Neurocognitive plasticity in verb bias learning in children and adults

Verb-specific preference for syntactic structure (verb bias) is considered as a critical parsing constraint that guides online sentence comprehension. Both adults and preschoolers show great sensitivity to verb bias in their temporary parsing commitment as sentences unfold in time. How do people learn verb bias in the first place? In natural language, frequency-sensitive verb argument structure is closely intertwined with the event information delivered by the verb and its argument, which raises complexity in teasing apart the information from linguistic co-occurrence frequency and the information from the event semantics. In this dissertation I began by examining the independent roles of each information source during the process of updating familiar verb bias. The rest of the study focused on the verb bias learning without event cues from verb semantics. Two parallel approaches were applied to explore the details of the learning mechanisms. One set of studies used eye tracking to monitor the time course of online usage of newly learned verb bias during sentence ambiguity resolution across different age ranges. The other set of studies examined the neural stages of verb bias learning as well as the individual differences of verb bias retrieval during online sentence reading with event-related brain potential (ERP) techniques. I demonstrated with very brief training paradigm in both listening and reading modality that children and adults were capable of quickly adapting to new information about verb-specific structural preference from the dynamic language input. The results provided evidence for a central role of linguistic distributional information in verb bias learning. Newly learned verb bias plays a similar role as the existing verb bias knowledge in affecting language users’ parsing commitment and online ambiguity resolution. In addition ERP results revealed separate neural stages that transits from semantic prediction to syntactic rule-based processing as learners continuously collected distributional information of verb-specific structural preference. Individuals who were highly sensitive to familiar verb bias also showed greater use of newly learned verb bias during conflict detection, further indicating the same mechanism underlying natural verb bias acquisition