A Study Of The Self-repairs In The Oral English Of Some Chinese Graduates Of English

Language teachers in their teaching practice frequently observe self-repair behavior of L2 learners in their L2 speech production. Speakers will repair the inappropriate information or correct errors in their speech spontaneously while they are speaking. It's quite a common feature shared both by L1 and L2. Researchers regard self-repair as a useful tool to produce more comprehensible output and an important factor indicating learning. Therefore, observing the feature of self-repair can facilitate us to understand the language production mechanism and self-monitoring in the spoken language.50 graduate students of Grade 3 from Beijing International Studies University (BISU) participated in this research as subjects. Their presentations directed by two IELTS topics are recorded by the recorder pen. Through the qualitative and quantitative analysis, the self-repair frequency, and relationship between types and structures of self-repairs are observed, which facilitates us to reveal the problems existing in the subjects'oral English.Levelt's Perceptual Loop Theory of Monitoring is an important theory for explaining the psychological mechanism of self-repair. It is developed on the basis of speech model. The speech model consists of five major components:conceptualizer, formulator, articulator, audition and speech comprehension system. According to this model, the speech processing is modular and the monitoring work is done by the conceptualizer. Monitoring can control production process, detect errors and initiate repairs. Self-repair is the overt manifestation of self-monitoring. This theory is necessary but not enough to explain self-repair. It should be completed with theory of attention. Attention is an important cognitive mechanism. The degree of attentional control influences speakers' ability to detect errors.The classification of self-repairs is built on Levelt's (1983,1989) category on the first language analysis combined with van Hest (1986) and Kormos (1999a)'s classification on self-repairs of the second language. Analysis on the frequency of self-repairs reveals that the frequency of appropriateness self-repair occupies the highest rate (55%), among which ambiguous reference repair takes up higher rate (23%) than the other two subtypes. The different repair ranks the second (28%), followed by error-repair (18%). We found that self-repairs in the graduates'spoken English shifts from form-focus to content-focus. It means these advanced learners are equipped with higher automatization mechanism to apply second language. But the frequency of error-repair is much higher than expected. In addition, the analysis from the perspective of automatization theory can help us clearly understand the underlying mechanism of self-repair.The studies abroad have gained fruitful results. But researchers in China did not begin to show their interests in this topic until recent years. Thus it is necessary to conduct a comprehensive study on this issue.