TRANSLATION PERFORMANCE OF GOOGLE TRANSLATE AND DEEPL IN TRANSLATING INDONESIAN SHORT STORIES INTO ENGLISH

Abstract

The prevalence of machine translation systems has increased in recent years due to their accessibility and the high demand for translation services worldwide. While translation tools and systems are useful for several fields and genres, their reliability for literary works continues to be a subject of debate. Critiques are often directed at the inadequate quality of literary texts that have been translated by machines. This research aimed to examine the performance of two neural machine translation systems,Google Translate and DeepL, in translating Indonesian short stories in the book “Cerita-Cerita Jakarta”. This qualitative research was designed to assess the quality of machine translation in translating literary work. The translation errors category proposed by Koponen (2010) was used as the theoretical framework. The findings revealed several errors in the literary translation by Google Translate and DeepL. The translation errors included untranslated concept, omitted concept, and mistranslated concept. Challenges were encountered by both Google Translate and DeepL in translating cultural terms, onomatopoeia, abbreviations, idiomatic expressions, slang words, and address terms.