A Multilevel Approach in the Study of Talk-In-Interaction

The beginning of research on the functioning of conversations and other forms of talkin-interaction can be dated back to the late 60's, a period where this subject became a more or less autonomous field of research. This research has now reached maturity (provided that one accepts this anthropomorphic metaphor...). Although I do not intend to give a 'state of the art' survey, I would like to consider here some of the recent developments in interactional linguistics. A preiiminary question is: given a definition of this field as covering any study approaching whatever form of talk-in-interaction on whatever perspective, what conventional term should be used for labelling this field? French literature commonly uses'analyse des interactions verbales'2 (in English: 'verbal interactions analysis', that is 'ViA'). It is doubtless, anyhow, that in our perspective 'conversation analysis' is too restrictive, for two reasons: considering the object of investigation, 'conversations' are only one of the numerous different types of verbal interactions (even though one can admit it to be prototypical)3; considering the methodological aspects, 'CA' refers to a particular approach which is well defined from both a historical and methodological point of view; but interaction analyses are also based on other descriptive traditiois than ethnomethodology let us quote, among others: symbolic interactionism (Goffman), the ethnography of communication (Hymes), interactional sociolinguistics (Gumperz), discourse analysis as represented by Labov & Fanshel, or Sinclair & Coulthard ('school of Birmingham'), and more recently by the 'school of Geneva' (the so cal led 'hierarchical model ' elaborated by E. Roulet). . . Verbal interactions analysis, when considered extensively, has been characterized from the very beginning by an extreme variety of the recommended approaches. This variety was later both reduced (since CA soon became the 'hard core' of this field in