Junji Eto, Satoru Ikehara, Naoshi Ikeda, Yoshihiko Nitta, Katsumasa Shibata, Masahiro Miyazaki & Satoshi Shirai, IPSJ SIG Notes, May 26-27, 2003

Semantic Classification of Conjunctive Expressions based on Semantic Typology

Junji Etoh^*1 Satoru Ikehara^*2 Masashi Saraki^*3 Masahiro Miyazaki^*4 Takashi Ikeda^*5 Yoshihiko Nitta^*6 Satoshi Shirai^*7 Katsumasa Shibata^*8

*1 Languageware Inc.
*2 Faculty of Engineering, Tottori University
*3 Nagasaki Junshin Catholic University
*4 Faculty of Engineering, Niigata University
*5 Faculty of Engineering, Gifu University
*6 College of Economics, College of Science and Technology, Nihon University
*7 Advanced Technology Co.Ltd.
*8 Faculty of Science, Fukuoka University

This paper describes the application of the Theory of Semantic Typology (TST) to the analysis of clause relations in complex and compound sentences. TST is one of the two principles of the new machine translation system under construction, through which semantically non-linear expressions can be extracted from the huge bilingual corpus and patternized by meaning. Another principle is the Analogical Mapping Theory (AMT), through which a set of meaning patterns of a source language is mapped to that of target language via common concept to both. Two principles set forth have been proposed by the special interest group of "Language, Cognition and Expression".

keyword:

Machine Translation, Semantic Typology, Analogical mapping, Complex Sentences, Logical Relation