| This dissertation sought to explore the effectiveness of metalinguistic corrective feedbacks(MCFs)and recasts on L2 learning of English articles in ten contexts,and how the effectiveness is correlated with participants’ language aptitude.The ten contexts investigated in this study include:(1)a in[-definite,+specific,+partitive]contexts,(2)the in taxonomic anaphoric conditions,(3)the in anaphoric bridging contexts,(4)the in the names of rivers and oceans,(5)the in the names of islands and mountains,(6)the in the names of geographical parts,(7)(?)in the names of lakes,(8)(?)in the names of squares and parks,(9)(?)in the names of canyons,falls and valleys,and(10)(?)in the names of streetsFollowing a pre-treatment-post-delayed research design,the current study recruited altogether fifty Chinese L2 English learners,with twenty-five in the metalinguistic group and twenty-five in the recast group.Each participant completed an untimed forced-choice elicitation task(FCET),as a measure of explicit knowledge,and a timed grammaticality judgement task(GJT),as a measure of implicit knowledge,in each of the pretest,immediate posttest,and delayed posttest stages MCFs and Recasts were provided for participants in a form of guided oral production task.Each participant completed LLAMA B,LLAMA D,LLAMA E,and LLAMA F,which measured different components of their language aptitudeResults indicated that in the untimed FCET,MCFs and Recasts were effective in the immediate posttest,but not in the delayed posttest.In contrast,they played a facilitative role in both the immediate posttest and the delayed posttest of the timed GJT.Between-group difference was only found in the pre-to-immediate gain scores in the untimed FCET,with the metalinguistic group outperforming the recast group With regard to the correlation between language aptitude and participants’performance in the two posttests,only scores of LLAMA E,a measure of phonemic coding ability,was found to be predictive of the posttest scores of the untimed FCET There was no interaction between feedback types and LLAMA E scores,indicating that the difference between the two feedback types is the same for participants with different LLAMA E scoresThe study further discusses the results.The enduring effectiveness of the two feedback types in the timed FCET suggested the stability of implicit knowledge.This might have something to do with the relatively high proficiency level of the participants,who tended to rely on their implicit knowledge rather than the explicit knowledge.This finding may lend some support to Bialystok(1990,1991,1994),which held that implicit knowledge developed earlier than explicit knowledge,with the former being the source for the latter.The superiority of MCFs to recasts demonstrated in pre-to-immediate gain scores of the untimed FCET,can be attributed to the explicitness of the MCFs,making the target forms more perceptually salient than recasts.Except for pre-to-immediate gain scores,the two types of feedback seemed to play an equal role.This could be attributed to the laboratory setting,in which distracting factors were well controlled In this sense,metalinguistic feedbacks,characterized by saliency and explicitness,did not display obvious advantages over recastsThe reason why LLAMA E scores could predict the performance of untimed FCET in the posttest,with the help of the two types of feedback,is that great phonemic coding ability allowed L2 learners to notice the articles,the unstressed feature of which makes it less salient to participants with lower phonemic coding ability.In addition,with outstanding phonemic coding ability,they were capable of deconstructing the articles into phonetic forms and morphological forms and analyzing forms and meaning separately.Therefore,they were more likely to establish form and function mappings and obtain more benefits from the feedback. |
