The effects of structural context on priming

Previous research has found that the recent processing of a linguistic form (e.g. word or syntactic pattern) facilitates its reuse. A separate line of research has found that the appearance of a linguistic form in certain structural contexts (e.g. the focus position of a cleft sentence) can increase the likelihood of a form's reuse. However, these two lines have not explored whether the structural context in which a recently-processed form occurred mediates the facilitatory effects of recent processing. I contend that such mediation exists. Specifically, I propose that the way a structural context is processed affects how memory represents the processing event and how the linguistic forms associated with that structural context are represented in memory. I further contend that differences in these representations affect the subsequent accessibility of the forms.;I present a series of priming studies that support this proposal by showing that the facilitatory effects of a form's recent processing are attenuated when the form occurred in particular structural contexts. By holding time constant and varying only the structural context in which a lexical or syntactic form occurred, I demonstrate that some structural contexts undermine forms' reuse. Specifically, speakers are slower to identify lexical primes occurring in the internal complements of nouns (e.g. the bolded word in "David knew the fact that the man kissed Sophia") relative to primes occurring in other structural contexts (e.g. a relative clause "David knew the man who kissed Sophia" or a main clause "As David knew, the man kissed Sophia"). Similarly, speakers exhibit less-stable structural priming for primes occurring in the internal complements of verbs ("David knew that the man kissed Sophia") relative to primes occurring in other structural contexts.;To clarify the source of structural context's affects on priming behavior, I present a novel activation-based model of language processing. My model describes how linguistic forms are retrieved and manipulated during processing and how the memory traces of linguistic forms are affected by structural contexts. During processing, the processor retrieves encoded memory traces for target linguistic forms. Features of these memory traces, such as how recently they were created and the number of forms (e.g. other words) associated with the memory trace, affect the processor's ability to reuse the target form. I argue that processing some structural contexts (e.g. those containing argument clauses such as the internal complements of verbs) leads to memory traces with more competing forms than the traces generated during the processing of other structural contexts (e.g. those containing adjunct clauses such as relative clauses). The differing number of competing forms associated with the memory trace stems from the processing of different structural contexts and ultimately affects the processor's ability to reuse recently encountered linguistic forms.