Extraction of Paralinguistic Information by Means of Speech Analysis of Teacher's Utterance in Class

Abstract

The aim of this research is to extract the paralinguistic information from the natural conversational speech using Maekawa's experimental method. The results of the experiment are as follows: (1) Maekawa's experimental method is also effective in extracting categories of paralinguistic information from spontaneous speech. (2) The categories of characteristic paralinguistic information extracted have not been defined in traditional categorical analysis. These categories are characterized by the tendency not to directly instruct, warn or hurry children, but rather to use softer, less direct expressions that take the child's independence into consideration. It may in the future be possible to develop category systems for speech analyses of class conversations.