Children ’ s Representation of Verbs : Evidence from Priming during Online Sentence Comprehension

categories (the Verb Island hypothesis; see Tomasello, 2000). Early evidence for this view came from children’s spontaneous production. For example, Tomasello (1992) found that his one-year-old daughter restricted most verbs to a single construction type, failing to use them in alternate permissible constructions which she had used with other verbs. For example, while draw was used with locative and benefactive prepositional phrases, cut appeared only in simple transitive sentences. Observational studies, however, cannot tell us whether restricted usage reflects the input that the child receives, differences in the meanings of the verbs, or the limited range of situations that the child wishes to discuss. These issues have been addressed in production experiments with novel words (see Tomasello, 2000 for a review). For example, Tomasello & Brooks (1998) exposed children to a novel verb in an intransitive construction while modeling an action (e.g., The sock is tamming). Subsequently, they modeled the same action and tried to elicit transitive constructions from the children (e.g., He’s tamming the car). While older children extended novel verbs to new constructions, 2-to-3-year-olds primarily used them in the constructions they were exposed to during training. The authors concluded