一 研究对象和研究目标

本书的研究对象是现代汉语形容词和“形名粘合结构”。

“形”指形容词，包括一般所说的性质形容词、状态形容词、非谓形容词（区别词）等，如“大、小、皑皑、广阔、国营、中级”等。“副词+形容词”在朱德熙《现代汉语形容词研究》（1956）中被视为状态形容词的一种，本书把它视为词组，因为“副词+形容词”直接修饰名词具有理论意义。

“名”指名词，包括一般名词、集体名词、物质名词、抽象名词等，以及指称化的谓词，有时还包括部分名词短语，如“桌子、班级、水、思想、战斗、石窟艺术宝库”等。每一种类型的名词都可以受形容词的修饰，但受形容词修饰的能力有相当的差别。

“形名粘合结构”指形容词和名词不用“的”所构成的名词性短语，如“小桌子”、“火热场面”、“温柔路线”等。朱德熙（1956）提出把形容词和名词的组配形式分为“甲₁式”、“甲₂式”和“乙式”三种，其中“甲₁式”指性质形容词不带“的”和名词构成名词性短语。后来朱德熙（1982：148—149）又把名词性偏正结构分为“组合式”和“粘合式”两种，其中形容词和名词的粘合式组配大致等同于前述“甲₁式”。朱先生的“粘合式”定义中除了形名组配不用“的”外，还有一个条件就是粘合式中形容词只限于性质形容词和非谓形容词（区别词）。我们这里采用比较宽松的定义，只要是形容词和名词组配构成名词性短语，就属于本书所讨论的“形名粘合结构”，不管该形容词是何种类型。因为形容词和名词采用粘合式组配虽然和形容词的类别有关，但并不必然。本研究主要考察一个形容词和一个名词的组配情况，在必要时考虑多个形容词或复杂形容词与名词的组配，如“大长耳朵”、“精瘦小老头”、“小小村官”、“特大好消息”之类。一般只考虑粘合式形名结构，在比较时也涉及部分组合式形名结构。为了对比或解释同类现象，本研究有时也涉及部分“名词+名词”、“动词+名词”结构。

构词法和造句法其实没有明晰的界限，特别是在汉语这样的“语素清晰、词不清晰”的语言中。沈家煊先生在中国社会科学院语言研究所建所60 周年学术讲座（2010年）上曾说：“一个世纪以来，从《马氏文通》开始，在探求自己的汉语的特点方面已经有了很多的成果，如朱德熙、陆志伟等，如陆志伟说‘汉语的构词法与其说是形态，不如说是造句法’，现在的新的国外理论中，如生成语法学派中乔姆斯基的合作者提出取消形态，和句法合并，词的形态就是句法。”无论是从共时还是历时看，汉语的合成词和短语之间的界限都是很模糊的。而且，形名粘合结构本来就被国内学者认为具有相当的“凝固性”，部分学者（特别是国外学者）更是认为形名粘合结构就是“复合词”，因为它们不能扩展。考虑到这种情况，本书所说的“形名粘合结构”，可能还应包括一些“形+名”的词，有时对词法结构和句法结构不作明确的区分。在论述组配原则和语义关系的时候，有可能也使用“形+名”词语作为对比材料或佐证材料。

本书讨论的“形名粘合结构”原则上只限于现代汉语普通话，除了用作比较、佐证外，一般不涉及古汉语和方言的情况。对古汉语和汉语方言中形容词和名词组配的研究很少，我们在进行本研究的过程中收集了部分文献资料和语料，将另作专门研究。

本书的研究目标主要有四：

a.现代汉语形容词的词类地位和特点。

b.跨语言视野下现代汉语形名粘合结构的性质和功能。

c.基于真实语料的现代汉语形名粘合结构组配规律。

d.现代汉语形名粘合结构的语义表达。

二 语料问题

如果不考虑单个组配和单个组配之间的频率和可接受性的差异，也不考虑组配的各种句法、语义和语篇以及认知上的动因差异，只单纯考虑组配的实际情况，那么现代汉语中形容词和名词在理论上的可能组配的集合是一个笛卡尔积（Cartesian product，符号化为：A×B= {（a，b）︱a∈A且b∈B}），该集合由所有有序对（a，b）组成，其中 a∈形容词，b∈名词。例如，设形容词集合 A= {大，红，优良}，名词集合 B= {人，苹果，传统}，则A×B= {（大，人），（大，苹果），（大，传统），（红，人），（红，苹果），（红，传统），（优良，人），（优良，苹果），（优良，传统）}，共 9种理论上的可能组配。如果按照这种算法，假设有 2000个形容词，5000个名词，其构成形名结构的笛卡尔积是 1000万个，如果再考虑专有名词、名词短语、多个形容词修饰名词性成分，那么组合的可能性就是一个让人头痛的巨大数字。要弄清现代汉语中的形容词和名词组配的实际规律，就必须对这些可能组合一一进行甄别归类，先后运用求异法和求同法，找出制约组配的规律。如果只考虑一个形容词和一个名词的二元组合，那么我们也必须在这个理论上动辄以千万计的可能组配中至少辨析出如下一些情况：

a.某些形容词和某些名词在一般情况下不能组合。如“红、传统”。

b.某些形容词有时可以和某些名词组合，但只能以组合式的方式组合。如“优良、人”。

c.某些形容词可以和某些名词组合，但只能以粘合式的方式组合。如“红、人”。

d.某些形容词可以和某些名词组合，既能以粘合式也能以组合式组合。如“优良、传统”。

e.某些形容词和某些名词一般情况下不能组合，但是借助某些特殊句法格式或者语境，可以组合。如“（这是汉民族的一）大传统”。

……

研究现代汉语的形名粘合结构，可能有的人认为只要考察上述 c、d、e三类的具体组合就足够了，甚至连d类中的组合式也不要考虑。这实际上就只考虑了两个类别：能构成形名粘合结构的和不能构成形名粘合结构的。我们认为这是很不够的，因为形名粘合结构的构成属于一个系统性的构造，肯定性组配和条件（能构成形名粘合的类型和条件）、否定性组配和条件（不能构成形名粘合结构的类型和条件）、相关性组配和条件（构成和形名粘合结构最相关的形名组合结构的类型和条件）在系统考察中都是必要的，我们不但要考察形名粘合组配“有哪些”，具有哪些特点；也要考察形名粘合组配“没有哪些”，不具有哪些特点；还需要考察形名粘合组配“和哪些组配最相近”，相互关系怎么样。因此，我们必须通过大量的反复检验，尽可能把辨析出来的各种组合类别进行归纳概括，最终全面、系统地发掘形名粘合组配的规律。

这种非常合乎理性的方法看似非常科学，但语言毕竟不是数学和逻辑，它是一个“复杂系统”。在上述的分类方法中，还是有多种情况是很难处理的：

a.词和短语的界限很难区别，如“红人”、“大人”都可以看做一个词，也可都看做短语（做词看时和做短语看时意思有差别，其中“红人”做短语看的情况需要特殊的语境）。

b.一个词可能有多个义项，不同义项的组配能力和表现是不一样的，其语境依赖性也有差别。如“老”在“年岁大”的意义上组合式和粘合式都可以，如“老画家”、“老的画家”；但在“关系持续时间长”这个意义上一般只适应粘合式，如“老同学、老朋友”不能说成“老的同学、老的朋友”。

c.一个组合可能有多种功能，不同功能状态制约组配的能力和类型，如所谓的离心结构“高速度、长距离”等的类推性。

d.个人的语感和对具体的组配语境的构想可能有很大差别，如“温柔女人”这样的组配在一些人那里可能就不成立。

e.最关键的一点，不同的形容词和不同的名词的组配实例的频率、可接受性、标记性程度等都是有差异的，一刀切是否忽视了应有的差别？

……

虽然有上述这些难以解决的问题，但是我们用相对简单的“能”或“否”来进行判断归类至少在现阶段还是可以得出一些规律的。因此，本研究对某些形容词和名词构成粘合结构能力的判断不可避免地存在简单化的倾向，这个不足我们会在其他相关论述中尽量弥补。

为了避免单一方法导致的明显失误，我们对于具体的形容词和名词的组配能力的断定，多依靠作为语言研究者的直觉判断和检索、收集实际语料验证相结合的手段。

用基于语言直觉的内省方法判断组配的可能，在典型的、常用的组配情况中是非常有效的，但是分析一些边沿情况就比较麻烦，往往会出现诸如“于右任的胡子”式的问题：

有一次，有个小孩摸着于右任的长胡子，好奇的问道：“于爷爷，请问您晚上睡觉时，这把长胡子是放在棉被里还是棉被外呢？”当时于右任答不出来。晚上他上床睡觉，不管把胡子放在棉被里还是棉被外，都觉得很不自在，整晚为这个问题辗转难眠。（魏悌香《点灯的心、点灯智慧：小故事中的大智慧与简单人生真理》，北京：中华工商联合出版社2001：153）

大多数人对于语言的认识，一般都是习焉不察的。吕叔湘先生（[1980]2006：5）也说：“语言，也就是说话，好像是极其稀松平常的事儿。可是仔细想想，实在是一件了不起的大事。正是因为说话跟吃饭、走路一样的平常，人们才不去想它究竟是怎么回事儿。”我们很多的语言研究，也会遇到于右任的胡子的情况，我们想对语言探个究竟，想把我们习焉不察的语言现象弄个清清楚楚明明白白，实际的结果可能是我们对自己熟悉的语言的本来状况都产生怀疑，各种理论、各种状况、各种变化、各种解释……让我们对自己关于语言的某些规律的价值或者断定感到怀疑，直至对语言研究本身都感到精神颓废甚至信心崩溃。到了形名粘合结构这个专题，我们可能要问别人，也可能要问自己：某个/某些形容词能不能够和某个/某些名词组合？形容词是以粘合式还是以组合式和名词组合？粘合式和组合式的形名结构究竟有哪些区别？……被问的人或者我们自己，都会对某些回答犹豫不决、自我辩驳甚至走向直觉的反面。在我们的日常语言使用中，可能就像于右任的胡子，根本就不在意，但是如果真要研究个一清二楚，反复地思考、比较、斟酌，最终可能会连我们自己都不知道究竟是不是真的是这样。

自省的语言材料往往会出现于右任的胡子式的结果，这可能是干扰我们语言研究的一大障碍。萧国政（2001：11）的一段话道出了这种困惑：“有一定语法研究经验的人，很多可能有这样的经历：有些说法，开始觉得不能说，可研究着研究着，慢慢觉得能说了。你说是刻意制造，确实说不过去，说不是刻意制造吧，它看起来又不常见，或者简直就不这么说。”他的解释有二：一是随着研究的深入，语言研究者对异样构造的排斥程度可能会降低，他说“研究者开始造句时用的是和大家相同的语言编码和语言鉴别的原始语法系统，可是在研究的过程中，由于同一现象的流水的慢慢冲洗，语言的排斥性警觉疲软了、妥协了，最后就让一些不大能接受的句子进入了合法的通道而被接受了。”这样的原因导致的内省语料可能是错误的。二是研究者补充语句成立的语境的能力比一般人强。他说：“一些句子成立的语境相对比较宽，而另一些句子的成立对语境的要求比较严，而成立要求比较严的远语境句，其语境的铺垫，是语言接收者瞬间所难以一下补充出来的，而长时间进行这种现象的研究，却能提高对特定远语境句的语境补充水平和补充速度，因而能使这种起始断错的所谓不合格句，此时能够……领到合法证书。”这种原因导致的内省语料是可接受的，但会有异样感。语言研究者通过内省收集的语料往往随思考和研究的深入，其可靠性都会有所降低。如果还涉及迁就理论，那么问题会更多，这可能也是国内部分学者对形式主义学派的某些论文的一些奇怪的基于自省的语言例子感到疑惑和不满的原因。

自省的材料可能不可靠，那么统计和收集的材料就可靠？收集到的书面语料、口头语料也会因说话人个体差异、语境差异甚至语速、遗漏、错讹等原因而出现“习非成是”的情况和“板上钉钉的误用”的情况。况且，结构主义的归纳方法早就受到转换生成语法学家们的驳斥：第一，结构主义的归纳法不可能完整地描写语言；第二，发现程序是单向的；第三，结构主义的归纳研究在本质上无非是列举观察到的语言事实，然后进行分类，研究的结果只是构建了一部分类语法（徐烈炯，1988）。当然，使用频率的统计和内省判断可以在一定程度上弥补这个缺陷。

用纯粹内省的方法或随机收集语料的方法得出的结论可能还是有不全面、不准确的地方，二者的结合可能会提高结论的普适性和科学性，“内省—查证”式或“内采外证”式（萧国政，2001：6、11）是非常可靠的语料收集和处理方法。我们对语料的处理基本上是“内省判断+真实语料查证”的模式。具体做法是：（1）手工随机收集大量语料，这些语料经过我们的选择，具有相当的代表性和典型性，可以作为立论的具体实例，并在初期作为提出形名结构分类和组配规律假设的基础。（2）依据《形容词用法词典》（郑怀德和孟庆海编，2003）每条形容词词项后边的说明和举例来初步确定形容词构成形名结构的情况。（3）通过我们基于普通话语感的自省以及请教他人的语感来判断形容词和名词构成粘合结构的可能性，进一步修正组配可能判断并统计。（4）利用网络的海量语言信息，通过网络搜索引擎对《形容词用法词典》中的形容词进行了一一检索，摘取了前20—50个该形容词和各种名词组配的实例。（5）利用较为成熟的语料库（如北京大学中国语言学研究中心的CCL平衡语料库和中国传媒大学“有声媒体语言文本语料库检索系统”），检索常用形容词和名词的组配的情况。五条途径收集、处理和验证语料，这样可以在一定程度上避免语料和结论的明显错误。

另外，我们的考察不可能是穷尽性的，因为我们对全部可能组合进行一一辨别是不现实的，至少在时间和精力上是不经济的。由于很多可能组合的表现具有“类”的特征，我们人类天生的类推思维能力可以让我们更加快捷地、概括性地掌握规律。因此，我们的做法类似于计算机自然语言处理的某些模式：对部分数据进行预处理—概括规律—运用推广（效果测评）。简而言之，在使用语料的问题上，我们利用精心挑选的代表性语料和我们的语言直觉，概括出各类形容词和名词组合的规律及其类型，然后再进行真实语料的大范围验证，最后提出自己的看法。

三 研究思路

刘丹青（2003：21—29）用三个发音接近的英语单词test（测试）、text（语篇）、attest（验证）来概括当代语言学的形式主义、功能主义和语言类型学三大范式的研究方法特点，其中的“测试”指的是依据母语者的内省来说明语言的结构规律，一般不考虑不直接相关的语言成分，内省是形式主义学派多采用的主要研究方法，部分结构主义语言学家也经常采用这种方法；“语篇”是指依据真实语料（语篇）中的实际用法来概括语言规律，经常采用统计量化的方式来确定倾向性规律，注重语法单位和语境的关系，这是功能学派经常采用的研究方法；而“验证”则是当代语言类型学采用的主要研究方法，选取一些传统的参项，依据跨语言的证据来证明，多考察某种句法结构在不同的语言/方言中的的分布和表现，验证人类语言变异的极限，藉此发现语言的共性。三大范式各有优势，本研究在思路上致力于尽量吸收它们的优势。

本书的基本研究思路是：

（一）以跨语言视野考察现代汉语形名粘合结构的性质、功能及研究意义

形容词问题在当代语言类型学中颇受关注，我们在尽可能占有类型学相关研究成果的基础上，以跨语言的视野来探讨汉语的形容词问题，最后确定现代汉语形名粘合结构在语言类型学上的意义，利用跨语言的证据证明现代汉语形名粘合结构的性质，并解释形名粘合结构的组配规律和语义特点。

在当代语言类型学中，形容词的相关句法属性是一个主要关注的对象。首先，当代语言类型学在开创时期就非常关注形容词问题。在格林伯格（Greenberg，[1963]1966）的那篇经典论文中，所列举的总共 45 条共性中，有共性 5、17、18、19、20、21、22、24、31、40（原文序号）等 10条论及形容词，主要是语序方面的，占了 22%，这说明形容词和名词的语序在格林伯格的心目中还是占据很大分量的。格林伯格（Greenberg，[1963]1966）把修饰形容词（qualifying Adjectives）和相关联的名词的位置视为语序类型研究中除主语和宾语的“第三个分类的基础”。但它们有不同的语序表现。虽然后来的研究成果证明，形容词和名词的语序并非像那些同样可能充当名词核心的定语的指示词、冠词、数词、量化词（如 some、all）等的语序那样，和动词、宾语等的语序具有较大关联性，但形容词和名词语序在多项蕴涵式中还是具有很重要的作用。其次，在大量的语言类型学研究成果中，有关形容词的语言共性数量很多。我们于2009年12月9日检索了德国康斯坦茨大学（Konstanz University）共性档案库（The Universals Archive）2679756，该档案库几乎囊括了所有当代类型学的研究成果，我们在这个档案库中找到与“形容词（Adjective）”有关的共性总共有 127条，涉及形容词的词类性质、类别、句法功能等方面，该共性库不但列出了共性，还对共性研究的发展（如反例）进行了简单的描述。再次，形容词的词类地位和跨语言识别等问题是语言类型的研究热点。近二十多年来，人类语言的词类问题引起了很多关注，如有关跨语言词类判定的亨格维尔德（Hengeveld）、瑞杰可夫（Rijkhoff）等的“阿姆斯特丹模型”（Amsterdam model）、克罗夫特（Croft）的“类型学的典型”、迪克森（Dixon）的“核心语义类型—主要句法功能—主要词类”配对模型等等，各种理论都首先需要面对和解决形容词词类的跨语言的复杂情况，各家观点的分歧大多就在形容词的判定上，形容词问题成了语言类型学争议的热点，有的观点甚至针锋相对，个别学者在不同时间段还会对某种具体语言的形容词问题有截然不同的看法。形容词词类已成为检验类型学词类理论的试金石。

后面的研究将表明，现代汉语的形容词和名词构成名词性短语的时候，组合式和粘合式的区分与其他语言的有标记定语和无标记定语、前置定语和后置定语、并置定语和修饰定语、形容词定语和形容词谓语等区分手段具有部分功能上的一致性，粘合式和组合式的形式差异可以用来区分形容词和名词组配在形式、语义、句法和语用表达上的差异，如形容词的语义类别差异（如性质、状态等的区别）、形容词的外在形式差异（如音节、结构复杂度、成分构成等的区别）、名词的语义类别差异（如具体名词、抽象名词、专有名词等的区别）、形容词修饰名词的语义关系（如限制性、非限制性、内涵性等定语的区别）、形名结构的信息特点（如形容词所表属性的已知和未知的区别）、形名结构整体所处的句法环境（如整个形名结构充当定语、状语、宾语、谓语、主语等的区别）、形名结构的语用功能（如整个形名结构表示称谓、表示属性和表示归类等的区别）、形名结构组配的频率（如常见组配、偶见组配等）、形名粘合结构所处句子的功能类型（如陈述、祈使、感叹、疑问等的区别）等等。形容词和名词组配的粘合式和组合式是两种相对独立的功能性手段，采用粘合式或组合式是根据形式、语义和表达的需要进行选择的结果，和前述其他语言中形容词的不同句法形式的功能区别有着同样的价值。还可以进一步抽象，“直接粘合”和“使用功能词组合”就是两种不同的句法手段，在其他结构如“名词+名词”、“动词+名词”、“状语+中心语”中也同样可以因为功能差异而分别选择。

（二）注重真实语料和内省判断相结合，在小句语境下考察形容词和名词构成粘合结构的条件和类型

以往的大多数关于形容词和名词构成粘合结构条件的研究大都着眼于两个大的方面：（1）形容词本身的类别，如单音节形容词更倾向于构成粘合式形名结构，状态形容词和形容词短语更倾向于构成组合式形名结构等。（2）形容词和名词的语义关系，一般认为只有当形容词所表示的属性是名词所表示的事物的“本质属性”，并习惯上作为分类特征的时候，形容词才会和名词构成粘合结构。这类研究一般只需要考虑两个成分（即形容词和名词）以及二者的关系就可以了，研究的用例也多用短语形式。《现代汉语词典》、《形容词用法词典》等有关形容词和名词构成粘合结构的举例，也以短语居多，一般不涉及形名短语之外的其他句法成分和语篇环境。这种倾向于静态的分析优势是很明显的，我们按照“若无必要，勿增实体”的“奥卡姆剃刀”原则，在对形名粘合结构的组配规律进行分析时，如果能够在形容词和名词两个成分之间即可解决，则尽量不涉及其他相关因素。

用形容词和名词两个成分来考察形名粘合结构的组配规律，我们主要从如下几个方面进行：（1）形容词的次类对组配的影响。（2）名词的次类对组配的影响。（3）形容词和名词的音节数目对组配的影响。（4）认知语义对组配的影响。对于第4 方面，我们将做重点考察，认可部分学者从理想化认知模式、概念距离等角度处理形名粘合组配问题的合理性，并进一步引申和发挥。另外还特别强调与距离象似性（iconicity）原则相反的一类情况：认知处理的经济性也会把那些概念距离远的、语义复杂的“属性—主体”关系处理为形名粘合结构，概念距离与形名结构、概念复杂度与形名结构之间的标记颠倒现象大量存在。这一点也是本研究和其他研究不同的亮点之一。

静态的分析方法只考虑形容词和名词两个，其他不直接相关的成分一般都不予考虑，然后通过分析形容词、名词的形式和语义特点以及二者的语义关系来确定能否组配。这是一种拆零的研究方法，首先把形名组配从相关的语境中拆解下来，然后再把形名结构拆解成形容词和名词两个成分，接着对这两个构成成分以及它们的关系进行分析，得出结论，这个结论被推广到它们原先被拆解下来的类似完整语境中。这种拆零的研究方法是比较成熟的科学研究方法，但早已受到部分学者的批评。如美国著名未来学家阿尔文·托夫勒在为普里戈金（诺贝尔奖获得者）和斯唐热的著作《从混沌到有序》所写的前言“科学和变化”中说：

在当代西方文明中得到最高发展的技巧之一就是拆零，即把问题分解成尽可能小的一部分。我们非常擅长此技，以致我们竟时常忘记把这些细部重新装到一起。这种技巧也许是科学中最受过精心磨练的技巧。在科学中，我们不仅习惯于把问题划分成许多的细部，我们还常常用一种有用的技法把这些细部的每一个从其周围环境中孤立出来。这种技法就是我们常说的certeris paribus，即“设其他情况都相同”。这样一来，我们的问题与宇宙其余部分之间的复杂的相互作用，就可以不去过问了。（Prigogine &Stengers，1987：5）

但是在心理学界，注重“整体”研究的理论和方法早已在 20世纪前期就开始流行。如由科勒（Wolfgang Kohler）、考夫卡（Kurt Koffka）、勒温（Kurt Lewin）和其他一些20世纪30年代移民到美国的欧洲心理学家创立并发展的“格式塔”心理学。格式塔心理学最重要的观点之一，就是认为人们对于环境和事件的认知并不是由许多独立成分组成，而是一个“动态的整体”。（Taylor，Peplau & Sears，2004）

显然，“拆零”方法的优势很明显，可操作性很高，但其研究结论可能回避了很多应该考虑的事实。康德说：“不要轻率地减少实体的多样性”，对形容词和名词组配规律的考察如果只使用拆零法，只依靠内省的静态材料，必然会忽视形名粘合组配的复杂多样性，必然会忽视小句中相关的其他成分、结构的整体功能、相关语境等在构成形名粘合结构中的重要作用。形容词和名词构成粘合结构，往往无法“设其他情况都相同”，注重该形名结构整体在相关句子、语境中的实际表现是探求其构造规律和语义表达的必由之路。

静态分析方法的优势是很明显的，便于把握，且规律相对较为显明。但是仅仅这样分析形名粘合组配的规律是不够的。静态能够构成形名粘合式组配的形容词和名词，在一定的句法环境中一般是能够组配的，但是部分静态不能构成形名粘合式组配的形容词和名词，在特定的句法环境中也是可以构成粘合式组配的。沈家煊（1999：290）就指出有的形名组合在静态组合中一般要使用组合式，但入句以后可以不加“的”，如“大方的举止”，单说时一般要加“的”，但在“她在酒会上的大方举止令人刮目相看”中就可以不加。赵元任（[1968][1979]2001：304）也认为朱德熙《现代汉语形容词研究》（1956）中所列的不能说的“重箱子”之类形名粘合结构在对话、对举及其他有标记场合就可以说。这是小句的语境导致了粘合组配的实现。

萨丕尔依据语言中句法手段的表现把语言分为“分析的”、“综合的”和“多重综合的”三类。其中汉语被他视为分析性语言的典型，他认为“在分析的语言里；句子永远是主要的，而词是次要的”（爱德华·萨丕尔，[1921]2007：114）。结合句子来考察汉语的语法规则是必然的，以小句作为汉语语法研究的操作平台是非常合适的，也是必要的。小句是最小的具有表述性和独立性的语法单位（邢福义，1996：13—15）。小句有内部的结构成分和成分关系，有语义关系，有语体特征，有功能特征，有频率特征，有更大的语篇环境（复句、句群、话语、篇章等），以小句为语法研究的切入点（本位），能够集中尽可能多地影响语法结构的因素，可以发现一些以前难以发现的语法规律，可以解释一些以前难以解释的现象。因为小句是具体语境中的小句，在一定程度上，可以把“小句本位”的语法分析视为基于真实语料的分析。小句语境对形名粘合结构的构成存在多方面的影响，如句法位置、成分数量、有标记性、句子功能（疑问、祈使、陈述、感叹）、肯定与否定、上下文语境、语体、频率等，都会影响形名粘合结构的实际构成。

本研究先从静态角度分析形容词和名词构成粘合结构的条件，然后在我们通过手工收集、语料库检索、网页搜索等手段积累的、含有形名结构的大量例句的基础上，结合小句环境分析形容词和名词构成粘合组配的实际情况和制约条件，拓展静态分析的结果，充分考虑形名粘合结构出现的语境，特别是小句的调节作用。并把静态分析的结果和入句动态分析的结果结合起来考察，尽量用统一的模式来进行解释。

（三）结合结构语义学、认知语义学的一些观点和方法，在句管控中解读形名粘合结构的语义表达

从国内外有关形容词语义的研究文献（特别是各种词典释义）看，形容词的语义和名词、动词的语义有相当大的不同，其具体的语义解读往往随语境而变化。而且，形名粘合结构的构造紧凑，相对于组合式形名结构而言，语义指数高，形容词和名词的语义关系类型复杂。形容词的具体语义解读虽然有一些默认的值，但经常容许他解，变动不居，随具体语境而定：一方面，特定形容词和不同名词组配时，语义会有不同的解读，如“好手机、好厨师、好老师”中，“好”的具体语义内涵就差别很大。另一方面，特定形容词和特定名词组配，其语义也会随语境而改变，如“小老板”就容许多种解释。但是，正如形容词和名词能否组配是有规律的，组配后的语义解读也是有规律的：首先，形容词和名词各自会基于理想化认知模式而自动寻求“适配”解读；其次，语境会对默认的语义解读进行强化或者压制。因此，为了操作的方便，我们主要以小句作为语境域来分析形名粘合结构的语义。并且坚持一个原则：如果形名粘合结构的语义可以在短语层面解读的，就不纳入小句层面进行解读。

对形名粘合结构进行语义解读，分析和综合的方法要结合使用（沈家煊，2005a）。有些结构可以简单地视为“整体等于部分之和”，整个结构的语义等于形容词语义和名词语义按照简单的“属性+主体”关系构成，如“小桌子”。这种结构的语义解读可以适应于结构语义学的方法，可用分析的方法。但更多的形名粘合结构语义是“整体大于部分之和”，整个结构的语义会大于形容词和名词的语义，是“属性+X+主体”式的，X表示联系属性和主体的关键性语义因素，如“热杯盖”，指的是“用来盛热饮料的杯子的盖子”。因此这类形名结构的语义分析需要借助认知语言学的某些理论和手段，需要使用综合的方法。沈家煊（2005a）认为：“研究语言，分析和综合两种方法不可或缺。”我们在分析形名粘合结构的语义的时候，分析方法和综合方法的运用视需要而定。

结构语义学和认知语义学的相关成果为形名粘合结构的语义分析提供了很多有价值的理论框架和分析实例，如语义场理论、物性结构理论、心理空间理论等，我们将在研究中尽量吸收。因为我们采取的是“实用”主义态度，不对多种理论本身做过多的探究，在具体的分析中会忽略多种理论之间的冲突。

我们对形名粘合结构的语义分析主要集中于四个方面：（1）形容词的语义功能类型，如限制性、非限制性、内涵性、同位性等类别，特别是非限制性形容词定语的各种次级类型。（2）形容词的具体语义解读和解读模式的构建。（3）依据“整体大于部分之和”的原理，分析整个形名粘合结构的语义特点及其类型。（4）通过对比相近或相关的词、词类、短语等，对形容词、粘合结构、形名粘合结构的语义特点进行理论思考。

（四）一般性研究和具体个案研究相结合，合理分工，各有侧重，多角度、多层次地探索形名粘合结构的组配规律和语义特点

在一般性研究中，我们重点解决如下问题：第一，依据构成形名粘合结构的不同能力给形容词分出不同的类，概括出每一个小类的特点。第二，结合真实语料，分析影响形名粘合结构构成的诸多肯定性条件，并试图构建一个判断的操作流程模型。第三，在静态语义分析的基础上，分析形名粘合结构的各种语义关系类型和具体语义内涵。而在具体个案研究中，我们选择了“大/小+名词”、“前/原+指人名词”两个代表性的自由组配格式和“数词+大+名词”、“序数+形容词+名词”两个具有相当固化程度的组配格式进行深入探讨。一般性的研究主要以形容词作为变量，重点分析不同类型的形容词会对粘合结构的形成和语义表达构成哪些影响。具体个案研究主要以名词作为变量，重点分析不同类型的名词会对粘合结构的形成和语义表达构成哪些影响。虽然事实上一般性研究中也涉及名词的差异，具体个案研究中也涉及形容词的差异，但我们有意在侧重点上进行适当的分工，这种分工只是一种倾向性。

（五）倾向于功能主义语言观，重视语义因素对句法结构的影响，追求倾向性规律

功能主义的语法分析具有如下特点：（1）突破形式研究中只以内省的句式为研究对象的做法，更多地注意各种与语言行为有关的因素对话语组织的影响；（2）突破了形式研究中只把注意力集中在类型异同的做法，而较多地注意实例的多寡所反映出的倾向性的规律；（3）突破形式研究中把对象看成一个静态成品的做法，而较多地当做一个动态过程看待，研究听说双方的语言认知策略；（4）突破形式语法孤立地看待句子（甚至只是一个结构）的做法，十分重视联系语境进行分析（张伯江和方梅，1996：1）。本研究注重真实语料的分析，注重形名粘合结构在小句中的动态构建，注重联系语境分析该结构的语义，这种研究思路都是倾向于功能主义的。

结构主义、转换生成语法都希望在研究上撇除语义对句法结构的影响，但事实上很难做到。在最需要纯粹形式化的计算机语言信息处理中，语义因素也不能被截然抛开，著名的人工智能哲学家博登（Boden，[1990]2001：8）认为：“即使最简单的程序也并不是纯形式主义的，而是具有某种相当本原的语义特性，所以从根本上讲，计算理论并非不能解释意义。只要大脑生成意向性的能力是清楚的，而不是完全反直觉的，这种认识所采用的信息加工方式同样可以用于计算机。”计算机也能模拟人脑的意向性处理方式，也必须这样。不涉及历时的语义来源于人类大脑的意向性特征，而人类大脑的意向性是科学界关于“决定论”和“随机论”争论的焦点。因为有了意向性，世界才不像是一个庞大的自动机，我们已经认识到我们是生活在一个多元论的世界之中（Prigogine & Stengers，1987：26—27）。因此，在语言分析中充分考虑和依赖语义和功能是可取的。

科学家们都呼吁“应当注意潜伏在一些简单模型中的令人惊奇的复杂行为”（詹姆斯·格莱克，1990：4）。勒内·托姆（1989：396）认为：“如果系统行为不能用明确地叙述出来的简单定律来刻划，那就首先试一试借助于倾向性来定性地刻划这一种行为。”正因为世界不是一个庞大的自动机，不一定完全受背后“规则”的制约，有时带有很大的随机性，因此，基于意向性的语言分析往往是不全面的，并非决定论者认为或者希望的那样：能够反复应用规则，能够清晰包括或者排除某些语言元素，能够拆散和组合，等等。就是决定论的坚定的支持者，也会对概率的作用有一个清晰的肯定认识，语言的规律往往只是一个“大致”的概率问题。这种概率倾向是因为人们使用语言手段的同一和表达意向性的同一，而这种语言表达是概率问题而非“整齐划一”，也同样是因为人们使用语言手段的选择性与表达意向性的选择性。

任何科学理论都是不完备的，语言学理论及其规律的总结也是不完备的。德国著名数学家哥德尔早在 1931年就提出并证明了“不完备性”定理，认为：任何无矛盾的公理体系，只要包含初等算术的陈述，则必定存在一个不可判定命题，用这组公理不能判定其真假。该定理最初发轫于数学领域，但其影响远远超出了数学的范围。该定理不仅使数学、逻辑学发生革命性的变化，引发了许多富有挑战性的问题，而且还涉及哲学、语言学和计算机科学等领域。2679757由不完备性定理可知，几乎所有的语法规律都存在漏洞。本研究作为倾向于功能语言学和认知语言学的研究，其结论的不完备性可能更加明显。马真（2004）就认为功能语法学“只能提供一种倾向性的解释，不能提供严格的规则或者说规律”。雷科夫（Lakoff，1973）认为所谓“合乎语法”与“不合乎语法”只不过是程度问题。不能找到关于形名粘合结构的普适的、绝对的规律是必然的，但这并不能成为停止探索的理由。我们的研究致力于寻求倾向性、概率性的规律。

四 研究的基本架构

除了“前言”部分，本研究的主体部分共七章。

第一章，汉语形容词与形名粘合结构；第二章，形容词构成形名粘合结构的能力及其组配规律；第三章，形名粘合结构的称谓性和语义分析；第四章，个案研究一：“大、小+NP”；第五章，个案研究二：“前、原+指人名词”；第六章，个案研究三：“数+大+NP”构式；第七章，个案研究四：“序数+形容词”（+名词）结构。

这七章可概括为两大块：前三章是有关汉语形容词和形名粘合的总体性研究，侧重于不同形容词构成形名粘合结构的能力；后四章是有关形名粘合结构的个案研究，侧重于不同名词和特定形容词构成形名粘合结构的选择性及其语义表达。

这两大块内容内部又有所分工。在第一大块内容中：第一章是从语言类型学的高度来看汉语的形容词和形名粘合结构，在人类语言的形容词和形容词修饰名词结构的共性背景下准确定位汉语形容词和形名粘合结构；第二、三章是在准确认识汉语形名粘合结构性质的基础上，对汉语形容词构成形名粘合结构的能力、组配规律和语义语用功能进行具体研究。从这个角度看，第一章和第二、三章是“总—分”或“一般—具体”的关系。在第二大块内容中：第四、五章侧重于特定形容词和名词的自由组配，选取“大、小”和“前2679758、原”这两对常见的、组配范围广的形容词进行细致研究，这两对形容词构成粘合结构的情况具有相当的代表性；第六、七章则选取两个含有形名粘合组配的类固化构式分别研究，形容词和名词的自由粘合组配和固化构式中的粘合组配在组配可能性、句法性质、语义表达和语用功能上都存在较大的差异。“数词+大+名词”中的“大”和名词、“序数+形容词+名词”中形容词和名词的关系都比较特殊，这两个格式中形容词的功能也比较特殊。线性形式上它们也是形容词和名词组配在一起，一般不能带“的”，和一般的形名粘合结构具有相似性，正是因为这样的句法构式才使它们能够组配在一起。因此，它们作为比较特殊的一部分纳入本书的研究范围。

五 其他问题

关于行文中语料出处的标注问题，我们基本上采用这个模式：在文学作品、新闻、网页等中手工收集的例句，一般标注具体的完整出处信息，来自网页的一般还标注了网址及查阅时间，查阅时间置于方括号内（如[2001-11-18]）；来自语料库和网页检索（Google 或百度）的例句，一般不标注完整出处（仅标作品名或作者+作品名）或不标注；自造的例句，不标注出处。行文中使用的非“完整句”语料（主要是词、短语，以及部分对比性的简单句）不标注出处。语料出处一般使用圆括号，置于语例之后。

关于行文中语料的标序问题，我们基本上只对完整的例句进行标序，序号置于例句之前，每一节单独排序。单纯的词、短语等语料不标序号。

行文中涉及的英文作者名、作品名、地名、机构名、语系名、语族名、语言名，一律音译为汉语，第一次出现时一般把相应英语原文附注在括号中，大多数人名翻译参考了已出版文献的通行译法，个别翻译系笔者自译。因笔者英语水平有限，对部分英文人名特别是非英语国家人名的实际读法把握不太准，因此在书后附录中列了个人名对译表，以期能弥补可能出现的翻译谬误。

I. Research Objects and Research Objectives The object of study in this book is modern Chinese adjectives and "shape-name adhesive structures". "Shape" refers to adjectives, including generally known nature adjectives, state adjectives, non-predicate adjectives (distinguishing words), etc., such as "large, small, broad, state-run, intermediate", etc. "Adverb + adjective" is regarded as a type of state adjective in Zhu Dexi's "Study of Adjectives in Modern Chinese" (1956), and this book treats it as a phrase because "adverb + adjective" directly modifies the noun with theoretical significance. "Name" refers to nouns, including general nouns, collective nouns, material nouns, abstract nouns, etc., as well as denotative predicates, and sometimes partial noun phrases such as "table, class, water, thought, battle, treasure trove of cave art", etc. Each type of noun can be modified by an adjective, but the ability to be modified by an adjective varies considerably. "Shape name adhesive structure" refers to noun phrases formed by adjectives and nouns without "of", such as "small table", "hot scene", "gentle route", etc. Zhu Dexi (1956) proposed to divide the combination of adjectives and nouns into three types: "A 1 form", "A 2 form" and "B form", where "A 1 form" refers to the nature adjective without "the" and nouns forming noun phrases. Later, Zhu Dexi (1982:148-149) divided the noun partial structure into two types: "combination" and "adhesive", in which the adhesive combination of adjectives and nouns is roughly equivalent to the aforementioned "A1 type". In Mr. Zhu's definition of "adhesive", in addition to the combination of form names without "of", there is also a condition that adjectives in adhesive are limited to nature adjectives and non-predicate adjectives (distinguishing words). We use a looser definition here, as long as the adjective and noun combination form a noun phrase, it belongs to the "form-name adhesive structure" discussed in this book, regardless of the type of adjective. Because adjectives and nouns are bonded combinations, although related to the category of adjectives, they are not necessarily related. This study mainly examines the combination of an adjective and a noun, and considers the combination of multiple adjectives or complex adjectives and nouns when necessary, such as "big long ears", "lean little old man", "small village official", "great good news" and so on. Generally, only the bonded name structure is considered, and some combined name structures are also involved in the comparison. In order to compare or explain similar phenomena, this study sometimes involves some "noun + noun" and "verb + noun" structures. There is actually no clear boundary between word formation and syntax, especially in a language like Chinese that is "clear morpheme, unclear words". Mr. Shen Jiaxuan said at the academic lecture on the 60th anniversary of the establishment of the Institute of Linguistics of the Chinese Academy of Social Sciences (2010): "Over the past century, starting from the Ma Shi Wentong, there have been many achievements in exploring the characteristics of their own Chinese, such as Zhu Dexi, Lu Zhiwei, etc., such as Lu Zhiwei said that 'the word formation of Chinese is not so much a form, but a syntax', and now in the new foreign theory, such as Chomsky's collaborators in the generative grammar school proposed to cancel the form, merge with syntax, the form of words is syntax. "Whether in synchronic or diachronic perspective, the boundary between Chinese compound words and phrases is very blurred. Moreover, the shape-name adhesive structure is originally considered to be quite "solidified" by domestic scholars, and some scholars (especially foreign scholars) believe that the shape-name adhesive structure is a "compound word" because they cannot be expanded. Considering this situation, the "shape-name glue structure" mentioned in this book may also include some "shape + name" words, and sometimes there is no clear distinction between lexical structure and syntactic structure. When discussing the principle of composition and semantic relations, it is possible to also use the word "shape + name" as comparative or supporting material. The "shape-name bonding structure" discussed in this book is in principle limited to modern Mandarin Chinese, and generally does not cover the situation of ancient Chinese and dialects except for comparison and corroboration. There has been little research on adjectives and noun combinations in Old Chinese and Chinese dialects, and we have collected some literature and corpus in the course of this study, which will be studied separately. The research objectives of this book are mainly four: a. The status and characteristics of modern Chinese adjectives. b. The nature and function of modern Chinese form name bonding structure from a cross-lingual perspective. c. Composition law of modern Chinese form name adhesive structure based on real corpus. d. Semantic expression of modern Chinese form name glue structure. If the corpus problem does not consider the difference in frequency and acceptability between a single combination and a single combination, nor does it consider the various syntactic, semantic and discourse differences of the combination and cognitive motivations, and only considers the actual situation of the combination, then the set of possible combinations of adjectives and nouns in modern Chinese is a Cartesian product (symbolized as: A×B = {(a,b)—a∈A and b∈B}), which consists of all ordered pairs (a, b). , where a ∈ adjective and b ∈ noun. For example, if the adjective set A= {big, red, excellent}, and the noun set B= {man, apple, traditional}, then A×B= {(big, person), (big, apple), (large, traditional), (red, people), (red, apple), (red, traditional), (good, human), (excellent, apple), (good, traditional)}, a total of 9 theoretically possible combinations. If according to this algorithm, assuming that there are 2,000 adjectives and 5,000 nouns, the Cartesian product that constitutes the structure of the form name is 10 million, and if you consider proper nouns, noun phrases, and multiple adjectives modifying the noun component, then the possibility of combination is a huge number that causes headaches. In order to understand the actual rules of adjective and noun combinations in modern Chinese, it is necessary to identify and classify these possible combinations one by one, and successively use the method of seeking differences and seeking similarities to find out the laws that restrict the combination. If only the binary combination of an adjective and a noun is considered, then we must also distinguish at least some of the following cases among the tens of millions of possible combinations in this theory: a. Some adjectives and certain nouns cannot be combined in general. Such as "red, traditional". Some adjectives can sometimes be combined with certain nouns, but only in combination. Such as "excellent, people". c. Some adjectives can be combined with certain nouns, but only in an adhesive way. Such as "red, people". d. Some adjectives can be combined with certain nouns, either in an adhesive or combined manner. Such as "excellent, traditional". e. Some adjectives and certain nouns cannot be combined in general, but they can be combined with the help of certain special syntactic formats or contexts. For example, "(This is a great tradition of the Han people)". ...... To study the shape name bonding structure of modern Chinese, some people may think that it is enough to investigate the specific combinations of the above three categories of c, d, and e, and even the combination in class d should not be considered. This really only considers two categories: those that can form a name-name adhesive structure and those that can't. We believe that this is very insufficient, because the composition of the shape name adhesive structure belongs to a systematic structure, and the affirmative combination and conditions (the types and conditions that can constitute the shape name bonding), the negative combination and conditions (the types and conditions that cannot constitute the shape name adhesive structure), the correlation combination and conditions (the types and conditions that constitute the shape name combination structure that are most related to the shape name adhesive structure) are all necessary in the systematic investigation, we must not only examine the "what" and what characteristics the shape name adhesive combination is; It is also necessary to examine what the shape name adhesive combination does not have, and what characteristics it does not have; It is also necessary to examine which combinations are "most similar" and how they relate to each other. Therefore, we must summarize the various combination categories identified as much as possible through a large number of repeated tests, and finally comprehensively and systematically explore the law of shape name bonding combination. This very rational approach may seem very scientific, but language is not mathematics and logic after all, it is a "complex system". In the above classification method, there are still a variety of situations that are difficult to handle: a. The boundary between words and phrases is difficult to distinguish, such as "red person" and "adult" can be regarded as a word, or they can be regarded as a phrase (there is a difference in meaning when looking at words and looking at phrases, among which "red people" need special context when looking at phrases). b. A word may have multiple meanings, and the composition ability and performance of different meanings are different, and their contextual dependence is also different. For example, "old" can be combined and glued in the sense of "old", such as "old painter" and "old painter"; However, in the sense of "long relationship duration", it is generally only suitable for glue, such as "old classmates, old friends" can not be said to be "old classmates, old friends". c. A combination may have multiple functions, and different functional states limit the ability and type of combination, such as the so-called analogy of centrifugal structures "high speed, long distance", etc. d. Personal sense of language and conception of specific composition context may be very different, such as "gentle woman" may not be valid in some people. e. The most critical point is that the frequency, acceptability, and markability of different adjectives and different nouns are different, and does a one-size-fits-all ignore the difference? ...... Although there are these difficult problems to solve, we use relatively simple "can" or "no" to judge and categorize, at least at this stage, we can still draw some rules. Therefore, the judgment of the ability of certain adjectives and nouns to form adhesive structures in this study inevitably tends to be simplistic, and we will try to make up for this deficiency in other related discussions. In order to avoid obvious errors caused by a single method, our determination of the ability to combine specific adjectives and nouns relies on the combination of intuitive judgment and retrieval and collection of actual corpus verification as a language researcher. Using an introspective method based on language intuition to judge the possibility of combination is very effective in typical and commonly used combination situations, but it is more troublesome to analyze some marginal situations, and problems such as "Yu Youren's beard" often appear: Once, a child touched Yu Youren's long beard and asked curiously: "Grandpa Yu, when you sleep at night, is this long beard placed in or outside the quilt?" At that time, Yu Youren couldn't answer. He went to bed at night, whether he put his beard in or outside the quilt, he felt very uncomfortable, and he couldn't sleep all night about this problem. (Wei Tixiang, "The Heart of the Lamp, the Wisdom of the Lamp: The Great Wisdom and the Truth of Simple Life in Small Stories", Beijing: China United Press of Industry and Commerce 2001: 153) Most people's understanding of language is generally unobserved. Mr. Lu Shuxiang ([1980] 2006:5) also said: "Language, that is to say, seems to be an extremely ordinary thing. But if you think about it, it's a great deal. It is precisely because talking is as common as eating and walking that people do not think about what it is all about. "A lot of our language research will also encounter the situation of Yu Youren's beard, we want to explore the language, we want to understand the linguistic phenomena that we are accustomed to without noticing, the actual result may be that we have doubts about the original condition of the language we are familiar with, various theories, various situations, various changes, various interpretations... Let us doubt the value or assumptions about certain laws of language, until we feel decadence or even a breakdown of confidence in the study of language itself. When it comes to the topic of shape name bonding structure, we may have to ask others, or we may have to ask ourselves: Can a certain adjective be combined with a certain or certain noun? Are adjectives bonded or combined with nouns? What is the difference between the bonded and combined form name structure? ...... The person being asked, or ourselves, will hesitate, self-justify, or even go against the intuition of certain answers. In our daily language use, it may be like Yu Youren's beard, and we don't care at all, but if we really want to study everything clearly, think, compare, and consider it repeatedly, in the end, we may not even know whether this is really the case. Introspective linguistic material tends to appear as a result of the right beard, which can be a major obstacle to our language research. A passage from Xiao Guozheng (2001:11) expresses this confusion: "Many people who have some experience in grammar research may have such experience: some sayings, at first, feel that they cannot say, but they study and study, and slowly feel that they can say." You say it's deliberately manufactured, but you can't say it, it's not deliberately manufactured, it doesn't look common, or it's not at all. His explanation is twofold: First, as the research deepens, the degree of rejection of heterogeneous constructs by language researchers may decrease, he said: "Researchers began to make sentences using the same language coding and original grammar system of language identification as everyone, but in the process of research, due to the slow flushing of the flow of the same phenomenon, the repulsive vigilance of language weakened and compromised, and finally some unacceptable sentences entered the legal channel and were accepted." "The introspective corpus for such reasons may be wrong. Second, the ability of researchers to supplement the context in which the sentence is established is stronger than that of ordinary people. He said: "The context in which some sentences are established is relatively wide, while the establishment of other sentences has stricter requirements for context, and the establishment of distant context sentences with stricter requirements, the foreshadowing of its context, is difficult for the language receiver to supplement at once, and long-term research on this phenomenon can improve the level of contextual supplementation and the speed of supplementation of specific distant context sentences, so that the so-called unqualified sentences that start to break can be made at this time... Receive a legal certificate. "The introspective corpus for this reason is acceptable, but there will be a sense of strangeness. The corpus collected by language researchers through introspection tends to decrease in reliability as they think and study deepen. If the theory of accommodation is also involved, then there will be more problems, which may be the reason why some scholars in China are puzzled and dissatisfied with some strange introspection-based linguistic examples of some papers of the formalist school. Introspective material may not be reliable, so is statistics and collected material? The collected written corpus and oral corpus will also be "learned or misused" due to individual differences in speakers, differences in context, and even speech speed, omission, mistakes, etc. Moreover, structuralist inductive methods have long been refuted by transformational generative grammarists: first, structuralist inductive methods cannot describe language completely; Second, the discovery procedure is unidirectional; Third, structuralist inductive research is essentially nothing more than enumerating observed linguistic facts and then classifying them, and the result of the study is only a part of the class grammar (Xu Liejiong, 1988). Of course, statistics and introspective judgments of frequency of use can compensate for this deficiency to some extent. The combination of the two may improve the universality and scientificity of the conclusions, and the "introspection-verification" or "internal and external evidence" (Xiao Guozheng, 2001: 6, 11) is a very reliable corpus collection and processing method. Our treatment of the corpus is basically a model of "introspective judgment + real corpus verification". The specific methods are: (1) Manually collect a large number of corpus, which are quite representative and typical after our selection, which can be used as a concrete example of the thesis, and as the basis for proposing the hypothesis of morphological structure classification and assembly law in the early stage. (2) According to the description and examples after each adjective entry in the Adjective Usage Dictionary (edited by Zheng Huaide and Meng Qinghai, 2003), the adjective formation structure is preliminarily determined. (3) We judge the possibility of adjectives and nouns forming a bonding structure based on our introspection based on the sense of Mandarin discourse and the sense of language of others, and further modify the judgment and statistics of the possible combination. (4) Using the massive linguistic information of the Internet, the adjectives in the "Adjective Usage Dictionary" were searched one by one through the Internet search engine, and the first 20-50 examples of the adjective and various noun combinations were extracted. (5) Using relatively mature corpus (such as the CCL balanced corpus of the Center for Chinese Chinese Linguistics of Peking University and the "Audio Media Language Text Corpus Retrieval System" of Communication University of China) to retrieve the combination of commonly used adjectives and nouns. Five ways to collect, process, and verify the corpus, which can avoid obvious errors in the corpus and conclusions to some extent. In addition, our investigation cannot be exhaustive, because it is unrealistic for us to discern all possible combinations, at least in terms of time and effort. Since many possible combinations of expressions have the characteristics of "class", our innate ability to think by analogy can allow us to grasp the laws more quickly and generally. Therefore, our approach is similar to some models of computer natural language processing: preprocessing of some data - generalization of laws - application promotion (effect measurement). In short, on the issue of using corpus, we use carefully selected representative corpus and our linguistic intuition to summarize the rules and types of various adjectives and noun combinations, and then conduct a large-scale verification of the real corpus, and finally put forward our own opinion. Three research ideas Liu Danqing (2003:21-29) uses three pronunciation close English words test (test), text (discourse), attest (verification) to summarize the characteristics of the three paradigms of contemporary linguistics formalism, functionalism and language typology, in which "test" refers to the introspection of native speakers to explain the structural law of language, generally does not consider the language components that are not directly related, introspection is the main research method used by formalist schools. This approach is also often used by some structuralist linguists; "Discourse" refers to summarizing linguistic laws according to the actual usage in the real corpus (discourse), often using statistical quantification to determine the tendency law, and paying attention to the relationship between grammatical units and context, which is a research method often adopted by functional schools; "Verification" is the main research method adopted by contemporary language typology, which selects some traditional parameters and proves based on cross-lingual evidence, investigates the distribution and performance of a certain syntactic structure in different languages/dialects, verifies the limits of human language variation, and discovers the commonality of language. Each of the three paradigms has its own advantages, and this study aims to absorb their advantages as much as possible. The basic research ideas of this book are: (1) to investigate the nature, function and research significance of modern Chinese form name adhesive structure from a cross-language perspective, the adjective problem has attracted much attention in contemporary language typology, we occupy the relevant research results of typology as much as possible, explore the adjective problem of Chinese from a cross-language perspective, and finally determine the significance of modern Chinese form name adhesive structure in language typologies, and use cross-language evidence to prove the nature of modern Chinese form name adhesive structure. The composition law and semantic characteristics of the shape-name adhesive structure are explained. In contemporary linguistic typology, the relevant syntactic properties of adjectives are a major object of concern. First, contemporary linguistic typology has been very concerned with adjectives since its inception. In Greenberg's classic paper [1963] 1966), among the 45 commonalities listed, there are 10 items such as commonalities 5, 17, 18, 19, 20, 21, 22, 24, 31, 40 (original serial number), mainly word order, accounting for 22%, which shows that the word order of adjectives and nouns still occupies a lot of weight in Greenberg's mind. Greenberg (1963]1966) regarded the position of qualifying adjectives and associated nouns as the "basis of the third classification" in word order type studies in addition to subject and object. But they have different word order representations. Although later research results have proved that the word order of adjectives and nouns is not as related to the word order of verbs and objects as the word order of verbs, objects, etc., like the word order of determiners, articles, numerals, quantifiers (such as some, all), etc., which may also serve as the core of nouns, adjectives and noun order still play an important role in multiple implications. Second, in a large number of linguistic typology research results, there are many linguistic commonalities about adjectives. We searched the Konstanz University Universals Archive on December 9, 2009, 2679756 included almost all contemporary typological research, and found a total of 127 commonalities related to "adjectives" in this archive, covering the part-of-speech nature, category, syntactic function of adjectives, etc. This library of commonalities not only lists commonalities, but also provides a brief description of the development of commonality research (such as counterexamples). Thirdly, the part-of-speech status of adjectives and cross-language recognition are the research hotspots of language types. In the past two decades, the problem of parts of speech in human language has attracted a lot of attention, such as the "Amsterdam model" of Hengeveld and Rijkhoff on cross-language word class determination, Croft's "typology typical", Dixon's "core semantic type-main syntactic function-main word class" pairing model, etc. Various theories first need to face and solve the cross-language complexity of adjectives, most of the differences of views are in the determination of adjectives, adjective issues have become a hot spot in linguistic typology disputes, some views are even-for-tat, individual scholars will have completely different views on the adjective problem of a specific language at different times. Adjective parts of speech have become the touchstone for testing the theory of typology. Subsequent research will show that when adjectives and nouns in modern Chinese form noun phrases, the distinction between combination and glue is partially functional consistency with other languages' distinguishing means such as marked and unmarked determiners, prepositional definite and postpositional determiner, conjuxative determiner and modifier, adjective determiner and adjective predicate, and the formal difference between adhesive and combined can be used to distinguish the difference in form, semantics, syntax and pragmatic expression between adjectives and noun combinations. Such as the difference in the semantic category of adjectives (such as the difference in nature, state, etc.), the difference in the external form of the adjective (such as the difference in syllables, structural complexity, component composition, etc.), the difference in the semantic category of nouns (such as the difference between concrete nouns, abstract nouns, proper nouns, etc.), the semantic relationship of adjectives modifying nouns (such as the difference between restrictive, non-restrictive, connotative and other determiners), the information characteristics of the structure of the form name (such as the known and unknown differences between the properties of the adjective), The syntactic environment in which the shape name structure as a whole (such as the entire form name structure acts as a distinction between determiner, adverb, object, predicate, subject, etc.), the pragmatic function of the form name structure (such as the difference between the entire form name structure to represent titles, representation attributes and representation classifications, etc.), the frequency of the form name structure assembly (such as common combinations, occasional combinations, etc.), the functional types of sentences in which the shape name adhesive structure is located (such as the difference between statements, imperatives, exclamations, questions, etc.), and so on. Adjectives and noun combinations are two relatively independent functional means, and the use of adjectives or combinations is the result of selection according to the needs of form, semantics and expression, and has the same value as the functional differences of different syntactic forms of adjectives in other languages. It can be further abstracted that "direct glue" and "use functional word combination" are two different syntactic means, which can also be selected in other structures such as "noun + noun", "verb + noun", and "adverb + center" due to functional differences. (2) Pay attention to the combination of real corpus and introspective judgment, and investigate the conditions and types of adjectives and nouns forming adhesive structures in the context of small sentences, and most of the previous studies on the conditions for adjectives and nouns to form adhesive structures have mostly focused on two major aspects: (1) the category of adjectives themselves, such as monosyllabic adjectives are more likely to form adhesive form name structures, state adjectives and adjective phrases are more likely to form combined form name structures, etc. (2) The semantic relationship between adjectives and nouns, it is generally believed that adjectives and nouns will form an adhesive structure only when the attributes represented by adjectives are the "essential properties" of things represented by nouns and are customarily used as categorical features. This type of research generally only needs to consider two components (i.e., adjectives and nouns) and the relationship between the two, and the use cases studied are mostly in the form of phrases. Examples of adjectives and nouns forming bonded structures, such as Modern Chinese Dictionary and Adjective Usage Dictionary, are also mostly phrases, and generally do not involve other syntactic components and discourse environments other than formal name phrases. The advantage of this tendency to static analysis is obvious, we follow the "Occam's razor" principle of "do not add entities if necessary", when analyzing the composition law of the shape-name adhesive structure, if it can be solved between the two components of adjective and noun, try not to involve other relevant factors. Using the two components of adjectives and nouns to investigate the composition law of the form-name adhesive structure, we mainly proceed from the following aspects: (1) The influence of the subclass of adjectives on the composition. (2) The effect of the subclass of the noun on the composition. (3) The effect of the number of syllables of adjectives and nouns on composition. (4) The influence of cognitive semantics on composition. For the fourth aspect, we will make a key investigation, recognize the rationality of some scholars to deal with the problem of shape-name bonding from the perspectives of idealized cognitive mode and conceptual distance, and further extend and exert it. In addition, special emphasis is placed on the opposite of the principle of iconicity: the economics of cognitive processing will also treat those "attribute-subject" relationships with long conceptual distance and complex semantics as shape-name gluing structures, and there are a large number of markup reversals between conceptual distance and shape-name structure, conceptual complexity and shape-name structure. This is also one of the highlights that sets this study apart from other studies. Static analysis methods only consider adjectives and nouns, and other non-directly related components are generally not considered, and then determine whether they can be combined by analyzing the formal and semantic characteristics of adjectives and nouns and the semantic relationship between the two. This is a zero-breaking research method, first dismantling the form-name combination from the relevant context, and then dismantling the form-name structure into two components, adjective and noun, and then analyzing these two components and their relationships, and concluding that this conclusion is generalized to the similar complete context in which they were originally disassembled. This zeroing research method is a relatively mature scientific research method, but it has long been criticized by some scholars. For example, the famous American futurist Alvin Toffler wrote in the foreword "Science and Change" for Prigogin (Nobel Prize winner) and Stanger's book "From Chaos to Order": One of the most developed techniques in contemporary Western civilization is to break the zero, that is, to break down the problem into the smallest possible parts. We are so good at this that we often forget to put the details back together. This technique is perhaps the most well-honed in science. In science, not only are we accustomed to dividing problems into many details, we often use a useful technique to isolate each of these details from its surroundings. This technique is what we often call certeris paribus, which means "all else is the same". In this way, the complex interaction between our problems and the rest of the universe can be ignored. (Prigogine & Stengers, 1987: 5) But in psychology, theories and methods that focus on "holistic" research have been popular since the early 20th century. Such as "Gestalt" psychology, founded and developed by Wolfgang Kohler, Kurt Koffka, Kurt Lewin and other European psychologists who immigrated to the United States in the 2030s. One of the most important ideas of Gestalt psychology is that people's perception of the environment and events is not composed of many independent components, but a "dynamic whole." (Taylor, Peplau & Sears, 2004) Obviously, the advantages of the "zeroing" method are obvious and highly operable, but its conclusions may avoid many facts that should be considered. Kant said: "Do not rashly reduce the diversity of entities", if the investigation of the adjective and noun combination law only uses the zero method and only relies on introspective static materials, it will inevitably ignore the complex diversity of the shape name bonding composition, and inevitably ignore the important role of other components related in the small sentence, the overall function of the structure, the relevant context, etc. in the composition of the shape name adhesive structure. Adjectives and nouns form an adhesive structure, which often cannot be "set to be the same in other situations", and paying attention to the actual performance of the structure as a whole in related sentences and contexts is the only way to explore its construction law and semantic expression. The advantages of static analysis methods are obvious, easy to grasp, and the rules are relatively obvious. However, it is not enough to analyze the law of shape name bonding assembly in this way. Adjectives and nouns that can be composed of static can form name adhesive combination, and can generally be combined in a certain syntactic environment, but some static can not constitute adjectives and nouns of form name adhesive combination, and can also form adhesive combination in a specific syntactic environment. Shen Jiaxuan (1999:290) pointed out that some combinations of shapes and names generally use the combination in static combinations, but after entering the sentence, you can not add "of", such as "generous manners", and generally add "of" when saying it alone, but you can not add it in "her generous behavior at the cocktail party is impressive". Zhao Yuanren ([1968][1979]2001:304) also believes that the unspeakable "heavy box" and other name-like adhesive structures listed in Zhu Dexi's "Study of Modern Chinese Adjectives" (1956) can be said in dialogue, dubs, and other marked occasions. It is the context of the small sentence that leads to the implementation of glue combination. Sapir divides language into three categories: "analytical", "synthetic" and "multi-synthetic" according to the performance of syntactic devices in language. Among them, Chinese is regarded by him as a typical example of analytical language, which he believes is "in the language of analysis; Sentences are always primary, while words are secondary" (Edward Sapir, [1921] 2007: 114). It is inevitable to examine the grammar rules of Chinese in combination with sentences, and it is very appropriate and necessary to use small sentences as an operating platform for Chinese grammar research. Small sentences are the smallest expressive and independent grammatical units (Xing Fuyi, 1996: 13-15). Small sentences have internal structural components and component relationships, semantic relationships, stylistic characteristics, functional characteristics, frequency characteristics, a larger discourse environment (compound sentences, sentence groups, discourses, chapters, etc.), with small sentences as the entry point (standard) for grammar research, can concentrate as many factors that affect the grammatical structure as much as possible, can find some previously difficult to find grammatical rules, can explain some previously difficult to explain phenomena. Because small sentences are small sentences in a specific context, to a certain extent, the grammatical analysis of "small sentences" can be regarded as an analysis based on the real corpus. There are many influences of sentence context on the composition of the shape name adhesive structure, such as syntactic position, number of components, markability, sentence function (question, imperative, statement, exclamation), affirmation and negation, context context, language style, frequency, etc., which will affect the actual composition of the shape name adhesive structure. In this study, the conditions for adjectives and nouns to form an adhesive structure are analyzed from a static perspective, and then on the basis of a large number of example sentences containing morphonomic structures accumulated by us through manual collection, corpus retrieval, web search, etc., the actual situation and constraints of adjectives and nouns forming an adhesive combination are analyzed in combination with the context of the clause environment, and the results of static analysis are expanded, and the context in which the form-name adhesive structure appears, especially the regulating effect of small sentences. And combine the results of static analysis with the results of dynamic analysis of sentences, and try to use a unified pattern to explain. (3) Combined with some views and methods of structural semantics and cognitive semantics, the semantic expression of the shape name adhesive structure is interpreted in sentence control From the research literature on adjective semantics at home and abroad (especially the interpretation of various dictionaries), the semantics of adjectives are quite different from the semantics of nouns and verbs, and their specific semantic interpretations often change with the context. Moreover, the structure of the shape-name bonding structure is compact, and compared with the combined name-based structure, the semantic index is high, and the semantic relationship types of adjectives and nouns are complex. Although the specific semantic interpretation of adjectives has some default values, it often allows other interpretations and changes, depending on the specific context: on the one hand, when specific adjectives and different nouns are combined, the semantics will have different interpretations, such as "good mobile phone, good cook, good teacher", the specific semantic connotation of "good" is very different. On the other hand, the semantics of specific adjectives and specific noun combinations will also change depending on the context, such as "small boss" allows for many interpretations. However, just as adjectives and nouns can be combined regularly, so is the semantic interpretation after combination: first, adjectives and nouns will automatically seek "adaptable" interpretations based on idealized cognitive patterns; Second, context strengthens or suppresses the default semantic interpretation. Therefore, for the convenience of operation, we mainly use small sentences as the context to analyze the semantics of the shape-name glue structure. And adhere to a principle: if the semantics of the shape name glue structure can be interpreted at the phrase level, it will not be interpreted at the clause level. The semantic interpretation, analysis and synthesis of the shape-name adhesive structure should be combined (Shen Jiaxuan, 2005a). Some structures can be simply thought of as "the whole is equal to the sum of the parts", and the semantics of the entire structure are equal to the adjective semantics and noun semantics are formed according to a simple "attribute + subject" relationship, such as "small table". The semantic interpretation of this structure can be adapted to the methods of structural semantics, which can be used to analyze the methods. But more shape name adhesive structure semantics is "the whole is greater than the sum of parts", the semantics of the entire structure will be greater than the semantics of adjectives and nouns, is "attribute + X + subject" style, X represents the key semantic factors that connect attributes and subjects, such as "hot cup lid", which refers to "the lid of the cup used to hold hot drinks". Therefore, the semantic analysis of such name structure requires the help of some theories and methods of cognitive linguistics, and requires the use of comprehensive methods. Shen Jiaxuan (2005a) argues: "The study of language, analysis and synthesis are indispensable. "When we analyze the semantics of shape-name adhesive structures, the use of analytical methods and synthesis methods depends on the needs. The relevant achievements of structural semantics and cognitive semantics provide many valuable theoretical frameworks and analysis examples for the semantic analysis of shape-name bonded structures, such as semantic field theory, physical structure theory, psychospatial theory, etc., which we will absorb as much as possible in our research. Because we adopt a "pragmatic" attitude, we do not explore too much into the multiple theories themselves, and we will ignore the conflicts between multiple theories in the specific analysis. Our semantic analysis of the structure of form name bonding mainly focuses on four aspects: (1) the semantic function types of adjectives, such as restrictive, non-restrictive, connotative, isotopic and other categories, especially the various subtypes of non-restrictive adjective determiners. (2) The specific semantic interpretation of adjectives and the construction of interpretation patterns. (3) According to the principle of "the whole is greater than the sum of the parts", the semantic characteristics and types of the entire shape-name adhesive structure are analyzed. (4) Theoretical consideration of the semantic characteristics of adjectives, adhesive structures, and form-name adhesive structures by comparing similar or related words, parts of speech, phrases, etc. (4) Combining general research and specific case studies, rational division of labor, each with its own emphasis, multi-angle and multi-level exploration of the composition law and semantic characteristics of shape-name adhesive structures In general research, we focus on solving the following problems: First, according to the different abilities that make up the shape-name adhesive structure, adjectives are divided into different classes, and the characteristics of each subclass are summarized. Second, combined with the real corpus, this paper analyzes many affirmative conditions affecting the composition of the shape name adhesive structure, and attempts to construct a judgment operation process model. Third, on the basis of static semantic analysis, various semantic relationship types and specific semantic connotations of shape-name bonding structures are analyzed. In the specific case study, we selected two representative free combination formats of "big/small + noun" and "front/original + personal noun" and two combination formats with a considerable degree of solidification of "number + large + noun" and "ordinal + adjective + noun" for in-depth discussion. General research mainly uses adjectives as variables, focusing on the analysis of how different types of adjectives affect the formation of adhesive structures and semantic expression. Specific case studies mainly use nouns as variables, focusing on the impact of different types of nouns on the formation of adhesive structures and semantic expression. Although in fact noun differences are also involved in general studies and adjective differences in specific case studies, we intend to make an appropriate division of labor in focus, which is only a tendency. (5) Tends to functionalist view of language, attaches importance to the influence of semantic factors on syntactic structure, and pursues the grammatical analysis of tendentious regularity functionalism has the following characteristics: (1) Break through the practice of only taking introspective sentence patterns as the research object in formal research, and pay more attention to the influence of various factors related to language behavior on discourse organization; (2) It breaks through the practice of focusing only on the similarities and differences of types in formal research, and pays more attention to the law of tendencies reflected by the number of examples; (3) Break through the practice of treating the object as a static finished product in formal research, and more as a dynamic process, and study the language cognitive strategies of both listening and speaking; (4) The practice of looking at sentences (or even just a structure) in isolation from formal grammar and attaching great importance to analysis in relation to context (Zhang Bojiang and Fang Mei, 1996: 1). This study focuses on the analysis of real corpus, the dynamic construction of shape-name bonding structures in small sentences, and the analysis of the semantics of the structure in relation to context, all of which tend to be functional. Structuralism and transformational generative grammar both hope to study the influence of semantics on syntactic structure, but in fact it is difficult to do so. In the most purely formal computer language information processing, semantic factors can not be completely abandoned, the famous artificial intelligence philosopher Boden ([1990] 2001: 8) argued: "Even the simplest programs are not purely formalistic, but have some fairly primitive semantic properties, so fundamentally speaking, computational theory is not inexplainable meaning." As long as the brain's ability to generate intentionality is clear and not completely counterintuitive, the information processing used for this understanding can also be applied to computers. "Computers can also simulate the intentional processing of the human brain, and they must do the same. Semantics that do not involve duration are derived from the intentional characteristics of the human brain, which is the focus of the scientific debate about "determinism" and "randomism". Because of intentionality, the world does not look like a giant automata, and we have come to realize that we live in a pluralistic world (Prigogine & Stengers, 1987: 26-27). Therefore, it is desirable to fully consider and rely on semantics and functionality in language analysis. Scientists have called for "attention to the surprisingly complex behavior lurking in some simple models" (James Glake, 1990: 4). René Thom (1989:396) argued: "If systematic behavior cannot be characterized by a simple law clearly stated, then first try to characterize this behavior qualitatively with the help of tendency." Precisely because the world is not a huge automata, not necessarily completely constrained by the "rules" behind it, and sometimes with a great deal of randomness, linguistic analysis based on intentionality is often incomplete, not what determinists think or hope: the ability to repeatedly apply rules, the ability to explicitly include or exclude certain linguistic elements, the ability to disassemble and combine, and so on. Even staunch proponents of determinism will have a clear affirmative understanding of the role of probability, and the laws of language are often just a "rough" probability problem. This probability tendency is because of the identity of people's use of linguistic means and the identity of expression intentionality, and this linguistic expression is a matter of probability rather than "uniformity", and it is also because of the selectivity of people's use of linguistic means and the selectivity of expression intentionality. Any scientific theory is incomplete, and so is the summary of linguistic theories and their laws. The famous German mathematician Gödel proposed and proved the "incompleteness" theorem as early as 1931, arguing that any axiom system without contradictions, as long as it contains statements of elementary arithmetic, must have an undecidable proposition, and this set of axioms cannot determine its truth or falsehood. The theorem originally developed in the field of mathematics, but its impact extended far beyond mathematics. The theorem not only revolutionized mathematics and logic, raising many challenging questions, but also involved philosophy, linguistics and computer science. 2679757 from the incompleteness theorem, almost all grammatical laws have loopholes. As a study that tends to functional linguistics and cognitive linguistics, the incompleteness of the conclusions may be more pronounced. Ma Zhen (2004) argues that functional grammar "can only provide a tendentious explanation, not strict rules or rules". Lakoff (1973) argues that "grammatical" and "ungrammatical" are simply a matter of degree. It is inevitable that a universal, absolute law of the bonding structure of form names cannot be found, but this is not a reason to stop exploring. Our research is dedicated to finding tendencies and probabilistic laws. IV. The basic structure of the study, in addition to the "Preface" part, the main part of the study consists of seven chapters. Chapter 1, Chinese adjectives and form names bonding structures; Chapter 2, the ability of adjectives to form form names to bond structures and their composition rules; Chapter 3, Nomenclature and Semantic Analysis of Form Name Bonding Structures; Chapter 4, Case Study I: "Big and Small + NP"; Chapter 5, Case Study II: "Former, Original + Referential Noun"; Chapter 6, Case Study III: "Number + Large + NP" Configuration; Chapter 7, Case Study IV: "Ordinal + Adjective" + Noun Structure. These seven chapters can be summarized into two main blocks: the first three chapters are an overall study of the gluing of Chinese adjectives and form names, focusing on the ability of different adjectives to form a form name adhesive structure; The last four chapters are case studies on form-name bonding structures, focusing on the selectivity and semantic expression of shape-name bonding structures formed by different nouns and specific adjectives. These two large pieces of content have a division of labor internally. In the first major piece of content: the first chapter is to look at the adjective and form name bonding structure of Chinese from the height of linguistic typologies, and accurately locate the adjective and form-name bonding structure of Chinese adjectives and form-name in the context of the commonality of adjectives and adjectives modifying noun structures in human language; The second and third chapters are based on the accurate understanding of the properties of Chinese form name adhesive structure, and the specific research on the ability, composition law and semantic pragmatic function of Chinese adjectives to form name adhesive structure. From this point of view, the first chapter and the second and third chapters are "total-score" or "general-specific" relationship. In the second block: chapters 4 and 5 focus on the free combination of specific adjectives and nouns, and select two common pairs of adjectives "large, small" and "first 2679758, original", which are quite representative of the situation in which these two pairs of adjectives form an adhesive structure; Chapters 6 and 7 select two curing-like configurations containing form-name bonding combinations to study separately, and there are great differences in composition possibility, syntactic properties, semantic expression and pragmatic functions in the free adhesion combination of adjectives and nouns and the bonding combination in the curing configuration. The relationship between "big" and noun in "number + big + noun", and adjective and noun in "ordinal + adjective + noun" are more special, and the function of adjectives in these two formats is also more special. Linear forms they are also adjectives and noun groups together, generally not with "of", and have similarity with general form-name adhesive structures, and it is precisely because of this syntactic construction that they can be combined together. Therefore, they are included as a relatively special part of the scope of research in this book. V. Other QuestionsRegarding the marking of the source of the corpus in the text, we basically adopt this model: example sentences collected manually in literary works, news, web pages, etc., generally marked with specific and complete source information, and generally marked with the URL and access time from the web page, and the time of access is placed in square brackets (e.g. [2001-11-18]); Example sentences from corpus and web search (Google or Baidu) are generally not marked with the full source (only the name of the work or the author + the name of the work) or not; Self-made example sentences without attribution. Non-"complete sentence" corpus (mainly words, phrases, and partially contrasting simple sentences) used in the text are not attributed. Parentheses are generally used for corpus sources, placed after the example. Regarding the order of the corpus in the text, we basically only label the complete example sentences, the ordinal number is placed before the example sentence, and each section is sorted separately. Mere words, phrases, etc. are not marked with sequential numbers. The English author's name, work name, place name, institution name, language family name, language family name, and language name involved in the text are all transliterated into Chinese, and the corresponding English original text is generally annotated in parentheses when it first appears, and most of the name translations refer to the common translation methods of published literature, and some translations are self-translated by the author. Due to the author's limited English proficiency and inaccurate grasp of the actual pronunciation of some English names, especially those of non-English-speaking countries, I have listed a translation table of personal names in the appendix at the end of the book in order to make up for possible translation errors.（AI翻译）