Working memory is considered to be a crucial cognitive mechanism for the simultaneous interpreting process, but this assumption has not been substantially supported empirically. Previous research has mostly investigated working memory capacity differences between interpreters and non-interpreters. Little work has been done on relating working memory capacity directly to interpreting performance. The aim of the present exploratory correlational study was to test whether a relationship exists between working memory capacity and simultaneous interpreting performance measures in a sample of professional interpreters. Twenty-eight professional interpreters, aged 25-55, were tested on their working memory capacity (letter span, Corsi task, complex span) and on several measures of interpreting performance (lexical, semantic and syntactic processing, temporal delay, vocabulary richness and dealing with speed). Additionally, measures of general cognitive ability, age and interpreting experience were considered. There are two main findings. First, working memory performance in this sample shows predictable patterns in the structure of interpreters' working memory: there was a dissociation between verbal and spatial memory, and a negative relationship between age on the one hand and working memory measures and general cognitive ability on the other. This negative relationship provides evidence which is difficult to reconcile with the hypothesis of superior working memory in interpreters, as tested by previous research. Secondly, working memory measures were only marginally significantly related to simultaneous interpreting measures, and then only to those which have a predictable high memory component, i.e. lexical processing consisting of interpretation of isolated lexical items, such as figures and lists of nouns. The results suggest that working memory capacity, where the focus is on the memory component of working memory, may not be as important for professional simultaneous interpreting as previously thought. Working memory is probably the single most often researched isolated cognitive component in interpreting studies. The primary theoretical starting point has been the multimodal working memory model by Baddeley and Hitch (1974), although more recent work considers other models, too. More specifically, Cowan's (1988) model of activated long-term memory (Mizuno, 2005) and Ericsson and Kintsch's (1995) long-term working memory models (Pöchhacker, 2004) were suggested as being useful in the context of interpreting. Additionally, the role of working memory was modelled in several models of the interpreting process (for a review of general cognitive and interpreting-specific working
