Feature 9: Discourses
Definition and Analysis Method
My particular research interest was in news media representations of the BP events, and the ideologies that were important to my account were those that informed the ways in which the story of the Deepwater Horizon events was represented to the reading public. In investigating significant discourses, I sought patterns and characteristics of the representation of the BP events which suggested shared ways of thinking, as well as changes in thinking over the time period of the data. Discourses have been found problematic to analyse systematically, and the work of Foucault does not provide specific methodological guidance (Graham, 2005). Sunderland points out (2006: 166) that discourses are “partial”, “non-finite” and “non-ubiquitous”—and subject to the perspective of the individual analyst.
Different discourses are accordingly likely to be “spotted” by different social groups of readers and analysts—for example those who favour a feminist perspective and those with a more traditionalist perspective—even when looking at exactly the same textual set of linguistic traces.
Sunderland defines a number of linguistic features as relevant to discourse identification, including lexical choices, verb forms and moods, speech acts and collocations. She points out that the analyst should be alert to what may be omitted as well as what is present. Fairclough’s (1989, 1992a, 1995a) approach to identifying the presence of ideologies focuses strongly on analyses within the framework of Systemic Functional Grammar, where linguistic choices concerning, for example, actor, mood and word order are held to be significant in positioning speakers and audience in certain desired relationships. A great diversity of individual and mixed method approaches have been used in the identification and analysis of discourses, including content analysis (Potter & Reicher, 1987), grammatical analysis (Fairclough, 1989), conversation analysis (Speer, 2001), poststructuralist approaches (Baxter, 2008; Wetherell, 1998) and corpus methods. Many analysts propose a mixed method approach in their selec?tion of tools for analysing discourses (e.g. Baxter, 2010; Cook, Robbins, & Pieri, 2006).
To identify patterns in the BP data relating to the ninth feature—discourses—I drew on my findings for the other eight discursive features. In my analysis so far, I had already identified a number of recurrent themes. I refer to these now as evidence of four representational discourses.