Preferences for a Verb′s Complements and Their Use in Sentence Processing

Abstract We explore how subject′s preferences for complements of particular verbs affect the on-line processing of sentences. We conduct three experiments: The first experiment gathers binary preference choices for several verb types and their complements: the second and third experiments use these preference choices and find that a subject′s preference for a verb′s complements affects an initial parse in sentences that do not contain obvious structural ambiguities. These data are discussed in terms of the lexical entry-driven and phrase structure-driven models of sentence processing. We conclude that lexical information is used during the initial analysis of a sentence to set up possible paths the parser might take; when the structure of the sentence is incongruent with a subject′s preference, sentence processing performance is disrupted and reanalysis must occur