Uniformity of pronoun case errors in typical development: The association between children's first person and third person case errors in a longitudinal study.

摘要

The purpose of this study was to determine if grammatical case is acquired as a unified system or if it is acquired in a piecemeal fashion. Case is a property of sentences that codes subjects and objects for their roles in sentences. In the course of typical English language acquisition, some children make errors in marking case on subject position pronouns (e.g., Me do it, Him like it; Huxley, 1970; Rispoli, 1994; Schutze & Wexler, 1996; Vainikka, 1993). It is an assumption of generative linguistic accounts that the case system is unified with case assignment uniform across grammatical person. A prominent account of the phenomenon posits a link between case and finiteness (i.e., tense and agreement; Schutze, 1997; Schutze & Wexler, 1996). Cognitive linguistic accounts do not posit uniformity and have claimed there is a causal link between the errors and a pattern found in input, in which subject pronouns in embedded clauses are standardly assigned object case (e.g., Let me do it; Kirjavainen et al., 2009; Pine et al., 2005). Previous literature had not established whether the same children who make first person errors (i.e., me or my used for I) also make third person errors (i.e., him, her, them, for he, she, they), a pattern that would be congruent with generativist linguistic accounts, but not necessarily with cognitive linguistic accounts. Previous literature had not established whether errors in first person pronouns are made at the same point in development as errors with third person pronouns, which would be expected if case were uniform across person. Previous literature had not examined whether first person pronoun case errors or finiteness could predict the presence of third person pronoun case errors.;Spontaneous pronoun case errors were collected from language samples of 43 typically developing toddlers interacting with their primary caregivers and examiners. Language samples were collected at 21, 24, 27, 30, 33 and 36 months of age. Association of errors across person was tested with a chi-square test using data from the entire longitudinal sample by asking whether children were more likely to make both types of errors. The uniformity of the system was further investigated by asking whether errors in first and third person occurred at the same time using a Wilcoxon signed-ranks test. Logistic regression was used to test the predictive value of first person error, vocabulary size, and finiteness development on third person pronoun case error.;Most children treated case uniformly in the first and third persons, producing both first and third person pronoun case errors or producing no case errors at all, resulting in a significant association. Of the 43 children, 23 children produced both first and third person errors, seven children produced only first person errors, five children produced only third person errors and eight produced neither. Additionally, errors were not significantly different in timing. Neither the peak number of first person errors nor their duration between 21 and 30 months predicted the presence of third person errors from 30 to 36 months. Likewise, the number of different words at 30 months of age, a measure of vocabulary size, did not predict the presence of third person error. However, tense/agreement accuracy at 30 months, a measure of finiteness development, was a significant predictor. Children with lower tense/agreement at 30 months were more likely to produce a third person pronoun case error between 30 and 36 months.;The results of this investigation lend support to the assumption of generative linguistic accounts that case is a unified system. Errors are associated across person, and errors in first and third person overlap in development. Pronoun case is not acquired in a piecemeal fashion with case being acquired separately for each grammatical person. Further, a deep connection between finiteness and third person case errors was confirmed. Taken together, these results implicate the existence of abstract grammatical features early in development.