The project is motivated by a shared demand observed between translation technology and corpus-based translation studies. The former traditionally prefers parallel corpora but in reality often needs to work with comparable corpora, and recent studies have shown that sentiment knowledge could improve machine translation performance. The latter unavoidably deals mostly with comparable corpora, whether it is for contrastive linguistic analysis, or comparative analysis of one or more translations of a literary work. Corpus tools for this kind of analysis would also benefit from more semantic information to go beyond the lexical level. This project thus aims at building a prototype English-Chinese comparable corpus with sentiment annotation for pilot explorations of the potential applications of bilingual sentiment corpora in various areas of translation-related research.
|