【作者】 王琳；

【导师】 唐孝威；

【作者基本信息】 浙江大学 ， 语言学及应用语言学， 2013， 博士

【摘要】 语码转换是语言接触现象的重要产物之一。近年来,语码转换的研究已经从语言学领域发展到语言学与认知科学相交叉的跨学科研究。研究者们广泛关注语码转换的定义、分类、句法结构、社会功能甚至认知加工机制等问题。本文采用句法和认知的双重视角,以真实语料为基础建立汉英语码转换的依存树库,运用自控步速阅读实验和眼动阅读实验的技术手段,对汉英语码转换现象进行计量和认知分析。论文的第一部分重点探讨汉英语码转换的句法变异问题及产生的原因,第二部分考察引起汉英语码转换句法变异的主要原因之一——语码转换的边缘性特征在心理实验中的体现并验证第一部分得出的结论。关于语码转换的句法变异问题,本文提出了具有普适性的研究模式。根据依存语法,基于19766词次的汉英语码转换的依存树库,本文以依存距离和依存方向为测量指标研究语码转换的句法变异问题。从句法上讲,与汉语和英语语料库相比,语码转换树库存在依存距离变异和以依存方向为测量指标的词序变异。从认知上讲,依存距离作为词语法中记忆承载的测量指标,本研究发现混杂依存比相应的单语依存的平均依存距离长。主要语法关系(主语、宾语、定语和状语)和语码转换的某些特征(边缘性、语流断续标记和移位)是导致依存距离变异的主要原因。依存方向在混杂依存中分布不同于单语依存导致了语码转换语料库中的词序变异。主要句法关系在单语依存和混杂依存中的不同分布导致了混杂依存中依存方向的分布区别于单语依存。为了更加细致和深入地探究引起汉英语码转换句法变异的原因,本文从动词的结合力入手,精确地计算并概括汉英语码转换中动词的概率配价模式,统计和比较动词支配不同词类所形成的三种主要句法关系在单语依存和混杂依存中的概率分布以及相应的平均依存距离。通过计量分析,不仅可以清晰地展现和对比汉英语码转换树库中汉语和英语动词的结合力,即动词支配其他词类和受其他词类支配形成各种句法关系的能力,而且能够发现引起汉英语码转换依存距离变异的词类分布状况。本文尝试性地扩展了概率配价模式理论的应用范围,为语言学和认知实验相结合的研究提供了一定的理论基础。在第一大部分中,基于语码转换依存树库的计量分析,我们发现引起语码转换依存距离变异,进而引起认知加工差异的主要原因之一是语码转换的边缘性特征,为了通过心理实验进一步验证以上结论,论文的第二大部分采用汉英语码转换树库中的真实语料作为实验材料,在自控步速阅读实验和眼动阅读实验两种不同实验情境下,考察语码混杂句子和汉语句子中主语的位置效应,实验结果证实了语码转换的边缘性特征能够引起认知加工难度的变化。更多还原

【Abstract】 Code-switching is one of the important language-contact phenomena. In recent years, the study on code-switching has involved in not only the field of linguistics but also the interdisciplinary researches. The definitions, classifications, syntactic structures, social functions and even the cognitive processing mechanism have been widely focused and studied. From the syntactic and cognitive perspectives, building a dependency treebank of Chinese-English code-switching based on natural language data and applying self-paced reading and eye-movement reading experiments, the paper takes an quantitative and cognitive analysis on Chinese-English code-switching. The first part mainly discusses the syntactic variations of Chinese-English code-switching and the causes. The second part analyzes the presentation of the peripherality of code-switchingwhich which is one of the major causes of syntactic variations in Chinese-English code-switching in psychological experiments and verifies the conclusions drawn from the first part.The paper tentatively proposes a universal model to analyze syntactic variations of code-switching. According to Dependency Grammar, based on a dependency treebank of Chinese-English code-switching containing19766tokens, the paper takes an analysis on syntactic variations in Chinese-English code-switching with the criteria of the depenendcy distance and the dependency direction. Syntactically, we find that code-switching affects language syntactic structures in terms of the dependency distance and the dependency direction. In other words, compared with the monolingual Chinese and English corpora, there are variations in dependency distances and word-order variation in dependency direction in the code-switching treebank. Cognitively, by comparing the dependency distance, a measure of memory load in Word Grammar, it is mainly found that mixed dependencies present longer distances than corresponding monolingual ones, thus verifying the alleged processing cost of code-switching. Specifically speaking, major grammatical functions (subject, object, attribute and adverbial) play important roles in syntactic variations; MDDs of main grammatical relations in mixed dependencies are, on average, longer than those in monolingual ones. The data in the corpus further confirm that some properties of code-switching such as peripherality, dislocation and flagging contribute to the longer dependency distances of mixed dependencies. To test the word-order variation, we analyze statistically the distributions of head-initial and head-final dependencies in Chinese, English, and mixed corpora. Distributions of the dependency direction in mixed dependencies are different from those of monolingual ones, which usually cause the word-order variation of the mixed corpus. It is the different distributions of main grammatical functions in monolingual and mixed dependencies that make the dependency direction of the mixed dependencies different.For a more detailed and in-depth analysis on causes of syntactic variations in Chinese-English code-switching, this paper starts from the binding forces of the verb, accurately calculates and summarizes the Probabilistic Valency Pattern of verbs in Chinese-English code-switching, statistically compares the distributions of three main grammatical relations formed by verbs and their dependents of different part of speech in monolingual and mixed dependencies. Through the quantitative analysis, it can not only clearly display and compare the binding forces of Chinese and English verbs in Chinese-English code-switching treebank, namely forces of the verbs dominating or dominated by other parts of speech to form various syntactic relations, but also find the distributions of different parts of speech that result in syntactic variations in dependency distances. This paper tentatively expands the application of Probabilistic Valency Pattern theory and provides a theoretical basis for the comprehensive study of linguistics and cognitive experiments.In the first part, based on the quantitative analysis in Chinese-English code-switching treebank, it is found that one of major causes of syntactic variations in dependency distances and cognitive processing difficulties is the peripherality of code-switching, in order to verify the above conclusion through the psychological experiments, in the second part we adopt the natural language data in code-switching treebank and test the position effects of subjects in code-mixed sentences and Chinese sentences. It is verified that the peripherality of code-switching can result in the changes of cognitve processing difficulty.更多还原