| This paper is intended to provide a relatively comprehensive discussion of aspects of error analysis, and to examine in the meantime to what extent the common practices and findings in the field are feasible when applied to Chinese sophomore learners of English. This reading is achieved by a comparison of what has been observed in the field of EA with my own performance of investigations. By this it is hoped to obtain a better understanding of their implications for researchers, learners and teachers in particultar.Error analysis, replacing contrastive analysis in the late 1960s, reached its highest popularity throughout 1970s.Since then extensive studies have been carried out and voluminous resulting papers have been published. However, there is little literature on EA involving college Chines learners of English. Therefore, further studies are called for to make the matter more transparent.Recording to Corder, error is evident of a deficiency in learner's competencewhile mistake is suggestive of a lack of skill or strategy in performance. To follow the common practices in teaching, error is defined in this paper as any utterance deviant from the grammar of native adults.The concrete anaylizing procedure consists of several successive steps. The first step discussed in the present paper is to identify errors. Corder provides an elaborate procedure for recognizing errors. Following his model, the research I have done on erroneous data from learner's diary shows that 90% of their errors can be tracked down by means of a normal interpretaion, leaving the remaining 7% and 3% to plausible and authentic interpretaion respectively.The second step is to calssify identified errors. The common way is to divide errors into five subgroups: addtion, omission, selection, order and blend. When this criterion is applied to 648 errors collected from students' compositions, selection alone accounts for more that a half of the total errors, followed by omission, addition, order and blend. The study suggests that errors are most likely in selection while least possible in order and blend.The next step is description. Description tries to answer the question of "how" learner's language if defferent from the traget language norm. When applying the criterion of linguistic categories to 648 errors from students' compositions, I have observed that errors concerning verbs will account for 47.2%. Learners' errors in verbs seem to be most prevalent. The least likely places of error are where articles and conjuctions are involved.The forth step is error explanation. Explanation tries to answer the question of "why" learners commit errors. Three error sources have been discussed at length in this paper: interlingual, intralingual and unique. My studies on errors from sphomore learners' translation, diary and composition suggest that the mean percentage figure for transfer errors is around 30%. This is, to my surprise, very close to the ones suggested by Ellis and James. Intralingual errors are twice as many as that of transfer error, leaving less than 10% to iduced errors. These studeis do not point to the claim that, of all the first languages examined, Chinese is the most inhibitory in learning English.I also include error evaluation and error correction here, for I think they have much relavance with teaching. Ther are not as random as some people imagine. In fact, there are some principles at work.Despite what has been said, EA is still an area permeated with uncertainties. There are two major problems: one is methodology and the other is limited scope.Imperfect as EA stands now, its implications for teachers, students and researchers are not to play down.To sum up, so long as learners continue to commit errors, teachers will likewise go on with the business of EA. With sophomores as subjects, few teachers can afford to ignore the interference of mother tongue, which is responsible for about 30% of the total errors. However, greater attention should be paid to developmental errors rather , than to... |
PDF Full Text Request