The debate over the (un)availability of uninterpretable features in L2 acquisition is examined here (see, e.g., Dekydtspotter and Renaud, 2009; Hawkins and Casillas, 2008; Lardiere, 2009). Because the gender feature is not instantiated in the L1 grammar of the participants (English), this allows testing of whether learners process the interpretable and the uninterpretable gender feature similarly. This study also sought to find out whether learners rely on grammatical information when confronted with the co-reference of a pronoun with synonyms of different gender. Four groups of participants--French natives (n=15), and second-semester (n=24), fourth-semester (n=26), and advanced (n=15) American learners of French--completed a judgment task in a self-paced moving window format. Experimental items investigated two conditions: co-reference of pronouns with a noun available in the discourse and adjective agreement, targeting the interpretable and the uninterpretable features respectively. Each item consisted of a context and a subsequent segmented sentence in which the gender of the pronoun or the adjective was manipulated. Grammaticality judgments and reading times (RTs) on the critical segments (i.e., the pronoun and the adjective) as well as the following segments were analyzed. Overall, L2 learners were found to rely on UG-constrained operations indicative of the computation of the uninterpretable feature, and evidence in support of a feature-geometric account (e.g., Harley and Ritter, 2002) was also obtained.
Proceedings of the 11th Generative Approaches to Second Language Acquisition Conference (GASLA 2011)
edited by Julia Herschensohn and Darren Tanner
Table of contents