Inhibitory semantic priming: does syntactic class play a role in determining competitor status?

Most current models of lexical or 'lemma' selection propose the parallel activation of a set of possible candidates where candidates compete to meet a 'best match' criterion. One method for testing the competitive activation hypothesis is to manipulate the recency with which a likely competitor word has been produced by priming it. Under certain assumptions one would predict that the point at which a target reaches criterion will be delayed if a competitor has been given such a head start. Implicit in the idea of competition among lemmas is the idea that the activation cohort is limited to items that share both semantic and syntactic features with the target item, as lemmas are argued to specify semantic and syntactic features of lexical items. We set out to explore the importance of syntactic class in determining the composition of the competitor cohort by comparing priming effects for semantically related items that either shared syntactic class-a noun priming a noun-or were from different syntactic classes-a verb priming a noun. We show that the competitor effect is a general one not limited to items drawn from the same syntactic class: giving either a related noun or a related verb a head start increases the time taken to produce the name of a pictured object.