Hodzik, E. .
Bogazici University
This study takes an interdisciplinary approach applying psycholinguistic methodology to the investigation of predictive processes during simultaneous interpreting (SI), i.e. the oral translation of speech at a very close latency, as an instance of spoken language processing. The importance of prediction has been emphasized during SI between two languages with asymmetrical syntactic structures, such as German, a head-final language, and English, a head-initial lan-guage (Wilss, 1978; Jörg, 1995). Predictive processes have also been investigated during visual and spoken language processing as part of language comprehension, where words are predicted as a result of contextual constraint, i.e. semantic and syntactic cues available in the context, and transitional probability (TP), i.e. the statistical likelihood with which words appear together in language (McDonald and Shillcock, 2003a, 2003b; Frisson et al., 2005). Effects of contextual constraint and TP on predictive processes were investigated by measuring the speech latency between the input and output during i) shadowing in German, a within-language repetition task, and during ii) SI from German into English. Speech latency measures revealed that contextual constraint affects prediction during shadowing and SI, while TP triggered prediction during shadowing but not during SI. However, when syntactic structure was manipulated in a second set of experiments involving i) SI from head-final German into head-initial English sentences and ii) SI from head-initial German into head-initial English sentences results revealed no effect of TP on SI latency between asymmetrical sentence structures, but an effect of TP was observed during SI between syntactically symmetrical sentence structures. TP also affected shadowing latency. The present findings provide support for language-specific processing of the input during SI, showing that an online investigation of prediction as an integral part of language comprehension can help reveal something about the processes underlying a complex cognitive task like SI.