COMPARATIVE ANALYSIS OF MACHINE TRANSLATION TECHNIQUES PERFORMED BY SYSTRAN, O.TRANSLATOR AND M-TRANSLATE

Abstract

The article examines the accuracy of both free – M-Translate – and professional and paid machine translation systems – Systran, O.Translator – in the complex context of legal translation. The paper analyses machine translation performed by Systran, O.Translator and M-Translate on a large sample of English legal vocabulary items derived from a long text of a judgment issued by the Court of Justice of the European Union. Prior to this study, the same text was translated by a professional translator without MT; the human's translation was used as a reference for comparison with the translation performed by Systran, O.Translator and M-Translate machine translation systems. The study involved a statistical analysis of the corpus using Sketchengine. It was found that many of the words and phrases of the translation options selected for the above vocabulary items are mainly available in the EU case law databases in English, while Ukrainian versions of such lexical items are not available, and therefore it was impossible to select them manually without applying the translation and then analysis procedure. The article argues that the use of MT systems in the context of legal translation can be useful if the purpose of such translation is to provide a general familiarisation with the text; it is also emphasised that translation performed by MT systems can consistently provide a reasonable number of accurate translations of the types of vocabulary items that translators in this context often need to research before being able to translate them effectively. For this purpose, a sample of such vocabulary items was selected and presented in a table. It is concluded that MT systems in the context of legal translation can also be used to teach law students, since, in recent years, curricula are often based on the analysis of the EU legal framework and judgments of the European Court of Justice. In addition, MT systems can be used as an immediate countertext to dictionary items as a way of trying to identify the correctness of a translation.