A Corpus-based Analysis Of Chinese EFL Learners' Syntactic Performance In Written Production

This paper reports on a study that investigated the characteristics and development of Chinese EFL learners' syntactic performance in written production. The target subjects were high school students, lower-grade university students and high-grade university students. Syntactic performance was measured by accuracy and complexity. Both syntactic complexity and accuracy were calculated through frequencies and ratios. Complexity frequencies included T-units and subordinate clauses. Complexity ratios included words per T-unit and subordinate clauses per T-unit. From the subordinate clause frequency derived another two frequencies, finite subordinate clauses and nonfinite subordinate clauses, and from subordinate clauses per T-unit derived another two ratios, called finite subordinate clauses per T-unit and nonfinite subordinate clauses per T-unit. The accuracy frequency referred to the number of errors, including the number of within-clause errors and cross-clause errors. And the accuracy ratio referred to the rate of errors per T-unit, including within-clause errors per T-unit and cross-clause per T-unit. Three results were found, concerning the relationship between syntactic accuracy and complexity, the development of syntactic accuracy, and the development of syntactic complexity respectively. First, measured by frequencies, there was a general negative correlation between syntactic accuracy and complexity in terms of the number of T-unit and errors, whether within-clause errors or cross-clause errors. Though no correlation was found between subordinate clauses and errors, there was correlation between subordinate clauses and cross-clause errors. Turning to each proficiency levels, similar negative correlation was found between accuracy and complexity in terms of T-units and errors only at the high schoollevel and the high-grade university level. Furthermore, the correlation mainly came from within-clause errors. Measured by ratios, as a whole, there was negative correlation between accuracy and complexity in terms of words per T-unit and errors per T-unit, especially cross-clause errors. However, positive correlation was found either between subordinate clauses per T-unit and within-clause errors per T-unit or between finite subordinate clauses per T-unit and within-clause errors per T-unit. When the individual proficiency level was concerned, at the high school level, some correlation was found. For example, there was negative correlation between within-clauses per T-unit and cross-clause per T-unit as well as between finite subordinate clauses per T-unit and cross-clause errors per T-unit. No correlation was found at the other two levels. Second, there was general growth in syntactic complexity by proficiency levels while the growth from lower-grade university level to the high-grade university level did not reach a significant level. Third, coming to the development of accuracy, general decrease was observed. Still, no significant was found from the lower-grade university level and the high-grade university levels. A combination of all these results suggested that studies which investigate Chinese EFL learners as a whole should be differentiated from those investigating learners at certain proficiency level.