Delayed Priming of the Pronunciation of Inconsistent Words and Pseudowords

Four experiments investigated priming by inconsistent word-body neighbors in pronunciation. Experiment 1 replicated the retardation of naming that is observed for a regular inconsistent word (MUSH) when it is preceded by an exception neighbor (BUSH), compared with a regular consistent control prime. The priming effect was comparable when the prime occurred 1 or 10 trials before the target. Experiment 2 showed priming interference at the 10-trial prime-target delay when the target was an exception word and the prime was a regular inconsistent neighbor. Experiment 3 replicated at the 10-trial prime-target delay the bias in pronunciation of a pseudoword (FUSH) by regular versus exception inconsistent word neighbors. In Experiment 4, 24 subjects read aloud a block of regular and exception prime words and 24 subjects pronounced these primes when cued by sentences from which the noninitial letters of the prime had been deleted. Subjects reading the primes showed significantly greater bias in subsequent pseudoword pronunciation than did subjects generating the primes. The results of the four experiments are compatible with an interpretation of inconsistency priming as weight changes on links between sublexical letter clusters and a phonological representation, rather than as ephemeral activation effects of the prime