Referent resolution for zero pronouns in Thai

1. Introduction Resolving zero pronouns is a major problem in developing a natural language understanding (NLU) system for Thai. Since subject and object pronouns in Thai can be omitted from a sentence, an NLU system must be capable of identifying the missing subjects or objects in the sentence. This process of identifying referents for zero pronouns, which is a part of referent resolution 1 process, is the concern of this paper Basically, I assume that referent resolution for zero pronouns can be done at two levels: the sentence level and the discourse level. Some zero pronouns can be resolved on the basis of sentence grammar principles. These principles are implemented as a part of syntactic/semantic parser. As for zero pronouns which cannot be resolved by a sentence grammar, discourse principles will be used. The sentence grammar that is adopted in this paper is that of government and binding theory (Chomsky 1981, 1982, 1986a, 1986b). In this framework, zero pronouns are analyzed as empty categories. An overview of empty categories and related principles are reviewed in section 2. In section 3, zero pronouns in different syntactic structures in Thai and the domain in which the government and binding theory is applicable are discussed. Section 4 is concerned with the centering theory which is adopted as the basic discourse principle for resolving zero pronouns at the discourse level.