Towards Multimodal Analysis of Dialogic Moments in Storytelling-based Discourse

Abstract

The key to an effective communication method arguably lies in its ability to facilitate moments of high mutual understanding (dialogic moments). It would, therefore, be useful to identify these moments and perhaps facilitate them. In this pilot study, we present a multimodal analysis of dialogic moments in storytelling-based discussions. We collected skin conductance, heartrate, speaking turns, relative body position, conversation transcripts, and subjective experience. This multimodal data corpus enables the computational study of these highly subjective moments and the potential creation of digital communication aids. Preliminary results show that there might be subcategories of dialogic moments that were previously unidentified.