Effect of recent L1 exposure on Spanish attrition : an eye-tracking study

Abstract

Previous research has shown L1 attrition to be selective (Gürel 2004) and often restricted to structures at the interfaces between syntax and context/pragmatics, but not to occur with syntactic properties that do not involve such interfaces (Interface Hypothesis, Sorace & Filiaci 2006). This is supported by many studies exploring cross-linguistic influence effects in interface structures, such as the production and/or interpretation of null versus overt pronominal subjects, not only in L1 attriters (Tsimpli et al. 2004, Montrul 2004) but also in other bilingual groups with different language combinations, such as early bilinguals (Paradis & Navarro 2003, Sorace et al. 2009), and advanced late bilinguals (Belletti et al. 2007, Rothman 2009). The current hypothesis is that individual L1 attrition affects only the ability to process interface structures but not knowledge representations themselves (Sorace 2011). In this thesis, we first compared a well-studied syntax-pragmatics interface phenomenon (pronominal subjects in Spanish) with a non-interface structure (the Spanish personal preposition a, also known as Differential Object Marking, DOM). In Spanish, the distribution of null and overt subject pronouns is pragmatically constrained, whereas the presence of the preposition just depends on the animacy and specificity of the direct object. Participants included a group of attrited speakers of L1 Spanish who had been living in the UK for a minimum of 5 years, and a group of Spanish monolinguals. Using a naturalness judgment task and eye tracking while reading, participants were presented with anaphoric sentences in which number cues matched or mismatched predicted antecedent preferences (i.e. null pronoun: subject preference; overt pronoun: object preference). The DOM study also used a mismatch paradigm, crossing preposition presence (al vs. el) with animacy, where an animate object requires the prepositional form al and an inanimate object requires the article el. Offline ratings revealed equal mismatch sensitivity for both groups of participants with both structures. However, eye-tracking measures showed that monolinguals were reliably more sensitive than attriters to the pronoun mismatch, while both groups showed equal on-line sensitivity to the DOM mismatch, which reveals that attrition affects interface structures, but not non-interface structures. Second, we investigated the effects of recent (re)exposure to L1 input on attrition. A second group of attriters carried out the same experiment after having been exposed exclusively to Spanish in a monolingual Spanish-speaking environment for a minimum of a week. Their eye-tracking results patterned with the monolingual group. This novel manipulation shows that attrition effects decrease as a result of L1 exposure, which reveals that bilinguals are sensitive to input changes and that attrition affects online sensitivity rather than causing a permanent change in speakers’ L1 grammatical representations.