| Combining Script-logging Software with stimulated recall, the present study analyzes revisions in English writing processes of 20 non-English major NNES Chinese freshmen, arguing that patterns of revision are in correlation with score-measured text quality which reflects the writers'EFL proficiency on a multi-level basis. Drawing on both the qualitative and quantitative methods, the revision data are investigated within an analytical framework that integrates four dimensions, i.e., objective, syntactic level, location and revision-related pausing. The results show that the high-score group (HSG) and the low-score group (LSG) share similarities with each other respecting global revision distribution, where cases of revision suspense and absence are present in all the essay samples. The two groups differ in the following respects:1) The grammatical monitoring of the HSG is more steady and lasting than that of the LSG; 2) pause-free revisions and lexical-diversifying revisions occupy a larger proportion of the total phrasing revisions in the HSG indicating that the HSG is faster when retrieving linguistic resources and pays more attention to lexical diversity; 3) As to content revisions, the LSG has made more'delimit'revisions triggered by production failures but produced clauses less diversified in terms of syntactic structure, suggesting that the LSG suffers from a conflict between working memory capacity and the cognitive load of retrieving restricted linguistic resource more intense than that of the HSG; 4) the HSG has fewer and smaller revision-proof domains. Negative transfer is traceable in revisions to specific linguistic expressions, varying from writer to writer. The HSG takes more active and effective actions to search for counterparts in English so as to circumvent difficulties in direct translation while the LSG shows a stronger tendency of giving up the production plan in question. In the end, pedagogical implications are proposed based on the findings above. |
PDF Full Text Request