HPSG理论中“自己”和量化名词短语的语用解释的约束条件体系

论文摘要

本文的主要目标是在中心词驱动的短语结构语法理论框架内建立一个汉语中反身代词“自己”和量化名词短语的语用解释的约束条件体系。约束条件是指能够对某些语言现象进行歧义消解的语言规律;约束条件体系通过这些约束条件的共同作用来达到这个目的。首先,论文先分别定义了“自己”和量化名词短语的语用解释的种类:“自己”有2种(特指与泛指);量化名词短语有6种(定指、特指、分指、全指、互指和泛指)。我们先从自然语料中选出含有研究对象的部分分别建立针对两个研究对象的专门语料库,通过语料分析发现并验证可能的约束条件,如:表达形式及约束条件值的分布、特殊词汇、修饰词的数量、句法位置、时态、体、研究对象和先行词或同指成分的距离、文本类别等。如果约束条件证明有效,那么将对其进行评估和量权。研究建立了两个约束条件体系,分别包含8个关于“自己”的约束条件和23个关于量化名词短语的约束条件。然后根据每个约束条件对语用解释影响力的大小,分别给它们施加一定的点数。把在具体的研究对象上发现的约束条件对应的点数相加得到一个点数和。这个点数和是确定语用解释的最终依据。完整的约束体系将应用于自然语料分析,验证其对研究对象在语篇中的语用解释的预测力,并最终确定具有最佳预测结果的点数和作为判断标准。研究表明两个约束条件体系都能成功预测出绝大部分的语用解释,其中关于“自己”的判断成功率为91%,量化名词短语为80%。本文还分别研究了应用中心词驱动的短语结构语法理论分析“自己”和量化名词短语的相关文献,并把两个约束体系嵌入到该理论框架内。之所以采用这个框架,主要是基于三个原因:首先,该理论把各个层次的信息统一纳入到一个特征结构中描写,这符合约束条件体系的性质;其次,汉语中词汇包含了大量决定语法和语义结构的信息,该理论的词汇主义特色也比较符合这个特点;最后,该理论的模块性结构为约束体系的计算实现提供了便利。本文最后还分别提出了语言学研究中约束体系的产生和它的计算实现的流程图。

论文目录

• Acknowledgements
• Abstract
• 摘要
• Chapter One INTRODUCTION
• 1.1 Theoretical preliminaries
• 1.1.1 Linguistic form and its pragmatic interpretation
• 1.1.2 HPSG and its application to Chinese
• 1.1.2.1 HPSG
• 1.1.2.2 HPSG application to Chinese
• 1.1.3 Expert System （ES）
• 1.1.3.1 ES in Linguistics
• 1.1.3.1.1 Fundamental belief
• 1.1.3.1.2 Some implemented linguistic ESs
• 1.1.3.2 ES in Computer Science
• 1.1.3.2.1 Definition of ES
• 1.1.3.2.2 A brief history of ES
• 1.1.3.2.3 Limitations of ES
• 1.2 Research design and its rationale
• 1.2.1 Why HPSG
• 1.2.2 Link between the linguistic and computational ESs
• 1.2.3 Integrated language model
• 1.3 Research Methodology
• Chapter Two RELATED WORK
• 2.1 Unification and combination
• 2.1.1 Unification of simple AVMs
• 2.1.2 Unification of tagged AVMs
• 2.1.3 Combinatory rules
• 2.2 Semantic computability based on HowNet
• 2.2.1 Introduction to HowNet
• 2.2.1.1 Structure and relations in HowNet
• 2.2.1.2 Limitations of HowNet
• 2.2.2 Lexical semantic similarity computation （LSSC）
• 2.2.2.1 A simple model
• 2.2.2.2 Word similarity computing model
• 2.2.2.3 A model for content words
• Chapter Three EXPERT SYSTEM OF THE CHINESE REFLEXIVE ZIJI
• 3.1 A brief diachronic survey of ziji
• 3.2 A brief review of anaphoric studies
• 3.3 Noun Phrase Accessibility （NPA）
• 3.4 Pragmatic interpretations of ziji in discourse
• 3.5 Discussions on some experts
• 3.5.1 Formal types of ziji and its antecedent
• 3.5.2 Syntactic positions of ziji and its antecedent
• 3.5.3 Text Style
• 3.5.4 Distance between ziji and its antecedent
• 3.5.5 Potential experts
• 3.6 Database construction
• 3.7 Discussion and findings
• 3.7.1 Distributions and their hints
• 3.7.2 Correlative experts
• 3.7.2.1 Correlation between the form and number of ziji’s antecedent
• 3.7.2.2 Correlation between distance and text styles
• 3.7.2.3 Correlation between syntactic positions of ziji and its antecedent
• 3.7.3 The role of semantic computation in ziji’s pragmatic interpretation
• 3.7.4 Votes assignment and tally computation
• 3.7.4.1 Votes assignment
• 3.7.4.2 Tally computation
• 3.7.5 Hints from failure analyses
• 3.8 Ziji’s ES in HPSG
• 3.8.1 Ziji and binding theory in HPSG
• 3.8.2 Studies on ziji in HPSG framework
• 3.8.3 ES of ziji in HPSG
• Chapter Four EXPERT SYSTEM OF CHINESE QNPS
• 4.1 Classifications of Qs
• 4.1.1 Grammatical classifications of English Qs
• 4.1.2 Classifications of Chinese Qs
• 4.1.3 Formal Types of Chinese Qs
• 4.2 Semantic Relations between Qs in English and Chinese
• 4.2.1 Research Methodology
• 4.2.2 Analysis
• 4.2.2.1 From English Qs to Chinese Qs
• 4.2.2.2 From Chinese Qs to English Qs
• 4.2.3 Conclusions
• 4.3 Some approaches to Q interpretations
• 4.3.1 Q interpretations by standard logic
• 4.3.2 Q interpretations by syntactic theories
• 4.4 Expert System of Qs
• 4.4.1 Pragmatic interpretations of QNPs in discourse
• 4.4.2 A Functional Account of Q Scope in English: Expert System
• 4.4.3 ES of Chinese quantified sentence
• 4.4.4 Q experts
• 4.4.4.1 Tense
• 4.4.4.2 Aspect
• 4.4.4.3 Voice
• 4.4.4.4 Vocabulary
• 4.4.4.5 Syntactic Structures
• 4.4.4.6 Other influential factors
• 4.5 Database construction
• 4.5.1 General Principles
• 4.5.2 Q database construction
• 4.5.3 Overview of the database structure batch one
• 4.6 Discussions and findings based on database batch one
• 4.6.1 Merge and division of experts
• 4.6.2 Rearrangement of existing experts
• 4.6.3 Cancellations and additions of experts
• 4.6.3.1 Cancellations of experts
• 4.6.3.2 The additions of experts
• 4.6.4 About the values of Q’s formal types
• 4.6.5 About values of a QNP’s pragmatic interpretations
• 4.6.6 About syntactic positions
• 4.6.7 About input form
• 4.6.8 More information to be collected
• 4.6.9 Suggestions for order of operation in computing
• 4.6.10 Overview of Q database batch two
• 4.7 Discussions and findings based on database batch two
• 4.7.1 Overall statistics
• 4.7.2 Word segmentation and construction filtering
• 4.7.3 Tally difference in dual-Q sentences
• 4.7.4 Influence of tense and aspect experts
• 4.7.5 Newly discovered experts
• 4.7.5.1 Intensifier 是
• 4.7.5.2 Text type
• 4.7.6 Average tallies of Qs with each pragmatic interpretation
• 4.7.7 Hints from failure analyses
• 4.8 Q ES in HPSG framework
• 4.8.1 Quantification in HPSG
• 4.8.2 ES in HPSG
• Chapter Five LINGUISTIC AND COMPUTATIONAL MODEL OF ES CONSTRUCTION
• 5.1 Linguistic ES construction
• 5.2 Computational ES construction
• Chapter Six CONCLUDING REMARKS
• 6.1 What we have achieved
• 6.2 Limitations and suggestions for further research
• References
• Appendix 1 Hierarchical Structure in HPSG
• Appendix 2 Relations in Hownet
• Appendix 3 [CODE] of Dictionaries Referred to
• Appendix 4 Generated data from Q database batch 2