【作者】 刘陈艳；

【导师】 俞东明；

【作者基本信息】 上海外国语大学 ， 英语语言文学， 2014， 博士

【副题名】基于非英语专业大学生的语料

【摘要】 为了确定中国大学阶段英语学习者的语用能力，本文选取他们请求习得所处的语用阶段作为切入点，建立了一个中国英语学习者实施请求言语行为的语料库。语料收集方法为口语语篇补全测试，研究对象为山西某高校112名大学生，调查使用的26个涉及请求言语行为的情景并非作者选定，而是三次问卷调查的结果，涵盖了学习者在教室和宿舍实施请求的基本需求。本文用于分析的学习者语料只涉及其中四级组20名、六级组20名的请求表达。本族语语料则来自美国两所大学20名本族语者。两个语料库研究对象年龄相当，受教育程度接近。研究方法包括错误分析、对比分析、基于语料库的研究等。研究工具有GoldWave录音软件，SPSS分析软件和AntConc检索软件。主要研究结果如下：一、学习者实施请求时的偏误类型。中国学习者在校园内实施请求时出现偏误的句子超过四分之一，其偏误类型大致可以分为语法错误（句法错误、词汇语法错误）与语用失误（语用语言失误和社会语用失误）两大类。语用失误占到总偏误的78.9%，其中又以社会语用失误居多，高达58.6%。不同情景中学习者的偏误次数不同，其中说话人和听话人之间权势关系对偏误次数影响最大，学习者在等级体系中出错频率明显增加，主要表现在招呼语的错误使用上。不同水平学习者在语法错误方面没有差异，均占总偏误的10.5%，语用失误方面，四级组高出六级组，前者47.3%，后者31.6%，差异主要体现在四级组招呼语错误使用一项就高达67次，占总偏误的23.5%，高出六级组14%。二、学习者内部请求言语行为使用特征从组成请求的中心言语行为、招呼语和修饰语三个方面讨论。在中心言语行为的三大策略——直接策略、暗示和规约类间接策略的使用上，研究表明他们偏好规约类间接请求，占总数的84.5%，倾向于使用固定的结构，具体表现为“情态动词+代词+动词”的情态序列来实施请求言语行为；学习者不能从功能出发，没有将形式与功能结合起来考虑以达成交际目的。情态动词使用先后顺序为can/could、will/would、may。六级组和四级组在情态序列选择的范围上没有变化，不能体现语用语言学习得的差异；任务的不同强加程度、场景的转换和交谈对象的变化对于四级组和六级组学习者偏好的情态序列没有影响，两组学生在社会语用能力方面没有明显区别。不同水平学习者招呼语使用基本相同。权势关系对学习者招呼语的选择影响很大，不仅种类上有差别，数量上也比较悬殊。交谈对象权势关系高，称呼语使用数量明显增多。和谐体系中四级组与六级组使用频率较高的招呼语前后顺序与等级体系有所不同，和谐体系中交谈双方彼此熟悉，招呼语使用次数降低。就调节请求的内部修饰语而言，学习者使用的词汇手段一共五类，其中以礼貌标记语“please”为主，占80%，主观表达次之，占到总数的10.5%，低调陈述语、恳求语和协商性表达总共约8%。这五类内部修饰语在不同礼貌体系中出现频率各不相同。任务的不同强加程度、权势关系和社会距离对不同水平学习者内部修饰语的选择没有产生影响。就缓解类辅助语（外部修饰语）而言，学习者语料中一共出现五类，它们分别是：准备语、提供理由、许诺给予回报、消除拒绝和降低强加程度。其中偏好的主要形式均为“提供理由”，占所有辅助语的86%，其他四类总共只占到14%，这五类辅助语在不同礼貌体系中频率也不相同。社会距离和权势关系对学习者偏好的辅助语产生一定影响，不过依然是以“提供理由”为主，其他形式出现频率偏低；但是他们在交谈对象身份和社会距离发生变化的情况下知道借助不同辅助语形式的来达成请求言语行为目的的实现。不同任务强加程度下学习者并不善于通过缓解语的调整实施请求。不同水平学习者之间在辅助语使用方面并不存在显著性差异。三、对比研究结果表明：1）学习者在招呼语使用的数量上与本族语者存在显著性差异。其中使用引起注意词语的数量在四级组、六级组和本族语者两两之间均存在显著性差异，三者之间在数量上呈现递减；称呼语使用的数量在学习者内部没有差异，但是与本族语者相比存在显著性差异。在称呼语种类的使用上，学习者和本族语者之间几乎找不到共同之处。2）在中心言语行为的三大策略——直接策略、规约类间接策略和暗示上，学习者与本族语者之间均没有显著性差异，三组研究对象使用间接策略达到80%以上。就其中的间接询问策略而言，三组虽均偏好能力类情态序列，但是在其他三类——允许类、意愿类和可能性的偏好顺序上有所不同；两组学习者在能力类和可能性间接询问策略的使用上接近本族语者的水平，在允许类上四级组与本族语者之间存在显著性差异，在情愿类上四级组和其他两组（即六级组和本族语组）之间均存在显著性差异。在句法上，学习者在能力类和允许类情态序列的使用与本族语者接近，在情愿类和可能性上本族语者语料中出现一些学习者没有使用的句型，这些句型大都更加委婉、礼貌，给对方留下更大的选择余地；另外本族语语料中出现了大量的独特动词，相比较而言，学习者使用的动词非常单一。3）修饰语的对比从内部修饰语和外部修饰语两个方面展开。内部修饰语方面，本族语语料中出现了句法缓解表达，而学习者语料中只出现词汇缓解语，学习者一共使用五类词汇缓解语，本族语语料中没有出现恳求语；就词汇缓解语而言，学习者和本族语者在数量上没有显著性差异，但是在偏好的顺序上略有不同，除了三组调查对象均大量使用礼貌标记语“please”之外，本族语者偏好“低调陈述”，而学习者则喜欢“主观表达”。外部修饰语方面，学习者和本族语者语料中均出现五类，除了在偏好顺序上略有不同之外，学习者和本族语者在外部修饰语使用的数量上存在显著性差异，其中六级组大量使用“提供理由”来达成请求的实施，与本族语者之间差异显著。另外，本族语语料中出现了低调陈述语“a bit”。综合以上研究结果和Kasper&Rose（2002：140）的语用习得五阶段论，本论文认为我国在校非英语专业大学生语用习得大致处于第三阶段和第四阶段，即变通惯用法阶段(主要使用规约类间接表达)和语用拓展阶段（出现复杂句型，开始使用修饰语）。在分析语料和解释现象的过程中，本文在理论上做了两个尝试：用“标记—偏误假设”解释学习者语法错误的发生的条件；提出外语学习者社会语用知识结构三分图剖析学习者称呼语使用中母语负迁移发生的原因。对于中国英语学习者大学阶段在实施请求过程中偏离本族语者表达的现状，本文从五个角度（可获得的语用输入、课堂教学的影响、学习者的语言水平、目的语国家的生活经历和母语迁移）中选取语用输入和母语迁移进行较为深入的讨论，目的是帮助学习者克服母语负迁移的消极影响，合理利用已有的语言资源，同时加强有效的语用输入，着眼于新的知识体系的建构。更多还原

【Abstract】 In order to find out the pragmatic competence of Chinese learners of English，ourimportance is attached to their acquisition-phase when making requests. A learners’corpus was created, subjects being112college students and their performance to makerequests in26situations recorded via Goldwave, one of the software for recording, andthen transcribed in order to avoid being criticized of collecting the oral data by writtenOral Discourse-Completion Tests. The26situations were decided on by learnersthemselves rather than by researchers after three-round questionnaires. Data of Englishnative speakers when making requests have been collected in two different universitiesto compare with the40learners group (20CET-4group and20CET-6group in thecorpus). The research methods adopted include: error analysis, contrastive analysis andcorpus-based research methods. Research tools involved are recording software----Goldwave, analysis software----SPSS, and search software----AntConc.The main findings presented are as follows:1. Error analysis shows that Chinese learners of English, when making requestson campus, make errors almost in more one fourth of their utterances. The errors theymake fall into two categories: grammatical mistakes and pragmatic failures. Thegrammatical mistakes consist of syntactic mistakes and gramma-lexical mistakes andthe pragmatic failures cover pragma-linguistic failures and socio-pragmatic failures.The data shows pragmatic failures constitute78.9%of the overall errors, of whichsocio-pragmatic failures alone are58.6%.The number of errors is influenced by both the social status of listeners andspeakers’ achievements. The number of errors, particularly that of the address forms,increases in hierarchical system according to Scollon&Scollon (2001). Even thoughCET-4and CET-6groups show no differences in terms of grammatical mistakes (bothaccounts for10.5%of the overall mistakes), the former made more pragmatic failuresthan the latter (47.3%>31.6%), out of them wrong address forms alone make up23.5%in CET-4groups and9.5%in CET-6groups respectively. 2. The learners’features were analyzed from the three perspectives: alerters, headacts and modifications, which make up request sequences.Among direct requests, hints and conventional indirect requests (three strategiesof head acts), learners prefer the last one, accounting for84.5%of the total requests.Researches on indirect requests show that the query preparatory is the learners’ priority.The most popular pattern for learners is the modal sequence----“modal verb+pron.+verb”, and favorite modal verbs are can/could, will/would, may. CET-4and CET-6groups made no difference on the choice of range of modal sequences when makingrequests, showing no pragma-linguistic development; their preference for modalsequence remains the same when dealing with different imposition of tasks withlisteners from different social distances in different situations, showing no socio-pragmatic development too.Learners from two groups tend to adopt the same kinds of alerters. Social powerseem to exert great influence on their preference of alerters in quantities as well as incategories. The higher the listeners’social status are, the greater the number of alertersbecome. Learners’ preference for alerters shows difference when they make requests indifferent politeness systems. The more familiar they are, the less alerters they turn to.With regards to internal modification, learners adopt five mitigating lexicons, ofwhich politeness marker “please” compose of80%, the subjectivezers10.5%,understaters, appealers and consultative devices altogether approximately8%. Thefrequency of the five mitigating lexicons shows differently in different politenesssystems. The imposition of tasks and learners’ differences among groups have no effecton learners’ range of internal modification.Five external modifications (supportive moves) are evidenced in learners’ corpus,that is, preparator, providing grounds, promise of reward, disarmer and impositionminimizer. Learners prefer “providing grounds”, alone accounting for87%and theother four14%. The frequency of the five supportive moves shows differently indifferent politeness systems. Power relations and social distances imposition of tasksmake an effect on learners’ choice of supportive moves but “providing grounds” arealways their priority; while they do not know turn to diverse supportive moves when dealing with varied imposition of tasks, and no significant differences are found in twogroups’ performance.3. The results of contrastive analysis shows that:1) Significant differences are found in the number of alerters between learners andnative speakers. There has been progressive decrease in the use of attention-gettersamong CET-4groups, CET-6group and native speakers; learners and native speakershave nearly nothing in common in the way to address others, and they show differencein the quantities of address forms.2) There is no striking difference between learners and native speakers in terms ofthree strategies related to head acts, namely, direct strategy, conventional indirectstrategy and hint. The subjects tend to use indirect strategy, accounting for over80%ofthe three strategies.As to the four types of query preparatory (conventional indirect strategies consistsof query preparatory and suggestory formula, and query preparatory is made up ofability, possibility, permission and willingness), three groups of the subjects preferability modal sequences, but they show different preference order concerning the otherthree types. About the ability and possibility query preparatory，two learners groupsperform almost the same as native speakers. As for permission，CET-4group showsgreat deviation from native-learner group. Regarding willingness, there is a greatvariation between CET-4group and the other two groups (CET-6group and nativespeakers group).When it comes to syntax, learners apply almost the same expressions as nativespeakers in the use of ability modal sequences and permission modal sequences. As forwillingness and possibility modal sequence, some patterns that have never been usedby learners can be found in native speakers’ corpus, most of them being more polite andmoderate. In addition, native speakers use many a special verb far beyond learner’sreach while learners focus on limited frequently-used ones.3) The contrastive study of modifications has been made in two aspects: internalmodifications and external modifications (or supportive moves). Regarding internal modification, native learners know to use mitigating syntacticforms with no appealers while learners only use mitigating lexicons instead. Learnersuse five types of mitigating lexicon altogether, the number of which approximatelyreaches native speakers’. But they show a bit difference on preference order----inaddition to their priority of politeness markers, native speakers’ second choice go to“understaters”, while learners stick to “subjectivezers”.Both learners and native speakers use five types of external modifications. A greatdifference appears in the number of external modifications adopted by learners andnative speakers. CET-6group uses “grounders” in a large number to make requests,varying considerably from the group of native speakers. Moreover,“a bit” isevidenced in native speakers’corpus.According to the five-phase division of pragmatic development proposed byKasper&Rose (2002:140), two phrases can be traced in learners corpus, namely thethird phase----unpacking formulas, which is featured by turning to indirect requestsand the fourth phase----pragmatic expansion, in which complicated syntactic formsare found and supportive moves come up.Two theoretical attempts are made in the process of explanations: one ismarkedness-errors hypothesis to explain in which situation learners make errors whenmaking requests, and the other is tri-parts of socio-pragmatic knowledge of foreignlanguage learners to explain why they transfer wrong address forms to foreign language.Out of five factors (available pragmatic input, classroom instruction, learners’level of proficiency, length of stay in target-language country and mother-tonguetransfer) involved, two are discussed in depth afterwards, that is, mother-tongue transferand the available pragmatic input.更多还原