A parsing algorithm for Korean implemented in Prolog

Abstract

This paper presents a parsing algorithm for a partial freeword-order language, Korean. It is based on dependency grammar, unification-based grammar, and implemented in Prolog. The dependency grammar accepts free order of the words in a sentence. Unification-based features separate the grammar from the parsing algorithm, and as a result, the parser can easily be applied to parse other languages by modifying only the lexicon and the grammar. Also, the features check syntactic, morphological, and semantic dependency. The parser provides a structure of a natural language sentence of Korean with dependency pairs. It constrains totally free-word-order languages, recognizes subcategorization features, restricts the order of dependents for a single head, and matches long-distance dependencies.