For studying the processes involved in translation and in interpreting, retrospection is one of the few research methods equally suitable for both areas. At the first workshop on research methods in process oriented research, in Graz 2009, we presented the results of a pilot study of retrospection as a research method, published as Englund Dimitrova and Tiselius (2009). The study involved data from 2 groups (15 years of professional experience versus no professional experience), each with 3+3 subjects, interpreter subjects versus translator subjects, all with Swedish as their L1. The source text was a 10 minutes plenary speech in English from the European Parliament, interpreted simultaneously into Swedish. For the translation data, the translator subjects translated the original European parliament transcript of the speech, 1093 words, writing in Translog. After the task, subjects did immediate retrospection. The first analysis of the data indicated that a challenge when using retrospection is that subjects tend to report having forgotten about some of their processes.
In this paper we report an analysis of the process data in relation to the retrospective protocols. Our focus is on reported problems, and the occurrences of problem indicators in the process. It was found that most reported problems are confirmed by the presence of problem indicators in the process. However, the majority of problem indicators found in the process do not correspond to any reported problem. Hence, subjects' problem reports can only explain a limited number of the potential problems in the process. The need for further research into retrospection as a research method in translation studies is pointed out.
