PHONEMIC ACTIVATION DURING THE 1ST 40 MS OF WORD IDENTIFICATION - EVIDENCE FROM BACKWARD-MASKING AND PRIMING

When a briefly exposed target word is followed by a pseudoword mask, the disruptive masking effect is reduced when the mask shares graphemes with the target word and is further reduced when the mask shares phonemes with the target word (Perfetti, Bell, & Delaney, 1988, Journal of Memory and Language 27, 59-70). These results implicate early (prelexical) phonemic activation during word identification. In Experiment 1, this phonemic masking effect was found to hold through target exposure durations ranging from 35 to 55 ms and to be independent of word frequency and spelling pattern consistency. In Experiment 3, completely convergent results were found in a masked priming paradigm that varied the exposure of a masked pseudoword prime from 25 to 65 ms. Phonemic priming effects were found by 45 ms correcting earlier conclusions favoring direct visual access based on this paradigm. Together, the masking and priming results strongly support a process of early phonemic activation prior to word identification. The results are most naturally handled by a single-mechanism activation model, but are also consistent with restricted versions of dual route theories. © 1991.