Comprehension of Indirect Utterances in Japanese conversation: Comparison among Japanese Children, Foreign Learners of Japanese and Artificial Intelligence

Project Leader
MATSUI Tomoko (Tokyo Gakugei University)


Background and Purpose

In recent years the need for research into language development with respect to language communication that takes into account context-dependent understanding (pragmatics) has increased world-wide; The applicant for the present research project, together with other researchers, has organized special editions of international journals (Ifantidou & Matsui 2013; Matsui 2014) on the subject. Furthermore, the importance of comparative studies of development looking at different language cultures and their respective styles of communication has been indicated (Fitneva & Matsui 2014). At present, with methodologies emerging in an interdisciplinary field that includes linguistics, psychology and cognitive science, and with this kind of international trend as background, it can be said that the time has come for comprehensive research on acquisition focusing on the understanding of indirect speech acts.

Objectives and Methods

The frequent appearance of ambiguous expressions (expressions that can be interpreted in more than one way) and circumlocutions is a characteristic peculiar to conversation. In order to understand what a speaker has meant to convey, a listener has to use context to extract a meaning. In order to shed light on the question of how the faculty for using context to interpret the indirect speech acts of others is acquired, this research compares learners’ first language, their second language, and artificial intelligence. The following two points comprise the pillars of this research, the results of which will serve as an indicator for teaching that supports learners who have difficulty understanding contexts.

  1. By experimentally probing the understanding of indirect expressions which allow more than one interpretation (irony, circumlocution, etc.), this research will reveal first language speakers’ powers of understanding and learning mechanisms with respect to indirect speech acts. Furthermore, by drawing comparisons with speakers of first languages, the characteristics of conversational abilities that are required for understanding of indirect speech acts by speakers of second languages can be examined, and the learning mechanism can be probed.
  2. A calculation model of the process of understanding indirect expressions such as irony by artificial intelligence will be constructed, and that model will be compared with and measured against the acquisition of the human power of understanding indirect speech.