互文性（语篇间性）是指在一个语篇中包含着来自不同来源的多个声音。这一概念由Kristeva首先在文学领域内提出来，后来得到包括语言学在内的其他领域的研究。虽然迄今已有不少研究成果，但是仍有很多问题还没有得到很深入的研究。本文拟从系统功能语言学的角度对互文性与语境的关系、互文性所传递的概念意义和人际意义以及不同声音间的关系等问题进行研究。书面广告语篇中的互文现象比较突出，但却尚未得到深入的研究。鉴于此， 本论文以书面广告语篇作为语料。由于互文性是一种常见现象，本研究从互文性的角度提出了一个新的语篇概念：语篇是包含一个或多个声音的服务于一定交际目的形式。基于这个概念，本研究从系统功能语言学的角度提出了一个研究互文性的语篇分析模型。该模型表示1）语篇内包含一个或多个声音，声音间存在着一定的对话关系，不同的对话关系会产生不同的总体意义；2）语篇中声音的选择及声音的体现形式受语境影响；3）各个声音所传递的概念意义、人际意义以及源于声音间对话关系的总体意义反映和构建语境；4）各个声音在语篇中的安排方式及对话关系应该确保语篇的连贯性。本研究在本模型下展开。研究表明互文类型的选择及各个声音的特点确实受社会文化语境和情景语境的影响。本论文研究了各个声音传递的概念意义及其关系。概念意义主要通过词汇和及物性手段得以传递。不同的声音间存在着一定的对话关系。多数声音相互补充，而不是相互对抗。除了补充关系外，声音间还存在投射的关系，即一个声音的意义被投射到另一个声音上去。在一个存在多个声音的语篇中，通常某些声音被凸显，某些声音被淡化。特别是各个声音所传递的意义彼此冲突时，某些声音会通过排版等手段得到凸显。本论文还研究了各个声音所表达的人际意义。每个声音都会为语篇的参与者构建角色关系。在多数情况下，各个声音所构建的角色关系是不一样的。有些角色关系被凸显，而有些则被置于背景中。通常是那些对语篇作者不利的角色关系被置于背景中。本研究还表明只要不同的声音服务于相同的交际目的和谈论相同的所指，它们所表达的评价意义通常互相补充而不是互相抵触。本研究还发现了各个声音在语篇中的连接机制。首先，各个声音能被安置于样类结构潜势（generic structure potential）中的某个位置，即不同声音在语篇中的安排要符合结构统一性的原则。第二，不同的声音可以通过衔接链连接起来。第三，各个声音可以通过语法结构连接起来。第四，不同的声音间存在着主题接续。所有的声音都服务于相同的大主题。每个声音还往往包含一些子主题，这些子主题能够被大主题所涵盖，不同声音的子主题间往往存在交叠的现象。第五，在语篇中，不同的声音间存在着一定的语义关系。总之，本研究表明系统功能语言学可以用于互文性研究，可以帮助解决有关互文性的一些问题。

Intertextuality, which means the co-presence of one or more Voices from different sources in a discourse, is an important feature of the discourse. This concept is first presented by Kristeva in literary studies. Later it found its way into other areas, including linguistics. Even though some linguistic studies of it have been conducted, some issues still call for further researches, such as the relationship between context and intertextuality, the functions of intertextuality, the interaction among different Voices in a discourse, the coherence of a discourse involving intertextuality and so on. This research tries to study such issues from the perspective of Systemic-Functional Linguistics.The feature of intertextuality in the print advertising discourse is quite striking, however, the study of it in advertising discourses has been scarcely conducted. For this reason, this dissertation takes the print advertising discourse as the data.As intertextuality is a significant feature of discourse, this research presents a new definition of discourse in terms of intertextuality: a discourse is a communicative occurrence which contains one or more Voices. Based on this definition, a model for the study of intertextuality is presented. According to the model, a discourse contains one or more Voices which interact with one another. There are dialogical relationships among these Voices. Different relationships among them can lead to different combined meanings. The choice of intertextuality types and the features of each Voice are influenced by the contextual factors. The ideational and interpersonal meanings expressed in every Voice as well as the combined meanings resulting from the dialogical relationships among them not only represent but also construct context. The Voices are not arranged at random in a discourse. To constitute a coherent discourse rather than a non-text, there must be some connections among different Voices. The whole study is conducted under the guidance of this model.This dissertation shows that there do exist relationships between context and intertextuality. The choice of a certain type rather than another type of intertextuality can be attributable to both situational and social-cultural contextual factors. Besides, the specific features of a certain Voice can also be accounted for by the contextual factors.This dissertation also shows that there are dialogical relationships between different Voices. Most Voices complement rather than contradict one another. Besides the relationship of complementation, there is the relationship of projection, that is, the meaning of one Voice is projected onto another. Moreover, some Voices can help to foreground the meaning expressed in another one. The relationship may be realized through sentence structure or graphic arrangement. In a discourse involving more than one Voice, some Voices are mainly intended and they are usually conspicuously positioned.The study shows that participant relationships can be established in every Voice. Some participant relationships are highlighted, while some others are put in the background. Usually the participant relationships unfavorable to the discourse producer will recede to the background. The study also reveals that in a discourse involving intertextuality, the appraisals in different Voices complement rather than counteract one another, so long as they describe the same entity and serve the same general purpose. This dissertation also identifies some means of linking different Voices in a discourse. Firstly, all the Voices in a coherent discourse can be allotted to certain slots in the generic structure potential of a genre. All the Voices are arranged in a discourse following the principle of structural unity. Secondly, the Voices in a discourse involving intertextuality can be linked by the similarity chains or identity chains of cohesive devices. The chains running across different Voices often interact with one another. Thirdly, in some discourses, the two Voices can be connected through structural means. Fourthly, there is usually topic continuity between different Voices. All the Voices contribute to the elaboration of the same global topic from different angles. There are often overlapping among the subtopics in different Voices. Fifthly, some kinds of semantic relationship can be observed among different Voices. In short, Systemic-Functional Linguistics can be used in the study of intertextuality. It can contribute to solving some questions related to intertextuality.