Do part-whole relations produce facilitation in the picture-word interference task?

Three experiments used the picture-word interference task to evaluate competing models of lexical access in spoken word production. Both the presence of a part-whole relation and association between the target and the interfering word were manipulated. Part terms associated with targets produced facilitation at early stimulus onset asynchronies (SOAs; -300 ms in Experiment 1, -300 and -150 ms in Experiment 3), but not at SOA 0 ms. Otherwise, part terms tended to produce interference, with unassociated part terms producing a significant semantic interference effect (SIE) at SOA of 0 ms in Experiment 1, and a similar trend in Experiment 3. Experiment 2 replicated the materials and procedure of Costa, Alario, and Caramazza ( 2005 , Experiment 2. On the categorical nature of the semantic interference effect in the picture-word interference paradigm. Psychonomic Bulletin & Review, 12(1), 125-131), yet failed to find any semantic facilitation at SOA 0 ms. We propose that these findings are consistent with lexical competition accounts of SIE but difficult to explain in terms of the plausibility of the interfering words as responses to the target.