Literary Machine Translation: Towards a Post-editing Tool

Abstract

Given the remark able development that took place in machine translation (MT) in recent years, people all over the world tended to rely on MT since it provides a faster result than human translation (HT) and saves effort. Even with these merits, there is still a difference of opinion on whether MT output has reached human equivalence. This study was provoked by the following question "Does post-editing (PE) improve MT output?" The study attempts to answer the question by providing a theoretical approach to PE methods, with practical examples. Data was taken from the Arabicnovelفرانكشتاين في بغداد,Frankenstein in Baghdad and the application of Google Translate (GT) was chosen for the automatic translation of the data. A professional translator with experience in literary translation was asked to do the task of PE Google output. Then an analysis of the grammatical and semantic differences - that have been found between GT and the PE version- was introduced. The research findings showed that conducting PE enhanced the efficiency of GT output, especially on the semantic level where the percentage of PE on the meanings (85%) was much more than on grammar (15%).