earticle

영어

This study explores how Korean-to-English machine translation systems (KEMSs) (e.g., Google Translate and NAVER Papago) deal with Korean relative clauses (RCs). Previous works have supported the claim that case markers in Korean take critical roles in planning Korean RCs. Cross-linguistically, Korean and English RCs are different in terms of the positions of the head nouns and gap positions in unfolding clauses. Because of the transparent case marking differences among types of Korean RCs, this study performed the translation processing of various Korean RCs from reliable publications in the fields of language and linguistics. The findings show that differences between the two languages are not fully reflected in KEMS. In particular, when KEMSs translate Korean direct and indirect object RCs, many errors are generated because of their insensitivity to the Korean case marking system.