DP and other forms of discourse analysis
The term ’discourse analysis’ (DA) is used to cover a range of approaches across the social sciences and the different ways in which it is used can be confusing and overwhelming. It is not always easy to know which form of DA you are reading about and which one is suitable for your own project or data analysis. For example, some of the terms that you might find when searching for DA are: computer-mediated discourse analysis, conversation analysis, corpus-based discourse analysis, critical discourse analysis, critical discursive psychology, dialogical analysis, discursive psychology, ethnography of communication, feminist post-structuralist discourse analysis, Foucauldian discourse analysis, functional grammar, linguistics, pragmatics, rhetorical analysis, sociolinguistics and textual analysis. No wonder it is a confusing field. The aim of this chapter, therefore, is to provide a brief and simple introduction to five of the main forms of DA that are commonly used in research across the social sciences: conversation analysis (CA), critical discourse analysis (CDA), critical discursive psychology (CDP), discursive psychology (DP) and Foucauldian discourse analysis (FDA). In addition to DP — which is the focus of this book — the other four have been chosen because they are most closely related to, and sometimes confused with, DP. The chapter will begin with a brief discussion of how we might approach this field before considering each of these five forms of DA in terms of what makes them theoretically and analytically distinct.
Seeing life through a lens: the world of discourse analysis
Discourse analysis, in all its forms, seeks to understand the role of discourse in the construction of our social world. It works a little like a video camera: a mechanism to examine discursive practices and social interaction. We might, then, be documentary makers, eager to record some aspect of the social world and thus render it observable and analysable. Using this analogy, we can consider each form of discourse analysis to be like a different camera lens that enables us to take a broader or narrower view of the setting in front of us. For example, we might use a fish-eye lens (panoramic, ultra-wide angle: like critical discourse analysis and Foucauldian discourse analysis), a pancake lens (moderately wide-angle: like critical discursive psychology), or a zoom lens (narrow focus: like discursive psychology and conversation analysis), depending on what features or discourses we are interested in. Each lens renders visible a different perspective, whether that is a panoramic view to provide a broad overview of a scene without focusing on one specific area in particular, or a zoomed-in image, capturing the detail of one aspect, but, potentially, missing the bigger picture. The camera lens analogy also means that we cannot use more than one form of DA at a time. We must make a choice as to what we want to focus on and then choose the appropriate form of DA to perform that analysis. There are also problems with each lens; they cannot alone perform every function, and inevitably we are accountable for the choices we make. Table 2.1 provides an overview of some of the differences between these five forms of DA.
Before we begin this journey into DA territory, however, three caveats are in order. First, although this book is focused on discursive psychology, it is important to be aware of what questions it cannot answer and how it relates to other forms of DA. There is no hierarchy in discourse analysis; each has its own advantages and challenges. While there may be tensions between the different forms (and, indeed, between researchers themselves), these tensions should be used to stimulate academic debate and to enable each form of DA to evolve and adapt to new research insights. No form of DA is perfect, nor should it be.
Second, it will become apparent that each of these five forms of DA has its own subtle variations. Each form is a vast world in itself, and it is hard to do justice here to the range and depth of research carried out by researchers in each of these fields. Since the early 2000s, for example, the differences between the forms of DA have become more emphasised, as each strand developed empirically and theoretically with new research and writing. In that sense, you may come across earlier writings (i.e., from the late 1980s or 1990s) that position these forms of DA slightly differently. So the distinction between the five forms of DA provided here is historically bound and these distinctions will shift again in the future.
Finally, and true to a social constructionist book, it should also be noted that the distinction between these five forms of DA — and my description of each — is itself a construction. Those working in discourse research may well find some descriptions inadequate or misleading. Some researchers have argued that we should not try to distinguish between different forms of discourse analysis, and that this might be divisive between researchers and be unhelpful analytically (see Box 2.1). Others argue that conversation analysis is not a form of discourse analysis, although I include it in my list here. There are also likely to be a range of research areas that I have overlooked or neglected to cover. It is, as they say, impossible to please everyone. What I hope to have done, however, is to provide a brief and illustrative introduction to the field of DA and to show how DP, specifically, can be distinguished from other forms of DA.
Box 2.1: Bottom-up or top-down: do we need to choose one or the other?
There are other ways of representing forms of DA in addition to the camera lens analogy used in this chapter, such as a distinction between ’micro’ and ’macro’ DA, and bottom-up (data-driven) or top-down (theory-driven) analyses. The first (micro/bottom-up) typically defines those forms of DA which are focused on the turn-by-turn organisation of talk and which situate the analytical context within the conversation itself (i.e., CA and DP). These might also be referred to as ’fine-grained’ approaches to discourse analysis. The second (macro/top-down) tends to be associated with forms of DA that engage with ideological aspects of discourses, that define discourse as more than talking or writing, including issues such as subjectivity (ways of being in the world) and the socio-historical context of discourse (i.e., FDA and CDA). Somewhere in the middle — combining the micro and macro elements — is CDP. These ways of categorising DA may help us to recognise and distinguish between different approaches, but they are also limiting in the way that they put forms of DA into boxes with apparently clear boundaries. They can also be divisive, setting researchers against each other in a bid to defend their own territory. There have been persuasive arguments, therefore, for the use of synthesised or multi-level DA, which draws across these classifications and aims to engage in a theoretical and analytical dialogue. While this book focuses on one form of DA (discursive psychology) and as such attempts to demarcate its theoretical and analytical principles, it is important to remember that this is just one approach, one way of analysing the discursive and social world. For examples of papers arguing for, or using, synthesised forms of DA, see Korobov (2001), Riley et al. (2010), and Wetherell (1998, 2007).
Comparing the five forms: theoretical and analytical differences in discourse analysis
The following five sections of this chapter will discuss each of these forms of DA in the following order: conversation analysis, discursive psychology, critical discursive psychology, Foucauldian discourse analysis and critical discourse analysis. The order proceeds from the narrow zoom-lens to the wider panoramic approaches. Each section will cover: (1) theory, (2) history, (3) how we can use it, (4) how it compares with DP (note that the DP section will obviously miss this out), (5) research questions, (6) transcripts and (7) analytical tools. To exemplify the differences for the last three of these points, each has been applied to a single piece of transcribed data, which is detailed below. This is not intended to provide a detailed analysis of the data, nor a methods-guide to using these forms of DA. The descriptions provided here are necessarily limited and are used to highlight the comparisons with DP.
The piece of transcribed data that will be used for this section is taken from a video-recorded family meal, involving a married couple (Bob, aged 54 years, and Linda, aged 55 years), their 26-year-old daughter (Lesley), and Bob’s mother, Edith (aged 84 years); see Figure 2.1 below for a still image from the video. Pseudonyms have been used here to protect the individuals’ real names. The four-minute video can be accessed directly on YouTube (www.youtube.com/watch?v=OtKaXw6WqYM).
The full length of the meal was just over one hour long, and it was recorded in early 2015 in Bob and Linda’s home in the west of Scotland. Full written permission was granted to use the anonymised transcript and still images from this video for this book, and for the section of video to be published on the internet for teaching and learning purposes (see Chapter 8 for more discussion on using images, videos and transcripts in written work). The family talked about a range of issues during the mealtime, but the section that we will focus on here involves the family members talking about the steak that they are eating, as well as shopping for, and eating, different kinds of fish.
Conversation analysis (CA) is one of the zoom-lens approaches discussed in this book, being the most detailed and narrow of the five forms of DA in its focus on discourse. CA argues that talk is primarily about actions; that when we talk we are accomplishing different activities, such as complaining, giving advice or accounting for our behaviour. It focuses on the structure and organisation of talk and on how actions are achieved through the careful arrangement of talk, gesture, eye gaze and objects. CA, then, is primarily used to analyse talk-in-interaction rather than written texts, though recently it has been applied to asynchronous and synchronous interaction in online environments. This form of DA is the one that captures the most detail in transcripts because it is argued that all aspects of talk (even our audible breaths, hesitations, ’mms’ and ’ohs’) are organised and consequential for interaction. For CA, then, the aim is to understand the mechanics of interaction, such as how we start a conversation, how we make assessments, and how refusals or offers are made. By addressing these ’building blocks’ of interaction, CA can then be used to understand how social order and social norms — regular and normative ways of behaving — are accomplished by people in everyday social interaction.
CA has its roots in ethnomethodology and was developed by Harvey Sacks, Emmanuel Schegloff and Gail Jefferson in the 1960s and 1970s. Sacks was perhaps one of the few academics whose lectures have been turned into a bestselling book (Sacks, 1992), and Jefferson developed the transcription system used by CA and DP (see Chapter 5). Sacks and Schegloff were students of Erving Goffman, and later also worked with Harold Garfinkel, the latter two being founding figures in the world of ethnomethodology (see Chapter 1). One of the pioneering aspects of Garfinkel’s and Goffman’s work was to demonstrate the taken-for-granted practices that make up everyday common sense. For instance, Garfinkel’s (1967) breaching experiments demonstrated that by small changes in everyday social practices we can expose the norms and rules that underpin social structures. Just try queue-jumping in the UK, for instance, to see what happens when norms are transgressed. CA emerged, then, from an intellectual environment that focused on mundane, everyday interaction and highlighted the normative aspects of our behaviour. Alongside the technological development of the tape-recorder at the time, these influences provided the stimulus for Sacks to develop an approach that would enable him to analyse people’s sense-making practices in careful detail: he could tape-record conversations and play them back repeatedly (see Silverman, 1998, for a very accessible book on Sacks). Since the early work of CA, there has been extensive debate about interpretations of Sacks’ work, with some areas of CA highlighting the importance of categories and identity work in the form of Membership Categorisation Analysis (Fitzgerald & Housley, 2015; Stokoe, 2012), others advocating ’applied’ CA (Antaki, 2011; Richards & Seedhouse, 2004) and the inclusion of more linguistically-orientated work such as that on prosody (Szczepek Reed, 2010).
How we can use CA
We can use CA when we are interested in how particular activities are accomplished in interaction, and how talk, eye gaze, gestures and objects are co-ordinated in order to carry out these actions. This is sometimes referred to as multimodality: the analysis of different aspects of interaction alongside speech (see later how this compares to multimodal critical discourse analysis). CA can be used within everyday or institutional interaction, such as medical settings (Heritage & Maynard, 2006) or in the workplace (Drew & Heritage, 1992). Some of the areas of research that have been developed using CA are issues around embodiment and gesture (e.g., Goodwin, 2000; Heath et al., 2010; Streeck et al., 2011), interaction with objects (Haddington et al., 2014; Nevile et al., 2014), communication issues in training (Stokoe, 2014) or where individuals have specific communication difficulties (Finlay et al., 2008; Goodwin, 1995; Wilkinson, 2015).
How it differs from DP
While there are extensive overlaps between CA and DP (see Chapter 1), there are also two important differences. First, whereas CA is primarily focused on the social organisation of talk and of how people make sense of each other in interaction, DP has an explicitly constructionist focus. That is, DP is concerned with the way in which categories are produced and performed in discourse; with the versions of reality that are invoked and made available for psychological and social actions. So while CA might examine how talk is socially organised, DP would examine how this particular version of talk is socially organised. A second difference is DP’s anti-cognitivist stance, in which there is an explicit rejection of the theory that cognitions can be ’revealed’ in talk and that the organisation of discursive practices can be reduced to intra-individual states of mind. CA is less directly concerned with cognition and, while most CA work does not make any assumptions about cognitive structures underlying talk, there is some work that does imply particular cognitive underpinnings of interaction. See Wooffitt (2005) for a detailed comparison of CA and DP.
The main aims of CA are to examine social actions and the sequential and organisational features of talk that accomplish these actions. As such, working with video-recorded family mealtimes (as an example of naturally occurring interaction) is familiar territory for CA because it enables researchers to examine a range of issues, from when food assessments occur during a family meal (Mondada, 2009) to how parents use directives (such as ’eat that up!’) towards children at mealtimes (Kent, 2011) or when children report bodily expressions at mealtimes (Jenkins, 2015). For CA, the food and eating practices during the meal do not need to be the focus of analysis, however; the interaction during the meal in itself provides a rich source of data of everyday (i.e., mundane, rather than institutional) talk-in-interaction. So we could examine how assessments (of food or other objects) are produced in the interaction and how these are structured, what follows them, and how, in making an assessment, speakers are also demonstrating an entitlement to assess the object (Pomerantz, 1984). We might also examine how membership of particular categories is used to invoke identities during a meal, such as those that distinguish between family member statuses (Butler & Fitzgerald, 2010). In the transcript opposite, for example, we could focus on how and when family members draw on category entitlements, and how these might be challenged or resisted by others, to make decisions about what food will be eaten in the family home.
It was noted earlier that CA provides us with a tool to examine the messiness of ordinary interaction, with all the sighs, hesitations, in-breaths and pauses that litter our everyday talk. As such, a transcript produced for CA would need to include as much detail as possible, because it is argued that each and every aspect of interaction is organisationally important. Line numbers are used to help identify the sequentiality of talk, clearly showing what happens in the order in which it occurs in the interaction. Speakers (and they are typically referred to as speakers in CA, rather than participants) are often denoted by a pseudonym; this can be as a first name (e.g., ’Bob’) or title (e.g., ’Dr Larsson’), depending on whether the interaction is mundane or institutional. The transcript opposite shows 21 seconds of talk from this clip, marking out as many details in the delivery of the talk as possible, including rising or falling pitch, loudness, extended or emphasised speech, and length of pauses between speakers. For example, CA distinguishes between gaps and pauses. A gap is a silence that occurs between ’turn completion units’ (TCUs: a point at which the speaker completes a turn at talk) and these are placed on a new line. A pause is a silence within a TCU and these are placed within the same line as the talk (Hepburn & Bolden, 2013). There are also ways to transcribe eye gaze and gesture that can capture additional and important aspects of interaction and include these in the written transcript. We can also make greater use of ’screen grabs’ or single frames from a video, or cartoon sketches of gestures to highlight particular visual aspects of the interaction (see Streeck et al., 2014). The transcript below presents a form that is consistent with CA research, though does not include eye gaze or visual detail.
Extract 2.1: Example CA transcript (see Box 5.4 for a transcription key)
CA typically works by identifying a particular phenomenon — a social action of some kind — in the data corpus, and then assembling a collection of instances that contain this phenomenon. This stage of building up a data set is often iterative; sometimes in those early stages it is not always obvious what it is that you are searching for (e.g., at what point does a raised hand become a pointed finger or a move to scratch one’s nose?). We noted above that one of the research questions might be to focus on the use of assessments during the meal. One of the analytic tools of CA is to examine sequence organisation and turn-taking; to identify the turn that precedes or initiates the action and then to identify what action follows it. We can see how Bob’s assessment ’I don’t like steak’ (line 3) is followed immediately by Lesley’s ’no don’t do either’ (line 5) as a second assessment. This has already been found in previous research to be a common pattern where an assessment is typically followed by another from a different speaker (Pomerantz, 1984). While Bob continues with further details about his assessment (lines 6—7), Lesley then offers another assessment (’I like fish’, line 8) which then serves to shift the topic to a different food, and to potentially making a complaint about not having enough fish. So we would begin CA by identifying the different actions (first and second assessments, making a complaint, and so on) in this particular sequence, and unpacking those further to examine how they actually work. For instance, Lesley’s ’I like fish’ is used as a preface to the complaint; it works to make relevant the complaint at just this point in the conversation (given that this is a considerable jump from the discussion about steak) as well as Lesley’s entitlement to make the complaint as a family member (note the shift from ’I’ to ’we’) within this turn. We could also then add in more details about multimodal features of the interaction — how eye gaze and facial expressions are co-ordinated with the assessments, and the footing shifts from ’I’ to ’we’ to ’you’ — alongside the business of eating food at the same time.
Box 2.2: Exploring CA further
One book chapter
Ten Have, P. (2007). Doing conversation analysis: A practical guide. London: Sage. Chapter 1 for an introduction to CA.
One journal article
Raymond, G. & Heritage, J. (2006). The epistemics of social relations: Owning grandchildren. Language in Society, 35(5), 677—705.
Discursive psychology (DP) is the second zoom-lens approach to DA. As noted in Chapter 1, and elsewhere throughout this book, DP is focused on the construction of psychological issues within discursive practices, and the consequences of these constructions for both theory and practice in psychology. For instance, DP might examine how emotion terms are used not as evidence of physiological or cognitive states, but in particular interactional settings to undermine or bolster accounts of people’s activities (Edwards, 1999). Rather than treating psychology as an individual issue, therefore, DP treats psychology as first and foremost an interactional concern. Like CA, DP also treats social interaction as the primary focus for research, and there is a concern with both the structure and the content of interaction (how, when and what is said). So DP will examine how people and their identities, responsibilities and behaviours are produced in particular ways in talk, and the implications of these constructions for that specific context. A classic focus of DP has been to consider how attitudes are formulated in talk: their format and rhetorical impact, rather than their status as a cognitive concept (Potter, 1998; Wiggins, 2015). So DP treats psychological concepts (attitudes, emotions, identities, for instance) as public and practical, rather than private, concerns.
DP emerged from within psychology in the 1980s and was inspired by research and theory within CA, ethnomethodology and post-structuralism, the work of Wittgenstein (1953) and Austin (1962) and the sociology of scientific knowledge (e.g., Gilbert & Mulkay, 1984). It was developed partly out of a critique of cognitivist approaches to psychology which were dominant at the time, in terms of how behaviour was being reduced to cognitive states, how people’s psychologies and subjectivities were regulated by mainstream psychology (Henriques et al., 1984) and by the ’turn to discourse’ more broadly in the social sciences (Parker, 1989). The form of DP advocated in this book is that developed by Derek Edwards and Jonathan Potter in the late 1980s and 1990s. As with the other forms of discourse analysis, DP also has its tensions and variations, such as in the work of Rom Harré (Harré & Gillett, 1994) and Ian Parker (1992). Harré’s work, for instance, advocates what he calls a ’new cognitivism’. Rather than eschewing cognitive theory, this work locates discursive psychology alongside processes of perception and intention (e.g., Davies & Harré, 1990) and deals with illustrated examples of discourse rather than empirical data. Parker’s work defies simple categorisation in one area or another, for his theories draw on psychoanalysis, the work of Lacan, and some aspects of political theory. While he has used the term ’discursive psychology’, he has also characterised his work as critical discursive psychology and one might even see his work fitting in with Foucauldian DA. These different strands of DP research often revolve around the status of cognition and subjectivity in discursive practices, and highlight the contested nature of these concepts.
How we can use DP
We can use DP when we want to focus on psychological issues in terms of when and how these emerge as part of everyday social interaction. So we stay focused on the detail of talk and text (how words are said, the organisation and sequential nature of talk) as well as how psychological concepts are produced and made relevant in particular ways. DP has been used to examine classic psychological topics such as attitudes and assessments (Potter, 1998; Puchta & Potter, 2002), cognitions (Antaki, 2006; Auburn, 2005; MacMartin & LeBaron, 2007), emotions (Edwards, 1999), identity (Eriksson & Aronsson, 2005; Lamerichs & te Molder, 2003), memory (Edwards & Potter, 1992), and racism and prejudice (Augoustinos & Every, 2007; Tileagǎ, 2005). Due to its inclusion of a broader range of discursive practices, DP has also been applied to interaction in focus groups (Goodman & Burke, 2010; Puchta & Potter, 2004), online interaction (Burke & Goodman, 2012; Horne & Wiggins, 2009; Lamerichs & te Molder, 2003), media texts (MacMillan & Edwards, 1999), interviews (Locke, 2004), and everyday face-to-face interaction (Wiggins, 2013). Note that the use of different kinds of ’researcher generated’ data is not without its criticisms (see Box 4.2 for further discussion on this matter).
The main aim of DP is to examine how psychological constructs are enacted and made relevant in interaction and the implications of these for social practices. Like CA, DP is predominantly used with video or audio recordings of everyday or institutional life. Using the meal example here, DP could be used to examine how the constructs of food preference (e.g., liking or not liking foods) or disgust (e.g., the facial expressions alongside ’chemically incarcerated’) are produced in particular ways and for particular purposes in the interaction (Wiggins, 2013, 2014). So the topic of eating and food practices within the mealtime could be a particular focus. Alternatively, we could examine other psychological issues, such as those around identity and responsibility: how choices are made (in choosing food), who has entitlement to say what gets eaten, and whose responsibility it is to ensure that those choices are carried out. These issues could be tackled from a DP perspective, specifically in relation to accountability and fact construction: how are people held accountable for food choices or food preferences? How do speakers orientate to food preferences as being ’real’ and relevant at particular moments in the mealtime?
As the form of discourse analysis that is most closely aligned with conversation analysis, the type of transcript that is typically used in DP research can look very similar, if not identical. What may differ is the way in which words are written. In CA, words are typically transcribed phonetically (as they sound) rather than using standard spelling. By contrast, DP transcripts use either phonetic or standard spelling. Line numbers are used to highlight the focus on sequence and organisation, and to enable us to readily point to specific parts of the data in the written analysis. The participants (sometimes referred to as speakers, as in CA) are denoted by pseudonyms, and again there is variation in terms of whether informal or formal names are used. One of the issues around labelling participants even at the stage of transcription is that if we refer to someone as ’Mum’ rather than ’Linda’ then the implication might be that all their talk should be interpreted in terms of their identity as the mother in this context. Given DP’s social constructionist assumptions, this can be problematic, so in many cases, DP transcripts use a pseudonym that does not infer a particular role. Like CA transcripts, DP transcripts also aim to capture as much detail about the delivery of the talk as possible, though details relating to eye gaze or gesture may be characterised by labels rather than with particular transcription symbols.
Extract 2.2: Example DP transcript
As will be detailed in Chapter 6, DP typically focuses on how and when psychological concepts are produced in interaction, and the consequences of these constructions for social practices. DP draws on many of the tools of CA, including focusing on turn-taking and sequential organisation, but rather than starting with social actions, DP focuses first on a specific psychological issue or the production of categories in talk. We noted earlier that food preference might be a relevant topic for this mealtime example, so we could focus on lines 3, 5 and 8, where Bob and Lesley both make claims about their food preferences. Here, we could examine how they are formulated, as ’I like/I don’t like’, rather than a more extreme form (e.g., I love, I hate) or a softened version (e.g., I prefer, I’m not so keen on) of similar assessments. So we could examine the specific way in which they have been stated, and also how this ’locates’ the preference in the individual. For instance, if we look at lines 9 to 17, we see how Bob is describing the fish in a particular way; this constructs the food as having particular characteristics (’incarcerated’, ’doesn’t look like a fish’) that are separate from, and apparently independent of, Bob’s own food preferences or taste experiences. We could then continue to examine how assessments of food are enacted differently throughout the mealtime — as individual preference, as a claim about the food as an object — and how these not only have consequences for how we understand food preference (that if an adult makes a claim about an individual food preference, that it cannot be challenged, for instance), but also for the current interaction (that food preferences can be used to make claims about what foods should or should not be eaten during family meals).
Box 2.3: Exploring DP further
One book chapter
Edwards, C. (2005b). Discursive psychology. In K. L. Fitch & R. E. Sanders (Eds.), Handbook of language and social interaction. Mahwah, NJ: Lawrence Erlbaum Associates, pp. 257—273.
One journal article
Stokoe, E. & Hepburn, A. (2005). ’You can hear a lot through the walls’: Noise formulations in neighbour complaints. Discourse & Society, 16(5), 647—674.
Critical discursive psychology
As we move from the zoom-lens to the wide-angle forms of DA, critical discursive psychology (CDP) is the one which sits most readily in the intervening space, capturing some of the detail of discourse while also addressing broader cultural issues. Crucially, it is argued that these issues — such as gender, class, sexuality — cannot be reduced to a sequential analysis of talk (as would be argued by CA and DP). While CDP might analyse everyday interaction, for instance, it would also examine this in light of how that interaction is located within a particular social, cultural and historical setting. So for CDP, discourse is understood as the intersection between the everyday and the cultural; that one can see cultural influences in the way we talk and write, and that these are influences in that they shape our discourses. It does this through the concepts of ideological dilemmas, interpretative repertoires and subject positions. Ideological dilemmas (Billig et al., 1988) are contradictory or oppositional ways of understanding the same concept, but unlike the hidden ideologies of critical discourse analysis (see later), CDP argues that people can actively use these dilemmas to argue for different positions. Similarly, interpretative repertoires (Potter & Wetherell, 1987, inspired by Gilbert & Mulkay, 1984) are coherent sets of ways of talking or writing about an issue — patterns in discourse, as it were. They are flexible discursive resources in that they can be used to argue for or against a particular issue, and there can be opposing repertoires used for the same issue. For example, you might draw on a romantic repertoire to justify a marriage that lasts for a lifetime or a realist repertoire to justify marriage that ends in divorce after a few years (Lawes, 1999). CDP argues, then, that we can draw on a range of repertoires at any time, but that some repertoires are more culturally dominant than others. These dominant repertoires are established over time, and can become so normative and established in a culture that they become naturalised and understood to be common sense or ’fact’. Interpretative repertoires can include bodily practices (e.g., ways of eating) and ways of dressing, walking or behaving (see Edley, 2001 for some useful examples in relation to masculinity and normative practices) and they make available different subject positions. So discourses not only impact on how we talk, but also how we behave in particular ways. For instance, repertoires around eating can shape the types of food we treat as ’normal’ or ’traditional’ food, as well as how often and when we eat, and what we talk about when we’re eating.
CDP developed within psychology primarily through the work of Margaret Wetherell with Jonathan Potter, and later with Nigel Edley — out of a concern that the zoom-lens approaches to discourse (CA and DP) were missing historical context, and that the wide-angle lens approaches (Foucauldian DA (FDA) and critical DA) were missing an empirical rigour (Wetherell, 1998). So it can be understood as a synthesis between DP and FDA, and the concept of subject positions used in CDP is very similar to the subjectivities referred to in FDA. CDP draws on many of the same influences as DP, then, because it has the same intellectual roots. Where it differs is a movement away from CA and ethnomethodology towards a stronger post-structuralist focus. There are other strands of CDP, most notably that referred to by Ian Parker (see also DP section above), which adopts a more political (and therefore, arguably, a more critical) approach to CDP.
How we can use CDP
We can use CDP when we want to understand how a particular issue, such as gender or sexuality, is understood in a cultural context and how this translates into people’s discourses about that issue. So we can begin to see some of the ’critical’ side of discourse analysis here — hence the ’C’ in CDP — in that it seeks to identify the ways in which people are positioned in particular ways and how repertoires are reproduced and held to be common sense. CDP has been used to examine issues of parenting (Edley & Wetherell, 1999, 2001; Locke, 2015), gender (Reynolds, 2013; Seymour-Smith et al., 2002; Seymour-Smith & Wetherell, 2006), emotions and gender (Walton, Coyle & Lyons, 2004), racism (Wetherell & Potter, 1988) and citizenship (Gibson, 2009).
How it differs from DP
The overlap between DP and CDP should therefore be apparent from the similar roots of both forms of discourse analysis and the collaborative work of Margaret Wetherell and Jonathan Potter, which was the foundation of these two approaches. The main point of tension between DP and CDP is, however, what has become known as the context debate: should we focus only on which topics are raised within a conversation explicitly by participants (what has been termed participants’ orientations: Schegloff, 1997) or which topics are thought to be relevant from a broader cultural or social perspective (Billig, 1999). CDP is argued to bridge this theoretical and analytical gap, bringing to bear both a broader understanding as well as the empirical focus of what is actually said (Wetherell, 1998). CDP is less concerned with the sequential aspects of talk (as DP is) and more concerned with the broader patterns of talk across a particular data set. What is more important is how these words draw on wider social meanings and make them relevant in the here and now. CDP also assumes that discourses have an action orientation; that when we talk or write, these discourses have specific functions that are grounded in particular contexts. This means that a phrase such as ’I love broccoli’ can do different things depending on when it is spoken (or written). So the meanings of words are not abstract and universal, nor are they formed through the speaker’s intentions (cf. critical discourse analysis), but they derive meaning in interaction itself.
The main aim of CDP is to examine the repertoires, dilemmas and subject positions that shape our discourses in particular ways. As such, CDP can make use of data from naturally occurring interaction (such as a family meal), but might also draw on interviews and focus groups, where there can be greater opportunity to access the repertoires that people use (i.e., one can ask direct questions in an interview or focus group, and so we can ensure that the topics we want to explore are covered in sufficient detail). A research question for CDP might then be to examine the dilemmas around choice of food for the family: how individual versus collective preferences are organised, how issues of gender, age or family status are used to negotiate meal choices in family settings. Like DP, CDP research might be interested in how choices are negotiated in the family home, but would examine how these relate to culturally available repertoires about choice (e.g., choice as an individual right) rather than focusing on how choices are formulated in a specific conversation in one family meal.
As CDP is concerned with the patterns in the way in which we talk or write, there is less of a concern with capturing the detail of intonations and the sequential features of the talk; what is more important is that the transcript captures what is said. So a transcript typically features the words spoken, any noticeable but un-timed pauses (depicted as (.) in the transcript), and any emphasised words (noted by underlining in the transcript). As with DP and CA, CDP does not typically use punctuation in its grammatical use. Line numbers can be included but are not necessary because the analysis often pulls across broader stretches of text. We are less focused on specific lines of data and more focused on words and phrases that feature across a conversation. As a consequence of including fewer phonetic details, a transcript can cover a greater length of the interaction in the number of lines of transcript. The transcript in Extract 2.3, for example, captures 31 seconds (compared with 21 seconds) of this mealtime in about the same number of lines as the transcript shown for DP above.
Extract 2.3: Example CDP transcript
Bob: mine’s is quite a lot of- (.) I don’t like steak
Lesley: no don’t do either
Bob: it doesn’t fire my rockets it just doesn’t
Lesley: I like fish we never have enough fish
Bob: well see that fish you buy (.) don’t buy that see that thing that gets like incarcerated in some kinda (.) chemically compound thing you get it like it doesn’t look like a fish that thing we had the other night
Lesley: that battered thing no I don’t like battered fish I mean proper fish
Bob: it wasn’t
Linda: it was plaice (.) it was plaice out of Moors and Shawlands
Bob: it was horrendous
There are three core analytic concepts within CDP — interpretative repertoires, ideological dilemmas and subject positions — each one draws on the others, making a link between broader social or cultural concepts and situated discourses. A CDP analysis would therefore begin by searching through the data corpus for ways of talking about a particular issue. In the meal example here, that might be food preferences, food quality, or choice in food purchasing. For instance, in the first part of Extract 2.3, Bob qualifies his statement about not liking the steak by also stating that ’it doesn’t fire my rockets’. This frames it as if it were a biological or chemical reaction, and perhaps also that food can (and should) provide pleasure from eating, and that food tastes are important when considering choice of food (rather than, for example, nutritional requirements or cost-efficiency). So we might begin to search for other ways of talking about food preference to see if there is a pattern here and a possible interpretative repertoire. Alternatively, we might focus on the latter part of this extract where the quality of the fish is being disputed. Here, we see Bob describing it in chemical terms and as being ’horrendous’, Lesley also talks about ’proper fish’, and Linda draws on a particular high street shop (Moors and Shawlands) that suggests that this is superior to fish from other shops. There may be, then, an interpretative repertoire about what counts as ’proper’ food and better quality; this may also relate to the subject positions of the speakers. Linda, for example, appears to be criticised for buying the ’wrong’ kind of fish, and thus is positioned as unable to judge the quality of food. In response, she defends her subject position through reference to not only the particular type of fish (plaice) but also the associations of a particular brand of shop, and, by corollary, her subject position as a good judge of food through her shopping choices.
Box 2.4: Exploring CDP further
One book chapter
Edley, N. (2001). Analysing masculinity: Interpretative repertoires, ideological dilemmas and subject positions. In M. Wetherell, S. Taylor & S. J. Yates (Eds.), Discourse as data: A guide for analysis. London: Sage. Chapter 5.
One journal article
Reynolds, J. & Wetherell, M. (2003). The discursive climate of singleness: The consequences for women’s negotiation of a single identity. Feminism & Psychology, 13(4), 489—510.
Foucauldian discourse analysis
The first of our wide-angle lens approaches to DA is Foucauldian discourse analysis (FDA). The core focus for FDA is to examine the relationship between discourse and subjectivity. Discourses are understood not only as constructing a particular version of the social world, but also as having an impact on subjectivities and practices; that discourses shape how we think and feel (our subjectivities) and how we can talk and behave (our practices). For example, discourses construct ’objects’ in the social world, such as intelligence, and these objects have real consequences for people: our intelligence can be measured and we can be separated in education or workplaces on the basis of such measurements. Discourses therefore produce truths about the world (which have real effects on people), and in turn these truths influence what we can say, think and do. This is where power comes into play, in that discourses privilege some ways of being over others. Some discourses have become so embedded within a culture that they are considered to be common sense, and they can be resistant to change. Within FDA, however, power is understood as being a productive or circular (rather than linear, top-down) force, in that it produces ’truths’, but it is also maintained by the ways in which discourses are used by people. For example, discourses produce certain social structures (such as intelligence and educational institutions), and these structures in turn validate the discourses (the grading of student work, for instance, supports the discourse of intelligence). Discourses can also change over time. Truths about sexualities and genders, for example, have changed considerably over the last century. People can also draw on different discourses at different times — even over the course of a day — and so subjectivity is understood to be a fluid and dynamic process of the subject positions that are available to us.
FDA has its roots in the work of French philosopher Michel Foucault and post-structuralism, which challenged assumptions about a simple structural relationship between language and culture. Developed in the late 1970s, it draws heavily on Foucault’s classic works, such as those on sexuality (Foucault, 1979), and the historical changes within discourses; that what we understand to be truth here-and-now is very different from what was truth twenty years ago, or in a different society. FDA was also influenced by Henriques et al.’s Changing the subject (1984), as was DP and CDP, and with how people are positioned or constrained by particular social practices. FDA works across the social sciences, due to the influence of Foucault in these areas, and perhaps because of this there are variations and tensions in how FDA is understood (see, for example, Hook, 2001). Some variations emphasise the historical analysis of discourse (e.g., Carabine, 2001), others draw on psychoanalysis (e.g., Hollway, 1989), for example.
How we can use FDA
We can use FDA to examine how discourses make available different ways of being and speaking in the world. This focus on discourses and their impact on subjectivities makes it ideal to examine issues where people’s own bodies are regulated in some way, such as gender and sexuality (Bernasconi, 2010), child-bearing (Barcelos, 2014), children and health (Gibson & Dempsey, 2015; Walters et al., 2015) and fashion (Jackson et al., 2012). FDA has also been developed in slightly different ways across the globe, such as in gender studies in the USA (Butler, 1990, 1993), sociology of knowledge work in Germany (Keller, 2007), and the related work of Laclau and Mouffe (1985) in what has become known as the Essex School of discourse analysis.
How it differs from DP
There are a few main differences between DP and FDA (see also Willig, 2013). First, FDA is interested in the influences that discourses have on people’s subjectivities and practices; on how they shape our ways of experiencing the world. So while DP would analyse subjectivity in terms of how people account for their experiences, or construct sensations, feelings or emotions in particular ways in interaction, FDA would treat subjectivity as more than just a discursive construction. Second, FDA takes a different approach to agency. Whereas DP mostly treats people as active users of discourse, FDA theorises people as being both shaped by discourse and having the ability to make choices over which discourses are used. The process of data analysis and illustrating the consequences of discourses can thus help to enable people to make different choices about the discourses that they use. Finally, FDA situates discourse within a wider context than DP: it examines the discourse not only in terms of what is said, but how it relates to social, cultural or historical issues. Discourses are understood, then, like a bridge between social structures and subjectivity; that our words have meaning beyond the local context of the conversation or interaction.
FDA would typically collect samples of discourse that focus around a particular ’discursive object’, so it would not necessarily use a video-recorded family mealtime for the same reasons as CDP: if we are interested in a particular topic this may not readily occur in everyday or institutional talk. Using our meal example, instead of recording family meals, what might be more appropriate would be semi-structured interviews with adults about the process of buying and choosing food, for instance, or government policy documents about recommendations for healthy eating. In contrast to CDP, therefore, the emphasis with FDA would be on how discourse constructs particular knowledge about ’recommended’ or ’correct’ foods to eat, and about typical (or ’normal’) family arrangements; so not just identifying ways of talking about the topic, but how subject positions and possibilities for agency and choice are limited through the construction of the topic in a particular way. In the example data, we can see the discussion at the family meal beginning to cover issues around what the family members (Lesley and Bob) like or don’t like about the food. As we will see in the transcript below, the conversation soon moves onto issues of choice and money: who buys the food, whose money is used, and who gets to choose the type of food eaten by the family.
In order to identify subject positions and the construction of the discursive ’object’ (in this case, food choice), we will need to include a little more transcript than those shown above. This is consistent with the wide-angle lens analogy; we need to look more broadly across the text to understand how discourses weave through at different points, rather than to focus on just one section of the interaction. We pick up the interaction just as the discussion about the choice of fish begins. Note that the transcript here does not have line numbers because the focus is more on discourses across a corpus rather than with the sequentiality of the interaction, though line numbers can be used in FDA research. In fact, due to the overlaps between CDP and FDA, there is considerable variability in how transcripts from these two DA forms are presented, and the transcript shown below is just one example of how this might be represented. In practice, an FDA transcript can look very similar to a CDP transcript because they are interested in similar issues (just as DP and CA transcripts are often very similar). Speaker names have also been replaced with family status labels (i.e., Mum, Dad) to highlight their position with respect to a family structure; given that FDA is focused on social structures (such as the family as an institution), it is relevant and important for FDA to know which family member is speaking.
Extract 2.4: Example FDA transcript
Dad: well see that fish you buy? (.) don’t buy that. see that thing that gets like incarcerated in some kinda (.) chemically compound thing
Dad: you get it like it doesn’t look like a fish that thing we had the other night
Daughter: that battered thing (.) no- I don’t like battered fish I mean proper fish
Dad: it wasn’t-
Mum: it was plaice (.)
Dad: listen- (.) plaice is-
Mum: it was plaice out of Moors and Shawlands
Dad: it was horrendous
Daughter: yeah but I’m not talking about battered fish I’m talking about fresh fish
Dad: fresh, fresh, fresh-
Daughter: why do we never have some sea bream or sea bass or some tuna steak? (.)
Mum: there’s some in the fridge
Dad: listen, everything’s out of a tin in this house (laughs)
Daughter: when I contribute I’ll be inputting in my demands for my food
Gran: oh my (.) oh my
Mum: she was trying to put demands in the other day, because I had gone to Mindletons and bought prawns
Daughter: they were bowfin! ((disgusting))
Mum: and she went ’these are dry (.) I’ve told you, don’t buy Mindletons’ and I said ’see when you contribute (.) you say to me where I can go to buy’
Dad: no don’t bother (.) listen, you’ll just contribute and get no choice same as I contribute and get no choice. Unless you get the messages ((food shopping)) then you can go and buy
FDA is concerned with the ways in which discourses produce particular ways of seeing and ways of being in the world; how subject positions and subjectivities are produced in discourse. Instead of focusing on the interaction between the family members during the mealtime, therefore, we could use FDA to examine the subject positions that revolve around the complex issues of money, choice and food. We have already seen with the CDP analysis that the mention of particular shops (’Moors and Shawlands’) can be interpreted as invoking the status of both the food and the person doing the shopping. With FDA, we might also interpret the different claims about foods here — plaice from Moors and Shawlands versus prawns from Mindletons — as setting up a dichotomy between ’good’ and ’bad’ foods. The family members could therefore be interpreted as drawing on discourses around quality of foods through reference to particular shops. The daughter also mentions particular types of fish that are likely to be less common in the average household in the UK, so there is a status of food being invoked here; it is not just ’fresh’ or ’proper’ fish that is requested, but a less common and possibly more expensive type of fish. The daughter then raises the issue of contributing money to the family income as a means of justifying her ’demands’, and both Mum and Dad take this further: stating that contributing money does not necessarily give you choice in what the family eats. So there are culturally relevant topics here about who pays into the family income, whether or not this entitles that person to make food choices, and expectations about the standard of food that the family should be eating. This in turn might lead to analyses around the family as an institution, and how different ways of being a family are made relevant and available through different discourses.
Box 2.5: Taking FDA further
One book chapter
Willig, C. (2013). Introducing qualitative research in psychology: Adventures in theory and method. Maidenhead: McGraw-Hill Education (UK). Chapter 11.
One journal article
Riley, S., Thompson, J. & Griffin, C. (2010). Turn on, tune in, but don’t drop out: The impact of neo-liberalism on magic mushroom users’ (in)ability to imagine collectivist social worlds. International Journal of Drug Policy, 21(6), 445—451.
Critical discourse analysis
Critical discourse analysis (hereafter, CDA) is perhaps the most critical of the five forms of DA discussed in this chapter as it takes an explicitly political stance and argues first and foremost for the emancipation of marginalised groups and individuals in society. The aim of CDA research is to reveal the ideologies that underpin oppressive discourses, and to highlight subversive or resistant discourses that can be used to liberate people from oppression. The starting point is typically a social problem or issue relating to power or political inequality. CDA aims to reveal the underlying ideologies behind discourses, what kind of power is being used, and how this is achieved through text and images. Discourses are not, therefore, separate to issues of power: they are the means through which power is exerted in society. For example, language choice is regarded as a political act because it has implications for people and their actions. Discourses are also understood in terms of their broader context — hence the wide-angle lens — not just the immediate context of what is said, and to whom, but how the discourses draw on the socio-political context and cultural context and have implications for a broad spectrum of people in society. For example, if a newspaper reports on immigration as being problematic and makes a distinction between the ’we’ who are being ’invaded’ and the ’they’ who are coming into our country, then this has consequences for anyone who comes to live in that country from a different place. It doesn’t just effect whoever is reading the newspaper report; it has wider implications, like water ripples when a pebble is dropped into a lake.
The roots of CDA are embedded within critical linguistics, sociolinguistics and semiotics — and, to some extent, build on FDA — and so CDA is often focused on the ’meanings’ underlying language. Whereas other forms of DA might argue that meaning is constructed in interaction, as people are talking for example, CDA research often argues that meanings are hidden within the text or talk. That is, that meanings pre-exist in the words or images used. What unites researchers within the CDA approach is a focus on language, ideology and power, and here we see the main difference from FDA: critical discourse analysts are more concerned with the ideas or ideology that underpin discursive practices than with knowledge, which can in practice be ’owned’ or used by anyone in society. An ideology is the domain of the dominant group or groups in society, and therefore there is a political and motivated edge to CDA research. Various strands of CDA research exist, from the socio-cognitive model of Teun van Dijk (1993, 2001), the discourse-historical model of Ruth Wodak (1996), the dialectical-relational modal of Norman Fairclough (1989, 1995) to multimodal CDA (Kress, 2009).
How we can use CDA
We can use CDA when we want to address a social issue or problem, where individuals or groups of people are being marginalised or oppressed. The ’critical’ element of CDA is that it aims not only to illuminate how discourse creates certain types of reality, but also to highlight power inequalities in order that these might be subverted or removed. This can be particularly important for those researchers who want to make an impact: to analyse discourse in a way that can potentially have wide-reaching implications. CDA works for the oppressed and the dominated, and thus it is not sufficient merely to highlight dominant discourses; one must also use the research to make a political statement. CDA is therefore ideally suited to research on racism (e.g., Van Dijk, 1993; Wodak & Matouschek, 1993), political speeches and documents (Bhatia, 2006; Fairclough, 2001), and national identity and immigration (e.g., Banda & Mawadza, 2015; Paltridge et al., 2014).
How it differs from DP
The main differences between CDA and DP are similar to those between FDA and DP; both CDA and FDA treat issues such as power, agency and society as existing at a broader level than DP. That is, that they cannot — and should not — be reduced to a sequential analysis in a specific interactional context. While both FDA and CDA are concerned with power, however, CDA argues that power is controlled by dominant groups and organisations in society through the use of discourse. With FDA, power is a more flexible and bi-directional concept; power can come from using discourses in particular ways as well as being used by discourses. Another defining feature of CDA is that it treats language use as a matter of individual choice: that people deliberately make use of particular word choices in order to express an argument or communicate an idea. By contrast, DP is agnostic about word choice and does not make claims about whether people intentionally use particular discursive practices, focusing on the social actions performed in talk rather than cognitive processes of word selection.
The main aim of CDA is to begin with a social problem, and with our meal example this might be to consider the rising costs of food and the need for ’food banks’ to help support those families who are not able to buy enough food to eat. Like FDA, CDA would preferably work with texts that have an impact at a cultural or societal level, rather than focusing on what might be considered an idiosyncratic family situation. With a focus on revealing the oppressive discourses that underlie everyday texts, then, media advertisements about ’typical’ family meals or political speeches about the causes of food poverty might be more appropriate. While CDA might typically examine media texts, it can also be applied to everyday interaction because it is argued that individuals are as much a part of semiotics (making meaning) as public texts or images (Machin & Mayr, 2012). It can also be argued that this is one of the ways in which oppressive discourses can impact on individuals: through the way in which we use or re-make meanings to frame the world around us. If we were to use CDA to examine this family meal, we might therefore focus on how different family members use different frameworks of talking to justify their positions. For example, Dad talks about contributing to the family budget, and makes inferences about his own lack of agency and the importance of traditional working values.
Rather than focusing on the interaction between people at this family mealtime, CDA might consider instead the language used by one of the people present. To remain consistent with the earlier transcripts above, the extract here presents Dad’s (Bob’s) part of the discussion as a series of statements, to collect together the discourses he used and present them in a way that is easy to read and to search for key terms and word choices. Note how there are no line numbers here, and punctuation is used grammatically rather than to represent intonation. Again, there is variation in the transcripts used within CDA research and the example given here is just one possible representation of the data.
Extract 2.5: Example CDA transcript
“Well see that fish you buy? Don’t buy that. See that thing that gets like incarcerated in some kind of chemically compound thing. You get it, like it doesn’t look like a fish, that thing we had the other night. It was horrendous. Everything’s out of a tin in this house. You’ll just contribute and get no choice same as I contribute and get no choice. Unless you get the messages ((food shopping)) then you can go and buy.”
The starting point for CDA would be to focus on lexical choice, that is, what kinds of words have been chosen, what vocabulary has been used? The use of language in this way has been likened to marking out the territory on a map (Fowler, 1991). Certain features may be highlighted, boundaries may be established and other features made absent (or suppressed). A common type of lexical choice that a CDA analysis might focus on would be verb choice. Verbs are particularly useful because they can imply responsibility, authority, agency and formality (Caldas-Coulthard, 1994). In the extract above, then, we might examine how Dad uses the directive statement ’don’t buy that’ as a way of asserting an authoritative voice; someone who has control or expertise in this situation (as claiming greater knowledge about the chemical-compound covered fish; again, the use of more technical terms presents Dad in an authoritative position). The ’chemically incarcerated’ term may also be considered a type of ’ideological squaring’ (Van Dijk, 1998), where meanings (in this case, ’natural’ or ’free from additives’) are invoked through opposites. Note how Dad then presents his position as lacking agency despite having contributed to the family income, and asserts power over his daughter to be treated in the same way. It is also interesting to note that CDA makes use of some terms or concepts that appear in other DA forms (e.g., directives, CA; hyperbole, cf. Extreme case formulations, DP/CA; modal verbs, DP). The main point of difference here is that CDA would interpret their use as a deliberate choice; that they are being used by speakers to achieve a particular social goal, one that exerts power over others. CDA would also be more prepared to argue that issues that are not explicitly said can still be present in texts; that absences are not just analysable but necessary for understanding the discourse.
Box 2.6: Taking CDA further
One book chapter
Machin, D. & Mayr, A. (2012). How to do critical discourse analysis. London: Sage. Chapter 1.
One journal article
McGannon, K.R., Berry, T.R., Rodgers, W.M. & Spence, J.C. (2016). Breast cancer representations in Canadian news media: a critical discourse analysis of meanings and implications for identity. Qualitative Research in Psychology, 13(2), 188—207.
· There are many different forms of discourse analysis, though they all share the assumption that discourse constructs, rather than reflects, reality.
· The analogy of camera lenses can be used to help distinguish between those which focus on the detail of talk or pan out to incorporate a wider social context.
· Conversation analysis and discursive psychology are two of the ’zoom lens’ or fine-grained approaches to analyses of talk and interaction and focus on turn-taking and the immediate interactional context.
· Critical discursive psychology examines the detail of what was said in the broader context of social and cultural issues.
· Foucauldian discourse analysis and critical discourse analysis are like ’wide-angle’ lenses and study the ideological aspects of discourse and the influence of these on subjectivities or power relations.
· There is no hierarchy of discourse analysis approaches; each has its own advantages and should be chosen on the basis of the research questions that you want to address.
Wetherell, M., Taylor, S. & Yates, S. J. (2001). Discourse as data: A guide for analysis. London: Sage.
Wooffitt, R. (2005). Conversation analysis and discourse analysis: A comparative and critical introduction. London: Sage.