Lexicon-Driven Transfer In English-Chinese Machine Translation

This paper identifies the major differences between English and Chinese due to lexical idiosyncrasy and describes a proposed mechanism for bridging the differences in the generation phase of an English-Chinese machine translator. The method involves uniform intermediate representation for clauses and noun phrases, a minimal set of transfer operations and an active bilingual lexicon that encodes the needed transfer. Using the method, we are able to deal with lexical idiosyncrasy in both languages in a modular and efficient manner.