| This dissertation describes and evaluates a thesis about the means of identifying verbs early in learning a language, and a first language in particular. The thesis is presented briefly in the first section. The second section provides a critical review of theories about children's early part-of-speech identifications. Section 3 presents a new theory of verb identification. I argue that learners initially identify members of a category, predicator, that subsumes verbs and adjectives. Predicators have argument structures. Learners identify a predicator through an inference that the word must take noun-phrase arguments because the phrase containing the word is interpreted into a nonseparable phenomenon--a property or relation that exists or occurs only by virtue of one or more individuals (i.e., the bearers of the property, or the participants in the relation), the referent(s) of the argument(s). Actions are prototypical of that which is nonseparable (being dependent for their realisation upon one or more participants), and so words for actions will usually be identified as predicators. This tendency will be augmented when an unfamiliar predicator appears in an utterance with its one or more noun-phrase arguments, and the noun phrases are interpretable (by the learner) into the one or more individuals that are the participants in an ongoing action (or other nonseparable phenomenon); under these conditions, the learner should readily divine that the novel word is a predicator and the noun phrases are its arguments. These conjectures form the nonseparability hypothesis. To identify verbs in particular, a learner must first discover a distinction between verbs and adjectives, where it exists in a language, through distributional analyses within phrases. Subsequently, details of syntax and morphology will reveal to the learner a predicator's subcategory (verb or adjective). Section 4 contains reviews of literatures that provide support, in varying degree, for the theory's assumptions, proposals, and predictions. Section 5 lays out the major predictions generated by the nonseparability hypothesis. Section 6 describes the positive results of three experiments that tested these predictions. The final section presents a summary and my conclusions, which are generally positive about the prospects for the theory. |
