An analysis of discourse-pragmatic and grammatical constraints on the acquisition and development of referential choice in child English

This dissertation investigates how discourse-pragmatics informs children's choice of referential forms. Three studies utilizing naturalistic data examine the role that six discourse-pragmatic features play in the acquisition of referential forms in a non-null subject language, English. The influence of child-directed speech on development is also examined. Two accounts of omitted arguments in non-null subject languages (the competence-based account and the discourse-pragmatic account) are compared to discover how each contributes to a complete explanation.;Study 1 analyzes the production of overt versus null subjects for four children and their caregivers at two times (T1: 2;0-2;7; T2: 3;0-3;1). This study confirms that children at T1 omit more subjects and produce fewer pronouns than children at T2. Moreover, results of a baseline study indicate that a full range of forms must be analyzed for a clear picture of the acquisition of referential choice to emerge. Study 2 measures the cumulative effect of six accessibility features on whether children and caregivers select null, pronominal, demonstrative, or lexical forms. Caregivers and children use lexical forms when referents are highly inaccessible demonstrating that at T1, children are somewhat sensitive to accessibility, and this sensitivity increases at T2. Multinomial regression analyses demonstrate that three features, PRIOR MENTION, PHYSICAL PRESENCE, and JOINT ATTENTION, are stronger predictors than the others.;Study 3 explores explanations from the competence-based account and the discourse-pragmatic-based account for the well-established finding that children omit more subjects in non-finite contexts. Significantly more accessible subjects occur with non-finite verbs, revealing a previously unobserved interplay between discourse-pragmatic and grammatical effects, and providing further insight into the asymmetry of subject realization in finite and non-finite contexts. ANIMACY plays a particularly strong role, together with CONTEXTUAL and LINGUISTIC DISAMBIGUATION. Grammatical competence also plays a role; verb finiteness correlates with the realization of accessible subjects as pronouns (finite) or as null (non-finite). The evidence shows that each account complements rather than competes with the other.;The centrality of the interface between syntax and pragmatics to this one aspect of acquisition provides a glimpse into how different theoretical approaches can work together to provide a more accurate picture of language development.