An Optimality-Theoretic Model of Acquisition of Tense and Agreement in French

We present a novel theoretical model of multiple stages in the acquisition of tense and agreement in Child French. First, we show that tense and agreement inflection follow independent courses of acquisition. Over the three stages of development attested in the data, tense production starts and ends at near-adult levels, but suffers a “dip” in production at the second stage. Agreement develops linearly, going roughly from none to 100% over the same time. This profile suggests a competition between tense and agreement at the second stage which is naturally expressed in terms of constraint violability and constraint re-ranking (Optimality Theory, Prince & Smolensky, 1993). By incorporating the further mechanism of partial rankings of constraints, our analysis successfully predicts, over three stages, the frequency with which children use tensed, agreeing, and nonfinite verbs. The Attested Development of Tense and Agreement in French It is cross-linguistically well-attested that young children (around the age of 2) often produce simple sentences with a non-finite root (verb) form (NRFs), ungrammatical in the adult language, while also producing adult-like finite verbs with tense and agreement marking (Wexler, 1994, 1998, inter alia). What has been previously overlooked, however, is whether the distinct inflectional categories of tense and (person/number) agreement develop independently over time. A detailed analysis of spontaneous speech production data from three French children from the CHILDES Database (MacWhinney & Snow, 1985) provides strong evidence that the two categories indeed follow a different path of development. __________ 1 The data we report on in this paper is part of a larger project aimed at studying the acquisition of tense and agreement As a preliminary step of analysis the relevant files were analyzed by hand and classified into PLU (Predominant Length of Utterance) stages (Vainikka, Legendre & Todorova, 1999). This independent measure refines the traditional observation that children progress through one-word, two-word, and multi-word stages and has proven better suited to capture syntactic development than the well-known MLU measure (Brown, 1973). French presents specific challenges for a study of the development of finiteness because the overwhelming majority of verbs used by young children belong to the first conjugation class (‘-er verbs) which displays considerable homophony across morphological person inflections. In the absence of an overt subject (which is frequently omitted by young children) it cannot be determined whether a given phonetic form like [dãs] danse ‘dance’ in the present tense carries correct agreement in person and number. However, clitic subject pronouns in French (e.g., je ‘I’) provide a diagnostic crosslinguistically (Vainikka, Legendre & Todorova 1999). We have done in-depth work on transcripts from eight different children at this stage of development, covering English, French, Polish, Russian, and Swedish. Further analyses are underway as well. We should point out that while the conclusions drawn in this paper are made on the basis of data from (only) three children, we believe (following well-established tradition in the study of the acquisition of syntax by children) that examining a small number of subjects in detail allows us to uncover complexity that would be missed in a necessarily less detailed overview of a larger group of subjects. Furthermore, there i s strong evidence to suggest that syntactic acquisition proceeds in a highly constrained and species-universal manner. Given this, we do not expect to find a great deal of variation from child to child, increasing the likelihood that the results reached on the basis of these three children will generalize across French-