기사 헤드라인의 한-영 기계번역 고찰 - 헤드라인의 언어형식 측면을 중심으로

Abstract

The purpose of this study was to examine problems in machine translation output of newspaper headlines with a focus on linguistic features, in comparison with human translation output. For this purpose, 200 headlines from Infostock Daily (machine translation) and 200 headlines from Pulse, Maeil Business Newspaper’s English-language daily (human translation), were collected and analyzed in terms of five different linguistic features between Korean and English newspaper headlines. The analysis results showed a range of problems in machine-translated headlines. Ellipsis, quotation marks, commas used instead of subject case markers, and verb-less noun phrases – a common strategy for Korean newspaper headlines – negatively affected readability of Infostock Daily headlines, while unnecessary be-verbs, present perfect tense to report past events, and ‘will’ for future tense – a feature not common in English newspaper headlines – were frequently found in Infostock Daily. All in all, these findings suggest the need for light post-editing of machine-translated headlines, providing some useful insights into the post-editing guidelines for newspaper headlines.