Investigando processos de pós-edição e de tradução: uma análise cognitivo-pragmática da relação esforço/efeito no par linguístico japonês/português

Abstract

The present research presents a study of cognitive process and aims to investigate if there are anysimilarities and dissimilarities between translation process and post-editing process of Japaneseand Brazilian Portuguese language pair as well as to compare the distribution of cognitive effortin both tasks. This research basically focuses on the Alves and Gonçalves model (2013), whichdeals with the cognitive effort distributed on edits in conceptual and procedural encodings in L1and L2 translation tasks, based on relevance theory (TR), introduced by Sperber and Wilson(1986/1995). According to these theorists, the inference is important in linguistic communicationto maximize relevance with minimal cognitive effort. In order to achieve successfulcommunication, a set of conceptual and procedural encoding is used to facilitate inference to lookfor appropriate contexts for the decoding of explicatures and implicature derived from utterances.The same principle can be applied in translation, according to Gutt (1991/2000), since thisbilingual and written communication is the interpretive use of language. The same applies also incommunication performed by the combination of automatic translation and post-editing process,once it is a form of communication. In line with Alves and Gonçalves, our results show that thetranslators group manifests a distinct performance, exclusively in the translation task, observedin instruments such as key-logging and eye tracker. The results also show that there are more noneditedsegments (P0 and PE0) and fewer keystrokes, which are considered as an evidence of thedistinctive performance of translators in translation process, ensuring the durability of the targettext. With respect to cognitive effort we identify that the result of the mean fixation duration bythe number of edits, translators edit the texts in a considerably concentrated manner. Moreover,although it is the first experience of post-editing for all the participants, translators can transfertheir translation competence for this new task. Like Alves and Gonçalves results, in this research,the translators also show a greater number of edits in procedural encoding in both tasks, whichagain supports the postulation of Sperber and Wilson. The study also shows that in post-editing,students edited the conceptual encodings as much as procedural ones. From these results, it ispostulated that the post-editing is characterized by the manipulation of procedural encodings,interpreting the data obtained from the group of translators. We assume that the cognitive effortdistributed in edits of these categories point out the element of similarities observed in the tasks.With regard to the dissimilarity, the results demonstrate that the flow of reading, pauses and thesize of the translation unit segmentation develop differently in each task. The set of resultsrevealed in this study between translation and post-editing process will serve as a basis; we maybe able to develop further research with more specific details both quantitatively and qualitatively.