Tricklet: Translation Research in Corpora, Keystroke Logging and Eye-Tracking (DFG-Projekt Nr. NE 1822/2-1, Laufzeit 2013-2016)

Corpus studies have provided evidence for typical properties of translated texts. However, they do not allow claims about causes for these properties during the translation process. Using a novel combination of experimental and corpus-based methods, this project seeks a deepened understanding of the cognitive processing during translating of certain features which have been identified as displaying translation properties. The methodological combination of keystroke logging and eye-tracking established in experimental translation studies will be brought closer in line with mainstream psycholinguistic methods. To this end, three experiments with additional control over confounding factors by manipulating segments in the source text will be carried out, thus striking a new balance between ecological validity as required by translation studies and control over confounding factors ensuring reliable and valid claims. Observations will be linked to linguistic insight by deriving the manipulated features from theoretically informed concepts. The experimental data will be annotated corpus-linguistically thus facilitating a rich linguistic analysis and will be examined with respect to correlations between cognitive demand and changed complexity in the translation product, concentrating on three different forms of complexity. At the same time, classical product-based insight from translation corpora will be included in the interpretation. The research follows a strictly quantitative approach in order to yield stochastic generalisations. The project will not only provide new insight into the role of cognitive effort during translating but also narrow the methodological gap between experimental and corpus-based translation research by transforming the experimental data into a database which can then be queried as a corpus and subsequently providing causal explanations for corpus-based observations. The resulting detailed understanding of causes for specific properties of translated texts in the translation process will be conducive to the empirical modelling of translation. In a later second phase of the project this can then be elaborated towards an empirical theory of translation superseding existing impressionistic models.