Oral Corrective Feedback, Individual Differences, and L2 Acquisition of French Past Tenses

Abstract

The research reported in this thesis aimed at investigating the differential effects of two types of oral corrective feedback in a foreign language classroom: recasts as an input-providing strategy and clarification requests as an output-prompting strategy (Ellis, 2006; Ellis & Sheen, 2011). Drawing on the Interaction hypothesis (Long, 1981a, 1983b, 1996) both strategies can be considered a type of focus-on-form. The effectiveness of oral corrective feedback was examined in relation to the acquisition of two French past tenses - the passé composé and the imparfait, which have been documented among the most difficult language structures for French L2 learners (Bardovi-Harlig, 2000; Harley, 1989, 1993; Kaplan, 1987). The effects of corrective feedback were also examined in relation to the mediating role of the learners' individual differences in language analytic ability, working memory, and anxiety. The study was conducted with three intact classes involving 52 high school students learning French as a foreign language in New Zealand. A quasi-experimental design was employed, with a pre-test, treatment, immediate post-test and a delayed post-test. Oral corrective feedback was provided in the form of recasts in one class and clarification requests in the other class as the learners performed three picture-based, information-gap tasks designed to encourage communication and elicit the use of the two past tenses. Acquisition was measured by means of similar picture-based narrative tasks requiring relatively free oral and written production. The tasks were counterbalanced at the three testing times to ensure that there was no task-effect. The accuracy of use of the two tenses was scored by means of the targetlike use analysis (Pica, 1983) which takes into account the overuse of the target forms. Mixed design Repeated measures ANOVAs, followed by the ANCOVAs, were computed to examine the effects of the treatment. Overall, the results indicate that recasts were more effective than clarification requests for the acquisition of both target structures. Recasts resulted in a high level of uptake with repair but, by and large, this was not associated with learning, except for the long term gains in oral production of passé composé. The mediating role of individual learner differences was examined by means of Multiple regression analysis. Language analytic ability predicted the gains for both grammatical structures in oral iii production, and for passé composé in written production of the Clarification Request group. In the Recast group analytic ability predicted only the long-term gains for passé composé in oral production. Working memory measured by the Speaking span test predicted only the short-term gains in oral production of passé composé in the Recast group. Language learning anxiety did not appear to have any influence on the effects of corrective feedback on learning. The results are discussed in terms of cognitive-interactionist theories of L2 acquisition. In general they are supportive of the Interaction and Noticing hypotheses, and of the role played by input rather than of the skill-learning theory and the role of modified output. The findings suggest that teachers need not be afraid to make the corrective force of recasts clear to learners during communicative tasks.