Orientation Observation In-depth interviews Document analysis and semiology Conversation and discourse analysis Secondary Data Surveys Experiments Ethics Research outcomes



Social Research Glossary

About Researching the Real World



© Lee Harvey 2012–2017

Page updated 13 January, 2017

Citation reference: Harvey, L., 2012–2017, Researching the Real World, available at
All rights belong to author.


A Guide to Methodology

6. Conversation and discourse analysis

6.1 Introduction to discourse analysis

6.1.1 Early forms of discourse analysis
6.1.2 'Non-specific' discourse analysis

6.2 Conversation analysis
6.3 Genre analysis
6.4 Pragmatics
6.5 Discursive psychology
6.6 Interactional sociolinguistics
6.7 Ethnography of communication/speaking
6.8 Critical discourse analysis
6.9 Summary and conclusion

6.1 Introduction
This Section of the Guide explores the various forms of discourse analysis including one area, conversation analysis, that used to be regarded as distinct from discourse analysis but is increasingly viewed as a form of discourse analysis.

The term discourse analysis is thought to have first appeared in 1952 in the title of a paper by Zellig Harris. However, it was from the late 1960s that it emerged as a cross-disciplinary approach, coinciding in with the interest in semiotics, psycholinguistics and sociolinguistics.

Over a thirty year period, discourse analysis, which started as a specific approach, has become a more generic notion, with a range of approaches being added, often retrospectively, as forms of discourse analysis.

There are, thus, several different forms of discourse analysis, covering a wide range of communication, and, apart from conversation analysis, the following approaches are explained in some detail in this Guide: genre analysis, pragmatics, discursive psychology, interactional sociolinguistics, ethnography of communication/speaking and critical discourse analysis.

Other approaches are mentioned in passing, including: variation analysis, applied linguistics, cognitive psychology, rhetoric, stylistics, textual analysis, mediated discourse analysis, multimodal discourse analysis and corpus-assisted discourse analysis.

In essence, discourse analysis examines the use of language in communication and seeks out the underlying, concealed or implicit, meanings and associated presuppositions.

Conversation analysis and discourse analysis have much in common, albeit the former is rather more restricted, in that they explore language. Conversation analysis, in it's original conception, undertakes the close analysis of real conversations. At root, conversation analysis takes as axiomatic that conversation is central to social life and that meanings are constructed through conversation.

However, the main difference between conversation analysis and most forms of discourse analysis is that conversation analysis is primarily an empirical exercise while (some forms of) discourse analysis appeal to theory to dig beneath of the expressed meanings in communication to attempt to reveal underlying principles, ideologies or structures.

There is a similar fuzziness about the boundaries between discourse analysis and discursive psychology. The latter is usually seen as a form of discourse analysis.

Paul ten Have (2005) explains:

The term "discourse analysis" is much less clear than "conversation analysis," or rather, it is used in many different ways by different people, in different countries and in different contexts. On the one hand, it can serve as an overall blanket term for any and all efforts to analyze "discourse," texts, talk and so forth. But many people in the U.K. ...use it to indicate one particular, although diversified, research tradition which, nowadays, finds its most prominent expression in a branch of social psychology which is called "discursive psychology" and is associated with people like Derek Edwards and Jonathan Potter at Loughborough University in the UK.

Accounts of the genesis of discursive psychology indicate that it evolved from early discourse analysis, notably, Discourse and Social Psychology (Potter and Wetherell, 1987). Potter and Wetherall's version of discourse analysis had elements in common with what became discursive psychology. These included 'drawing a constructionist sociology of scientific knowledge', a focus on 'categories and descriptions and the way they are involved in actions' and 'both offer a respecification of basic psychological notions' (Wiggins and Potter, 2007, p. 75) such as the traditional conception of 'attitude'.

However, Potter and Wetherall's approach had two main features that differed from later developments: first, open-ended interviews, characteristic of early discourse analysis were dropped by discursive psychology because of the analytic problems; second, the notion of repertoire, central to Potter and Wetherall's approach, have been heavily criticised (see Section 6.5).


6.1.1 Early forms of discourse analysis
The book, Opening Pandora’s Box, by Nigel Gilbert and Michael Mulkay (1984), is frequently cited as an early example of a systematic discourse analytic study. In their study of a scientific dispute over chemiosmosis (the means whereby energy is transferred in cells), Gilbert and Mulkay analysed and revealed the different accounts of the dispute.

In this book we have approached the social world of science as a multiple reality….We claim to have shown that scientists use distinctive interpretative forms as they construe their actions and beliefs in different social contexts. We have made an attempt to capture various significant facets of these interpretative forms by devising the concepts of empiricist and contingent repertoires. These concepts have proved to be useful, not only in describing certain recurrent features of scientists' formal and informal discourse, but also in understanding interpretative phenomena which have no obvious connection with our initial observations on versions of action in research papers and interviews. Thus we showed that the two repertoires were used by participants as resources for constructing asymmetrical accounts of error and correct belief….We were able to show clearly that, although participants' substantive accounts of action and belief are highly diverse, they are constructed out of recurrent interpretative forms and repertoires which can be identified, described and documented by the analyst…. We focused on the supposedly collective phenomenon of cognitive consensus….Examination of participants' interpretative work showed unequivocally that a given collectivity at a given moment can be made to exhibit radically different 'degrees of consensus'. (Gilbert and Mulkay, 1984, pp. 188–9)

This may appear to be no more than an observation of the variability of accounts that occur in the everyday world. However, Gilbert and Mulkay went further, by rejecting what they regarded as the dominant mode of sociological method, which tends to ignore variability in the search for similarities. As Wilson (2005) noted:

Gilbert and Mulkay identified different ‘linguistic repertoires’ in the way scientists expressed themselves about their work. In public texts the vocabulary would paint a picture of an empirically knowable real world populated by knowable and secure facts (Gilbert and Mulkay called this the ‘empiricist repertoire’). In private, however, the scientists’ words would change to a ‘contingent repertoire’ that described a shifting world where things could have been otherwise and where facts were humanly constructed. Gilbert and Mulkay found that the contingent repertoire was used especially when things went wrong.

If another team of scientists failed to confirm their findings, it was because of ‘contingent’ things like the other laboratory’s poor procedure, or carelessness, or even cheating. It wasn’t ‘proper science’. Only if things went ‘right’ would the scientists talk about ‘facts’ and a regular, predictable universe. It sounds petty, but it’s more than that. The effect of this variability, Gilbert and Mulkay argued, was to maintain the idea of Science, and defend the principle that there is a knowable objective world. ‘Error’ is accounted for by human or other failings; ‘fact’ is arrived at by correct methods. The scientists’ discourses of ‘contingency’ and ‘empiricism are defending the very constitution of Science.

Gilbert and Mulkay (1984) thus offered an alternative approach to the analysis of scientific culture, moving away from previous studies that focused on scientists’ actions and beliefs to an analysis of scientific discourse. In demonstrating that scientists produce varying accounts of their actions and beliefs in different social situations they challenged the view of scientific consensus and objectivity. The use, by scientists, of factual or empiricist discourse in formal contexts like research reports is counterbalanced by subjectivist or contextual discourse used in gossip and informal interviews. Gilbert and Mulkay argued that variability of accounts is not something arbitrary: it is not the case that some accounts are truer than others. Both types have their function in different contexts.


6.1.2 'Non-specific' discourse analysis
Despite all the variants of discourse analysis outlined in Section 6.1, there are studies that just describe themselves as discourse analysis studies or refer to their methodology as 'discourse analysis'.

For example, Butt and Langdridge (2003) use discourse analysis to reveal how the gay actor Kenneth Williams, who did not ‘come out’, was reluctant to admit his sexuality to himself. They analysed his diaries to develop a self-theory of the actor, which they claim showed that he was ill at ease accepting his sexuality.

For example, Butt and Langdridge (2003, p. 487) wrote:

… we see Williams wrestling with his sexuality, affirming that celibacy was the solution for him.

Sexually, I am getting to feel much more settled. I used to think that the reason for my virginity was moral and physical cowardice: part of some psychological fear of responsibility. More & more however, I am coming to believe in the rightness of abstinence for me. I am convinced that celibacy is an essential quality in my own character. I must never allow myself to be vulnerable in the sexual sense. That kind of humiliation would be detrimental in every way. (5.2.54)

In this passage, he disavows his sexuality – one cannot choose one’s fantasies; they ‘happen’.

The study, though,goes beyond this accounting of Williams’ own tussle with his sexuality to the affirmation of an existentialist phenomenological view of the self and the unconscious. Drawing on Maurice Merleau-Ponty (1962), Butt and Langdridge claim that self-theory is not conceptualized as a set of internal cognitions, rather it is viewed as a narrative that the person uses in accounting for his or her action. Merleau-Ponty argued that people are immersed in a social world to which they respond prior to being in a position to use internalised reflective language. Thus, practical consciousness precedes discursive consciousness, we tend to respond to questions prior to reflecting on them first.

Butt and Langdridge (2003, pp. 489–90) state:

From the point of view of existential phenomenology, the unconscious is not a deep structure, lying beneath conscious activity, and inaccessible to it (Sartre, 1958; Spinelli, 1994)….Our engagement with the world is primarily pre-reflective, or unconscious. Whether we then become conscious of it depends on, firstly, our ability and, secondly, our willingness to reflect on our engagement.… Through a discourse analysis of a diary, we have suggested how a person may choose not to reflect on their sexual engagement….The diary affords an opportunity to see the fashioning of this self-theory, which in Williams’ case was that of a conservative moralist.….Of course, the private pain can never be strictly private in that when one reflects on it, it enters the realm of discourse and is shaped by the vocabulary to hand. Yet, as Merleau-Ponty (1962) argued, there is surely a pre-reflective engagement on which we reflect. And this engagement is of course, not within the person, but between people.

They concluded:

Traditional psychological approaches seem to ignore the social world entirely, over-emphasizing personal agency and positing essentialist private worlds that pre-date social construction. And in discursive psychology, the main standard bearer of social constructionism in the UK, the private sphere is so permeable to the public that all individual agency evaporates entirely, in what Burkitt (1999) refers to as ‘discursive reductionism’. Both pragmatism and existential phenomenology recognize the social construction of the individual, as well as an individual that once constructed is the centre of agency. Drawing on these traditions, we have advocated a dialectical relationship between a pre-reflective ‘practical consciousness’ and a discursively mediated self-theory. (Butt and Langdridge, 2003, p. 490)

Another example of a study described as discourse analysis was undertaken by Riggs and Palasinski (2011). They interviewed young men on the issue of carrying knives, with a view to informing policy on tackiling knife violence. As a rapid response in the British Medical Journal to a previously published article they suggest that their discourse analysis reveals different conceptions by knife carriers about carrying a knife to that ascribed by society, viz. assuming a link to immaturity, machismo and deviance. The contibution by Rigss and Palasinski can viewed in more detail in CASE STUDY knife carrying. There is a very thin line here between studies that are described as 'in-depth interview' or 'ethnographic interview' studies (Section 4.1) and the description of the Riggs and Palasinski study as 'discourse analysis'.

Another example of discourse analysis is a study of the language used by tabloid newspapers in the United Kingdom and the consequent notion of nation that it constructs and reinforces. Martin Conboy (2006) examined four tabloid newspapers and showed that they construct ideal readership communities through the use of particular language forms: rhetoric, irony, alliteration, rhyme or parallelism.

An essential part of tabloid news values is the exaggerated foregrounding of sensation and ‘human interest’. These features have the effect of structuring the world in a way which rejects fundamental political issues… (Conboy, 2006, p. 15)

Conboy’s approach is to use a sample of the four main tabloids published in the space of a month in 2004. His examination of the papers leads him to conclude that they focus on people ‘at the edge of human experience’ such as the aged, very young, mentally ill, very rich, criminals and so on. He lists examples of tabloid content to support this, for example:

Multi-millionaire internet tycoon….seriously ill last night after a road crash.
Evil drug dealers are swamping Britain…

After exploring how different language is used to appeal to a particular readership he then showed how rhetorical devices are deployed to ‘provide a relatively consistent view of the national community which it seeks to reinforce’ (Conboy, 2006, p. 46). He provided examples of how nationhood is referred to explicitly, how national traditions are flagged, how Europe is used as a common external opponent, how terrorism is another external threat to unite against, how football is used as a national metaphor, how the country is suffering national decline for a range of reasons from immigration to binge alcohol drinking.

The book undertakes no complex discourse analysis, just explores the rhetorical devices used by the tabloids and quotes brief sentences or headlines to illustrate the way the devices are used to construct a nation and appeal to a targeted readership. In many respects this is qualitative text analysis that reproduces ethnographic processes of analysis, theory building and presentation.

Astrid Ennslin (2011), on the other hand, took an eclectic approach to an impressive study of discourse used in gaming. She explains:

This book has sought to help develop an understanding of the language of gaming in two main respects…. [First] It has studies the ways in which videogames and their textual ecologies re[resent and convey meanings and social actions they represent. This has included critical analyses of race, gender, language and ideologies in games themselves and the paratexts created top-down….by industry professionals and media stakeholders… [Second it shows] the idiosyncratic ways in which gamers communicate (through buddylects, for example,) and how the discourse they engage with help them construct identities and group membership. To do so, several layers of discourse had to be examined: language about games and gaming used by gamer across different media and communication platforms; language about games and gameplay used y industry professionals…; language about games and gaming used by journalists, politicians, parents, activists and other media stakeholders; ‘scripted’ language used within games; and language used in instruction manulas, blurbs, advertising and other peritexts.

More specifically, I have focused on genre-analytical, medial, textual, leical, metaphorical, collocational, conversational, pragmatic, discursive, multimodal and narrative issues and have sought to merge them with ludological theories with a view to offering a broad interdisciplinary understanding of the language of gaming. (Ensslin, 2011, p. 158)

Ensslin identified ‘three dominant gamer discourses’, the ‘discourse of ‘cool’ of fun and of appreciation (Ensslin, 2011, p. 159) and concluded:

Throughout this book I have interspersed exemplary analyses to exhibit a critical stance towards the language of gaming…[including] hype and moral panic in the mass media in relation to the highly controversial Grand Theft Auto series, as well as linguicism and racism in Black and White II”. My investigations have… [revealed] that videogame makers seem to follow a conservative line with respect to representations of gender, race and language, gamers themselves engage in a multitude of subversive practices. (Ensslin, 2011, p. 160).

An example of the critical approach is her “Case Study 2: The lexical construction of gender in World of Warcraft and Diablo”. In this, Ensslin ‘merges corpus-base, lexical examination of gaming language with critical discourse analysis’ (Ensslin, 2011, p. 84). Referring to a study by Carrillo Masso (2009, 2011), which shows clear gender differences in games (World of Warcraft and Diablo), Ensslin noted that the needs of female gamers are sidelined by a male-dominated industry: ‘…videogame designers traditionally implement variations of types of female characters: the damsel in distress and the ‘foxy villainess’ or femme fatale’ (Ensslin, 2011, p. 84).

A quantitative and qualitative concordance analysis of the node words ‘he’ ‘she’ ‘him’ her’ ‘they’ … [indicate] three time as many instances of ‘he’ than ‘she’…which is particularly striking given that in World of Warcraft there is an equal number of male and female avatar types tp choose from….Further evidence suggests that ‘he’ proportionally collocates with a greater variety of verbs than ‘she’.


Next 6.2 Conversation analysis