This article reports on an empirical study on translation revision. With the aim of investigating the possible link between revision procedure and quality, the research correlates an indicator of quality, error detection, with revision procedure. Error detection and revision procedure were studied drawing on a convergent parallel mixed-methods research design involving three different sources of data. Nine subjects performed a revision task and thus produced text data; their activities on the computer screen were captured and saved as video fi les; and retrospective interviews were conducted with the revisers upon completion of the task. Results show that the highest error detection scores were linked with a variety of revision procedures, but with one common denominator: The target text was consistently the point of departure. Revisers with high error detection scores thus engaged in various different revision procedures, but their focus of attention in the initial operations was the translation rather than the source text in all cases. Conversely, the revisers whose initial attention was directed towards the source text received the lowest error detection scores in the revision task.
|Udgivet - 2016