MACHINE TRANSLATION IN LAW DOCUMENTS. RISKS AND CONSIDERATIONS

This paper explores the phenomenon of the widespread usage of machine translation. While this procedure may save a considerable amount of time in determining meanings between L1 and L2, it may be faulty depending on what languages it involves. This is an especially problematic situation if we consider a text that needs to represent the same pragmatic meaning both in L1 and in L2, such as, for example, a law document.This research is especially valuable because it lays the necessary groundwork to examine machine translation (from now on, MT), to outline and define the possible risks and legitimacy of the application of MT to documents pertaining to the law sphere. This article argues the legitimacy of unsupervised technology in the modern world, in order to determine whether or not they present a reliable, valid alternative to CAT Tools and to the art of professional translation. It also tries to determine the nature of the errors, if any, that MT offers in the suggested target text. This will be achieved by translating sections of the Russian Constitution [31] into English, through the lenses of different MT tools; the end result will be evaluated basing on how well preserved the original meaning in L1 is kept in the target text. We will also focus on pondering the difference between CAT tools and MT, on the method used to achieve this goal, and we will draw likely conclusions at the end of our analysis. The research main point of this study focuses on understanding how reliable MT is in a world that constantly shifts and changes through technology, and that may believe the art of professional translation to be perfectly replicable without exerting particular effort.Here we will consider the Russian Constitution as our prime example, seeing as it is a founding document for the country. We will first explain the difference between CAT tools and MT. Then we will shed light and justify our method of choice, the one of fidelity to the source text. We will then examine the articles of the Russian Constitution through three different MT tools, and we will determine the nature and the scope of the mistakes found in translation. We will employ Google Translate [39], Yandex Translate [40], and Bing Translator [41], three of the most used and best regarded free access MT tools available to the public. We have made this choice exactly because of their diffusion and advanced vocabulary, which makes them prime candidates in the eyes of individuals to translate any kind of document.Our findings and remarks, which constitute the closing part of the article, are as follows: MT is a valid instrument for many kinds of documents, but due to its lack of flexibility related to different language typologies, and due to its impossibility totranslate words with specific, culture bound meanings, it is to this day not suited for a translation process involving Russian and English.The paper focuses on a cultural, technological, and epistemological phenomenon that embraces the world at large, but that is not often considered as a source of comprehension difficulties.