Métodos tradutórios do Google Tradutor na resolução de questões de ambiguidade.

Abstract

Machine Translation as a tool occupies an important space in the present time since the advance of the new technologies of communication. Google Translate as one of the most widely used automatic translation systems nowadays (GOMES & SANTOS, 2015), for example, has played an exceptional role in the statistical machine translation field, although, in this regard, it still prsents cases of lexical ambiguity due to the translated textual genre, despite the constant system updates. The objective, therefore, of this paper is to investigate to what extent the latest updates of Google Translate have is able to resolve lexical ambiguity issues in the translated texts. Thus some popular proverbs were translated via Machine Translation, since this textual genre can serve as a considerable source of lexical ambiguity. Therefore, this paper follow some methodological steps: (1) a survey was made on the most popular proverbs used in the UFCG academic environment in the form of a research questionnaire; (2) use of Google Translate system for translating popular proverbs into English; (3) use of COCA to indicate the incidence of terms used in the target language; (4) analysis and discussion of texts translated automatically. In view of the latest updates, Google translation system can use the following statistical translation logic: the more translated texts, the more the system uses statistical method. The fewer translated texts, the more it uses the direct method.