【作者】 蔡艳玲；

【导师】 周国强；

【作者基本信息】 上海交通大学 ， 外国语言学及应用语言学， 2009， 博士

【副题名】认知负荷视角

【摘要】 论文从认知负荷视角探讨二语写作中任务相关变量对中国大学生英语写作的影响。影响二语写作的因素繁多,就写作任务而言也涉及到很多不同的变量。虽然人们对二语写作任务做了富有成果的研究,但很少从认知负荷视角进行,而且对任务的认知负荷缺乏测量。本文既研究二语写作的外部任务环境,也对其内部写作过程做了相关研究,具体研究的是写作任务呈现方式、写作体裁和任务完成方式,即写作过程中认知资源分配对写作成绩的影响,以及从认知负荷视角进行写作任务研究对二语写作的理论意义和教学指导意义。本研究以二语写作理论为研究背景,研究的理论框架是在认知负荷理论和任务复杂度理论研究基础之上建立起来的。认知负荷理论(Sweller,1988,1994,2006;Sweller, Van Merrienboer, & Paas, 1998; Van Merri?nboer & Sweller, 2005)从资源分配的角度来研究学习尤其是复杂学习,认为学习与人类认知资源体系调整一致时能够达到最佳效果。认知体系包含有限的工作记忆和无限的长时记忆,图式习得和自动化可以克服工作记忆的有限性,减轻工作记忆负荷即认知负荷,提高学习效率。认知负荷受到学习任务内在特性(内部负荷)、任务呈现方式(外部负荷)和学习者在图式构建中的认知投入(有效负荷)等影响。任务复杂度是引起认知负荷三个主要因素之一(其它两个为学习者、任务与学习者的关系),也是最为关键的因素(Kirschner,2002)。许多研究者对任务复杂度进行了研究,两个最有影响的任务复杂度模型是:Skehan与Foster (1999, 2001)提出的“注意力容量有限模型”和Robinson (2001, 2003, 2005, 2007)　的“认知假设模型”。两种模型都对任务复杂度一系列相关因素进行具体分类,预测了任务复杂度变化对学习结果的影响,但是他们的预测有明显差异。本研究依据三种认知负荷对任务变量进行分类设计,考察它们对二语写作的影响。研究的问题包括:1.对中国大学生英语写作而言,任务的两种呈现方式(图片型和文字型)带给他们的认知负荷是否有差异?如果有,它们对不同水平学生的英语写作产生什么影响?　2.记叙文和议论文写作对中国大学生的认知需求是否有差异?如果在不同阅读条件下写不同体裁的文章,对他们英语写作有什么影响?3.通过写作指导对英语写作过程干预对中国大学生英语写作有什么影响?为了研究这些问题,我们根据相关认知负荷理论作出假设,设计三个量化研究实验,分别请了不同的被试(93人/120人/64人)用英语来写作,以验证假设,测试任务对二语写作的影响。前两个实验是对二语写作过程外部任务环境的研究,都采用2*2析因实验方差分析,以考察写作外部负荷和内部负荷相关任务变量对EFL写作的影响。第三个实验是对二语写作过程内部认知操作的研究,采用t检验的方法,探索同一写作任务不同写作过程对EFL写作的影响。由于很多认知负荷相关研究缺乏对学习输入和输出的中介变量——认知负荷进行测量而受到质疑,我们在前两个实验中都采用多维度主观认知负荷测量法(NASA-TLX)对特定的写作任务与学习成绩变化的中介变量(认知负荷)作了客观定量分析,来体现不同写作任务对学生的认知需求,第三个实验被试完成的是同一写作题目,其认知负荷总体上一致,因而没有测量。　　　　本研究对任务负荷测量结果、写作成绩和写作文本的语言表现(准确性,复杂性和流利性)进行统计分析。实验结果基本证实了假设,不同写作任务类型对英语写作都有显著影响,说明写作任务相关变量及认知负荷变化的确导致了不同的写作成绩和语言表现。实验一证明,与文字型任务相比,图片型任务认知负荷较低。两种任务复杂度沿资源分散变量方向变化,图片型任务具有清晰的故事结构,写作构思所需注意力资源较少,有助于写作转译阶段的认知投入。文字型任务则需要在构思阶段投入较多的认知资源,因其对写作具体内容限制较少,有助于学生发挥想象力。结果还表明任务呈现方式与大学生英语水平具有交互作用,即文字型任务对高水平学生英语写作有显著促进作用,而图片型任务则有助于低水平学生的英语写作。　　　　　实验二结果证明,议论文内在认知需求大于记叙文。两种任务复杂度沿资源指向变量方向变化,记叙文用第一人称来写,而议论文写作则需要更多逻辑性的推理。值得注意的是,在两种阅读条件下,中国大学生英语写作议论文成绩好于记叙文。这一结果虽然与先前多数母语和二语写作研究结果相反,但它基本支持了Robinson的预测,即任务复杂度沿资源指向变量增加,会提高学习者语言表达的复杂性和准确性,从而得到更好的学习结果。实验还表明,单纯阅读相近主题的文章对英语写作没有产生显著性的影响,而在做笔记阅读的情况下,文章结构及表达在英语写作中的迁移明显增加,对英语写作有显著的促进作用。实验三证实,不同任务完成方式对中国大学生英语写作影响显著,即注意力分配干预对英语写作有显著的促进作用。认知负荷研究表明,总认知负荷不变的情况下,不同的认知操作会使其内、外负荷和有效负荷有所浮动,影响写作成绩。在英语写作过程中,注意力资源首先指向写作高水平认知过程(如内容构思、整体结构布局等),有助于增加有效负荷,促进英语写作。写作低水平认知过程(如句子转译、书写等)对注意力资源的过多消耗会牵制有效负荷。研究凸显了二语写作的独特性,语言问题不仅从整体上增加了二语写作的额外负荷,还影响二语写作过程和作文的语言表达。本研究具有一定的理论意义和实际意义。从理论意义上讲,二语写作理论研究应从不同角度多方位地考虑任务变量,因为研究结果表明特定的任务变量或条件会影响二语写作过程的注意力分配和语言加工;在二语写作领域,也为认知负荷理论提供一定实证支持。本研究具有非常重要的实际意义,为教师在写作教学实践中操纵任务有关变量、利用任务特点更好地调节学习者工作记忆资源、促进写作教学提供了有效的依据和范例。学习者也可以通过负荷量表了解任务认知需求与难度,认识二语写作任务的复杂性,有意识地培养自身写作监控和反馈机制;结合阅读与写作,学会分解文章结构,调节写作各分过程和知识运用,逐步提高二语写作成绩。更多还原

【Abstract】 This research explores task effects on EFL writing by Chinese college students in a cognitive load perspective. A variety of factors influence L2 writing. The writing task is just one factor but with many different task-related variables. Although many studies have obtained informative findings with regard to task effects, few studies have been found to be devoted to L2 writing task effects in a cognitive load perspective. There is a lack in measuring cognitive load levels of writing tasks. This dissertation investigates the external task environment and the internal writing process as well. It focuses on the effects on L2 writing of presentation modes of tasks, discourse modes called for by task writing, and ways of approaching a task, that is, the cognitive resources allocation in the writing process. This study provides empirical support for relevant theories and has pedagogical implications that could be taken advantage of by Chinese learners to improve their English writing skills.This dissertation carries out its research in the field of L2 writing. The research theoretical framework is built on Cognitive Load Theory (CLT) and task complexity theories. CLT (Sweller,1988,1994,2006;Sweller, Van Merrienboer & Paas, 1998; Van Merri?nboer & Sweller, 2005) is a learning theory which has its foundations in information processing theory. CLT suggests that learning, especially complex learning, could be optimized when it is aligned with human cognitive resource architecture. The cognitive resource architecture consists of limited working memory and unlimited long-term memory. Two kinds of mechanisms—scheme acquisition and automation—could overcome the limitation of working memory and relieve working memory load (i.e., cognitive load) so as to raise learning efficiency. Cognitive load may be affected by the element interactivity of the learning tasks themselves (intrinsic load), the manner in which the tasks are presented (extraneous load), or the amount of cognitive resources that learners willingly invested in schema construction and automation (germane load). They are additive in nature. Task complexity is one of the three major causal factors of cognitive load and the most crucial one (Kirschner, 2002). The other two are the learner and the relationship between the task and the learner. Many researchers have studied task complexity. The two most influential models of task complexity are the Limited Attentional Capacity Model developed by Skehan and Foster (Skehan, 1998; Skehan & Foster, 1999, 2001) and Robinson’s Cognition Hypothesis (Robinson, 2001a, 2001b, 2005, 2007). Both of the models identify a series of task-related factors that are relevant to different levels of task complexity, and predict the effects of task complexity on learning outcomes. However, their predictions are of distinct difference in some aspects.This study classifies the task-related variables based on the three kinds of cognitive load, and investigates their effects on L2 writing. The research questions are as follows: 1. a) Do writing tasks presented in picture and in text cause different levels of cognitive load for Chinese tertiary learners? b) If so, what effects do the writing tasks (the pictorial task and the text task) have on the EFL narrative writing of the learners with different English Levels? 2. a) Do narrative writing and argumentative writing impose different levels of cognitive load on EFL learners? b) If writing tasks of different discourse modes are assigned in different reading conditions on the related topic, what effects do they have on Chinese tertiary learners’EFL writing? 3. What effects do different ways of approaching a same writing topic intervened by writing instructions have on Chinese tertiary learners’EFL writing?In order to study these questions, we put forward the relevant hypotheses based on Cognitive Load Theory. Different subjects (93, 120, and 64 respectively) were invited to write in three experiments so as to test the task effects on EFL writing. The first two experiments intended to explore the external task environment of L2 writing process, which tested the effects of task-related variables classified by writing extraneous load and intrinsic load. The third experiment intended to investigate the internal process of L2 writing with cognitive resources allocation, which studied the effects of different approaches to the same writing task. As many cognitive load studies have been challenged for lack of measuring cognitive load, their interpretations of the effects of learning inputs on outputs are largely diminished. In this study, a subjective cognitive load measure, the multidimensional scale NASA-TLX, was selected and modified to be used for Chinese EFL learners in the first two experiments to reflect the cognitive demands of different writing tasks, but not in the third experiment because the same writing topic was assigned to both the experimental group and the control group. Before the application of subjective load measure in the formal experiments for Chinese college students, a pilot study was carried out to explore the structure of the scales. The results show that the questionnaire has good estimates in terms of reliability and validity.In Experiment 1, we investigated the effects of different writing task presentation modes on EFL writing by students of different English levels. A 2ⅹ2 factorial design was employed with two between-subjects factors: English proficiency level (high, low) and presentation mode (pictorial task, text task). The results show that the presentation mode makes a significant difference when interacting with English proficiency level. For the subjects with high and low levels, different patterns exist. For the subjects with higher level, text task promoted their L2 writing significantly than pictorial task. Whereas for the subjects with lower level, though it is not significant, pictorial task worked better because it helped them in translating ideas into written words. Students’assessment of cognitive load in the questionnaires indicates that the text task is higher in cognitive load and more difficult than the pictorial task. The text task resulted in significantly longer text length and more T-units. But it brought lower complexity for high level subjects. It also brought lower accuracy in terms of total errors, though further analysis indicated it was not in terms of errors per T-unit.In Experiment 2, we explored the effects of different discourse modes on students’EFL writing when they are assigned in different reading conditions. This was achieved by a 2ⅹ2 factorial design with two between-subjects factors: reading condition (with note taking, with no note taking) and discourse mode (narration, argumentation). It has been found that intrinsic cognitive load caused by the internal schemata of discourse modes exerts a significant effect on tertiary students’EFL writing. In both reading conditions, the subjects performed better in writing argumentation than narration. This supports the prediction of Robinson that better performance is expected of complex task which increases its complexity along the resources-directing variables. But, reading alone as an orientation task for writing was not effective. Taking notes on the sample text while reading did help students to write better in their own texts. The rating scales indicated that the argumentative writing possesses higher cognitive load than the narrative writing. Argumentative writing resulted in significantly higher syntactic complexity (but not accuracy) than narrative writing.In experiment 3, we examined the effects on EFL writing of different approaches to the same writing topic intervened by instructions. It was carried out by using an independent sample t-test design with two levels of intervention conditions (attention priority instruction versus no attention priority instruction). It is argued that different attention allocation during the writing process will cause fluctuation of intrinsic load and germane load inside the same total load in writing, thus affecting writing outputs. Results show that attention priority given by intervention to high-level writing processes and postponing linguistic consideration in the final draft affected EFL texts and resulted in great and impressive benefits to the writing process and written text quality. This contributes to the increase of germane load helpful for EFL writing. However, in the control group, low-level writing processes consumed great cognitive resources that would otherwise devote to high-level writing processes. This study highlights the uniqueness of L2 writing in that linguistic problems not only pose extra load on L2 writing in general, but also affect L2 writing processes and the linguistic expressions. The significant difference of text length exists in subjects’written texts. But the complexity (T-unit length) in the subjects’texts failed to be significant.In spite of the limitations of this study, the research findings have offered some insights for both theoretical research and pedagogical practice. Theoretically, this research not only provides empirical support for Cognitive Load Theory in the field of L2 writing, but also suggests that studies of L2 writing theories should take into consideration different task-related variables from various perspectives. Pedagogically, this research provides efficient paradigms for teaching and learning of L2 writing. Teachers could manipulate task variables to monitor learners’working memory resources to promote their L2 writing proficiency. Learners could learn more about task cognitive demands and difficulties through the cognitive load rating scales, raising their awareness of self-monitoring and feedback. What’s more, learning writing in combination with reading, they could learn to decompose the structure of a reading text, break the complex writing task into subtasks to lower cognitive load, and improve their L2 writing gradually.更多还原