【作者】 刘晓然；

【导师】 俞理明；

【作者基本信息】 四川大学 ， 汉语言文字学， 2007， 博士

【摘要】 上个世纪中叶，王力、张世禄等先生就指出，汉语双音词的主要来源是短语的词汇化。一直以来，这个观点在学界得到了较为广泛的认同，渐为成说。由于以往选取的研究材料都是随机性的，从来没有、也很难做到对一些特定的材料如断代的材料，专书的材料等，进行穷尽性的研究，因此，这个结论只是停留在观察印象的层面，它在多大程度、多大范围刻画了汉语词汇历史发展的本来面貌，还有待更加细致深入的论证。汉语词汇的双音化滥觞于先秦以远，迁延至从上古汉语到中古汉语转型期的东汉朝，双音化的步伐大大加快。本文截取这个汉语转型期的一段共时材料——道教元典《太平经》，以本经中高频复现的双音词和双音短语为研究对象(暂时主要以本经中复现率最高的前100个词例为研究对象)，对作为这些双音词历史来源的双音短语和本经中既有双音短语的词汇化流变进行历时的踏勘，在此基础上试图对双音短语词汇化在汉语发展史上的作用和影响做出客观评估，对双音短语词汇化的某些未曾引起学界重视的动因、机制和规律做出理论条陈。全文主要由两大部分组成。第一章和第二章是词例分析，穷尽考察本经中高频复现的100个词例。首先根据它们在本经中意义是否曲折化、词形凝固的等级、语法功能是否具有词的性质等，把它们区分为双音词、双音准词和双音短语，并参证《太平经》的共时语料，如《汉书》、《论衡》以及东汉佛典等，验证这些词例的性质；接着，对全部词例按照传统的句法结构关系分类，穷尽考察相关词例是否成词、成词的不同路径等基本事实。第三章是双音短语词汇化的理论阐释，主要比较词与非词、成词的短语路径与词法路径、准词和典型词在不同历史时段的数量差别，藉此评估双音短语词汇化这一语言现象在汉语词汇发展史上的地位和作用。研究发现，汉语双音词有两个主要来源：短语词汇化与词根复合法，而经由短语词汇化变来的双音词在词汇总量中只占很少的一部分，以本经100高频词为例，具体数据是：来自短语词汇化路径的双音词占本经100词例的15％；来自短语词汇化路径的双音词占可解释成五种基本句法关系的全部结构的25％；包括句法结构、跨层结构在内的来自词汇化路径的双音词占本经100词例的27％。这与以往认为双音词绝大部分是由短语固化而来的成说大相径庭。而且，有很多公认的来自短语的复合词，如“妻子”、“月亮”等，尽管这些语言单位在汉语史上确实有过用如短语的情形，但那些短语与同形的复合词之间并不存在衍生关系，它们只是偶然同形，这些词其实是通过词根复合法直接黏合而成的。词汇化理论可以涵盖句法结构(如“不及”、“不能”、“不足”等)、跨层结构(如“以为”、“所以”、“以来”等)的成词现象，但认为双音词的来源主要是短语的词汇化，则夸大了词汇化在汉语词汇发展史上所实际产生的作用。黄廖本《现代汉语》说，“汉语中运用复合法组合词根语素构成合成词的情况最多”，这是对现代汉语共时平面观察的结果，但于汉语史也是适合的：汉语的双音词，多数都是由两个词根语素直接黏合而成的，中间并没有经过短语的过渡阶段。本文还对五种不同类型短语发生词汇化的可能性进行了比较分析，研究发现，这个可能性存在很大的类型差异：凡是那些不能实现语义即时融合的结构，都无法通过词法路径在线(on-line)成词，它们都具有发生词汇化的可能性；同时又由于语义融合的困难，最后完成固化成词的终究只是少数。述宾结构、主谓结构和述补结构是谓词性的，具有较强的陈述性质。而陈述过程的完成，必须凭藉至少两个构成语素在语义上的关联才能实现，因此，这三类结构构成语素的指称性很难衰减，其短语性质很难改变，它们也就很难通过词法路径直接黏合成词，是倾向于通过词汇化路径成词的结构类型。联合短语中，倾向于选择词汇化路径成词的，只有不易即时黏合成词的反义联合类；类义联合类有少数情形，比如像量词短语等一些特殊类别的类义联合短语，可能发生词汇化；而同义联合结构能够即时黏合成词，不存在出现词汇化的情况。绝大部分双音偏正结构都是体词性的，它们在现实的语境中一经出现，多是即时“命名”的结果，应该视为表达一个专门概念的复合词，这部分词没有经过词汇化的改造：可以解释为定中式偏正结构的复合词，只有很少一部分是词汇化的结果；状中式偏正短语是谓词性的，这在偏正短语家族中只占很小的一部分，其中有一定数量倾向于发生词汇化。总之，所有经由词汇化路径衍生出来的双音词，都曾遭遇到这种必须选择困难路径的尴尬。生成双音词快捷路径优先的潜规则，最终决定了汉语不同来源词汇数量消长的基本格局：汉语的双音词，绝大多数都是词法路径的产物，其中主要的又是词根复合法的产物。更多还原

【Abstract】 The view that Chinese disyllabic words mainly come from lexicalization of phrases has been widely acknowledged in academic circles since it was proposed by scholars such as Wang Li and Zhang Shilu in mid 21st century. However, this conclusion is based on the observation of materials randomly collected and exhaustive study of materials in separated periods or of a certain book has never been conducted or stands hard to be conducted. As a result, to what extend and on what scale this conclusion has reflected the development of Chinese words still calls for further and more detailed study. Disyllablization of Chinese words started in the pre-Qin days and quickened its pace in the Eastern Han Dynasty when ancient Chinese language was transforming to medieval Chinese language. Taiping Jing, a Taoist classic, is written in this transitional period of Chinese language. This dissertation will study words and phrases frequently appear in Taiping Jing (the 100 words and phrases most frequently appear in this scripture for the time being) and explore the lexicaliztion process of disyllabic phrases that appear as disyllabic words in Taiping Jing and of the disyllabic phrases appear in this scripture. Based on this research, this dissertation attempts to objectively evaluate the role and influence of the lexicalization of disyllabic phrases in the development of Chinese language and then presents some of its motivation, mechanism and law that are largely ignored by the academic circles from the theoretical perspective.This dissertation consists of two major parts. The first and second chapters analyze 100 words and phrases most frequently appear in Taiping Jing. Based on whether the meanings of these examples are transparent in Taiping Jing, whether they have assumed set forms and whether they embody the grammatical function of words, these examples are divided into three categories: disyllabic words, disyllabic quasi-words and disyllabic phrases. These two chapters also verify the nature of these 100 examples by referring to the contemporary corpus of Taiping Jing such as Han Shu (History of the Han Dynasty, by Ban Gu ) , Lun Heng (By Wang chong) and Chinese translated Buddhist Scriptures in the Eastern Han Dynasty. Then these two chapters go on to classify all the examples in accordance with traditional syntactic structures and explore whether these examples are words and if they are, how they evolve into words. The third chapter presents theoretical explanation of the lexicalization of disyllabic phrases. It mainly compares the number of words and non-words, of words from lexicalization of phrases and those formed in accordance with syntax, and of quasi-words and typical words in different times, all of which will serve as a base for the evaluation of the status and role of lexicalization of phrases in the development of Chinese words.The study reveals that there are mainly two sources for Chinese disyllabic words: lexicalization of phrases and compounding, but words coming from lexicalization of phrases only account for a small part of the total. Take the 100 words most frequently appear in Taiping Jing for example, the disyllable words deriving from lexicalization of phrases make up 15 percent of the total and constitutes 25 percent of the structures that can be explained from the perspective of the five basic syntactic structure. 27 percent of the 100 words are disyllabic words formed through lexicalization of syntactic structure, cross-class structure, etc. These findings are widely divergent with the traditional view that the majority of disyllabic words evolve from phrases. What’s more, many words that has been widely accepted as deriving from phrases, such as "妻子" and "月亮", are in fact compounds. Though there were cases of such linguistic units being used as phrases, these phrases were merely taking on the same forms as those of certain compounds by accident and these compounds do not actually evolve from these phrases. The theory of lexicalization can explain words formed in compliance with syntactic structures (such as "不及" , "不能" and "不足" ) and those formed through cross-class combination (such as "以为" , "所以" and "以来" ), but the view that disyllabic words mainly come from lexicalization of phrases exaggerates the role of lexicalization in the development of Chinese words. In their book Modern Chinese Language, Huang Borong and Liao Xudong comment: "In Chinese, compounds formed by the binding of roots and morphemes dominate." Though their opinion is based on a synchronic observation of modern Chinese language, it can still be applied to the evolving process of Chinese language. Disyllabic words mainly come from the direct binding of a root and a morpheme rather than lexicalization of phrases.This dissertation also studies the five basic types of phrases in terms of the possibility of lexicalization and finds that they differ a great deal in this regard. If the meanings of morphemes of phrases cannot integrate when they first appear, these phrases cannot become online words in accordance with morphology. At the same time, though these phrases are likely to be lexicalized, only a small number of them can evolve into words due to difficulty in semantic integration. Verb-object structure, structure of predication and verb-complement structure are predicative and highly declarative, but this "declarativeness" cannot be achieved without connection between the meanings of at least two morphemes. Therefore the morphemes of each of these three types of structures are highly denotative and the nature of phrase embedded in these structures is hard to disappear. As a result, phrases of these three structures seldom develop into words through morphology but tend to evolve into words through lexicalization. Among coordinate phrases, those composed of morphemes with opposite meanings that cannot bind directly to form words involve into words in accordance with morphology in most cases. A small number of phrases, the meaning of whose morphemes are related to each other, also become words through lexicalization. For example, special phrases such as quantifier phrases are likely to be lexicalized. But phrases composed of morphemes with similar meanings can become words with their morphemes binding together instead of through lexicalization the first time they appear. The majority of disyllabic subordinate structures are substantive and the first time they appear they are to express a concrete concept, so they should be regarded as compounds and they have not gone through the transformation of lexicalization. Among those that can be interpreted as attributive-head compounds, only a small number result from lexicalization. Adverbial-head phrases are predicative and only account for a small part of subordinate phrases. Some of adverbial-head phrases tend to be lexicalized. To sum up, all disyllabic words deriving from lexicalization has undergone certain hard choices. The underlying rule that disyllabic words are formed through the most convenient way determines the basic pattern of the increase and decrease of Chinese vocabulary. Most of Chinese disyllabic words are formed through morphology, especially compounding of two roots.更多还原