O processo de (re)tradução : um estudo quali-quantitativo baseado em um experimento no par inglês-português

Abstract

This dissertation, carried out at the Laboratory for Experimentation in Translation (LETRA), at Faculdade de Letras (FALE), Universidade Federal de Minas Gerais (UFMG), is affiliated to the discipline of Translation Studies, more specifically to process-oriented, descriptive translation studies (HOLMES, 1972). It builds on an experimental study to approach (re)translation, i.e., a task of text production in a target language based on two types of inputs, a text in a source language and previous translation(s) of that text in the target language. This study draws on Malta’s (2015) seminal research which posits a (re)translation model based on the results of an aggregate analysis of data on the allocation of cognitive effort and processing flow across the inputs provided (i.e., source text, previous translation 1, and previous translation 2) and the target text that emerged during the (re)translation processes. This study aims to test Malta’s (re)translation model based on a sample involving the English - Brazilian Portuguese language pair and based on a qualitative and quantitative analysis of the participants’ aggregate and individual data. To this end, data collection and analysis targeted the performance/behavior and the allocation of cognitive effort data of eight professional translators and eight translation students while performing a (re)translation task. The methods and instruments of data collection included: eye tracking, keylogging, and free and guided retrospective protocols. The texts used as input were an excerpt from Jane Austen’s novel Emma (1815) and two of its published translations in Brazilian Portuguese. Aggregate and individual analyses targeted the participants’ distribution of visual attention, processing flow and allocation of cognitive effort based on metrics used in translation process research, namely: number of visits; percentage to total task time; visual transitions across areas of interest; and fixation counting, average duration and total time. The results partially corroborate Malta’s model for aggregate analysis of the data; however, the qualitative analysis of the non-aggregated data pointed to translator behaviors absent or even unaccounted for in his model. The aggregate analysis showed that visual attention and cognitive effort are higher for processing the target text, followed by the source text and the previous translations in this order, which is consistent with Malta (2015). However, the non-aggregate analysis pointed to the need to refine the (re)translation model, which originally starts from the source text; based on data of five participants, the previous translations proved to be as relevant as the source text in the information processing flow for (re)translation purposes. The results have the potential to shed light on both (re)translation and translation and revision processes in general.