The translation evaluation of google translation

Abstract

Machine translation (MT) offers great assistantship when people have difficulties in understanding or comprehend text of their non-native language both in spoken and written language. One kind of machine translation is Google Translation that can be got on their hand just by using mobile phone. This machine can help people to translate the text from one language to another one, in short time. This study aims at finding out the translation procedures found in the translation product by Google Translation and analyzing the errors occurred in Google Translation product. The data of this study are in the form of written data taken from a bilingual children story book of archipelago legend, which consists of 100 lines. There is a source language text in Bahasa Indonesia which is translated into English by using Google translation. The texts are chosen as it contains several sentences that are constructed by paying attention on the rhyme of ending sounds. The data of this study were analyzed through qualitative method. The translation results were analyzed based on translation procedures proposed by Vinay and Dalbernet (1989). Based on the analysis, it can be concluded that translation procedures used by Google Translation found in the data source are borrowing, calque, literal, transposition, modulation, equivalence, and adaptation. The most dominant type of translation procedure is literal translation. It has the highest frequency data; with a total of 50 data (50 %). Some errors are found in the result of Google translation. They are the errors in the use of pronoun, the use of plural and singular noun and inconsistency in the translation of name. Those errors are caused by limitation of machine in detecting the context of the text. Thus, the translated text is incoherent.