PARSING CHINESE NOMINALIZATIONS BASED ON HPSG

A Chinese sentence parsing system based on Head-driven Phrase Structure Grammar (HPSG) is proposed in this paper. It is designed as a module of CEMAT, a Chinese-to-English MAchine Translation system. Basically, it is a bottom-up data-driven chart parser with unification as its primary operation. We augment the unification process with structure sharing. The phrases or clauses included in this paper are nominalizations. A new feature func is proposed to specify declaratively the properties of the sentence constructions. Also a function-rule-firing mechanism is proposed to process constructions involving functional words, such as de5. Incorporating with Subcategorization Principle, Adjunct Principle, and Head Feature Principle, our parser is able to parse a lot of sentences.