Multimodal analysis of "well" as a discourse marker in conversation: A pilot study

Dec 2, 2013·

G Esfandiari Baiat

Matt Coler

Marissa Pullen

Sinenaat Tienkouw

László Hunyadi

· 1 min read

Cite IEEE Xplore

Abstract

Discourse markers, a group of pragmatic elements, are used widely in our daily communication. Among these markers, ‘well’ is found to be the most frequently used one in conversation. Although much research has been carried out on its theoretical bases such as its definition, functions and categorizations, few studies have taken a multimodal approach in analyzing this marker. The aim of the present study was to provide a multimodal description of the use of ‘well’ as a discourse marker in spontaneous speech. Firstly, we wanted to investigate whether the use of this marker in conversation, regardless of its function, was accompanied with any kind of non-verbal behavior (posture shift, head shift, gaze and eyebrow movement) and secondly, we wanted to determine whether any of these non-verbal cues (if any) were specific to one pragmatic function of ‘well’ and if they could be used as a signal to disambiguate the functions.

Type

Conference Paper

Publication

2013 IEEE 4th International Conference on Cognitive Infocommunications (CogInfoCom)

This pilot study takes a multimodal approach to analyzing the discourse marker “well” in spontaneous conversation. While “well” is one of the most frequently used discourse markers in English conversation, most research has focused on its linguistic and pragmatic functions without considering the accompanying non-verbal behaviors.

Our research examines whether specific non-verbal cues—such as posture shifts, head movements, gaze patterns, and eyebrow movements—consistently accompany the use of “well” in conversation. Furthermore, we investigate whether different pragmatic functions of “well” correlate with distinct non-verbal behaviors, potentially offering visual cues that could help disambiguate the marker’s various functions.

Using video data of natural conversations, we conducted a detailed analysis of both verbal and non-verbal components when speakers used “well.” This multimodal approach provides insights not only for discourse analysis but also for applications in human-computer interaction, where understanding both verbal and non-verbal aspects of communication is essential for developing more natural conversational systems.

The findings contribute to a better understanding of how verbal and non-verbal communication channels work together in conversation and have implications for improving multimodal dialogue systems.

Last updated on Dec 2, 2013