Mapping Word Senses of Middle Ancient Chinese to WordNet

Word Net is a widely used lexical knowledge for semantic information processing. Besides English, a lot of languages in the world have their own word nets. However building a new word net always takes tremendous manual labor. In this paper we present a strategy to automatically construct a word net for middle ancient Chinese (Mid acWordNet) which is mapped into the Princeton Word Net (PWN). For each sense of a middle ancient Chinese word, we first extract the headwords of its gloss. Then we compute the semantic distances among the headwords and choose the one most close to the others as our mapping result to PWN. We find that about 35% senses of middle ancient Chinese are common with English and the automatic mapping achieves a precision of over 80%.