Tracking conflict and emotions with a computational qualitative discourse analytic support approach.

Nikodem Rybak,Daniel J Angus

doi:10.1371/journal.pone.0251186

Nikodem Rybak, Daniel J Angus

Open Access

https://doi.org/10.1371/journal.pone.0251186

Copy DOI

Abstract

Accurate inferences of the emotional state of conversation participants can be critical in shaping analysis and interpretation of conversational exchanges. In qualitative analyses of discourse, most labelling of the perceived emotional state of conversation participants is performed by hand, and is limited to selected moments where an analyst may believe that emotional information is valuable for interpretation. This reliance on manual labelling processes can have implications for repeatability and objectivity, both in terms of accuracy, but also in terms of changes in emotional state that might go unnoticed. In this paper we introduce a qualitative discourse analytic support method intended to support the labelling of emotional state of conversational participants over time. We demonstrate the utility of the technique using a suite of well-studied broadcast interviews, taking a particular focus on identifying instances of inter-speaker conflict. Our findings indicate that this two-step machine learning approach can help decode how moments of conflict arise, sustain, and are resolved through the mapping of emotion over time. We show how such a method can provide useful evidence of the change in emotional state by interlocutors which could be useful to prompt and support further in-depth study.

Highlights

Human communication is a dynamic, complex and reciprocal process that relies on interlocutors encoding and decoding communicative acts via multiple modalities
In qualitative analyses of discourse, most labelling of the perceived emotional state of conversation participants is performed by hand, and is limited to selected moments where an analyst may believe that emotional information is valuable for interpretation
In this paper we introduce a qualitative discourse analytic support method intended to support the labelling of emotional state of conversational participants over time

Summary

Introduction

Human communication is a dynamic, complex and reciprocal process that relies on interlocutors encoding and decoding communicative acts via multiple modalities. Advances in computational processing in the previous decades have motivated interest in the development of computational techniques and tools for aiding the analysis of conversational exchanges, many belonging to a family of techniques called Computer Assisted Qualitative Data AnalysiS (CAQDAS) [1,2,3]. Most CAQDAS techniques do not seek to replace the need for human judgement or interpretation, rather instead to augment human judgements [4]. We propose a computational affect detection technique, aimed at assisting the analysis of human emotion in communication. Emotions play essential roles in the development, preservation, and evolvement of important social constructs, an example being conflict, which arise from human-to-human interactions on various time scales.

Methods

Results

Conclusion