Analyses De Contenu Et De Discours Dans La Recherche Sociologique Française: Pratiques Micro-Informatiques Actuelles Et Potentielles

Abstract

Since most French sociologists are unfamiliar with the acronym CAQDAS, this paper explains the logic behind this research practice that involves methods of analysis of textual data, both computer-based and in the process of being computerized. An initial awareness is needed of the debates regarding the choice of research approaches in sociology and the diversity and specificity of methods currently being used in the domain of textual data analysis in France.

In general, the influence of the French socio-linguistic tradition looms large. Readers will find, on the one hand, the older works of Michel Pécheux on the linguistic or discursive formations inspired by Marxist-Structuralism as well as Pécheux's 1969 AAD (Analyse Automatique du Discours) method; on the other hand, there is the perspective of the Ecole Française d'Analyse du Discours (AD) in Langages (see the 1969 and following issues), also expounded in the journal Langage et Société since 1971. Such a tradition - and here we are speaking of various versions of it, though all refer to the speech act concept (problematics and enunciation) - places an emphasis on the processes and relations pertaining to socio-discursive or socio-enunciative types of action of socially-located speakers, and simultaneously builds on a methodology akin to a classic, `theme based' content analysis - a methodology that is not (as yet?) translated into an operational software supporting its very specific requirements.

Having explained the context, the paper now addresses the major developments - theoretical or methodological - currently having an impact on practices of computer-based analysis of textual data in France. These developments take the following approaches:

lexicometric, using either conventional factor analysis or automatic data classification procedures such as those of Benzecri;

socio-semantic, using various statistical methods customized for a given analysis or conceptual grid;

specified in terms of networks of associated words, a perspective that already has available for its purposes a well-known program for codification and investigation of bibliographical databases;

representative of a propositional and predicative analysis of discourse, akin to paradigms in psycholinguistics;

representative of a kind of textual engineering;

offering sets of software specically designed for the management of sociological surveys and observations but very expensive and thus only accessible to public organizations - yet to be approved by `discourse analysts' who have a reputation for being very critical of computer technology;

representative of system and other experts working on research questions and specific theoretical frameworks, such as studies on `la grammaire des actes de civilité' and `la dissonance axio-idéologique'. In our view, these systems represent the most promising future methods of textual data.