Lou Boves, Louis ten Bosch & Mirjam Ernestus

Modelling spoken word processingCenter for Language Studies, Radboud University

We present DIANA, a model of spoken word processing that simulates participants in core psycholinguistic experiments such as lexical decision. DIANA takes speech signals as input, and generates decisions as well as reaction times with millisecond precision. Although great effort was spent to make DIANA as cognitively plausible as possible, we will explain the limitations and discuss ways for overcoming these. Human reaction times in psycholinguistic experiments have also been modelled with multi-level regression models. We will confront these regression models with the processing model DIANA, and discuss ways in which the two modelling approaches can contribute to each other. We end with a general discussion of the role(s) of computational modelling in psycholinguistics.