The purpose of the present study is to investigate effects of recasts on Korean language learner's linguistic accuracy that arose incidentally in response to during oral interaction between a Korean Language Teacher and a Korean language learner.
To...
The purpose of the present study is to investigate effects of recasts on Korean language learner's linguistic accuracy that arose incidentally in response to during oral interaction between a Korean Language Teacher and a Korean language learner.
To this end, the current study adopted quasi-experiment design with
38 Korean language learners who completed 800 hours Korean lessons were divided into 2 experiment groups and control group. All the participants had writing task and subsequent dyadic interaction with Korean language instructor on the their writing. Under the notion that recasts can be provided in an explicit or implicit manner, this study operationalized recasts as explicit type with prosodic emphasis and partial reformulation on learner's nontargetlike L2 production and implicit type with monotonous tone and full reformulation.
The data collection procedures for three testing components are as follows: Before the dyadic interaction, learners were asked to perform writing task which is sharing personal experiences type. When they complete their writing task, learners then participated in oral interaction on their composition. Immediately after interaction, each leaner was given back to the copy of their original composition and asked to identify errors and the correct them if possible. 4 week after immediate testing, they were asked to review their composition again and make any changes if they felt it is not correct.
The data analyses involved examining the immediate and delayed error identification/correction tasks to determine the extent to which leaners were able to identify and correct the errors that were in common between their written and oral performance and that were targeted by recasts during interaction. All errors were categorized into morphological and grammatical, the frequency of errors were analyzed by descriptive statistics and non-parametric statistics.
The study findings as follows: First, there were significant difference on error identification/correction of immediate post-interaction test between experiments and control group. Second, explicit recasts group significantly outperform implicit recast group on immediate test. Third, on delayed test, the significant difference among groups was measured in grammatical errors, whereas in vocabulary errors, there was no such difference. Forth, as mentioned already, there was significant difference among groups by error type. Fifth, individual factor such as leaner's development level did not affect the group performance on post tests. However there was tendency that the difference between more advanced group and less advanced group were increased after interaction. Sixth, there are strong and positive correlation between leaners' uptake and grammar error correction.
These findings show that recasts were effective corrective feedback and adding the element of perceptual saliency such as stress and partial reformulation to the construct of recasts increased the frequency of errors identification and grammar errors correction, while perceptually non saliency such as monotonous narration less effective. These findings show that recasts were effective corrective feedback and adding the element of perceptual saliency such as stress and partial reformulation to the construct of recast increased the frequency of errors identification and grammar errors correction, while perceptually non saliency such as monotonous and whole reformulation was less effective. Regarding the attrition of explicit knowledge on their own errors, however, implicit type of recasts were more effective than explicit one. In other word, it was effective on memory attrition. Based on this result, there is potential that implicit recasts might be helpful for long-term basis. The current study suggest that there is need to make further investigation for the role of implicit recasts through longitudinal study.