also address how discourse is shaped by non-discursive aspects of the social world, which can be accessed through a qualitative approach. The combined approach also allows us to ground BAT discourse in the social process of accounting for BAT. This avoids characterising BAT—in an abstract and reified manner—as a discursively constructed norm concept. Participants in the BAT definition process do not invoke a BAT concept. Instead they are engaged in a process of describing BAT in which the meaning of BAT appears to be elusive.

Research methods should also fit the actual characteristics of the social process being studied. Two impressions from the initial fieldwork phase seemed to suggest that combining a qualitative and discourse analytical approach would be fruitful. The BAT determination process gave rise to alliances which crossed traditional interest group boundaries. Sometimes, the BAT discourse seemed to suggest that individual social actors influenced significantly BAT determinations. Hence, this process could not be fully captured through reference to pre-given interests and lobbying by distinct social groups, such as ‘regulators’, ‘industry’ and ‘environmental NGOs’. Furthermore there was an opaque and labyrinthine system of consultations over a period of time which provided a number of formal and informal opportunities for various actors to express what they considered as BAT. Hence, the exercise of political power in the BAT determination process seemed complex and more fully captured through Foucault’s notion of the microphysics of power. One of the key aspects of this concept of power is that it departs from the idea that power can be possessed, for example by the state or various social actors. Instead power is best understood as a strategy and its effects arise from ‘small-scale manoeuvres, tactics, techniques and functionings’.30 Not conscious intentions, the interests of groups or individuals, but detailed practices are fore grounded in this analysis of power.31 In fact categories, such as ‘individual’, ‘group’ or ‘social actor’ are perceived as effects of power. Hence, Foucault’s concept of power helps to move away from ideas of power as institution or social structure. It even considers various points of resistance as an integral part of power. An understanding of the microphysics of power, however, is not restricted to a ground level perspective on power. It can also assist analysis of macro-level manifestations of power by rendering visible how small-scale tactics of power can be appropriated for its more large-scale exercise.32

While discourse is clearly central to the construction of BAT normativity, initial fieldwork also suggested that social life—beyond the words— seemed to matter for understanding BAT law in action. Surprisingly, some important sources of BAT discourse, such as written records, contained few references to the costs and benefits of technologies and their cumulated effect upon all three environmental media, land, water and air. According to the text of the IPPC Directive, these were, however, key criteria for defining BAT. Hence, it seemed important to consider a broader conception of the social world—accessible through qualitative methods—in order to explain this silence of the BAT discourse.

Thirdly, the idea of combining qualitative and discourse analytical perspectives also developed in response to access opportunities and problems.33 A focus on selected elements of BAT discourse helped to compensate for restrictions on traditional qualitative observational data about oral BAT negotiations in technical working group meetings (TWG). In these meetings delegates from EU Member State regulators, industry and environmental NGOs would debate what should be considered as BAT for a specific industrial sector. I was refused access to these meetings, but was provided with an official audiotape recording of a past TWG meeting which, in transcribed form, provided a rich source for discourse analysis. So far I have argued that the combination of a discourse analytical and qualitative approach is particularly suited to answering the research questions of the IPPC project. It needs to be further explained, however, in what way I consider these two approaches as different.


Conventionally the research methods literature does not clearly distinguish between discourse analysis and qualitative research. The important role of language in social life is recognised in both approaches. Some textbooks even consider two approaches towards discourse analysis—conversation analysis and the ethnography of speaking—as a form of qualitative research.34 Discourse analysis, however, is not necessarily a form of qualitative research. One variant of it—content analysis—relies on quantitative methods.35 Furthermore some discourse analytical and qualitative methods work with different concepts of agency, social action and the relationship between discursive and non-discursive elements in the social world.

Especially qualitative approaches informed by hermeneutics perceive social actors as ‘conscious, individual, meaning-giving subjects’.36 They are independent language users and language facilitates their agency. Social actors can be ‘speakers’, ‘hearers’ and unaddressed third parties— ‘bystanders’—in a conversational encounter.37 In contrast to this, discourse analysis perceives social actors as constructed through discourse.38 Subjects can not be outside discourse and they have little control over it.39 Descriptions of social life can ‘become established as solid, real and independent of the speaker’.40 Hence ‘subjects’ in discourse analysis personify the particular forms of knowledge which the discourse produces: ‘the human voice is conceived merely as another means for registering differences’.41 These two different conceptions of agency have implications for the two perspectives’ respective concepts of social action.

From a discourse analytical perspective discourse itself accomplishes social action.42 It can ‘order, request, persuade, accuse, take sides and disclaim responsibility’.43 When discourse generates social action it acquires a material quality and no longer exists just in a realm of ideas. From a qualitative perspective language fulfils merely a representational function. Texts can ‘open the door to an understanding of the social world’, but can not be equated with it.44 Whether language is seen as constituting or as merely representing the social world also influences these two approaches’ different views of the relationship between discursive and non-discursive aspects of the social world.

Qualitative approaches see discursive and non-discursive elements as linked through a process of interpretation. From a hermeneutical perspective, discourse can be understood and explained through reference to a ‘horizon of intelligibility’, a field of shared social practices.45 Social actors, including researchers, look for what is ‘underneath’ the use of language, in order to understand how meaning in the social world is achieved. Hence, the non-discursive world becomes an important resource for understanding discourse.46 For qualitative researchers the discursive and non-discursive world can also be linked through a process of causation. Non-discursive factors, such as pregiven, separate interests are sometimes considered as explanations for linguistic phenomena.47

In contrast to this, Foucault does not intend to explain discursive elements in terms of non-discursive ones.48 A Foucauldian approach describes the surface details of a discourse.49 It does not assume that discourse could only be rendered meaningful through reference to exterior shared social practices.50 There is no ‘deep truth behind experience’ and hence interpretation is considered as an arbitrary and groundless process.51 Discourse becomes decontextualised.52 Non-discursive social practices, however, still inform discourse: ‘[…] what gets said depends on something other than itself, discourse, so to speak, dictates the terms of this dependence’.53

Hence, the non-discursive sphere can provide the conditions of existence for a discourse and form the objects of discourse.54 But it is discourse, rather than non-discursive elements, which generate real, material effects in the social world. In fact, Foucault suggests that there is a circular and complex relationship between the non-discursive and discursive world. He distinguishes between primary and secondary relations. Primary relations occur between institutions, techniques, social forms and other elements which make up the non-discursive world. Secondary relations describe how ‘practising subjects reflectively define their own behaviour’.55 Foucault calls relationships between primary and secondary relations discursive practices. They determine ‘who has the right to make statements, from what site statements emanate, and what position the subject of discourse occupies’.56 Foucault also suggests that relationships between a discursive and a non-discursive sphere vary with the organisation of a particular discourse.57 Hence, the archaeological method of discourse analysis searches for the way in which discourse is ‘articulated’ with the non-discursive world.58 Given these different perspectives which underpin discourse and qualitative research the question arises how the two approaches, and in particular their criteria for what constitutes ‘good’ data and analysis procedures, can be combined in practice.


1. Constructing Criteria for What Constitutes ‘Good’ Data

Validity is a key criterion for evaluating any research data. Do the data shed light on the research question accurately? Sample size can influence the validity of the data and for qualitative researchers sample size matters more than for discourse analysts. From a qualitative perspective data refer to a separate, external social realm. It therefore needs to be considered how much and what type of data are needed in order to generate a valid description of this world. In contrast to this, discourse analysis focuses on a detailed, in-depth analysis of the construction of the discourse itself and hence only a small sample can be entirely sufficient.