1 When is a Text? (Keynote Paper)

Each discipline or theoretical tradition in the field of discourse analysis takes up a particular set of relationships around discourse and analysis, text and context, text and commentary, etc. These in turn construe or constitute the field in specific ways and produce specific categories of analysis. What is not often attended to, however, are the terms of these determinations. For example, how is a text defined? On what basis is it selected? What are the possible relationships between a corpus of data, and the production of specific categories or units of analysis? What is the relationship between the word and the world? When is a text? This paper asks questions of the conditions under which specific categories of analysis such as „text‟, „context ‟ or „discourse ‟ are determined within specific traditions of discourse analysis, the epistemological assumptions underpinning these, and the effects. It argues that these units are artifacts of the analytic disciplines themselves and that the terms and conditions of their production need to be reflexively accounted for. It takes up Bourdieu‟s framing of the problem of limits, in order to demonstrate major epistemological challenges within the project of discourse analysis that continue to require attention. It then considers a brief personal history of practical engagement with the theoretical questions of text-context and txt-commentary relations within discourse-analytic research The motivation for this paper is the pedagogical imperative of working with doctoral students and other researchers who seek to understand and use the tools of discourse analysis in a range of sites of research practice. It seeks to move discussion to pragmatic considerations concerning how, when and according to what principles texts come to be constituted through selection as objects for inquiry and analysis