A Corpus-based Study Of The Acquisition Of English Relative Clauses By Chinese EFL Learners

The present study was intended to explore the acquisition of English relative clauses by Chinese EFL learners. The study first described the types of the English RCs found in Chinese EFL learners'CET6 writings together with an analysis of the overall frequency distributions of the RCs produced in those test writings, and then investigated the differences between high-achievers and low-achievers in terms of the types and frequencies of the RCs they produced in their writings, followed by the analysis of the correlations between Chinese EFL learners'proficiency levels and their accuracy of the RCs they produced. Finally the study examined the developmental pattern for RC acquisition characteristic of Chinese EFL learners in their CET6 writings.300 pieces of discourse were randomly picked up from Chinese Learner English Corpus (CLEC), sub-corpus ST4, which has been built upon the CET6 Writing Tests. Those texts contain 264 RCs, therefore, they were chosen as the samples of investigation in the study. The data analysis generated the following findings:First, in terms of grammatical functions, there are five types of RCs found in Chinese EFL learners' test writings:SU, DO, GEN, OPREP and OCOMP. The double-object-structure IO type is not included, possibly because the style or content involved in such test writings does not need such RC structures and the students do not have opportunities to use this type of RC.Second, of the collected 264 RC samples, the frequency distributions of those five RC types produced by Chinese EFL learners are ranked as:SU>DO>OPREP> OCOMP>GEN (">" means larger in number), with their respective percentage:59.1%, 31.4%,5.7%,2.7%and 1.1%. However, the overall frequency distributions of English RCs in the selected sample sentences are rather low, with only 6.7%in the total number of sentences in the randomly selected pieces of discourse. The possible reason is that typological differences pose great difficulties for Chinese EFL learners; as a result, they might substitute other simpler structures for the RCs or resort to the strategy of avoidance. This confirms the hypothesis by Schachter (1974):a low frequency of RC use with learners whose L1 does not have RCs and a high frequency of RC use with learners whose L1 have RCs.Third, the five types of RCs are all found in both HG and LG's CET6 writings. The cross-group comparison of RC frequency distributions reveals:HG exceeds LG in reference to DO, OPREP and OCOMP types; by contrast, LG exceeds HG only in the case of SU and GEN types, which is the most striking disparity between the two groups.Fourth, the high-proficiency students perform better than the low-proficiency ones in the use of RCs. In other words, there is a significant positive correlation between the proficiency levels and the RC accuracy. To be specific, Chinese EFL learners' acquisition of RCs is a continuous process along with the improvement of their target language proficiency levels. This accords with Chiang's finding (1980) that the production of English RCs is strongly determined by learners'overall language proficiency.Fifth, the developmental pattern represented by the accuracy order of RC use by Chinese EFL learners can be summarized as:SU>DO>OPREP>OCOMP>GEN>IO (">" means more easily acquired). This basically agrees with the accessibility hierarchy proposed by Ellis (2000) but disagrees with it in the case of the 10 and OCOMP types. The possible reason for the higher accuracy rate in the production of SU and DO types of RCs may be that learners place greater emphasis on those types and have learnt them more systematically or that learners avoid more marked structures, choosing to use SU or DO structures instead. The reason for such disagreement may be that:Ellis'model is based upon the data collected from the target language English and may not be totally applicable to the Chinese EFL learners. Another possible reason for this partial inconsistency may involve other factors, such as negative language transfer, the differences of topics, contents and styles required for test writings, and learners' processing strategies; all these, to some extent, interact with each other to affect Chinese EFL learners'acquisition of RCs.The present study has the following pedagogical implications:First, teachers are expected to help students weaken the impact of source-language negative transfer by raising the learners'awareness of target-language thinking pattern and strictly obeying the organizing rules of the target language sentences; Second, teachers are supposed to ensure a very high-frequency of the RC structure input and output in their teaching practice; Third, teachers should set more reasonable learning sub-goals in different periods and follow certain developmental pattern of RC acquisition systematically in their scheduling of the instruction.