Discourse analysis for considering emotional effects on others
Hiroki Yamaguchi, Yukio Ohsawa, and Yoko Nishihara · 2010
We developed a system for extracting Discounted Utterances(DUs) and evaluating them. We defined the DU as “Utterance which was ignored by others in a conversation in spite of its potential.” We aim to mine participants' unique stories and specialized knowledge based on their experiences from DUs. The proposed system is designed for considering emotional effects on others in conversations. We analyzed a meeting of toilet design as a case study. Each participant here has his/her own specialty. The system extracted DUs as triggers for mining the unique stories. We employed double helical model in Chance Discovery to the system. The study shows the significance of the system, and makes it clear that the DU in this paper is not just ignored in the conversation and includes speakers' unspoken stories based on speakers' own experiences. These stories are useful for the development of systems. Besides, some of DUs indicated the sense of existing values. By utilizing DUs, the system can also provide opportunities to improve existing systems.