Philipp, A. M. & Koch, I.
Department of Psychology, RWTH Aachen University, Aachen, Germany
When people switch between languages, language-switch costs occur. Empirically, language-switch costs are measured as the performance difference between switch trials, in which an object has to be named in a different language as the object in the previous trial, and repeat trials, in which the same language is relevant in two successive trials. In the present experiment, participants had to name digits in either German (L1) or English (L2). As an additional manipulation, a go/no-go paradigm was used. That is, 25% of all trials were no-go trials, in which no verbal response was requested. We observed substantial switch costs in go trials following go trials but not in go trials following no-go trials. Even when the no-go signal was presented 1500 ms after the digit and the language cue, so that participants were able to engage in lexical selection (i.e. select the correct word in the correct language), we observed no language-switch costs. This finding indicates that late, response-related processes, presumably phonological encoding, speech production, and/or articulation, play a crucial role for the occurrence of language-switch costs.