Exploring Discourse Corpora Using Process Mining Techniques

In our paper we will introduce and discuss Process Mining (PM) as a means for conducting conversational analytic, linguistic and rhetorical investigations into discourse processes. PM is a technique for automatically deriving and further analyzing process models from event data. It is mainly applied for the analysis of Business Processes. We will argue that conversations and other kinds of discourse can be treated as processes too, and that PM enables us to systematically investigate large quantities of discourse transcripts. It would be possible to examine many different linguistic research questions, such as examining the conditions of turn-taking in dialogue or the pragmasemantics of discourse particles like English "well"or German "halt", among others.