A Methodology for the Identification of Topic Transitions in Interaction

Many analysts use the notion of “topic” to describe segments of discourse, whether written or spoken, and many regret that even though a considerable amount of work has been devoted to understanding and defining the notion of discourse topic, the analyst is often left without a robust methodology to apply to their own data. This paper uses a mixed-methods approach that draws on the qualitative-oriented theoretical frameworks of Conversational Analysis and Interactional Linguistics, and combines them with quantitative methods used in other sub-fields of linguistics, such as the coding schemes and inter-rater agreement measures used in Corpus Linguistics. The goal of this study is not to provide a new understanding of “topic”, but rather to propose 1) a rich definition compiled from various earlier studies and suited to the analysis of talk-in-interaction, and 2) a systematic way to apply it to new data. The ambition of the paper is to provide a methodology for a qualitative, quantitative, or mixed-methods analysis of topic-related phenomena in interaction, as one step among many in various research protocols. It provides empirical grounds for the claim that a research methodology based on the analyst’s intuitions may be a valid and robust way to identify topic transitions – if a number of precautionary steps are taken. It also proposes a practical guide to the systematic analysis of topic in interaction.