【作者】 张秋杭；

【导师】 许余龙；

【作者基本信息】 上海外国语大学 ， 外国语言学及应用语言学， 2014， 博士

【副题名】以汉语书面叙事体篇章为例

【摘要】 本文以汉语书面叙事语体关系从句为研究对象，以功能主义的语法观为理论指导，采用语篇功能的研究视角，在验证已有的类型学假设的基础上，从汉语自身的特点出发，来探索汉语关系从句的规律。关系从句是典型的复杂结构，是语言研究中的一个重要课题，因而吸引了大量不同角度的研究。而汉语传统语法长期缺乏“关系从句”的概念，关系从句这类现象通常被看成是偏正结构的一个类型，即一种复杂的名词性偏正结构。20世纪80年代随着当代类型学的研究理念和研究方法不断引入我国，从语言共性的角度，来重新审视这一普遍性的语言现象，才明确提出“关系从句”的概念。可见，“关系从句”是源于印欧语的概念，如何立足汉语自身的特点来加强汉语关系从句的研究是亟待解决的问题。目前国内对关系从句的研究存在重短语轻小句、重形式轻功能的状况；而国外对汉语关系从句的研究主要是把汉语关系从句作为一种特殊的语言材料来加以考察，并将其实验结果与印欧语言进行对比，而对汉语关系从句本身的理论体系建设，关注并不多。综观目前国内外此课题的研究现状，启发我们从语篇功能的视角来加强汉语关系从句的研究。本文的研究内容主要包括两个方面：一是通过汉语关系从句验证已有的类型学假设；二是立足汉语自身的特点来加强汉语关系从句研究。通过汉语关系从句验证已有的类型学假设，包括：（1）验证AH假设和通格假设。AH假设作为影响最大的可及性假设，在母语习得和二语习得研究领域得到了广泛的应用和检验，而与之相对的通格假设同样具有跨语言的普遍解释力。基于汉语关系从句来验证AH假设和通格假设，分析两大假设的不同预测能力。研究结果表明从频率分析的角度AH假设和通格假设并不是互相排斥的，两者在汉语书面关系从句中均得到了支持，但通格假设较之AH假设能更好地解释汉语的关系化现象。（2）验证优先论元结构假设。优先论元结构强调语法和语篇的互动关系，而且该假设在信息压力较大的语境中才能更好地发挥其预期功能，所以我们选择书面叙事语体是比较理想的检验材料，而且关系化并不会导致“优先论元结构”的中断。研究结果表明汉语关系从句中S-RC和O-RC基本遵循PAS的制约,但是A-RC严重违背PAS的制约，再次说明了语体差异对语法现象的影响。立足汉语自身的特点加强汉语关系从句研究，包括：（1）汉语关系从句的谓词类型与论元分布。从汉语关系从句的谓词类型入手，引入事件谓语和属性谓语的概念，提出鉴别形容词作关系小句的句法标准，揭示不同谓词类型的关系从句在句法语义结构和语篇分布方面表现出的不同特点，并从语篇功能的角度给予初步的解释，从而为汉语关系从句范围的划定提供事实依据。（2）汉语关系从句的指称类型与语篇功能。从核心名词的指称类型入手，改变以往多从“有定”和“无定”范畴讨论核心名词的指称类型，而是采用单指、类指和无指的划分方法，讨论不同类型的汉语关系从句在语义属性和语篇功能方面的差异性，并结合语义和语篇功能来进一步解释相应的句法现象。研究结果表明单指关系从句主要起叙述功能，而类指和无指关系从句主要起描写功能，前景主要由表叙述功能的单指关系从句构成，汉语关系从句也可以表达前景功能。在主句分布上，三类关系从句都集中出现在主句宾语位置。（3）汉语关系从句中指数量标记分布的语篇功能解释。从指数量标记入手，讨论指数量标记有无的对立，前后顺序的对立以及内部小类在语篇分布上的差异，研究结果支持指数量标记“前景化”功能假设；指数量标记具有前置倾向，关系化的对象和关系小句的语篇功能在一定程度上影响指数量标记的前后分布，而生命性因素、核心名词的信息地位以及语篇显著性并不影响指数量标记的前后分布，而且指数量标记内部，即指量结构和数量结构前后分布也存在一定的差异性；标记内部“一+量”分布最广，“那”的用法要多于“这”。这些现象说明指数量标记的使用并不是由句法语义层面决定的，而是由篇章层面的因素决定的。全文始终贯彻语法研究的语篇意识和语体意识。本文受到功能主义特别是篇章语法研究理念的影响，讲究完全从真实语篇中研究语法，重视句法现象背后的语篇功能动因，立足语篇对句法现象作出解释。语法研究必须以具体的语体为中心（陶红印1999），任何语法研究都是依托一定语体的语法研究，所得出的规律也是基于某一具体语体的语法规律。本文的研究是基于书面叙事语体，所获得的研究发现只是关系从句在汉语书面叙事语体的表现和规律。虽然局限一具体语体，但这种探索也有助于汉语关系从句整个课题的研究。更多还原

【Abstract】 This dissertation studies the relative clauses in written Chinese narrative. From afunctionalist view of grammar, and a discourse-functional perspective in particular,the author examines the validity of the existing typological hypotheses with Chineselanguage, and then explores the patterns of relative clauses in Chinese by payingspecial attention to the characteristics of Chinese language itself.The relative clause, notorious for its complex structure, has become an importanttopic in linguistic studies and has been discussed from various perspectives. However,the term “relative clause” was seldom mentioned in traditional Chinese linguisticstudies. Instead, the linguistic phenomena covered by relative clauses were usuallyviewed as a subtype of attributive structure, a NP formed by adding a complicatedmodifier to a noun. It is not until the1980s when typology began to be introduced intoChina that this structure is reexamined, leading to the establishment of the notion of“relative clause” in Chinese language.In this sense,“relative clause” is a notion originated from Indo-Europeanlanguages. Therefore, a problem that has to be addressed is how to conduct similartypological studies on Chinese relative clauses within the Chinese grammaticalsystem. The domestic studies on this topic are mostly discussions which view thistype of structures as phrases rather than clauses, and more attention has been paid tothe formal aspect rather than the functional perspective. The studies of Chineserelative clauses by foreign researchers are usually contrastive studies in which theresults from Chinese data is compared with the results from western languages, anddo not help much to the systematic and theoretical studies of Chinese relative clausesitself. A sketch of the current studying situation on Chinese relative clauses seems toindicate that approaching the Chinese relative clauses from the discourse-functionalperspective might shed more lights on this topic.The present research includes two parts:(1) the examination of existingtypological hypothesis with the data obtained from Chinese relative clauses;(2) a systematic study of the Chinese relative clauses which accords with the characteristicsof Chinese language in general.The typological hypotheses that are examined with the data obtained fromChinese relative clauses include:(1) AH Hypothesis and Absolutive Hypothesis.AH Hypothesis is the most influential theory in the study of relativization, and hasbeen placed under extensive tests as well as into various applications in fields of bothfirst and second language acquisition. The Absolutive Hypothesis is believed to haveequally strong interpretive power in many cross-language studies. The two hypothesesare examined with Chinese relative clauses and are analyzed from the perspective ofpredicative power. The result from frequency statistics shows that AH Hypothesis andAbsolutive Hypothesis are not incompatible because both are supported by our datafrom written Chinese relative clauses. However, comparatively speaking, AbsolutiveHypothesis provides a better framework than the AH Hypothesis to account for theusage of Chinese relative clauses.(2) the Hypothesis of preferred argument structure.This theory stresses the dynamic relationship between grammar and discourse. Itspredicative power is best observed in situations where information load is heavy.Therefore, our choice of the written narrative as the language data seems to beappropriate and the relativization does not interrupt the “preferred argument structure”.The result demonstrates that in Chinese relative clauses, S-RC and O-RC basicallyobserve the PAS constraint, while A-RC does not. This again points to the influence ofgenre difference on grammatical relationship.The second part attempts at a systematic study of the Chinese relative clauseswhich accords with the characteristics of Chinese language. This includes:(1) ananalysis of the predicate types of relative clauses and their argument distribution.Beginning with the classification of different predicate types, the author introducesthe notion of stage-level predicate and individual-level predicate, and proposes thesyntactic criteria for recognizing adjectives as relative clauses. Then the syntactic,semantic and distributional features of relative clauses with different types ofpredicates are discussed. A tentative interpretation from the perspective of discoursefunction is also provided so that it will offer factual basis for circumscribing the Chinese relative clauses.(2) The referential types and discourse function of Chineserelative clauses. Different from most existing researches which discuss the referentialtype of head nouns in terms of “definite vs. indefinite” contrast, the authordistinguishes three types (individual, generic and non-referential type), and discussesthe differences of their corresponding relative clauses in terms of semantic attributesand discourse function, so that the related grammatical phenomena can beappropriately explained. Our research shows that individual-referential relativeclauses mainly serve the narrative function, while generic and non-referential relativeclauses mainly serve the descriptive function. Therefore, foregrounding is realized asthe individual referential relative clauses (which is also a proof that Chinese relativeclauses can also be used for foregrounding).With regard to their roles in main clauses,all the three types of relative clauses appear mainly in the object positions.(3)Thediscourse-functional study of the distribution of deictic, quantitative and classifiermarkers (DQC markers) in relative clauses. The study focuses on the presence andabsence of DQC markers that modify the head NP, and their word order variation(before or after the relative clauses) and the distribution of different subcategories.Our data support the foregrounding hypothesis of DQC markers. The DQC markerstend to precede the relative clause. The word order variation is to a certain extentinfluenced by the syntactic roles of the relativized NPs and the discourse function ofrelative clauses. Our research also demonstrates that the animacy of the head NPs, theinformation status of the head NPs, as well as discourse salience do not influence theword variation of DQC markers. Besides, the distribution patterns of deictic classifiermarkers and the distribution patterns of quantitative classifier markers are alsodifferent. Among the DQC markers, yi (one)+classifier has the widest distribution,and the occurrence of na (there) is more frequent than zhe (here).These show that theusage of DQC markers is not determined by factors at the grammatical or semanticlevel, but rather the factors at the discourse level.The whole dissertation sticks to the principle that grammar should be studied inrelation to factors in discourse and genre. Guided by the functionalist approach andthe notions of discourse grammar in particular, the present research stresses the importance of studying grammar in real discourses, and aims at finding out thediscourse motivations behind grammatical phenomena. We endorse the idea that anaccount of grammatical phenomena should be given on the basis of analyzing factorsat discourse level, and that grammatical study should center around some specificgenre and have it as the basis of analysis. The grammatical rules thus obtained are therules that ground on that genre. This dissertation is based on the written Chinesenarrative, and therefore the findings are but the manifestation and rules of the Chineselanguage in this genre. Though limited within this genre alone, the exploration is stilla useful attempt in the systematic study of Chinese relative clauses in general.更多还原