AWEJ. Special Issue on Translation No.3 May, 2014 Pp. 182-211
Language Errors in Machine Translation of Encyclopedic Texts from English into Arabic: the case of Google Translate
Ahmad Muhammed Al-Samawi
College of Education, Sana’a University, Yemen
Al Ain University of Science & Technology, UAE
Machine translation has facilitated the way for professional translators as well as ordinary people. Google Translate is undoubtedly the most popular machine translation program today. However, when translating texts from English into Arabic, the language errors that are produced by Google Translate are shocking, especially in encyclopedic texts. This article aimed at examining Google Translate errors in 10 encyclopedic texts of 10 sentences each. These texts were taken randomly from the free encyclopedia (Wikipedia) to represent 10 different academic disciplines. They were inputted into Google Translate to convert them from English into Arabic. The researcher traced 366 language errors in these texts (3.66 errors per sentence). Based on error analysis, these errors were categorized into: syntactic errors (55 errors), grammatical errors (174 errors) and semantic errors (137 errors). These errors are signaling alarm to Google Translate team to improve the program of translation from English into Arabic and to find solutions to its current deficiency.
Keywords: English-Arabic translation, Google Translate errors, language errors, machine translation errors