【作者】 徐健；

【导师】 朱永生；

【作者基本信息】 复旦大学 ， 英语语言文学， 2004， 博士

【摘要】 本文试图探讨衔接、语篇组织和连贯的本质，以及三者在语篇生成和理解过程中的相互关系。 自从Halliday和Hasan合著的《英语中的衔接》一书出版以来，语言学家尤其是语篇分析学家们对衔接理论产生了极大的兴趣并进行了大量富有成果的研究(Hoey 1991；Martin 1992；McCarthy & Carter 1994；Widdowson 1978；Brown & Yule 1983；胡壮麟1994；孙玉1997；张德禄2000，2001；张德禄、刘汝山2003；朱永生1995，1996，1997a，1997b；朱永生、郑立信、苗兴伟2001；徐玉臣1996；2003a，2003b，左岩1995等)。但是几乎所有的讨论都围绕衔接与连贯的关系展开。我们认为，Halliday和Hasan衔接理论中语篇组织(texture)这一重要概念长期以来一直没有能够得到足够的重视；对于衔接理论的批评在很大程度上源于对语篇组织这一概念的忽视或曲解。在Halliday看来，衔接只是语篇组织的一个方面，它与主位结构以及信息结构共同作用形成语篇的语篇组织。换言之，Halliday关注的是衔接与语篇组织之间而不是衔接与连贯之间的关系。这也许是Halliday在他的衔接理论中只明确定义语篇组织而没有给连贯一个具体定义的原因。 在《英语中的衔接》一书中，Halliday & Hasan指出“衔接是一个语义概念；它是指存在于语篇内部的、使其成为语篇的意义关系”；“衔接是指[语篇中]存在的把其中某成分与上文联系起来的一切可能性”。为了澄清这一概念的内涵，他们又作了如下的阐述：“有一种意义关系对语篇特征至关重要，就是语篇中某个成分的解释要参照另一成分。衔接正是关系到语篇成分的意义是如何解释的。只要话语中的一个成分的解释要参照话语中的其他成分，就产生了衔接。”可以看出，Halliday & Hasan认为语篇衔接在本质上是一种存在于语篇内部的语义关系。他们还特别强调《英语中的衔接》关心的是英语语篇所特有的语言因素，但不包括情景特征(situational properties)(Halliday and Hasan，1976：20-21)。对Halliday衔接理论提出批评的语言学家们对于这一点似乎没有给予足够的重视。 对于Halliday衔接理论的诸多批评从一定程度上说明该理论确实还存在一些需要改进的方面。我们试图从两个方面对这一理论作出完善。第一，我们认为衔接机制应该考虑语篇的情景因素，因此语篇内部的意义联系的范围应该拓展从而把凭借语用意义而建立的联系也包括在内。第二，Halliday认为衔接手段是非结构性的语言手段，但我们认为这些衔接手段构成了若干语篇语义结构，包括身份确定(identification)、磋商(negotiation)、连接(coniunction)以及概念(ideation)四个组成部分。 我们还探讨语篇特征的另外两个构成成分，主位结构和信息结构。 在肯定Halliday主位选择时语篇发展的作用的同时，我们注意到他对述位的语篇功能没有给予足够的重视。我们认为，Dane蓉的主位推进理论更好揭示了主位和述位所具有的语篇功能.并借助于Thompson对主位功能的细分，完善了Dane亏的主位推进理论。 关于信息分布的语篇功能，我们首先肯定Halliday把主位与旧信息以及述位与旧信息区分开来的正确性。但同时指出，新旧信息的区分取决于多方面因素，远比Halliday所阐述的情况复杂，我们提出了二分法，即区分两类新旧信.息，新信息有语篇新信息(Diseourse书ew)和听者新信息(Hearer一New)，旧信息有语篇旧信息(oiseourse一old)和听者旧信息(H eare卜old).有些信息对于语篇而言是新信.息，但对于听话人未必如此。具体地说，旧信息是说话人讲话时认为听话人已知的信息，这种信息不一定存在于语篇之中，而新信息则为讲话人想要传递给听话人的信息，语篇中已经出现的信息说话人仍然有可能把它作为新信息传递给听话人。 主位结构、信息结构和街接共同作用形成了语篇的语篇组织。我们认为，语篇特征是语篇的语言特征，但是具备了语篇组织的语篇，对于听者和读者未必是连贯的。这一点可以从认知的角度作出解释。 一个语篇对于一位接受者来说是否连贯取决于他的认知能力。如果他能够想象出语篇发生的语境，那么这一语篇对于他来说就有意义，因此是连贯的。然而，不同的人有着不同的经历和背景，因此语篇内部的联系对于一个人来说是有意义的或者说是连贯，但对于另外一个人来说则有可能是不连贯的。这一点我们从认知模式的角度作出了解释。 总之，Halliday的街接理论更多地关注的是语篇本身的语言特征，而对于连贯概念的探讨则必须超越语言系统本身，涉及到人类的认知领域。更多还原

【Abstract】 The present research is an attempt to explore the nature of cohesion, texture and coherence and the ways they relate to each other in the production and comprehension of a text.Since M.A.K Halliday and Ruqaiya Hasan published their seminal book Cohesion in English in 1976, there has been increasing interest in the concept of cohesion among linguists (Hoey 1991; Martin 1992; McCarthy & Carter 1994; Widdowson 1978; Brown & Yule 1983;胡壮麟1994;孙玉1997;张德禄2000，2001;张德禄、刘汝山2003;朱永生1995，1996，1997a. 1997b;朱永生、郑立信、苗兴伟2001;徐玉臣1996;2003a，2003b，左岩1995等)。Although hailed as a milestone in the domain of discourse analysis, this theory has given rise to a lot of controversies, centring on the relation between cohesion and coherence.However, we find that texture, as a key notion in the Hallidayan model, has been largely neglected by the critics of this model. Many linguists have taken it as a notion equal to coherence, which has been the root of misunderstanding about the model that Halliday and Hasan has proposed. On the other hand, the model itself leaves much to be desired.As we see it, cohesion is only one part of the study of texture, which considers the interaction of cohesion with thematic structure and information structure. Texture, in turn, is one aspect of the study of coherence, which takes the social context of texture into consideration and which should be seen as the contribution that texture makes to our sense that a text hangs together.Halliday and Hasan (1976) define cohesion as a semantic concept referring to the relations of meaning within a text, and have focused on the linguistic means whereby the cohesive relations are expressed or realized, holding that all the cohesive relations can be realized as lexical and grammatical items. They made it explicit in Cohesion in English that they are concerned only with the linguistic factors that are characteristic of texts in English in their discussion of cohesion in that book, excluding the situational properties. This is another key point that has failed to receive its due attention from all the linguists arguing against the framework. Thesecounter arguments, although groundless, indicate that the Hallidayan model does need to be improved. We have tried to do this in two aspects. First, we have proposed that the situational factors of the text be considered in the model and therefore the scope of meaning relationship within the text be extended to include relationship of pragmatic meaning. Second, unlike Halliday, who sees cohesive devices as a set of non-structural resources hi the textual metafunction, we, drawing on Martin (1992, 2001), have attempted to reformulate the notion of cohesive ties as discourse semantic structure. Halliday’s non-structural textual resources have been reworked as semantic systems concerned with discourse structure, comprising: identification, negotiation, conjunction and ideation.Apart from cohesion, we have also given due attention to the other two textual components of text: thematic and information structures.Realizing the neglect on Halliday’s part of the role Rhemes can play in developing a text, I have, drawing on Danes and Thompson, proposed a modified model of thematic progression as a practical tool for textual analysis.While agreeing with Halliday that the pattern of focus contributes to the creation of a meaningful text, I have pointed out the determination of the givenness of a piece of information is more difficult than it seems to be. I argue for a cognitive perspective of context in deciding on the givenness or newness of information in the course of communication.Thematic structure, information structure, and cohesion work together to contribute to the texture of a text. However, it is my belief that the texture of a text alone does not ensure its coherence to a hearer or reader. This position of mine about the coherence needs explanation from a cognitive perspective.Although repeatedly emphasizing the meaning relationship between segments of a discourse while theo更多还原