Analysing discourse: Textual analysis for social research

Analysing discourse: Textual analysis for social research. By Norman Fairclough. London: Routledge, 2003. Pp. 288. ISBN 0415258936. $39.95.

Reviewed by Bingyun Li, Fujian Normal University

In this well-written and accessible book, Norman Fairclough convincingly shows how a social perspective can be successfully combined into ‘real’ language analysis. The book is divided into four parts. Part 1 (‘Social analysis, discourse analysis, text analysis’) contains three chapters. In Ch. 1 (1–18), F distinguishes text analysis from discourse analysis, touching upon the causal effects of the textual elements of social events on social life. Ch. 2 (21–38) treats the relationship between texts, social events, social practices, and social structures. For F, ‘Any social practice is an articulation of the following elements: action and interaction, social relations, persons, the material world and, discourse’ (25). F also discusses three major types of text meaning: action, representation, and identification. Seeing texts as part of social events, F argues for a relational approach to text analysis. Ch. 3 (39–61) deals with intertextuality and assumptions, pointing out that ‘intertextuality is inevitably selective with respect to what is included and what is excluded from the events and texts represented’ (55). F distinguishes, among other things, three types of assumptions: existential, propositional, and value. I agree with F that ‘making assumptions is one way of being intertextual’ (17). The ideological aspects of assumptions are also discussed.

In Part 2 (‘Genres and action’), Ch. 4 (65–86) looks at three types of genre: dialogue, argument, and narrative, attempting to link the analysis of genres to a number of themes in social research. For F, the individual genres of a text or interaction can be analyzed in terms of activity, social relations, and communication technology. In Ch. 5 (87–104), F, making use of a lot of ‘real’ examples, focuses on meaning relations between sentences and between clauses within sentences. Ch. 6 (105–19) distinguishes two types of speech exchanges (knowledge exchanges and activity exchanges) and three main grammatical moods (declarative, interrogative, and imperative).

In Part 3 (‘Discourses and representations’), Ch. 7 (123–33) focuses on discourses, in which F argues that discourses are ways of representing the world. F also talks about how to differentiate discourses. Ch. 8 elaborates on representations of social events. In Part 4 (‘Styles and identities’), Ch. 9 (159–63) looks at styles, while Ch. 10 (164–90) discusses modality and evaluation, focusing on categories of explicit and implicit evaluation and showing how these two analytical perspectives can be used to address a range of social issues. Ch. 11 (191–211) summarizes the major ideas in the preceding chapters in the form of questions and presents a short ‘manifesto’ for the critical discourse analysis research program.

At the very beginning of each chapter there are boxes to highlight the important text-analysis issues and social-research issues covered in the chapter, and at the end of each chapter there is a summary. The book ends with glossaries of key terms and key theorists, an appendix of texts, references, and an index. In this sense, the book is very reader-friendly. Another strong point is that F makes use of a variety of ‘real’ language data (advertisements in newspapers, interviews, personal speeches, radio news reports, TV debates, etc.). This book should be welcomed by those looking for ways to analyze real language data without neglecting the social outlook. However, I find it odd that Ch. 9 on style is unusually short, as the essential role of style should never be underestimated. The second thing I would like to point out is that although discourses are ways of representing the world, as discussed in Ch. 7, the world represented by discourses is not necessarily the one we have. Third, is it really easy and necessary to make a distinction between text analysis and discourse? Finally, one has to wonder what is meant by ‘real language data’, especially since errors and verbal missteps abound in actual usage.