Robinson Hall A, #447
November 17, 2017, 11:00 AM to 01:00 PM
A prominent theory of variability in L2 learners’ use of inflectional morphology is the Missing Surface Inflection Hypothesis (MSIH). The MSIH argues that morphological variability results from ‘performance limitations’ - particularly when a learner is under communication pressure (Prevost&White, 2000). Such pressure is expected to be highest during production, and lower during receptive tasks, and thus many studies of L2 variability have focused exclusively on production. Recent work has shown that variable comprehension of agreement also occurs in second language learners (L2ers) (McCarthy,2008). This suggests that receptive tasks can help adjudicate among theories of morphological variability and into the mechanisms responsible. McCarthy (2007) also suggests that L2ers resort to the underspecified form when making production or comprehension errors; Morphological Underspecification Hypothesis (MUSH). Nevertheless, studies comparing production and comprehension are few and are restricted mainly to the Romance languages. This dissertation investigates morphological variability in gender and number agreement in English-speaking L2 learners’ production and comprehension of Arabic NP agreement and DO clitic agreement. The focus on Arabic is motivated by two important gaps: first, Arabic L2 acquisition is relatively understudied; second, Arabic has full agreement in gender and number, with a three-way number distinction (singular/dual/plural) in NPs and DO clitics. The acquisition of dual by L2ers not encoding this distinction is not well understood. Here I use L1 speakers of English; a language that does not encode number or gender agreement among nominal constituents. Data from 3 cross-sectional experiments at 2 proficiency levels were collected to test the predictions of the MSIH (Prevost and White, 2000) and MUSH (McCarthy, 2007). Results of the experiments suggest (1) Morphological variability is a persistent problem for L2ers even at advanced proficiency levels; (2) Morphological Variability extends to comprehension, however, it decreases in comprehension; (3) Animacy plays a role in the acquisition of agreement; where human targets were numerically higher in agreement accuracy than non-human targets; (4) The use of feminine adjectives in masculine contexts suggests that feminine is the default for errors in Arabic. On the other hand, since variability clearly decreased in comprehension compared to production, these results confirm predictions of the MSIH, that decreasing communication pressure decreases variability.