Doctoral Dissertation Research: Learning attitude verb meanings

University Of Maryland, College Park, College Park MD

Investigators

Abstract

Having strong linguistic ability is crucial for becoming a productive member of society. However, there remain many unanswered questions regarding how young children, who start out with no language whatsoever, learn this complex and crucial ability. The answer to these questions not only has educational ramifications but may also provide important insights into the root causes of such matters as the increased prevalence of cognitive-linguistic disorders and socioeconomic-status-based differences in learning outcomes. The issues examined in this dissertation project relate to how young children learn the meanings of verbs: Verbs constitute the core of a language's sentences, so understanding how verbs are learned is key to understanding how children learn language more generally. Some verbs are harder for children to learn than others. For instance, action verbs like "run" and "hit" are learned earlier than mental verbs like "believe" and "want". One reason "believe" and "want" might be learned later is that, whereas we can see and hear running and hitting events, we can't see or hear thinking and wanting. Children nevertheless learn these verbs, so a route other than the senses must exist. This research investigates what that route is using methods from linguistic theory, cognitive science, and computer science to examine one promising proposal: that children use the sentence contexts a verb occurs in to infer its meaning. The research will examine theories of how children leverage the fact that similarity in sentence context seems to correspond to similarity in verb meaning. To answer this question, this research will exploit experimental methods for measuring the effect of contextual information on verb learning. The investigators will measure these effects on performance in artificial grammar learning experiments using Mechanical Turk for data collection. The researchers will incorporate the findings from these experiments into a computational model of the cognitive processes that allow children to learn words. This second step is useful both scientifically and practically. On the one hand, it will result in a deeper understanding of the mental processes that underlie language learning. On the other, it will provide a model of cognitive development that will help predict the effect that differences in the language spoken to children--for instance, those conditioned by socioeconomic status--have on language learning outcomes.