Most computational models of visual word recognition incorporate letter-position coding schemes according to which transposed-letter (TL) nonwords (e.g., jugde) are only moderately similar to their base words (e.g., judge). Nonetheless, TL nonwords are often misperceived as their base words, especially when the transposition of the letters occurs in middle positions (Chambers, 1979; O'Connor & Forster, 1981). To examine these issues further, we conducted three masked form priming experiments using a lexical decision task. In Experiment 1, form-related primes could be TL-internal nonwords, TL-final nonwords or the corresponding orthographic controls (e.g, uhser vs. ufner, ushre vs. ushno; the target would be USHER). Masked TL-internal primes produced a significant form priming effect (30 ms) relative to their orthographic control condition, an effect that was greater than the priming effect for TL-final primes (13 ms). Experiments 2 and 3 replicated these findings with a different set of prime-target conditions and a different set of items, respectively. Taken together, the results show that form priming effects can be found for TL primes relative to the appropriate orthographic control condition and that these effects are larger when the letter transposition is internal to the word than when it involves the final two letters. We discuss the implications of these results for the choice of a letter-position coding scheme in visual word recognition models.
|Title of host publication||Masked Priming: The State of the Art|
|Number of pages||14|
|ISBN (Print)||0203502841, 9780203502846|
|Publication status||Published - 19 Jun 2003|