| It is well acknowledged that English writing is a must-master skill for second language learners. Compared with other skills, like listening, speaking, reading and translation, writing is a tougher task for learners to grasp as it requires the author to be capable of not only expressing ideas clearly, writing fluently and logically, but also keeping the idea relevant to the topic and making it coherent. In this way, it not only poses great difficulty to current Chinese students to reach the standard but also perplexes English teachers for better teaching.Reference, an important cohesive device, plays an indispensable role in ensuring the cohesion of the text. Since the book Cohesion in English was published, there have emerged many studies on reference in English compositions both at home and abroad. However, these studies all confined themselves in clauses-based reference analysis and deviated from the text. Based on Martin’s discourse semantics idea, the present study attempts to analyze reference from the perspective of discourse semantics.In Martin’s Identification system, reference is divided into three types—presenting reference, presuming reference, and comparative reference. Presenting reference is selected when new participants are introduced to the discourse, while presuming reference is selected when participants are to be tracked in the discourse. And comparative reference is used as a resource either for introducing participants or tracking participants. Adopting Martin’s idea on reference from the perspective of discourse semantics, the paper aims to clarify the following questions:1. What is the distribution of reference use in College English Test compositions? What are major error types of reference in College English Test compositions?2. Is there any significant difference among reference error types at three English writing proficiency levels?3. Is there any significant correlation among reference use frequency, reference error frequency and students’ English writing proficiency?The study adopts a quantitative and qualitative approach. It firstly classifies 86 sample compositions into three levels according to their scores and then analyzes the misuses of reference in each sample. Finally, ANOVA test and Spearman correlation test are employed to test the data.Results firstly indicate that students use a large number of presuming references but quite a few presenting references and comparative references. And correspondingly their reference error distribution follows the same pattern of its use distribution. In terms of presuming reference errors, the error types can be further generalized as pronominal referential errors and nominal referential errors. Among pronominal reference errors, referential ambiguity and referential inconsistency are most salient. Results also show that students at three English writing proficiency levels show great difference in presuming reference errors and nuance in presenting reference errors and comparative reference errors. Students with high English writing proficiency make much less errors than students with low English writing proficiency. Finally, in all the sample compositions, results reveal that reference error frequency is negatively correlated with students’English writing proficiency while it is positively correlated with reference use frequency. But the case in compositions with high scores is a little different as their reference error frequency shows no correlation with their final scores.The research result is of significance to guide students for the effective use of reference in English writing, to deepen students and teachers’understanding of the role of reference in English writing and to enlighten the subsequent reference study from the perspective of discourse semantics. |
PDF Full Text Request