한영기계번역을 위한 았다 구문의 처리 방법

Abstract

This paper aims to introduce the processing methods of issta constructions for Korean-English machine translation. Semantic ambiguities as well as many sentence patterns of issta constructions in Korean make them difficult to automatically translate into English ones (target language). My claim, in this paper, is as follows. 1) We need to differentiate issta that functions as predicate and one that does not. 2) We need to solve processing issta ambiguity in terms of using the case frame of issta constructions. 3) To analyze issta constructions, after searching features of predicate and analyzing case markers followed by arguments of the construction, we write its results in the form of table (Proposition Table). 4) Regarding multiple nominative (ka) constructions, we first search the dictionary information of the nouns following ka. In doing so, if identified as nouns [Possessor], then we assign [Agent] features to the nouns following ka; and if identified as nouns [Possessed], then we assign [Theme] features to the same ones. We think, the ways of dealing with semantic ambiguities and case analysis of issta constructions can surely be extended to other types of similar constructions in Korean.