The role of syntactic and discourse information in verb learning

Children use syntax in verb learning; this is syntactic bootstrapping (Gleitman, 1990; Naigles, 1990). This dissertation proposal investigates two research questions about syntactic bootstrapping: 1) how syntactic bootstrapping could begin, and 2) how syntactic bootstrapping could work despite noisy input due to argument omission. The first three sets of experiments (Chapters 2) showed that 15- to 19-month-old infants use the number of nouns in a sentence to differentiate novel transitive and intransitive verbs. The results confirm a key prediction of structure-mapping account (Fisher, 1996) suggesting that syntactic bootstrapping might begin with an unlearned bias to assign each noun in the sentence to a core participant-role in conceptual representations of events. Experiments 4 and 5 (Chapter 3) showed that Korean-learning 2-year-old children exploit distributional and discourse information to find true number of arguments of a new verb despite pervasive argument omission. Finally, Chapter 4 (Experiment 6, 7 and 8) asked whether a bias to discourse continuity could aid children’s verb interpretation even in English that allows noun omission only in restricted situations. The results showed that English-learning children also could use discourse continuity to interpret verbs with missing subjects. Taken together, the results of these studies support the conclusion that the inherent one-to-one mapping bias can guide children’s verb learning even in noisy input in any language, with children’s expectation for discourse continuity