A Corpus-based Study of Translation Universals in Korean-English Literary Translation

Abstract

Translation universals hypothesizes that translational languages have conventional yet unique features of translations, distinct from those of non-translations, and these features are universally present in all translations regardless of language pairs. However, the verification of translation universals has been restricted mainly to genetically close language pairs, such as English and closely related European languages. To confirm its universality, this article investigated Korean-into-English translations in the literary genre, which have remained mostly unexplored in translation universals studies. By constructing an approximately 500,000-word comparable corpus the Comparable Corpus of American and Korean Novels — —this study tested four major sub-hypotheses: simplification, explicitation, normalization, and leveling-out. Setting aside the lexical density of simplification indices, the results from all the other indices mean sentence length, proportions of high- and low-frequency — words, normalized frequencies of connectives and lexical bundles, and standard deviations of lexical density and mean sentence length supported the translation — universals hypothesis, indicating its high generalizability.