An Automatic Dialogue Representation with Reference to Real Spontaneous Korean Dialogue

An Automatic Dialogue Representation with Reference to Real Spontaneous Korean Dialogue

This paper argues that in order to deal with linguistic phenomena occurring frequently in Korean dialogue and construct its representation structure appropriately, the following information must be included in the representation structure: information about dialogue participants such as the speaker and the addressee of each utterance, information coming from utterances themselves, and information flow among utterances in a dialogue. A dialogue representation structure is constructed by modifying Discourse Representation Theory (DRT). In addition, this paper demonstrates a way to implement Korean dialogue representation structures on the computer. Furthermore, this paper shows that the dialogue representation system proposed here is very successful in constructing representation structures for utterances occurring in a real spontaneous face-to-face Korean dialogue.