【作者】 何青青；

【导师】 梁锦祥；

【作者基本信息】 华南师范大学 ， 英语语言文学， 2004， 硕士

【摘要】 关于VP-ellipsis的研究可追溯到二十世纪七十年代。先前的研究方法一般偏向于单纯的句法方面或语义方面，但较少解决方案能集两者于一身。实际分析显示出单纯的句法或语义方面的考虑都不能恰当地解释VP-ellipsis这种语言现象。基于此考虑，本文将提出一种把句法和语义结合起来的方法：[+tense]理论。该理论具体表述如下：VP-ellipsis存在的必要条件是省略成分之前的部分必须存在着指示语和中心语的呼应，而存在这种关系的前提是功能语类“时态”必须具有[+tense]特性，而具有这种特性的功能语类有两种：限定句中的时态和控制不定式中的to。本文与以往传统语言学家观点的一个不同之处在于：本文认为控制不定式中的to具有[+tense]特性，因此允许后面的VP补语被省略。 本文提出的理论建立在把三种理论综合的基础上，三种理论如下： 1、Fukui和Speas(1986)提出的关于功能语类的分类； 2、Saito和Murasugi(1990)提出只有当功能语类的指示语和中心语之间存在着呼应关系时，该中心语的补语才能被省略； 3、Stowell(1982)指出控制不定式和上升不定式之间的差异是由[±tense]特性决定的。 本文不但考虑到VP-ellipsis句法方面的特点，也会谈及其语义方面。本文将会采纳Ayumi Matsuo的观点：当我们解释目标句时，先行句中所隐含的时间变数必须得被复制到省略句中，该变数是用来表明事件发生的时间信息。关于如何理解VP-ellipsis的语义的方法，笔者认为LF copying比PF deletion更为恰当。 本文还说明了[+tense]理论在最简方案中也能得到合适的解释。接着，本文会浅略分析汉语中相似的省略现象。相对而言，汉语中相似的省略现象更为复杂，在VP-shell理论下可被分为两种：由汉语助动词“是”(相当于英语中的be／do)引起的省略，和汉语空宾语结构。前者可以通过本文所提倡的理论得到解释，但后者则更为复杂。通过VP-shell理论分析可看出汉语和英语的动词省略在结构上是不同的，在汉语空宾语结构中，ASyntaetie一semantiePersPeetivetoVP一elllPslsintheTransformational一GenerativeFrameworkvP一elliPsis实际上发生在主动词到轻动词的移动之后，省略部分紧跟着轻动词而不是助动词，而英语中的vP一elliPsis是包括轻动词在内的整个动词词组的省略。英语和汉语在动词短语省略上的差异可通过两种语言中主语的不同位置得以解释。更多还原

【Abstract】 The study of VP-ellipsis dates back to the 1970’s (e.g. Hankamer and Sag 1976). The previous approaches usually focus on either syntactic or semantic aspect but not both. It is suggested that VP-ellipsis cannot be adequately accounted for under either of them. In this dissertation, a syntactic-semantic approach will be presented, namely, the [+tense] theory: to allow VP-ellipsis, the part preceding the elliptical component must undergo specifier-head agreement. The prerequisite for this relation is that the functional category T is of the property [+tense]. There are two types of functional categories T of that feature: finite T and infinitive particle to in control infinitives. Particularly, it will be argued that infinitive particle to in control infinitives, different from what the conventional approaches have assumed, is of the property [+tense], and therefore, it allows ellipsis of VP complement.The proposed theory is put forward by synthesizing three theories, to be exact, the proposal of taxonomy of functional categories in Fukui & Speas (1986) , the argument in Lobeck (1991) & Saito & Murasugi (1990) that only functional categories that undergo specifier-head agreement can license ellipsis of their complement, the characterization of the interpretive difference in terms of the feature [+tense] between control infinitivals and raising infinitivals in Stowell (1982) .In addition to the syntactic feature of the VPE construction, the semantic aspect is also tackled. I follow Ayumi Matsuo and propose that the temporal variable contained in the antecedent is copied into the empty VP when the target clause is interpreted. The variable serves to convey to a VP the information at what time that event takes place. Besides, as far as the semantics of VP-ellipsis is concerned, it is claimedthat the LF copying approach is preferred to the PF deletion approach.It will be shown that the [+tense] theory also fits properly into the Minimalist Program. After the discussion of the [+tense] theory, a large proportion will be contributed to the more complex Chinese ellipsis constructions, which will be considered in two types under the VP-shell theory: the elided part preceded by auxiliaries shi ’be/do’ and the Chinese null object construction. The former type falls into the explaining scope of the proposed [+tense] theory, while the latter involves V-to-v raising plus deletion of the empty VP and therefore VPE takes place following a light verb instead of the Infl node as English VPE does. The difference between English and Chinese ellipsis of verbal phrases is argued to be attributed to the different positions of subjects in English and Chinese.更多还原