Similarity and Structural Priming

The increasing evidence that language processing is sensitive to lexical and structural co-occurrences at different levels of granularity and abstraction (Jurafsky, Bell, Gregory, & Raymond, 2001; Bybee, 2006) has led to the hypothesis that lexical and structural processing may be unified (MacDonald, Pearlmutter, & Seidenberg, 1994; Jurafsky, 1996). This paper examines the specific hypothesis that structural priming and lexical priming may be due to the same underlying mechanism. Lexical priming is known to exhibit sensitivity to the similarity between the prime and the target: the more similar the prime and target words, the greater the magnitude of the priming effect (Ratcliff & McKoon, 1981). Two corpus studies show evidence of an effect of similarity on structural priming. Structural and semantic similarity of the prime and target structures are modeled using a database of ditransitives extracted from the Switchboard corpus and a nearest-neighbor similarity metric. More similar prime and target structures are found to be more likely to occur in the same construction. This effect is in addition to the known similarity effect of verb identity (Pickering & Branigan, 1998), which is controlled through simultaneous multiple regression and model comparison. This suggests that lexical and structural priming could be the same process. Implications for models of representation and processing are discussed.