Locating the Source of Number Agreement Errors in Intermediate-to-Advanced Second Language Learners of English

This study explores the extent to which adult Second Language Learners (L2ers) of English are sensitive to Subject-Verb (S-V) agreement errors with thematic verbs and copulas. A group of intermediate-to-advanced Persian-English L2ers and a group of native English speakers (NSs) participated in an online Speeded Acceptability Judgment task. The results are as follows. Whereas NSs are sensitive to number agreement errors in both verb types, L2ers are not. For L2ers, the results reveal the following systematic errors. For agreement errors in thematic verbs, there is an omission-commission asymmetry; whereas L2ers are not sensitive to omission errors they are sensitive to commission errors in finite forms. For copulas, there is a singular-plural asymmetry indicating higher error rates in *plural subject NP + is configurations than in *singular subject NP+ are ones. Yet, proficiency seems to be a strong predictor of L2ers’ sensitivity to agreement errors. The results support the Morphological Underspecification Hypothesis (McCarthy, 2007; 2008; 2012).