Are Phonemic Effects in Backward Masking Evidence for Automatic Prelexical Phonemic Activation in Visual Word Recognition?

Abstract

Recently, several pieces of evidence have led to a resurgence in the popularity of phonologically based approaches to visual word recognition. We review one such piece of evidence: Backward masking with a homophonic mask induces better target recognition than backward masking with a graphemic mask that shares the same letters with the target word as the homophonic mask. From the advantage for words followed by homophonic masks, the existence of automatic prelexical phonemic activation has been inferred. The present study investigates whether this phonemic effect under backward masking conditions can be influenced by control strategies on the part of subjects. We demonstrate that the phonemic effect in backward masking is observed only when the procedure encourages the use of phonological information. The implications for claims of automaticity on the basis of visual masking and for theories of reading in general are discussed.