Consistency in the Evaluation Methods of Machine Translation Quality

Abstract

Machine translation is the general term for the programs concerning the automatic translation with or without human assistance. It is also an interdisciplinary research area with different questions yet to be answered. One of the fundamental questions of the area is related to the quality assessment of machine translation outputs. In addition, despite to its interdisciplinary nature, machine translation is rarely the research topic for Translation Studies, which focuses more on CAT tools. This study intends to fill this gap in the literature by focusing the evaluation of quality of MT programs available between English and Turkish, namely Google Translate, Proçeviri and Sametran.The study aims at exploring the consistency among two evaluation methods of machine translation quality from English into Turkish: micro evaluation and macro evaluation.