Translation Evaluation of Three Machine Translation Systems, with Special References to Idiomatic Expressions

Abstract

One of the major challenges posed by machine translation (MT) when translating between English and Arabic is idiomatic expressions since their meanings are not taken literally but have a cultural meaning behind them, and they are “all in all metaphorical and cannot be understood directly” (Ghazala, 2003). Therefore, the main aim of the present study is to evaluate and compare the performance of three online MT systems namely, Google, Microsoft Bing, and Systran in translating English idiomatic expressions into Arabic in online newspapers. To achieve the goals of the present study, twenty idiomatic expressions which are the most frequently used in six British and American online newspapers have been identified, fed, discussed, compared and analyzed qualitatively by the researchers. Findings of the present study show that although the three online MT systems were not able to provide accurate translations of the English idiomatic expressions into Arabic in six online newspapers, Microsoft Bing machine is rated the best and has a better performance. Whereas Googles’s translation is found the least accurate as some words were not translated at all, and others were mistranslated. The study concludes that MT systems can be improved to get a better quality and provide an accurate translation if they deal with the idiomatic expressions as a unit and not individually. Since these are fixed idiomatic expressions, they are usually resolved using a simple memory translation, which sets side-by-side the English and the Arabic expressions without much cognitive effort.