Comparative Study of Human Translation and Google Translation: Error Analysis of English into Pashto

Abstract

This paper aimed to compare Human translation with the machine translation, Google Translate. The primary purpose of the current study was to identify the types of errors made by Google Translate when translating sentences from the English-to-Pashto translation, of the book "The Pathan" written by the Pashto Poet Ghani Khan Baba. The data includes 50 random sentences that were translated into Pashto using Google Translate and were taken from the book "The Pathan". A mixed research methodology is employed to compare the capabilities of Human and Google translation in the context of Pashto language. The methodology of this proposed study is based on the research of Dzakiyyah (2017), who adopted the criteria of accuracy, acceptability, and readability suggested by Nababan (2012), as well as on the study of Al- Samawi (2014), who used a work sheet that categorises common translational errors. It was found that majority of the sentences were translated incorrectly by Google Translate. In addition, most of the sentences contained multiple types of translation errors. Only seventeen sentences were error-free. The remaining sentences, however, all contained one or more errors such as semantic error, grammatical error, syntactic error and translational error. After this study, it is claimed that Google Translate should not be considered the ultimate translation tool. Besides, human translation was found more accurate than the machine translation.