Direcionalidade em tradução: o papel da subcompetência bilíngue em tarefas de tradução L1 e L2

Abstract

This dissertation is linked to the empirical-experimental studies previously carried out at the Laboratory for Experimentation in Translation (LETRA) at Faculdade de Letras of Universidade Federal de Minas Gerais, Brazil. It investigates cognitive aspects observed in the translation process of eight professional translators while translating texts from a foreign language (English) into a native language (Portuguese) L1 translation and from the same mother tongue (Portuguese) into a foreign language (English) (L2 translation). The triangulation of translation process data methodology (JAKOBSEN, 1999; ALVES, 2001, 2003), was applied in order to combine different tools of data elicitation to investigate the translation process, in a complementary perspective. The quantitative analysis of translation process data drew on linear representations generated by means of keylogged files recorded with the software Translogc, and analyzed in conjunction with prospective questionnaires that provide information about the translator's profile, and the retrospective protocols that were recorded after each task. The direction of the translation (L1 and L2 translations) was the independent variable. The dependent variables are: i) time spent in each stage (orientation, drafting, and end-revision) and time spent in the task as a whole (JAKOBSEN, 2002; ALVES, 2005), (ii) pause patterns, (iii) recursiveness (BUCHWEITZ; ALVES, 2006), (iv) segmentation (DRAGSTED, 2004; RODRIGUES, 2009) and (v) retrospective protocols (PAVLOVIÆ, 2007). Data collection was carried out in two separate stages: data collection 1 and data collection 2. In data collection 1, participants translated two correlated scientific texts on sickle cell disease. Task order played an important role in the overall results, revealing the facilitation effect while carrying out the second task. In this sense, the study was replicated in data collection 2 and the participants translated non-correlated texts in an attempt to observe more clearly the impact of the directionality (L1 and L2) in the translation tasks. As a result, it was verified that the L2 translation presents a different configuration in comparison to L1 when the facilitating effect was eliminated or at least decreased when carrying out the second task. In this sense, the quantitative data pointed out that the L2 translation demanded more effort than the L1 translation task. This was corroborated when the qualitative data provided by the retrospectiveprotocols were analyzed, based on an adaptation of Pavloviæ (2007)'s study. The distribution of the protocols during the tasks pointed out that, when translating non-correlated texts (data collection 2), the access to the concepts and the lexical production are more limited during the L2 translation in comparison to the L1 translation, leading to two different models for explaining the distribution of the linguistic knowledge in the cognitive environment of the translators. During the L2 translation, lexical knowledge is not connected to the other types of knowledge with the same strength as they are in the L1 translation, resulting in more lexical problems and less solutions during the translation process. It was verified that the relation between those types of knowledge, which is connected to the level of bilingual knowledge of the participants, is a fundamental aspect during the translation process, and it presents more impact while translating into the non-mother language.