Variability and Systematicity in Interlanguage Development: An Analysis of Mode and its Effect on L2 Spanish Morphology

Abstract

The present study investigates the effect of mode on number and gender agreement in third-person direct object clitics produced by three different levels of nativeEnglish L2 learners of Spanish (N = 76). Previous research supports a performancedeficit account of variability, showing that proficiency level and working memory affect L2 morphological agreement (Sagarra 2007 andSagarra &Herschensohn 2010), and that L2 processing patterns are similar to those of the L1 (Hopp 2010), suggesting that gender inflection is attainable. The present study extends such research. Using film retell tasks, data deriving from oral and written production under identical task conditions were correlated with less and more access to working memory, respectively. Analyses revealed main effects for both variables; patterns of agreement in oral production were also evident in written production, but at earlier stages of development. Results support a performance-based account of variability, showing that mode affects gender agreement, but not number.

A cross-theoretical and cross-linguistic perspective on the L2 acquisition of case systems / L'acquisition de systèmes casuels en L2 : des études à travers plusieurs théories et langues

“…The listeners, assigned the role of employees at the cooking network, returned to the room after the video ended and were instructed to take notes on the information given by their partner. Both participants were provided with a list of vocabulary words (Malovrh 2008(Malovrh , 2014Malovrh & Lee 2013) to aid in task completion. Listeners were told that they could ask questions, but that the speakers were to do the majority of the speaking.…”

“…Malovrh 2008;Malovrh & Lee 2010McCarthy 2008;Zyzik 2004). Learners are more accurate producing clitics that are first person, masculine, and/or dative, than clitics that are third person, feminine, and/or accusative, and the overgeneralization of lo to all contexts is common (Franceschina 2001;Malovrh 2014;Malovrh & Lee 2010.…”

“…Learners' development, while systematic, also shows performance-based variability. Malovrh (2014) examined the effects of proficiency level and task type on gender/number agreement of accusative clitics. He discovered that learners follow the same developmental path (number being acquired before gender), but that task type played a role in how quickly different stages occurred.…”

“…While L2 clitics have been well-studied, Malovrh (2013Malovrh ( , 2014 states that there remains more to be done. For example, the morphological variability observed at all levels may be dependent on certain lexical, semantic, pragmatic, sociolinguistic, or contextual factors.…”

“…As Figure 3 shows, Level 4 learners struggled with morphology, more so than in Level 3. Several of the studies reviewed in Section 2 find that gender inaccuracies persist at high levels, and Malovrh (2014) surmises that the challenge of speaking may compromise learners' ability to produce accurate clitics. This is seen in a comment from a Level 4 participant, who recognized during the stimulated recall that she had made a gender agreement error in her narration.…”