人物专栏 | Yoshio Endo教授访谈（下）
点击上方蓝字关注我们
编者按
《理论语言学五道口站》（2022年第23期，总第226期）“人物专栏”继续与大家分享本站采编人员赵欣宇对Yoshio Endo教授进行采访的访谈录。Yoshio Endo，日本神田外语大学研究生院教授、北京语言大学语言学系国际教授委员会成员，曾在麻省理工学院担任访问学者。
本期访谈中，Yoshio Endo教授首先讨论了认知社会语言学以及社会语言认知学这两个术语，然后以英语连词为切入点比较了核心状语从句和边缘状语从句，接着举例论述了语言变体的重要性，最后解释了空算子和空从属中心语特征之间的区别。
Yoshio Endo教授简介可参考《理论语言学五道口站》“人物专栏”2022年第19期，总第222期。
访谈内容
06.
赵欣宇：变异社会语言学与认知科学融合产生了很多交叉学科,请问认知社会语言学(cognitive sociolinguistics)和社会语言认知学（sociolinguistic cognition）有何差异？
Yoshio Endo教授：我不是十分熟悉“社会语言认知学”这一术语概念，似乎这种说法并不常用。我认为唯一与之相关的可能是Labov教授的研究。Labov教授用两种不同的认知方式解释音变模式，认为音变链和合并取决于对音系范畴的认知，并且给出证据表明说话人拥有其所属群体特有的认知变体（用Labov的话来讲是社会认知）。
与之相比，“认知社会语言学”是更为常用的术语概念。认知社会语言学是语言学的一个新兴领域，旨在用认知解释框架理论去解释社会环境中的语言变体。其目标是建立一个与社会、个人、机构及其相互关系的心智模型。社会语言学领域最近成果颇丰，在句法方面也出现了很多新进展和新方法。Christensen教授2022年最新出版的书中对这些相关背景做了详细的介绍。
在认知社会语言学领域中，我们经常会使用认知语言学框架。认知语言学在意义解释方面可以发挥十分重要的作用。然而，在句法制图理论框架下也可以研究语言与社会的接口问题，而且对于某些现象来说，句法制图的呈现效果更好。让我们来看一些具体的例子。我在2021年的文章中提过，升调“わ”是日本东京及其他地区女性专用的句末语气词，其使用频率在昭和初期达到顶峰（1926~1988），后来逐渐减少，现在日本的年轻一代已经很少使用这一语气词。对日本动漫《海螺小姐》（1946年~1974年）中句末语气词的统计结果表明，在早期漫画章节中，女性专用的升调“わ”使用得更加频繁，后期的使用频率则逐渐减少，如下图所示。
为什么这个句末语气词不再常用了呢？其中一个原因是1940年日本继美国之后兴起的性别平等运动所产生的社会环境。如今，性别平等已经植根于日本文化中，因此当我问我的本科学生是否会使用升调句末语气词“わ”时，他们的回答都是否定的。然而，有趣的是，女生在一岁左右还会使用语气词“わ”，但是她们步入小学阶段后，就会逐渐停止使用“わ”。这是为什么呢？其主要原因在于日本的电视文化。日本电视上会广泛地播放动漫，儿童们就会经常使用从动漫中学习到的语言。那么，日本动漫的语言和特点为什么会鼓励年轻女孩使用现已不再使用的语气词“わ”呢？原因在于日语的特殊性。日本人会根据目前所处的社会环境改变字词，当日本人改变字词的时候，语言也会随之发生改变。例如，是便利店的店员的身份还是顾客的身份会改变我们的语言方式。他们还会根据是与家人还是朋友交谈来转变自己的角色身份，并使用适合该身份的语言。令人惊讶的是，根据身份的不同，日语中有十多种不同的第一人称英语“I”。身份变化可以从句末语气词中看出。幼儿在电视动漫中看到的受欢迎的角色大多是典型的女孩形象。而且由于某个角色使用了句末语气词“わ”，即使该句末语气词在日常生活中已经不再使用，但是儿童们尤其是女生还会使用“わ”。但当他们长大后发现周围的人不使用句末语气词“わ”的时候，他们也就不再继续使用了。
接下来，让我们看看这个句末语气词对句法结构的制图研究有何贡献。事实上，句末语气词“わ”曾是一个话题助词，后来被转换为动名词状语从句语气词“わ”，最后变成了句末语气词“わ”。在句法制图中，这种变化可以表达如下：
[Speech-actP wa [ModfierP wa [TopicP wa…
<--------------/ <-----------/
第一阶段，“わ”位于CP层TopicP的中心语位置，然后通过移位来到更高的ModP位置，最后到顶层的Speech-actP位置。在最后阶段，句末语气词“わ”在从句结构中占据最高位置，所以当句末语气词以这种方式自下而上发生变化改变时，它最后的命运只能是消失。最终，现代日语不再使用句末语气词“わ”。事实上，随着时间的流逝，言语行为句末语气词的消失是日语中的一种常见的模式。正如司富珍教授的研究，汉语的句末语气词似乎是由动词形成的，它们似乎也发生了自下而上的改变。换句话说，句末语气词可能因语言的不同而有所差异，但自下而上向 Speech-actP 位置发生改变的模式似乎是相同的。
07.
赵欣宇：Haegeman（1991 年及以后的文章）区分了两种状语从句（核心状语从句和边缘状语从句）。但是，一些英语连词可以同时引入这两种类型的从句，例如“as”，“since”，“if”等（Endo & Haegeman，2019）。以“if”为例，在您看来，“if1”（引入核心状语从句）和“if2”（引入边缘状语从句）之间的区别对“if”的制图研究有何影响？
Yoshio Endo教授：我先给大家介绍一下Haegeman教授对状语从句研究的理论背景。在Haegeman教授对状语从句的早期研究中，核心状语从句的结构很小，其中的成分只能投射到TP/IP层。另一方面，边缘状语从句的结构则很大，其中的成分可以投射到CP层。Haegeman教授通过截断发现了这种对比，即核心状语从句在IP层被截断，因此位置更高的成分不会出现在这里。相比之下，边缘状语从句根本没有被截断，因此可以出现在CP层。这个截断分析存在一个概念问题。那就是，我们还不清楚为何到了某一句法位置（如IP层）后，核心状语从句就不能继续投射了。为了解决这一问题，Haegeman教授对核心状语从句采取了另一种分析方式。她根据状语从句中空算子移位现象（Rajesh Bhatt，Richard K. Larson等），假设在核心状语从句中，算子从状语从句内部移位到了CP层。而移位之后，核心状语从句就不能再含有CP层的元素，因为根据Luigi Rizzi教授的“相对最简性”理论，CP层的其他元素会阻碍空算子进行移位。
不过令人遗憾的是这种移位分析也存有一个问题。核心状语从句与主句中位置较低的IP层的元素相关，而边缘状语从句却与主句中位置较高的CP层的元素相关（Werner Frey，2011）。这一现象可以简单用截断分析来解释：因为核心状语从句的范围较小而边缘状语从句的范围较大，所以我们认为核心状语从句的小范围与主句位置较低的IP层相关，而边缘状语从句的大范围则与主句位置较高的CP层相关。但是，在移位分析中，无论是核心状语从句还是边缘状语从句，在其内部结构都有着相同大小的范围，并且都有着同样的CP结构。二者唯一的区别在于，某一算子只有在核心状语从句中时，才会移位到IP层，因此我们难以得出前面的结论，即核心状语从句与主句IP层面的元素相关联，而边缘状语从句与主句CP层面的元素相关联。因此，我们需要找到一种机制，其可以证明：当一个算子从状语从句内部低层IP位置进行移位时，这一移位现象一定与主句IP层相关联。根据日语方面的研究数据，Endo&Haegeman（2019）利用Noam Chomsky教授的“标签”理论，提出了这样一种机制：整个状语从句的标签是由移位成分所决定的，例如如果移位发生在AspectP层，那么该状语从句的类型则为Aspect，移位成分就表明了整个状语从句的标签，并且该状语从句的Aspect标签与主句AspectP层相关联。（实际上，其中某些想法源自北京语言大学所办的句法制图国际研讨会上我的介绍海报中。）
出于这种考虑，我们再回溯到问题所在。问题中提到了两种类型的if从句，对于核心位置的“if”来讲，某一元素（在英语中称为算子，在日语中称为核心）会从Mood有关的功能范畴进行移位。因此，整个状语从句的标签就变为Mood类型，并且该状语从句就与主句中与Mood层相关联。而对于边缘位置的“if”来讲，移位会发生在更高的位置，因此整个状语从句的与主句更高位置的元素相关联。这样，整个状语从句的标签就变成了另一种更高的成分类型，而状语从句又被连接到主句的更高成分上（顺便说一下，Donati（2006）也提出了类似的标签算法，他认为在自由关系从句中，名词类型的wh移位元素决定了整个关系句的标签，即自由关系从句的名词性）。
08.
赵欣宇：在您看来，为什么应该关注语言变体？我们能从不同的语言变体中学到什么？
Yoshio Endo教授：重视变体的优势之一是，我们可以对理论进行细化，使之达到更高的精确度，从而做出更加准确的预测。我认为，只关注核心用法，但是忽略语言变体的理论，是不能准确反映实际语言使用情况的。诚然，关注语言的核心以获悉语言的本质也是一种想法。在我看来，关注哪些方面取决于每个研究者的背景和偏好。就我个人而言，我更喜欢句法制图的理念，它既重视语言的实际，也接受语言的变体。以去年我指导的一篇硕士论文为例，这位对社会语言学感兴趣的学生分析了在日本、欧洲和美国播出的相同广告，他注意到在该广告中使用的祈使句在日语中被表达为命令、祈祷/祈使句、恳求等，而这些表达方式在句末的功能层级结构中分处于不同位置。日语中这些含有功能层级的祈使句变体在英语中是不可见的。在此种情况下，英语可以被认为是一种“无声”的语言，因为日语祈使句所涉及的各种功能类别的语态在英语中都没有语音实现。通过假设英语中存在无声功能中心语，英语祈使句的含义就可以被表达得更加准确。Endo（2022）利用其他例子对这一方法进行了探讨。这篇论文已经开放获取，大家都可以免费阅读。
目前，补语“that”中出现了一种不涉及社会因素的语法变体。在我与Andrew Radford教授的一项调查中，我发现说话者之间存在差异，比如有些人在“how come”之后使用“that”，有些人不使用“that”，还有一些人在“how come”和“that”之间有状语成分时使用“that”。这些结果在Endo（2018）中有详细讨论。另外，在我最近的研究中，我在与Andrew Radford的电子邮件往来中发现，他可以使“wanna”收缩，即使在“want”和“to”之间存在wh移位的语迹。此外，他也没有“that”语迹效应。据我所知，大约有10%的人会使用这种语法。而且我还发现，有相当多儿童无限制地允许“wanna”收缩，即使在“want”和“to”之间存在wh移位的语迹，这与Getz教授（2020）严格操控下得出的实验结果一致。（我很感谢David Lightfoot教授让我知道她在北京的重要成果。）那么，在少数人的心中语法是什么？人们发现，在主语位置附近自由使用允准主语位置的名词性元素，会导致上述变化（详见Endo（2021））。这篇论文已经可以公开获取了，大家都可以免费阅读。
09.
赵欣宇：变异社会语言学的语言观命题是“语言是一个有序异质体”（Weinreich, et al.1968），这意味着语言以语言变体形式存在，但语言系统中的各种变体形式又是有规律的，您怎样看待这一现象？
Yoshio Endo教授：我同意这个观点。我认为Hagit Borer教授的想法与之类似，她认为语言之间的语法变体仅限于功能范畴。Chomsky教授后来也采纳了这一想法。在上文提到的Endo（2019）中，说话人之间的变化也是通过功能范畴来分析的。
10.
赵欣宇：请问空算子和空从属中心语特征之间有什么区别？
Yoshio Endo教授：Endo&Haegeman（2019）认为，在英语状语从句中移动的是空算子，在日语状语从句中移动的是功能中心语。这是因为日语状语从句比英语状语从句有更精细的内部元素分布限制，功能中心语的移动会比XP的移动受到更严格的核心移位限制。那么，汉语又是哪种情况呢？我在课堂上研究了日语和汉语中状语从句的内部结构。虽然我们发现了一些差异，但从整体来说，我们发现离动词近的状语从句（如时体型）内部包含的功能词比离动词远的状语从句（如原因型）少。例如，日语的“while”状语从句中不包含否定词，中文中似乎也不包含。当然，原因状语从句可以包含否定的成分。这与Endo&Haegeman（2019）的预测一致，但细节上的问题值得我们更进一步去研究！
(1) 猫一边没被摸着头，一边安静地待着
就句法制图而言，有趣的是原因状语从句。它有两种类型，比如英语中的“because”和“since”，但日语中可能在一种类型中包含情态元素，而在另一种类型中不包含。那么汉语呢？针对这一点进行研究将会十分有趣。任何的不同都会给我们提供灵感，以使我们能够更好地绘制汉语原因状语从句句法图。
最后，在句法制图领域还有待开拓的一点是：Cinque教授经研究得出副词的词序是固定的结论，在汉语中也是这样吗？据我所教的研究生班级所知，语言学家Noda认为对日语中副词排列的限制似乎也适用于汉语。例如，表示言语行为的副词出现在真假判断的副词之前，如下所示。
(2)a.简单来说[speech-act]，花子可能[adverbs of true or false judgments]要离职了
b.? 花子可能[adverbs of true or false judgments]，简单来说[speech-act]，要离职了
对这些要点的详细研究表明，汉语中也有可能画出一个详细的副词句法层级图。
参考文献
English Version
06.
Xinyu Zhao: The fusion of variationist sociolinguistics and cognitive science has produced many interdisciplinary disciplines, could you illustrate the difference between cognitive sociolinguistics and sociolinguistic cognition?
Prof. Yoshio Endo: I am not familiar with the term “sociolinguistic cognition”, which does not seem to be in common use. The only study that might be relevant is the study by Labov. Labov appealed to cognition in two different ways to explain sound change patterns, where he argued that chain shifts and mergers depend on cognition of phonological categories, and he presented evidence that speakers have knowledge (‘social cognition’, in his words) of which groups use particular variants.
In contrast, the term “cognitive sociolinguistics” is one I hear a lot these days. Cognitive sociolinguistics is a new field of linguistics that aims to account for linguistic variation in social settings with a cognitive explanatory framework. The goal of cognitive sociolinguists is to build a mental model of society, individuals, institutions and their relations to one another. Recent developments in sociolinguistics are dizzying, and there is a wide range of approaches, even within the syntactic field. The background to this is detailed in the introduction to the recently published book in Christensen (2022).
In the field of cognitive sociolinguistics, the cognitive linguistics framework seems to be often used. Cognitive linguistics is very useful for describing aspects of meaning. However, the interface between language and society can also be captured in the framework of the cartography of syntactic structures, and some phenomena can be seen better in cartography. Let's look at some specific examples. The Japanese has a female-specific final particle wa, as described by Endo (2021a), which is used in Tokyo and other parts of Japan, has a rising intonation and its use reached its peak in the early Showa period (1926~1988), gradually decreased and is now rarely used by the younger generation. For example, in the manga Sazae-san, which began in 1946 and ended in 1974, my study of the frequency of the final particle used in the series showed that the use of the female-only wa was more frequent in the early years of the comic and its use subsequently diminishes, as shown below:
Why is this final particle no longer used? One reason for this is due to the social context in which the movement for gender equality in Japan began to flourish around 1940, some years after the USA. These days, gender equality is already deeply rooted in Japan, so when I ask undergraduate students if they use the final particle wa with rising intonation, they say they don't. Interestingly, however, girls start using the final particle wa from around one year of age, but gradually stop using it when they start primary school. Why is this so? The main reason for this lies in the culture of television in Japan. Anime is widely broadcast on Japanese television, from which children often learn the language they use. So, what is it about the language and features of Japanese anime that encourage young girls to use the now-defunct final particle wa? The reason lies in the peculiarities of the Japanese language. The Japanese people change the character depending on the situation we are currently in and when the Japanese changes the character, we change our language, for example, depending on whether we are a clerk or a customer in a convenience store, for example. They also change their character depending on whether they are talking to family or friends, and use the language appropriate to that character. Surprisingly, there are more than ten different first-person English I's in Japanese, depending on the character. An important part of how the character changes is seen in the final particle. Some of the popular characters that young children see in TV cartoons are typical girl characters. And because that character uses the final particle wa, which is no longer used in the context of non-animated everyday life, small children, especially girls, also use wa. But as they grow up to find that people around them don't use the final particle wa, they stop using it themselves.
Next, let's look at what this final particle contributes to the study of the cartography of syntactic structures. In fact, the final particle wa was historically a topic particle, which was then converted into a gerundive adverbial clause particle wa and finally into the final particle wa. In cartography, this change can be expressed, as follows.
[Speech-actP wa [ModfierP wa [TopicP wa…
<--------------/ <-----------/
Here it is shown that in the first stage wa is in the head of TopicP in the CP region, then it is diverted to ModP, which is located higher up, and finally to Speech-actP, which is located at the top level. In the final stage, the final particle wa occupies the highest position in the clause structure, so when the final particle changed from below to above in this way, its fate can only be to disappear. As a result, the final particle wa is no longer used in modern Japanese. In fact, the disappearance of speech-act final particles over time is a common pattern in Japanese. As the study in Fuhzen Si shows that Chinese final particles seem to have been formed from verbs, here, too, they seem to have undergone a change from bottom to top. In other words, the content of what is used for the transition of the final particle may differ from language to language, but the pattern of change from bottom to top perhaps towards Speech-actP seems to be the same.
07.
Xinyu Zhao: Haegeman (1991 and later work) distinguish two kinds of adverbial clause (central adverbial clause and peripheral adverbial clause). However, some English conjunctions may serve to introduce both types of clauses such as as, since, if, etc(Endo & Haegeman, 2019). Take if for example, in your opinion, what’s the influence of the distinction between if1 (introducing central adverbial clause) and if2 (introducing peripheral adverbial clause) on the cartographic study of if?
Prof. Yoshio Endo: Let me give you first the background on the study of adverbial clauses by Haegeman. In early studies of adverbial clauses by Haegeman, the core adverbial clause has a small structure where elements are projected only up to TP/IP. On the other hand, peripheral adverbial clauses have a large structure where elements are projected up to CP. Haegeman captured this contrast by truncation, i.e. central adverbial clauses undergo truncation at IP and thus no higher element can occur there. In contrast, peripheral adverbial clauses do not undergo truncation at all, and thus CP elements can freely occur inside. There is a conceptual problem in this truncation analysis. The problem is that it is unclear why projection stops at a certain syntactic position like IP, and it is arbitrary. To solve this problem, Haegeman pursued another analysis on core adverbial clauses. Based on the fact that an empty operator moves in adverbial clauses (Bhatt, Larson, etc.), she assumes that an operator moves from inside to CP in the core adverbial clause. Then, core adverbial clauses cannot have CP elements because empty operator movement is blocked by CP element due to the general principle of Rizzi’s relativized minimality.
Unfortunately, there is a problem even with this movement analysis. The problem is that the core adverbial clauses are associated with a low IP element of the matrix clause, while peripheral adverbial clauses are associated with a high CP element of the matrix clause (Frey 2011). This fact can be easily expressed in a truncation analysis. Since the size of the core adverbial clause is small and that of peripheral adverbial clause is large, we can say that a core adverbial clause of the small size is associated with a low IP position in the main sentence, whereas a peripheral adverbial clause with a large size is associated with a high position in the main sentence. However, in the movement analysis, because both the core and peripheral adverbial clauses have the same size in its internal structure and both types of adverbial clause have the same CP structure, the only difference being that the operator is moved from IP only in the core adverbial clauses, we cannot capture the fact that the core adverbial clauses are associated with low IP elements of the matrix sentence, while peripheral adverbial clauses are associated with a high CP element of the matrix clause. So, what we need is the mechanism telling us that when an operator moves from a low IP position within an adverbial clause, it is connected to a low IP element in the matrix clause. Based on data from Japanese, Endo and Haegeman (2019) propose such a mechanism by using what Chomsky calls “label”. To be more specific, Endo and Haegeman claim that the label of an entire adverbial clauses is determined by a moving element, e.g., when an adverbial clause is aspect type like while movement takes place from a low AspectP, and this moving element serves as the label of the adverbial clause, and this label of aspect of the adverbial clause is connected with the same low aspect type element in the matrix clause. (Actually, some of this idea originates in Endo’s poster presentation at the international conference held at BLCU).
With this in mind, let us turn to your question. You asked about two types of if clauses. For core if, an element (operator in English, head in Japanese) moves from the part of the low functional category of mood. As a result, the label of the entire adverbial clause becomes a mood type and the adverbial clause is linked to the mood element of the main clause. On the other hand, in peripheral if, movement occurs from a higher position, so the adverbial clause is linked to a higher element of the main clause. As a result, the label of the entire adverbial clause becomes another higher element type and the adverbial clause is linked to a higher element of the main clause (Incidentally, a similar labeling algorithm is also proposed by Donati (2006), who claims that in free relative clauses, the moving wh-element of the nominal type determines the label of the entire relative clause, i.e. nominal status of free relatives).
08.
Xinyu Zhao: In your opinion, why is it important to pay attention to language variation? What can we know from different variations?
Prof. Yoshio Endo: One of the advantages of emphasizing variation is that the theory can be refined to achieve a higher level of precision that makes more accurate predictions. I believe that a theory that produces only core usages without variation will not accurately reflect the reality of the language as it is actually used. Of course, there is the idea of focusing on the core of the language to get to the essence of the language. What aspects to focus on will depend on the background and preferences of each researcher, in my opinion. Personally, I prefer the idea of cartography, which values the reality of the language, accepting variations as well. I will give you one example. Last year I supervised the master's thesis of a student interested in sociolinguistics. He analyzed identical commercials and advertisements aired in Japan, Europe and the US, where he noticed that imperative sentences used in its commercials and advertisements were expressed in Japanese as commands, prayers/optative sentence, solicitations, etc., each expressed differently in a hierarchical structure of function words at the end of sentences in Japanese. Here, variations on the Japanese imperative sentences with a hierarchical structure of function words are not visible in English. In such cases, English can be considered a silent language in the sense that various functional categories of mood involved in the interpretation of the various Japanese imperatives are not pronounced. By assuming silent functional heads in English, the meaning of English imperatives can be more accurately expressed. This approach is pursued in Endo (2022) using other examples. This paper has been made open access, so anyone can read it freely.
Now, a grammatical variation that does not involve social factors is also seen in the complementizer that. In a survey with Andrew Radford, I found that there is inter-speaker variation, such as those who use that after how come, those who do not use it, and those who can use it if an adverbial element intervenes between how come and that. The results are discussed in detail in Endo (2018). Also, in my recent study, I was in email correspondence with Andrew Radford and found that he could make "wanna" contraction even if there was a trace of wh-movement between want and to. In addition, he does not have that-trace effect either. There are about 10% of people with this kind of grammar, as far as I have been able to find out. And I also found that there are a significant number of small children who are free of wanna contraction even if there was a trace of wh-movement between want and to, as Getz (2020) found out in her carefully controlled experiments. (I am grateful to David Lightfoot for letting me know her important work in Beijing) So what is the grammar in the minds of such a minority? There, it was found that the liberal use of the nominal elements near the subject position that licenses the position of the subject can result in the variations described above. The details are given in Endo (2021). This paper has also been made open access, so anyone can read it freely.
09.
Xinyu Zhao: The language view of variationist sociolinguistics is “language is of orderly heterogeneity” (Weinreich, et al.1968), which means language exists in the form of linguistic variation, but the various variation forms in the language system follow orderly rules. What do you think of that?
Prof. Yoshio Endo: I agree with this idea. I think the ideas put forward by Hagit Borer are relevant here. She said that grammatical variation between languages is limited to functional categories. This idea was later adopted by Chomsky. In Endo (2019), mentioned above, inter-speaker variation is also analyzed by functional categories.
10.
Xinyu Zhao: What is the difference between an empty operator and an empty subordinate head feature?
Prof. Yoshio Endo: Endo and Haegeman (2019) considered that it is the empty operator that moves in the English adverbial clause and the functional head that moves in the Japanese adverbial clause. This is because Japanese adverbial clauses have more detailed internal element distribution restrictions than English adverbial clauses and the movement of the functional head rather than XP is subject to the stricter restriction of head movement constraint. Then, which is the Chinese language? I examined the internal structure of adverbial clauses in Japanese and Chinese together in class. Then although we found that there are some differences, in the big picture we found that adverbial clauses that are close to the verb (e.g., aspect type) contain fewer functional words within them than adverbial clauses that are farther away from the verb (e.g., reason type). For example, Japanese does not include negation in the “while” adverbial clause, and it does not seem to be included in Chinese either. Of course, the adverbial clause of reason may include a negative element. This is as Endo and Haegeman (2019) predict, but the details are worth pursuing in more detail!
(1)猫一边没被摸着头，一边安静地待着
Interesting in terms of cartography is the adverbial clause of reason. There are two types of these, such as because and since in English, but Japanese may include an element of mood in one type, but not in the other. What about Chinese? It would be interesting to study this point. If there are any differences, this gives us the possibility to draw a detailed map of adverbial clauses of reason in Chinese.
Finally, a point that has yet to be pioneered in the field of cartography: Cinque has worked out in detail that the word order of adverbs is fixed, but what about Chinese? As far as my graduate class has been able to determine, the restrictions on adverb arrangement in Japanese by a linguist named Noda seem to apply to Chinese as well. For example, adverbs of speech acts occur before adverbs of true or false judgments, as follows:
(2)a.简单来说[speech-act]，花子可能[adverbs of true or false judgments]要离职了
b.? 花子可能[adverbs of true or false judgments],简单来说[speech-act]，要离职了
A detailed study of these points suggests the possibility of drawing a detailed adverb map in Chinese as well.
往期推荐
本文版权归“理论语言学五道口站”所有，转载请联系本平台。
编辑：雷晨 赵欣宇 聂简荻 郭思源 丁子意
排版：雷晨 赵欣宇
审校：闫玉萌 田英慧 李芳芳
英文编审责任人：赵欣宇