earticle

영어

The paper analyzes differences in lexical bundles between translational and original Korean as identified by a 3-gram analysis of an English-Korean newspaper/magazine translation corpus and a comparable Korean news corpus. The n-gram analysis, with a cutoff of 20 times per million words, produced 84 and 44 lexical bundles for the translation and original Korean corpora respectively after excluding patterns containing numbers and proper nouns. A majority of the multiple word sequences related to verbal expressions expressing reported speech and modality of surety and ability. A comparison of the two lists revealed some patterns present on one list but absent on the other or with a marked difference in frequency. A further investigation of some of these patterns provided evidence indicating that the translational corpus exhibited some of the translational universals suggested in the literature, namely source language interference or shining-through, simplification and under-representation. The study also suggests that n-gram analysis can be a highly useful tool for corpus-based translation research.