The Equivalence Problems Produced by Machine Translation on A Literary Text: A Study on The Indonesian Translation of Harry Potter: The Order of Phoenix

Abstract

Google Translate (GT) is a machine translator (MT) this is powered by neural machine technology (NMT) which can produce generally fluent textual translation in more than 100 languages with 60% accuracy. Studies have been done to measure GT ability in translating texts and the quality of its translation products. The results showed that GT’s products are generally satisfying but they have inadequacies at some level of translation aspects. This study aims at finding the problems of equivalence that resulted from the process translating a Harry Potter novel, a literary work that is rich with cultural words and complex sentences, using GT from the English language into the Indonesian language. Using a descriptive qualitative method, the current study examines the problem of equivalence based on Mona Baker’s theory to categorize the translation errors found in the GT output text. This study showed that literary texts are still problematic for GT in terms of words that two languages do not share and sentences that contain several points of view. This paper suggest that GT has to update its database of Indonesian lexicons and any MT output has to go post-editing process in order to ensure its readability and naturalness to its target readers. The implication of this study emphasizes the need to concentrate on translation training programs in the post-editing work.