【作者】 易六德；

【导师】 金桂林；

【作者基本信息】 西南交通大学 ， 外国语言学及应用语言学， 2004， 硕士

【摘要】 自任务教学法出现以来，任务类型和任务条件对语言产出的影响受到越来越多的关注和研究。本文旨在研究三种不同的写作文体(叙述文，解说文，议论文)和不同的作文长度(一百五十字和三百字)对作文的复杂度和准确度的影响。 本文采用陈慧嫒和吴旭东的书面语产出模式分析了三种文体的难度，并对三种文体在两种条件下对复杂度和准确度可能产生的影响进行了预测。 一百二十四某校外语学院英语专业的学生参加了此次实验。他们的写作任务相同，但写作要求不同。两个班的学生要求每篇字数在一百五十字左右，用二十分钟完成。另两个班的学生要求三百字左右，用四十分钟完成。随后，从每个班随机抽取十名学生组成两组：长作文组和短作文组，每组各二十名学生。 实验结果表明，文体对复杂度产生了较大的影响。在两种条件下议论文的复杂度都是最高，叙述文其次，解说文最低。长度对复杂度也有影响。长度的增加能导致复杂度的增加。长度对复杂度的影响具有选择性。文体的难度越低，其复杂度受长度的影响越大。文体和长度对准确度有一定的影响，但其影响都很小。准确度并不总是表现出和复杂度反向的对应关系。长度的增加也并不一定导致准确度的降低。 本文最后部分对这些结果进行了解释，并对本研究对外语教学和研究以及外语测试的启示意义进行了讨论。更多还原

【Abstract】 EFL writing is an essential part in language tests in China. It is regarded as a useful tool to fully demonstrate one’s language ability. The influence of task type and task condition on writing performance has been investigated heavily in language teaching and language testing. As for task type, the previous researches have examined different types of writing tasks classified by different criteria according to the researchers’ preference. The researches on the influence of task condition mainly focus only on two conditions: playing vs. no-planning and time-constrained vs. time-unconstrained. This present study explores the influence of three discourse modes and length on Chinese students’ writing performance, which is assessed by measures of complexity and accuracy.One hundred and twenty-four English-major sophomores from four classes of a university participated in this research. Two classes are required to write about 150 words for each of the three tasks (narrative, expository, argument) within twenty minutes. The other two classes are required to write about 300 words for the same three tasks within forty minutes. Then ten students are randomly chosen from each class. These twenty students who write about 150 words form Group A and the other twenty students who write about 300 words form Group B. The complexity and the accuracy of the students’ written texts are measured by C/T (Clauses per T-unit) and EFC/C (Error-Free Clauses per Clause) respectively. T-tests will be performed to analyze the relationships between the variables. The significance level is at 0.05 level.When analyzing the difficulty of each task, Chen Huiyuan and Wu Xudong’s model of written language production is adopted. This model consists of three parts: conceptualizer, formulator, and transcriber. By this model, the cognitive load of each task in this study is discussed and their corresponding influences on complexity and accuracy are also predicted.The findings show that both discourse mode and length do have an influence on complexity. Either in the short or the long writing condition, different discourse modes display different complexity, with the argument taskbeing associated with the greatest complexity, the expository task the least complexity and the narrative task in between. The influence of length on complexity is that with the length increase of a written text, the complexity of the written text will also increase. This is especially evident in the less difficult discourse modes like the narrative or the expository.As for their influence on accuracy, neither discourse mode nor length has much influence. The accuracy of each discourse mode doesn’t always correspond negatively to its complexity in both conditions. With the length increase of a written text, the accuracy of the written text does not always decrease as expected, In all, accuracy remains stable either in different discourse modes or in different length conditions.At last, some implications of this study for language teaching and testing are also given in the last chapter.更多还原