Corpus-based Extraction of Chinese Historical Term Translation Equivalents.
Xiuying Li, Chao Che, Xiaoxia Liu, Hongfei Lin, Rongpei Wang · 2010
By closely examining the translation data of Chinese historical terms in some widely acclaimed English versions of Chinese historical classics, this paper has discovered a multi-equivalence phenomenon in term translation due to the polysemy of a source term and the synonymy of its potential equivalent concept in the target language as well as the different translation orientations by different translators in accordance with their considerations for the semantic and pragmatic dimensions of the translated term in its context for a particular readership. These findings exert a great challenge to the widely accepted basic assumptions underlying many statistical automatic terminological alignment models and the greedy algorithm applied to extract a valid translation pair from a bilingual parallel corpus. Therefore, this paper proposes to allow for the possibility of a multi-equivalent relationship between a source term and its potential target translation in the extraction algorithm design so that valid equivalent pairs would not be ruled out due to the exclusiveness of the algorithm.