How are discourse and rhetorical analysis similar and different?

The study of rhetoric was revived in the 1970s and 80s with a particular concern with the argumentative organization of texts and the different rhetorical forms used to make them persuasive (Perelman and Olbrecht-Tyteca, 1971).  Billig (1987) has highlighted the way rhetorical ideas can be used to reformulate thinking in psychology.  For example, that the metaphor of an argument can be used to make sense of thought processes; instead of viewing thought as the operation of some calculating mechanism on internally consistent systems of belief, thought can be seen as riven with argumentative dilemmas whose structure comes from the available interpretative repertoires of a culture (Billig, et al., 1988).  So while a more orthodox social psychologist might be concerned with an evaluative expression as an index of a person’s individual attitude, a rhetorical analyst might be concerned to reveal the way that evaluation is put together to counter an established alternative (Billig, 1988b).

Conversation and rhetorical analysis emphasise two different orders of relationship.  CA emphasises sequential organization across turns; rhetorical analysis emphasises the relationship between opposing argumentative positions.  These may themselves be sequentially organized, but are not necessarily so.  Sometimes they may be expressed as direct and explicit argumentative claims using the speech act vocabulary of argument (‘I don’t agree with that’); at other times rhetorical contrasts may be built implicitly, often through competing descriptions of some action or event; for instance at a rape trial: 

Counsel:         And during the evening, didn’t Mr. O [the defendant] come over to sit with you? 

Witness:          Sat at our table.

(See Drew, 1992; Edwards and Potter, 1992).  Discourse analytic studies sometimes collect a range of different analytic materials — newspaper reports, interactional materials, interviews, parliamentary records — to facilitate a rhetorical analysis of some domain.  In this way it becomes possible to identify the rhetorical targets and oppositions of particular arguments and descriptions.

What is the role of interviews in discourse analysis?

Interviews have been used extensively in discourse analytic work; however, they are construed in a novel manner.  Traditionally, the goal of an interview was to produce a piece of colourless, neutral interaction.  However, in practice interviews are as complex and vivid as any other type of interaction, and responses to answers which may seem neutral and noncommittal in the abstract may have an important impact on the trajectory of the interaction.  

In discourse research interviews have been used extensively because they allow a relatively standard range of themes to be addressed with different participants — something hard to achieve when collecting naturalistic materials.  They also allow a high degree of control over sampling.  Interviews are conceptualized as an arena for identifying and exploring participants interpretative practices rather than an instrument for accessing a veridical account of something that happened elsewhere, or a set of attitudes and beliefs (Mischler, 1986; Potter and Mulkay, 1985).  An interview can be a particularly effective way of getting at the range of interpretative repertoires that a participant has available as well as some of the uses to which those repertoires are put.  Billig (1992) and Wetherell and Potter (1992) are both extended discourse based studies which work principally from interview material and illustrate some of the analytic possibilities they provide.  Widdicombe and Wooffitt (1995) draw extensively on CA to show how interviews can be dealt with as an interaction rather than a research instrument, exploring the way that different social categories are worked up, used and avoided in the course of interview talk.

Despite the virtues of this use of interviews, there are problems in relating the practices that happen in interviews to what goes on elsewhere and avoiding the interaction being swamped by the interviewer’s own categories and constructions.  Even the most open ended of interviews is guided by a schedule which specifies topics and themes as important.  Moreover, even when an interview is understood as an interaction in its own right, the dominant question and answer format is not ideal for getting at the sorts of turn-by-turn display of action and understanding that conversation analysts have utilised so effectively.  Partly for these reasons, discourse analysts have been increasingly turning to the study of records of natural interaction.

What is the role of records of natural interaction?

Arguably, one of the most astonishing omissions in 20th century psychology has been the study of what people do; their interactions in the home and workplace.  The few attempts in this direction were marred by a simplistic behaviourism which ignored interaction or reduced it to brute movements.  Inspired in part by the success of CA in working with records of natural interaction, discourse analysts have started to work with transcripts of conversations, newspaper articles, recordings of counselling sessions and similar materials.  

The term ‘natural’ here should be taken contrastively.  These settings are made up of natural interaction in the sense that it is not ‘got up’ by the researcher.  The test is whether the interaction would have taken place, and in the form that it did, if the researcher had not been born.  Of course, the use of recording technology itself can have an impact on participants’ understandings of a situation and their actions.  However, in practice there are a range of techniques for minimizing the intrusive effects of recording, such as using a period of acclimatisation.  Practice suggests that such effects are often surprisingly small.  

In most cases, such records cannot be used directly — a tape, particularly a video tape, is a very clumsy way to deal with materials.  What is required is a transcription that turns the record into a form that can be read through quickly, that allows different sections to be compared, and that can easily be reproduced in research papers.  The transcript does not replace the tape — often it is most helpful to work in parallel with both.