The robustness of phonological effects in fast priming

Dual-route models such as the dual-route cascaded (DRC) model (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001) and strong phonological models (e.g., Frost, 1998) differ in the role they assign to prelexical phonological computation in the process of word recognition. Whereas the strong phonological theory regards prelexical phonological computation as the initial cognitive operation launched in the reading process, and considers it to be very fast, the DRC model specifically sets prelexical phonological computation to lag behind orthographic processing. Thus, the empirical debates between proponents of strong phonological models and dual-route models revolve around experimental paradigms that focus on very early and automatic processing of print, and involve fast priming. This chapter examines two main experimental paradigms that are traditionally used to examine fast automatic phonological computation: backward and forward masking. It focuses specifically on the pseudohomophone test in forward masking, and reviews recent results obtained in Hebrew. These results suggest that prelexical phonological computation is very fast, and occurs under masked presentation in which the primes are unavailable for report. Moreover, they demonstrate that the computed phonological code is initially impoverished and coarse-grained, as the strong phonological theory predicts.