심층 이해와 번역 - 학부생 번역과 기계번역을 중심으로 -

Abstract

Machine translation based on AI uses sentence as the unit of translation, not words or phrases, and by considering the correlation between each translation element within a sentence, it is possible to achieve more natural and accurate translation. However accurate meaning understanding is possible through inference because human language have many hidden messages. So translation requires an in-depth understanding. This study was conducted on three cases(metaphors and idiomatic expressions, pragmatic meaning, cognitive context) that require a deep understanding of the original text, targeting literary translation. Undergraduates have little awareness of deep understanding, so they mostly did literal translations, which did not differ much from the results of machine translation. However, after deep understanding was achieved, they showed clear differentiation and competitiveness from machine translation by doing creative and individualistic translations. Machine translation can sometimes produce fairly accurate translations due to the ‘Attention’ feature. However, this is limited to some cases and does not mean that it produces 100% accurate translations. The following conclusions can be reached when the above review results of this paper are summarized. First, there is little difference from machine translation and human translation without professional recognition of translation. Second, due to the limitations of machine translation there is still a human-only domain in foreign language communication. In this way, the field of foreign languages is a place where both crisis and possibility coexist, and bold innovation and new education are necessary to secure competitiveness superior to machine translation.