Dialogue Corpus Construction Considering Modality and Social Relationships in Building Common Ground

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 4088–4095 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 4088 Dialogue Corpus Construction Considering Modality and Social Relationships in Building Common Ground Yuki Furuya†, Koki Saito†, Kosuke Ogura†, Koh Mitsuda‡, Ryuichiro Higashinaka‡, and Kazunori Takashio† †Keio University, Japan ‡NTT Corporation, Japan †{furuyan, t19340ks, t19177ko, kazu}@sfc.keio.ac.jp ‡{koh.mitsuda.td, ryuichiro.higashinaka.tp}@hco.ntt.co.jp Abstract Building common ground with users is essential for dialogue agent systems and robots to interact naturally with people. While a few previous studies have investigated the process of building common ground in human-human dialogue, most of them have been conducted on the basis of text chat. In this study, we constructed a dialogue corpus to investigate the process of building common ground with a particular focus on the modality of dialogue and the social relationship between the participants in the process of building common ground, which are important but have not been investigated in the previous work. The results of our analysis suggest that adding the modality or developing the relationship between workers speeds up the building of common ground. Specifically, regarding the modality, the presence of video rather than only audio may unconsciously facilitate work, and as for the relationship, it is easier to convey information about emotions and turn-taking among friends than in first meetings. These findings and the corpus should prove useful for developing a system to support remote communication. Keywords: dialogue, corpus construction, voice/video chat, social relationship, collaborative task, common ground 1. Introduction In dialogue, information such as knowledge and beliefs that is shared among the participants is called common ground (Clark, 1996; Traum, 1994). Building common ground with the user in dialogue is essential for im- plementing dialogue agent systems and robots that per- form natural interaction with humans. Research that explicitly considers the common ground has been reported (Lee et al., 2011; Kiesler, 2005; Moratz and Tenbrink, 2006; Chai et al., 2017). These studies have analyzed the common ground in dialogues where workers collaboratively accomplish tasks. In addition, some studies have examined the process of building common ground in text chat (Udagawa and Aizawa, 2019; Mitsuda et al., 2022; Bara et al., 2021). Although nonverbal cues and relationships between interlocutors are essential in the process of common ground building, they have not been analyzed in depth. As nonverbal behavior between interlocutors is essen- tial for cooperative dialogue, nonverbal cues need to be taken into account (Carney and Harrigan, 2003). For example, nonverbal behaviors such as shaking one’s head without speaking can have a facilitating effect on the dialogue with the other party (Kleinke, 1986). The difference in modality between voice and video calls leads to a difference in the results of collaborative tasks through dialogue (Tomprou et al., 2021). By extending text chat to voice chat and video chat, interlocutors can smoothly build common ground based on the informa- tion in the other person’s voice and facial expressions. It is also crucial to consider social relationships, as the …

Content truncated to 500 words