Linguistic form follows pragmatic function: A multivariate developmental analysis of the determinants of referential form in young children's language

Several studies of children's early language have examined referential form in naturalistic speech (e.g., L. Bloom, 1991, P. Bloom, 1990) and through experimentally-elicited imitations (e.g., Gerken, 1994; Valian et al., 1996). The present studies constitute the first empirical analyses that directly address the extent to which discourse and pragmatic factors might contribute to whether young children produce pronominal or lexical forms or omit overt reference in productive language. In Study 1, 39 children's (ages 1;8–3;3) naturalistic utterances were analyzed. A variety of pragmatic and discourse factors were found to predict children's forms. Results suggest that young children are able to judge the communicative effectiveness of their utterances from the viewpoint of their interlocutors and modify their language when insufficient contextual information is present. These children were also able to track the discourse history of referents and make judgments regarding how topics should be re-introduced in conversation. In Studies 2a and 2b, 36 2- and 3-yr-old children participated in games in which the pragmatic and discourse context was manipulated. Results indicate that children were sensitive to subject and object identity, the constancy of subjects and objects, the way in which the subjects and objects were modeled by the experimenter, and the type of prompt used to elicit language from the child. Overall, these findings are the first to demonstrate that discourse and pragmatics help determine children's referential forms. These studies cast doubt on the assumptions of the null subject parameter account of subject omission and support a theoretical framework in which similar discourse and pragmatic factors influence children's subject and object production.