Pós-edição monolíngue: uma análise de indicadores de dispêndio de esforço temporal, técnico e cognitivo

Abstract

This study draws on Krings (2001) to investigate temporal, technical and cognitive effort during the performance of four monolingual post-editing tasks. Texts were machine-translated into Portuguese by Google Translate from three different source languages: Spanish, English and Chinese. Fifty-nine participants with some knowledge of post-editing performed the monolingual post-editing tasks without access to source texts, two participants performed translation and bilingual post-editing tasks in the language pairs English=>Portuguese and Spanish=>português, two participants performed bilingual post-editing tasks, and one participant performed a translation task in the language pair Chinese=>Portuguese. Data from these participants was collected using Translog© II, Tobii T60 eye tracker, free and guided written protocols, and TAPs (Think-aloud Protocols). The analysis of this study focused on indicators of temporal effort (task execution time, textual production time, total pause time, pause counting, non-textual production time), technical effort (number of insertion keys, deletion keys, navigation keys, return keys, copy/cut-and-paste keystrokes as well as the number of mouse operations and the total of keystrokes and mouse operations) and cognitive effort (average fixation duration, fixation count, total gaze time, average pupil size, duration of the longest fixation). Results indicate that temporal effort, technical effort and cognitive effort are significantly influenced by verbalization on the basis of 14 out of 17 indicators of effort. However, proximity between source and target languages does not seem to be a relevant factor in investigating temporal, technical and cognitive effort. In addition, it seems there is a tendency of progression of effort: monolingual post-editing bilingual post-editing translation, especially for temporal effort, thus indicating the impact of source text.