Abstract

ABSTRACT Comparative discourse analysis emerged in France, more precisely in the CEDISCOR laboratory of the Sorbonne Nouvelle University – Paris 3, around the year 2000. Currently, it is being implemented in Brazil within the Diálogo group (CNPq/USP). The purpose of this article is to reflect on the theoretical and methodological possibilities involved in the practise of Bakhtinian-inspired comparative discourse analysis in Brazil. To this end, we start from a theoretical reflection on the notion of category (of analysis). We investigate, on the one hand, how the notion of category was already present in the creation of CEDISCOR and comparative discourse analysis in France, when one sought to determine “comparable categories”; on the other hand, we show how the creation of a science by Bakhtin – metalinguistics – also starts from a revision of the notion of linguistic category, which has implications for the way Bakhtinian-inspired comparative discourse analysis can be practised in Brazil.