ABSTRACT
Virtually all current theories of visual word recognition are based on some version of a parallel activation model. Some are explicitly based on the original interactive activation model proposed by McClelland and Rumelhart (1981), which was itself based on ideas developed in the earlier logogen model (Morton, 1970), and before that, the Pandemonium model of pattern recognition (Selfridge, 1959). Foremost among these models are the dual-route cascaded (DRC) model (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001), MROM (Grainger & Jacobs, 1996), and BIA (Dijkstra & Van Heuven, 2002). Others are based on the parallel distributed processing (PDP) connectionist approach, such as Seidenberg and McClelland’s “single”-route model (Seidenberg & McClelland, 1989), Plaut (1997), and Kello’s junction model (this volume). Although these models differ in their basic assumptions and approach (e.g. whether word units exist), there is universal agreement on the basic algorithm that subserves the process of lexical access, namely, parallel activation.
