|Title:||Relating Parsers and Grammars: On the structure and real-time comprehension of English infinitival complements||Add Dissertation|
|Author:||Matthew Walenski||Update Dissertation|
|Email:||click here to access email|
|Institution:||University of California, San Diego, Cognitive Science|
|Linguistic Subfield(s):||Psycholinguistics; Syntax; Neurolinguistics;|
|Abstract:||The work in this dissertation explores the relation between theories of grammar and theories of parsing in the context of the real-time comprehension of English infinitival complements of subject control and subject raising predicates. Opinion is currently divided as to whether grammars and parsers can be related at all, whether these theories can be mutually constraining, or whether there are even two distinct types of theory at all. Additionally, representations of these infinitives differ across linguistic frameworks, and have been the source of conflicting neuro- and psycho-linguistic findings as well.
Linguistic evidence, linguistic analyses from four different linguistic frameworks, neurolinguistic and psycholinguistic evidence are considered as the basis for new experiments, using Event-related potentials (ERP), cross-modal priming, and cross-modal interference measures. Evidence from the ERP study presented here suggests that results from prior ERP and visual probe recognition (VPR) experiments, as well as the current ERP results, reflect differences in the expectancy of the infinitival complement for the two types of predicate, not differences in the structure of the complements.
New evidence from the cross-modal priming study presented here constitutes the first evidence of priming immediately at a purported NP-trace gap position (in a subject raising sentence), and provides evidence of priming after a short delay at a purported PRO-gap position (in a subject control sentence). This priming effect is hypothesized to result from a theta-checking strategy which ultimately has a grammatical source: the dependency between the matrix subject and the infinitive that is proposed by (at least) the four linguistic frameworks examined here. If a strict mapping between theories of grammar and theories of parsing representing the simplest relationship between them is imposed, the fact that no interference effect at the purported gap position supports only those theories of grammar that do not posit a structural difference between control and raising infinitives.
Based on these results, it is argued that by imposing the simplest mapping between theories of grammar and theories of parsing, they are found to be (at least) mutually constraining, if not actually reducible to a single theoretical system.