【作者】 周敏；

【导师】 韩景泉；

【作者基本信息】 中南大学 ， 英语语言文学， 2012， 博士

【摘要】 文献中对汉语关系结构的分类,特征描写颇有争议,对其句法生成机制的刻画更是分歧重大。本文以Chomsky(2000,2001,2005,2007,2008)的语段理论为理论框架,在充分探讨汉语关系结构的分类与特征的基础上,重新审视了现有的相关研究,分析其中存在的问题,并为汉语关系结构的生成提出了一种新的分析。本文首先简要概述了语段理论框架,介绍了其发展的历史进程,语言机制模式和句法运算系统。接着对汉语关系结构的特征及分类进行了更为全面细致的概括。汉语关系结构有六个基本特征：无关系代词,关系中心语可省略,复指代词能滞留,关系从句能独立使用,允许多项内嵌以及关系从句顺序自由化。汉语关系结构在句法上的分类主要根据三种特点：有无空缺,有无中心语以及有无标记词“的”。语义上可分为限制性、非限制性和扩大性关系结构(包括自由、关联和程度关系结构)。在回顾有关汉语关系结构的生成和标记词“的”的文献时,本文指出现有的移位分析不仅使部分语言事实如汉语无空缺和并列关系结构得不到合理解释,而且对诸如经济原则、空语类原则、孤岛限制以及领属条件等一些普遍语法理论原则构成严重挑战。非移位分析的问题主要体现在对关系从句中空语类的处理。关系从句中的空主语并不是必须遵守普遍控制原则的空代词,而空宾语也不是在生成中能改变身份的变项。标记词“的”虽然已得到广泛的研究,但对其句法范畴问题,至今仍未取得一致的认识,已有的C、D和De分析等都存在一定的问题。鉴于前人研究成果,本文对汉语关系结构的句法生成和“的”进行了重新分析。本文认为汉语关系从句中的空缺为特殊空语类,为满足关系从句中谓词的论元要求而基础生成,不属于Chomsky (1981,1982)归纳的四种空语类中的任何一种。这种特殊空语类的存在具有理论和经验上的支撑证据,其特点是只存在于论元空缺关系结构中,有语类和(?)特征,可被赋予题元角色和格,其指称可在句中或语境中得到识别。通过探讨“的”的语法化历程,本文认为“的”实则为一种语法化的修饰标志词,其为汉语所独有,只能标记为De。“的”与其前的关系从句构成中心语—补语关系,两者合并成DeP。因为DeP为修饰语,所以我们认为DeP左向附加于关系名词中心语,主要起语义限制作用而不参与主句的任何句法操作。关系名词中心语与关系从句中的论元空缺在(?)特征上存在一致关系,整个汉语关系化过程中不存在空算子和名词中心语的提升移位。至此,我们为汉语关系结构提出的新句法机制已勾画完成。这种机制主要采用经济的合并操作,而限制其生成能力的是我们为汉语关系结构归纳的语义允准条件。先前分析中某些句子的不合法也可在该条件下获得解释。为了检验新假设的合法性,本文以语段理论为框架,运用该假设对各种汉语关系结构如限制性、非限制性、自由、多项内嵌、并列、以及带“被”、“把”、双宾结构的关系结构一一进行了分析。为了进一步佐证该假设和深化对汉语关系结构的认识,文章最后对汉语关系结构的话题化,焦点化以及外置化现象进行了分析。鉴于汉语关系从句的独立使用特征和内部论元结构的完整性,本文认为汉语关系从句与一般句有相同的句法地位,可以有自己的话题和焦点。汉语关系从句可以先进行内部话题和焦点化,然后再来修饰中心语。汉语关系结构的外置化现象文献中鲜有讨论,我们认为这种现象确实存在,并且可以在我们的假设和Vries(2002)的合并省略分析下得到合理的解释。本文在合理解决各种争议问题的基础上,为汉语关系结构的句法生成提出了新的生成机制。与之前的分析相比,该研究具有理论与经验上的双层优势。从理论上讲,本文放弃移位而采用代价小的合并操作分析,既不违反孤岛和领属条件等生成语法普遍原则,又完全符合Chomsky (1995a)提出的合并高于移位的经济原则。确认汉语论元空缺关系结构中特殊空语类的存在,可以为汉语空语类的理论研究提供启示。从经验上讲,我们提出的新假设不仅可以合理解释汉语关系结构所表现的独特句法特征,如关系从句的独立使用和顺序自由化特征等,而且可以统一各类汉语关系结构的生成分析,弥补先前局限于限制性带中心语关系结构研究的缺陷。本文对汉语关系结构的话题化、焦点化和外置化现象的探讨能合理解释一些违反移位限制但又合法的话题和焦点化结构,可为今后这几类结构的研究提供诸多借鉴。本文属于新语段理论框架下汉语相关事实的实证研究,在检验语段理论有效性和解释力的同时,可加深对汉语事实的认识,为汉语生成语法研究的发展做出了一定的贡献。更多还原

【Abstract】 There have been great controversies over the general properties, classification and syntactic derivational mechanism of Chinese relative constructions (RCs). Within the framework of Phase Theory (Chomsky2000,2001,2005,2007,2008), this dissertation presents a detailed discussion on the properties and classification of Chinese RCs, and puts forth an alternative proposal on the basis of the critical review of the previous analyses.This dissertation begins with a brief outline of the theoretical framework of Phase Theory, introducing its historical development, language mechanism model and syntactic computational system. What follows is an explicit generalization of the properties and classification of Chinese RCs. It is assumed that Chinese RCs are characterized by the following basic properties:no relative pronoun, omissible noun heads, possible retention of resumptive pronouns, independent use of relative clauses, recursive and linear multiple embedding and free ordering of relative clauses. The syntactic classification of Chinese RCs is based on three criteria:presence or absence of a gap, a nominal head and the marker de. Semantically, Chinese RCs fall into three types:restrictive, nonrestrictive and maximalizing (including free, correlative and degree). In the literature review of the analyses on the derivation of Chinese RCs and marker de, this dissertation points out that the movement analysis not only leaves some linguistic facts, namely, Chinese gapless and coordinated RCs unexplained, but also poses serious challenges to the principles and theories of Universal Grammar, such as Economy Principle, Empty Category Principle, island constraints and Subjacency Condition. The major problem of the non-movement analysis lies in the treatment of empty categories in Chinese relative clauses. The null subject cannot be an empty pronominal that must observe Generalized Control Rule and the null object cannot be a variable which may change its status of identity in the process of syntactic derivation. Although the element de in Chinese has been thoroughly discussed in the literature, little consensus has been reached on its syntactic categorial status. The existing analyses such as C, D and De approaches are still problematic.Based on the previous research, the dissertation gives a reanalysis of the syntactic derivation of Chinese RCs and the marker de. We regard the gap in the relative clause as a special kind of empty category, which is generated to satisfy the argument requirement of the predicate in the clause and does not belong to any kind of the four types of empty categories specified by Chomsky (1981,1982). The existence of the special empty category can be proved on both theoretical and empirical grounds. Its properties may include:it only exists in the argument gapped RCs; it has category and φ features and can be assigned theta-role and Case; its reference may be identified in the sentence or in the context. Given the discussion on the grammaticalization process of the element de, it is proposed that de is a grammaticalized modifier-marker, which only exists in Chinese and can be marked as De. The marker de acts as a head to select the preceding relative clause as its complement, forming the phrase DeP. Due to the modifier status of DeP, we argue that it left adjoins to the relative nominal head. The adjunct DeP merely plays the role of a semantic restriction and does not participate in any computational operation of the matrix clause. The argument gap in the relative clause may agree with the relative nominal head in (p features. There are no movements of null operators and nominal heads in Chinese relativization. So far, our new syntactic mechanism proposed for Chinese RCs has been sketched. This mechanism mainly employs the economical merge operation and its derivational power may be restricted by the semantic licensing conditions we specified for Chinese RCs. The ungrammaticality of those examples in previous analyses can also be accounted for under the conditions. In order to testify our new proposal, we adopt it to analyze the syntactic derivation of different types of Chinese RCs under Phase Theory, including restrictive, nonrestrictive, free, multiple-embedded, coordinated RCs and those complicated RCs involved bei-, ba-and double-object constructions. With the purpose to further verify our new proposal and get a better understanding of Chinese RCs, the dissertation finally explores the topicalization, focalization and extraposition phenomena in Chinese RCs. Given the independent use property and complete internal argument structure of the Chinese relative clause, we postulate that a relative clause can have the same syntactic status as a normal sentence with topic and focus positions. The internal topicalization and focalization can be firstly allowed in the relative clause, and then the whole topicalized and focalized relative clause modifies the relative nominal head. There is little discussion on extraposition phenomena in Chinese RCs. However, we maintain that Chinese RCs indeed have extraposition constructions, which can be given a good account under our proposal and Vries’s (2002) specifying coordination plus ellipsis analysis.On the basis of a critical analysis of previous accounts, the dissertation puts forward an alternative account of the derivation of Chinese RCs. It can be shown that the proposed analysis has both theoretical and empirical merits. From a theoretical point of view, the analysis enables us to arrive at a more unitary theory of Phase building. In the dissertation we adopt the Merge operation approach rather than the movement analysis, which may not violate the universal principles in Generative Grammar such as island constraint and Subjacency Condition. Moreover, our analysis is more accordant with Chomsky’s (1995a) Economy Principle of Merge over Move. The finding of a special kind of empty category in Chinese argument gapped RCs gives insight into the theoretical research of Chinese empty categories. From an empirical point of view, the new analysis proposed in the dissertation not only can give a reasonable account for the distinctive syntactic properties of Chinese RCs, for instance, the independent use and free ordering properties of relative clauses, but also can provide a unified derivational account for different kinds of Chinese RCs, remedying the defect in previous research which confines the discussion to headed restrictive RCs. Our exploration on the topicalization, focalization and extraposition phenomena in Chinese RCs not only can give a feasible account for those grammatical topicalized and focalized constructions which seem to violate movement constraints, but also can provide some reference for the further research of these constructions. In analyzing relevant Chinese data under the recent framework of Phase Theory, we can test the validity and explanationatory power of Phase Theory, and at the same time, we can better understand the Chinese data and contribute to the development of generative research on Chinese grammar.更多还原