EL(LA) mapping: An integrated account of learning context, feedback and agreement morphology in the processing of OclVS sentences in advanced L2 Spanish

Previous literature has shown that beginning and intermediate English-speaking learners persistently misinterpret O-cliticVS sentences in Spanish, preferring word order over morphology when assigning semantic functions to the NPs of non-canonical sentences. Following the Competition Model (Bates & MacWhinney, 1982, 1989; MacWhinney, 2012), this dissertation investigated from an on-line and off-line account whether English-speaking advanced learners of Spanish are also prone to such misinterpretations, and whether their reconfiguration of L1 processing strategies benefits from: (a) mismatches in number agreement morphology, (b) immersion experience, and (c) computer-delivered feedback.;Three self-paced readings were conducted. In Experiment 1 (N=38), matching/mismatching number agreement between clitic and verb was manipulated to measure the usefulness of contrastive agreement to overcome word order bias. Experiment 2 (N=20) investigated whether a 5-week Study Abroad program promoted learners' OclVS sentence development. Experiment 3 (N=90) investigated potential interactions between agreement conditions and the provision or absence of computer-delivered feedback.;Results showed that advanced learners continued to rely on word order when interpreting OclVS sentences, but different agreement conditions were not processed alike. In Experiment 1, accuracy significantly improved in the mismatching condition in which the morphological cue was on the verb. Participants also tended to present longer reading times in the verb of Cl sgVplSpl structures. In Experiment 2, learners significantly improved at Week 5 and paid more attention to the verb and post-verbal subject, which are highly informative. In Experiment 3, the [+ Feedback] group outperformed the [- Feedback] group, which only received practice decoding manipulated input. Also, the [+ Feedback] group exhibited a speedup across experimental trials in the baseline condition (ClsgVsgS sg) whereas the [- Feedback] group exhibited a significant slowdown in the mismatching ClsgVplSpl condition.;This dissertation has implications for research on intra-subject L2 processing variation and emerging bilingualism. It also contributes to the debate on whether short-term studies overseas influence L2 grammar skills. Finally, the extent to which exposure to manipulated input alone helps learners reconfigure their L2 processing strategies is discussed along with feedback's potential role in enhancing this process.