马博森教授基于现场即席的话语口语的博士论文
现代汉语自然会话中的人物指称策略：文盲与非文盲话语的比较研究
中国社会科学院研究生院
摘 要
自然会话是一个不断发展变化的动态过程，现实世界的会话双方相互合作共同创建一个话语世界。在他们所创建的话语世界里，他们必然会谈论各种人物，而要谈论人物，指称行为便是不可避免的。为了更好地抓住会话的动态本质，了解现实世界的会话者是如何通过心理世界在话语世界里指称人物的，我们在本文首先提出了一个研究自然会话中人物指称的三分模式，然后以该模式为研究框架，借助语料库语言学方法，对比分析了文盲和非文盲在实施人物指称行为方面的异同。
全文分为六章。
第一章简要介绍了本文的研究目的、研究范围、研究方法及结构安排。
第二章从人称指示、社会指示和话语回指三个方面综述国内外相关研究成果，特别是一些海外学者用英文撰写的涉及现代汉语研究的论著。
第三章描述我们提出的研究自然会话中人物指称的三分模式。该模式由三部分组成：人物本体层、共享知识层和语言体现层。在人物本体层，我们首先区分了现实世界、话语世界和心理世界三个概念，然后以此为基础，探讨了会话中的人物类型。在共享知识层，我们主要探讨两个问题：一是分析会话双方在涉及各种人物时是如何估计对方的共享知识状态的；二是分析会话双方就对方共享知识状态所作的估计对他们选择合适的指称形式有何影响。在语言体现层，我们讨论指称话语世界各类人物的具体语言形式。为了较为全面地揭示会话双方如何在他们所创建的话语世界里指称各类人物，我们首先建构了一个人物指称语系统。该人物指称语系统下辖三个子系统：直接指称语系统、间接指称语系统和混合指称语系统。其中，前两个子系统又有各自的子系统。然后分门别类，逐一讨论了指称语系统中各个子系统的语言形式体现，绘制了一张人物指称语系统的语言形式体现图。
第四章是有关本研究语料采集和加工处理的详细描述，包括理论基础、语料的构成、语料的计算机数字化处理、本文的标注方案以及对语料中某些特殊情况的处理说明。
第五章对比分析语料标注结果，从不同侧面、不同角度量化比较文盲和非文盲两类人在实施人物指称行为方面的异同。比较工作围绕四大块进行：一块从人物指称语的分布状况角度加以比较；另一块从编码非现场人物时，人物指称语的分布模式角度进行比较；第三块从人物指称中的修正现象入手进行比较；最后一块属于综合对比分析。
第六章总结本文的主要发现，讨论本研究的理论与实践意义以及本文的局限和今后尚待研究的问题。我们的研究显示：
1、 指称现场人物时，文盲和非文盲两类人在使用人物指称语方面并无明显差异。
2、 编码非现场人物时，文盲和非文盲在实施人物指称行为方面并无明显区别，他们对人物指称语的选择均符合人物指称语分布的典型和非典型模式。
3、 在实施人物指称行为时，文盲出现指代不明现象的概率高于非文盲，不过，出现指代不明现象后，两类人采取的指称修正策略相同。
4、 无论指称现场人物还是非现场人物，文盲较之非文盲更偏爱于使用引语类指称语。
关键词：自然会话 人物指称 文盲 非文盲 对比分析 语料库方法
Abstract
Naturally occurring conversation is an ever-revolving dynamic process. Speakers and hearers in the actual world create a discourse world with joint effort. In their discourse world, they will surely make reference to human entities. To capture the dynamism of conversation and understand how interlocutors make reference to humans, this dissertation first proposes a tripartite approach to the study of reference establishment achieved by speakers. Using this as a framework, the paper then makes a contrastive study between illiterates and literates in terms of their referential behavior.
The dissertation consists of 6 chapters.
Chapter 1 briefly introduces the aim, scope, methodology and the organization of the paper.
Chapter 2 makes a survey of the research highlights done in the field in 3 aspects, namely, person deixis, social deixis and discourse anaphora, focusing its attention on some of the important research work on Chinese written in English published abroad.
Chapter 3 provides a detailed description of our tripartite approach which proposes a three-line analysis scheme: human entity analysis; mutual knowledge analysis and linguistic realization analysis. Along the line of human entity analysis, we first make a distinction among actual world, discourse world and mental world. On the basis of this, we distinguish different types of human entity. Along the line of mutual knowledge analysis, we concentrate on two issues. While the first one is concerned with how the speaker and the hearer in conversation make continuous assessments about each other’s mutual knowledge, the second one is about the effect these assessments may have on their referential choices concerning different types of human entity. The linguistic realization analysis is the actual analysis of the way interlocutors select appropriate forms to refer to different types of human entity. To have a comprehensive understanding of this, we establish a system of referential expressions for persons. It is composed of 3 sub-systems, namely, a sub-system of direct referential expressions, a sub-system of indirect referential expressions, and a sub-system of mixed referential expressions. The first two sub-systems in turn have their own sub-systems. Using this as a guide, we discuss the linguistic realizations of each sub-system in detail.
Chapter 4 gives a full account of the collection and processing of the corpus data we’re using in this study including its theoretical basis, the composition and computerization of the data, our annotation scheme as well as the explanatory remarks as to how to deal with some of the special cases in the corpus.
Chapter 5 analyzes the annotation results, comparing the referential strategies to persons employed by both illiterates and literates using a quantitative method of comparison. We first compare the distribution of referential expressions used by the two types of people as found in our corpus data. Then we make a comparison between the two in terms of the distributional patterns of their referential expressions and their referential repair behavior when they encode non-present human referents. Finally we conduct a contrastive study on a synthetic basis.
Chapter 6 concludes the paper by summing up its major findings, discussing its theoretical and practical implications, and making mention of its limitations as well as some of the topics for future research. Our research demonstrates that (1) With regard to the use of referential expressions, there is no real difference between illiterates and literates when they make reference to persons in the conversational setting; (2) In terms of their referential behavior, there is no noticeable difference between illiterates and literates when they encode non-present human referents. As a matter of fact, their choice of referential expressions conforms to both the prototypical and non-prototypical distributional patterns of referential forms; (3) Illiterates are more likely to make inadequate reference than literates when they refer to human entities. However, they adopt the same referential repair strategy once such problems occur; (4) Compared with literates, illiterates are more likely to make use of reported speech to refer to persons both in and absent from the conversational setting.
Key words: spontaneous conversation; human references; illiterates; literates; contrastive analysis; corpus-based approach
现代汉语自然会话中的人物指称策略：文盲与非文盲话语的比较研究
中国社会科学院研究生院
摘 要
自然会话是一个不断发展变化的动态过程，现实世界的会话双方相互合作共同创建一个话语世界。在他们所创建的话语世界里，他们必然会谈论各种人物，而要谈论人物，指称行为便是不可避免的。为了更好地抓住会话的动态本质，了解现实世界的会话者是如何通过心理世界在话语世界里指称人物的，我们在本文首先提出了一个研究自然会话中人物指称的三分模式，然后以该模式为研究框架，借助语料库语言学方法，对比分析了文盲和非文盲在实施人物指称行为方面的异同。
全文分为六章。
第一章简要介绍了本文的研究目的、研究范围、研究方法及结构安排。
第二章从人称指示、社会指示和话语回指三个方面综述国内外相关研究成果，特别是一些海外学者用英文撰写的涉及现代汉语研究的论著。
第三章描述我们提出的研究自然会话中人物指称的三分模式。该模式由三部分组成：人物本体层、共享知识层和语言体现层。在人物本体层，我们首先区分了现实世界、话语世界和心理世界三个概念，然后以此为基础，探讨了会话中的人物类型。在共享知识层，我们主要探讨两个问题：一是分析会话双方在涉及各种人物时是如何估计对方的共享知识状态的；二是分析会话双方就对方共享知识状态所作的估计对他们选择合适的指称形式有何影响。在语言体现层，我们讨论指称话语世界各类人物的具体语言形式。为了较为全面地揭示会话双方如何在他们所创建的话语世界里指称各类人物，我们首先建构了一个人物指称语系统。该人物指称语系统下辖三个子系统：直接指称语系统、间接指称语系统和混合指称语系统。其中，前两个子系统又有各自的子系统。然后分门别类，逐一讨论了指称语系统中各个子系统的语言形式体现，绘制了一张人物指称语系统的语言形式体现图。
第四章是有关本研究语料采集和加工处理的详细描述，包括理论基础、语料的构成、语料的计算机数字化处理、本文的标注方案以及对语料中某些特殊情况的处理说明。
第五章对比分析语料标注结果，从不同侧面、不同角度量化比较文盲和非文盲两类人在实施人物指称行为方面的异同。比较工作围绕四大块进行：一块从人物指称语的分布状况角度加以比较；另一块从编码非现场人物时，人物指称语的分布模式角度进行比较；第三块从人物指称中的修正现象入手进行比较；最后一块属于综合对比分析。
第六章总结本文的主要发现，讨论本研究的理论与实践意义以及本文的局限和今后尚待研究的问题。我们的研究显示：
1、 指称现场人物时，文盲和非文盲两类人在使用人物指称语方面并无明显差异。
2、 编码非现场人物时，文盲和非文盲在实施人物指称行为方面并无明显区别，他们对人物指称语的选择均符合人物指称语分布的典型和非典型模式。
3、 在实施人物指称行为时，文盲出现指代不明现象的概率高于非文盲，不过，出现指代不明现象后，两类人采取的指称修正策略相同。
4、 无论指称现场人物还是非现场人物，文盲较之非文盲更偏爱于使用引语类指称语。
关键词：自然会话 人物指称 文盲 非文盲 对比分析 语料库方法
Abstract
Naturally occurring conversation is an ever-revolving dynamic process. Speakers and hearers in the actual world create a discourse world with joint effort. In their discourse world, they will surely make reference to human entities. To capture the dynamism of conversation and understand how interlocutors make reference to humans, this dissertation first proposes a tripartite approach to the study of reference establishment achieved by speakers. Using this as a framework, the paper then makes a contrastive study between illiterates and literates in terms of their referential behavior.
The dissertation consists of 6 chapters.
Chapter 1 briefly introduces the aim, scope, methodology and the organization of the paper.
Chapter 2 makes a survey of the research highlights done in the field in 3 aspects, namely, person deixis, social deixis and discourse anaphora, focusing its attention on some of the important research work on Chinese written in English published abroad.
Chapter 3 provides a detailed description of our tripartite approach which proposes a three-line analysis scheme: human entity analysis; mutual knowledge analysis and linguistic realization analysis. Along the line of human entity analysis, we first make a distinction among actual world, discourse world and mental world. On the basis of this, we distinguish different types of human entity. Along the line of mutual knowledge analysis, we concentrate on two issues. While the first one is concerned with how the speaker and the hearer in conversation make continuous assessments about each other’s mutual knowledge, the second one is about the effect these assessments may have on their referential choices concerning different types of human entity. The linguistic realization analysis is the actual analysis of the way interlocutors select appropriate forms to refer to different types of human entity. To have a comprehensive understanding of this, we establish a system of referential expressions for persons. It is composed of 3 sub-systems, namely, a sub-system of direct referential expressions, a sub-system of indirect referential expressions, and a sub-system of mixed referential expressions. The first two sub-systems in turn have their own sub-systems. Using this as a guide, we discuss the linguistic realizations of each sub-system in detail.
Chapter 4 gives a full account of the collection and processing of the corpus data we’re using in this study including its theoretical basis, the composition and computerization of the data, our annotation scheme as well as the explanatory remarks as to how to deal with some of the special cases in the corpus.
Chapter 5 analyzes the annotation results, comparing the referential strategies to persons employed by both illiterates and literates using a quantitative method of comparison. We first compare the distribution of referential expressions used by the two types of people as found in our corpus data. Then we make a comparison between the two in terms of the distributional patterns of their referential expressions and their referential repair behavior when they encode non-present human referents. Finally we conduct a contrastive study on a synthetic basis.
Chapter 6 concludes the paper by summing up its major findings, discussing its theoretical and practical implications, and making mention of its limitations as well as some of the topics for future research. Our research demonstrates that (1) With regard to the use of referential expressions, there is no real difference between illiterates and literates when they make reference to persons in the conversational setting; (2) In terms of their referential behavior, there is no noticeable difference between illiterates and literates when they encode non-present human referents. As a matter of fact, their choice of referential expressions conforms to both the prototypical and non-prototypical distributional patterns of referential forms; (3) Illiterates are more likely to make inadequate reference than literates when they refer to human entities. However, they adopt the same referential repair strategy once such problems occur; (4) Compared with literates, illiterates are more likely to make use of reported speech to refer to persons both in and absent from the conversational setting.
Key words: spontaneous conversation; human references; illiterates; literates; contrastive analysis; corpus-based approach
