Lewis, James Robert

Contextual impenetrability is the measurable consequence of informational encapsulation, which is a key feature of input systems hypothesized to be modular. The data from the current psycholinguistic literature do not clearly converge on support for or refutation of the modularity hypothesis with regard to the existence of contextual impenetrability for lexical access and gap-filling. Most previous research has focused on the influence of global context on lexical access. The experiments in this dissertation explore a more local context imposed by the argument structure and lexical conceptual structure of a verb on its direct object. The primary goal for this dissertation was to develop a set of test sentences appropriate for study with both cross-modal lexical priming and word-by-word reading to connect key experiments in the lexical access and gap-filling literature, then to determine if the outcomes supported or failed to support the existence of contextual impenetrability for lexical access and gap-filling. The outcomes of these experiments supported the hypothesis of contextual impenetrability for these linguistic operations. A secondary goal of this dissertation was to address recent criticisms of studies that have used cross-modal lexical priming to study the contextual impenetrability of gap-filling. The cross-modal lexical priming experiments in this dissertation demonstrated that it is unreasonable to attribute priming effects at hypothesized gap locations to artifacts such as hypothesized "continuation priming." The dissertation concludes with discussions of the implications of these results for two competing theories of mental structure (the modularity hypothesis and interactionist hypotheses) and recommendations for the appropriate interpretation of various experimental tasks and additional experiments.