earticle

영어

The present paper introduces a computational approach to analyzing translating style based on Principal Component Analysis (PCA), a multivariate statistical method popularly used to classify texts and attribute authorship. The paper describes merits of using PCA for exploring translating style and explains the typical procedure involved in performing a PCA analysis on corpus data. As a case study for illustrating this approach, the paper applies PCA to investigate stylistic differences between two English translations of a Korean novel, using most frequent 1-gram, 2-grams and 3-grams as linguistic features. The results reveal a marked distinction across all three PCA analyses between the two translated versions, which include differences in vocabulary diversity, reference chains, syntactic structure and use of general nouns in referring to fictional characters.