Discourse Analysis

Terminological Issues

Different forms of DA have emerged in different disciplinary environments—linguistics and sociolinguistics, sociology and social psychology, philosophy, education, and so on. This can lead to confusion that can be compounded by the different senses in which the term discourse is used. At its most general, the term is used to refer to virtually any language use or even to related semiotic systems; however, other work uses the term specifically to refer to a linguistic object that can be described and counted.

Using the specific definition, a study might, for example, attempt to identify a discourse of medicine and a discourse of counseling operating side by side in a medical consultation. Under the broader definition, discourse analysis covers large areas of sociolinguistics and linguistics, much cognitive science concerned with language use, social semiotics, and work on educational interaction as well as areas of work such as discursive psychology, critical discourse analysis, Foucauldian discourse analysis, and simply discourse analysis.

There is also a strand of work associated with post-structuralist thinking, particularly the work of Michel Foucault, that is often referred to as continental discourse analysis (sometimes Roland Barthes and Jacques Derrida are included in this definition). Confusingly, this work is rather different from Foucauldian discourse analysis. DA has developed as a contested terrain where different books with “discourse” in their titles can exist with no overlap in content.

Historical Developments

The earliest use of the term discourse analysis is probably in Zelig Harris's linguistic research during the [Page 218]early 1950s that focused on the attempt to explicate sentence meaning in texts by relating the sentence to surrounding sentences. Since that time, the first sustained approach to explicitly use the term was developed by linguists John Sinclair and Malcolm Coulthard in studies of classroom interaction. Their aim was to build a systematic model of how interaction is organized in the classroom. The prototypical pattern they identified was the “initiation–response – feedback” sequence. This has the following form:

• Teacher: What are five fours? (initiation)
• Pupil: Twenty-four miss. (response)
• Teacher: Close, Julie. Have another go. (feedback)

Sinclair and Coulthard had the ambitious aim of explicating the interaction structures of different settings by comparative work. In this aim, it prefigured more recent work on institutional interaction by conversation analysts. Although the analytic materials are simplified and the analytic categories are relatively undeveloped, this work was a departure from a linguistic approach that worked with invented materials or written sentences.

A separate linguistic tradition focused on the way in which sentences are linked together in coherent discourse. One strand in this work involved studying the way in which terms such as however and but work to generate coherent discourse. This work has merged with a tradition of research called discourse processes that attempts to join linguistic and psychological concerns. This tradition considers questions about how psychological experiences become transformed when they are reconstructed into verbal narratives and, conversely, how mental scripts are used in the understanding of narrative. In its research practice, the work on linguistic coherence typically used standard linguistic methods, drawing on invented examples and considering whether they were “well formed” or “anomalous.” The more psychological work has drawn on a range of approaches; sometimes records of interaction are coded for statistical regularities and contrasts, and sometimes experimental manipulations are used. Work in this tradition is often published in the journal Discourse Processes.

...