人物专栏 | Andrea Moro 教授访谈（上）
点击上方蓝字关注我们
编者按
《理论语言学五道口站》（2021年第10期，总第144期）“人物专栏”与大家分享本站采编人员陈金玉对Andrea Moro教授进行采访的访谈录。Andrea Moro教授，意大利帕多瓦大学语言学博士，北京语言大学语言学系国际教授委员会成员，曾任美国麻省理工学院“语言学与哲学系”、哈佛大学“语言学系”客座研究员，意大利博洛尼亚大学语言学史副教授，意大利米兰圣拉斐尔大学普通语言学全职教授。
本期访谈中，Andrea Moro 教授首先就神经语言学与生物语言学之间的异同提出了深刻而独到的见解，随后从进化角度解释为什么人类具有语言能力，最后分享了对大脑语言处理以及生物学与语言学的相互关系的看法，带给我们许多启发。
后续采访内容将在人物专栏下一期的推送中继续与大家分享，敬请期待。
人物简介
Andrea Moro 教授
Andrea Moro 教授，1962年生于意大利帕维亚，毕业于帕维亚大学古典文学专业，获得帕多瓦大学语言学博士学位，并于1993年获得日内瓦大学“句法学与比较句法学理论”的“研究生学位（DES）”。Andrea Moro教授的主要研究领域有形式句法、句法语义关系和神经语言学。他的主要贡献包括倒装判断句的发现、动态反对称原则、“可能语言”与“不可能语言”神经生物学相关性的差异、否定对前运动皮层活动的影响以及内在言语的电生理表征。曾在麻省理工学院“语言学与哲学系”、哈佛大学“语言学系”担任客座研究员。曾任博洛尼亚大学语言学史副教授，米兰圣拉斐尔大学普通语言学全职教授，于1993年参与了米兰圣拉斐尔科学研究所认知科学系的筹备建设工作。同时， 他也是欧洲科学院(Academia Europaea)和罗马教皇文学与艺术学院(Pontifical Academy of Lettere ed Arti)的成员。
Brief Introduction
Andrea Moro (Pavia, 1962) graduated in Classical Literature at the Università di Pavia, and obtained a Ph.D. in Linguistics at Università di Padova and a Diplôme d'études supérieures (DES) in “Théorie de la syntaxe et syntaxe comparative” at the Université de Genève in 1993. His main fields of research includes theoretical syntax, the relation between syntax and semantics and neurolinguistics. He contributes a lot to the discovery of inverse copular sentences, the principle of Dynamic Antisymmetry, the neurobiological correlation of “possible” vs. “impossible languages”, the effect of negation on premotor cortex activity, the electrophysiological representation of inner speech. He has been visiting scientist at the “Department of Linguistics and Philosophy” at MIT and at the “Department of Linguistics” at Harvard University. He has been Associate Professor of History of Linguistics at the University of Bologna, Full Professor of General Linguistics at University San Raffaele in Milan and he contributed to the foundation of the Department of Cognitive Sciences at the Istituto Scientifico H “San Raffaele” in 1993 in Milan. He is a member of the Academia Europaea (London) and the Pontifical Academy of Lettere ed Arti (Vatican City).
访谈内容
01.
陈金玉：您认为生物语言学和神经语言学有什么区别和联系吗？应该如何从生物语言学的角度来解释语言的递归性呢？
Andrea Moro 教授：通常情况下，这两个术语被认为是同义词，但我认为它们应该属于不同的领域。从某种意义上说，定义“神经语言学”更容易，具体指的是语法的规律性与特定的神经活动相关联的领域，它们之间的相互关系构成了神经语言学的研究领域。“生物语言学”这个术语更笼统且带有隐喻性质（至少我个人是这样认为的）。一方面，它与人类语言是生物指令的产物这一一般假设相吻合，从这个意义上说，它与生成语言学约70年前所提出的理论完全一致；另一方面，它是将语言学视为与生物探索领域相融的方法，这些典型地反映在包括语言习得、临床研究、进化问题研究等的领域。从这个意义上来说，生物语言学是一种避免形式解的方法：形式解与生物指令的强制性限制条件形成鲜明对照。也就是说必须从生物语言学和神经语言学两个角度研究人类语言句法的根本属性——递归性，要获得递归性的神经关联，首先必须要有一个合理的生物学形式模型。实际上，正如Lenneberg所说，递归性的生物性质，相对于文化性和任意性的规约，有可能是句法获得的第一个神经语言学方面的证据。这是通过测试大脑对非递归性（即不可能语言）的反应，证明大脑并不认为那些基于线性顺序的规则是语言学指令。这也可以通过观察学习一种不可能语言来证明，这种语言并不涉及语言计算中所涉及的典型神经元网络。如今，研究正从我们称之为“where-problem”的证据转向“what-problem”的证据，前者旨在告诉我们大脑中某个被激活的区域，后者目的是解码神经元交换的电生理信息来详细阐述语言信息。这是一个更加雄心勃勃的计划，而我们目前只迈出了第一步。识别电生理信息方面的递归可以说是未来研究的基本步骤之一。（可参阅Moro, A. 2015. The Boundaries of Babel II edition, MIT press，Moro, A. 2017. Impossible Languages, MIT Press.）
02.
陈金玉：生成语言学认为人类语言具有生物学特征，并且生物语言学的创立标志着语言学研究重心向生物学特征偏移。从生物学特性的角度来看，在漫长的进化过程中，为什么只有人类才具备语言能力？
Andrea Moro 教授：对于这一重要问题，我认为还是像我第一个问题说的那样，首先得区分神经语言学与生物语言学，因为形式语言学与生物学范式在方法论方面相去甚远，比如说形式语言学几乎不采用定量分析。其次，我们必须理解“为什么”这个词的两种含义：一是何种生理机制让人类有说话的可能，二是何种原因导致在进化过程中只有人类这一“个体”学会了说话。对于上述两问，我想我们都没有确切答案，但是这两问的确有不同之处。实际上存在一些关于是什么导致了人类大脑不同的提议，比如说Angela Friederici的著作就有涉及，但对于进化方面的原因，似乎在现研究阶段，我们第一步应该做的是尝试排除不合理的答案。例如，对于我来说，很明显，人类语言并非被“设计”用来加强交流，就像手并非被“设计”用来写字一样。此外，有人试图将控制单词序列（语法）的规则缩减为与控制镜像神经元链接的动作序列的相同规则：这种相似性只能是隐喻性的，从某种意义上说，这是比较好的情况，否则我们应该解释为什么对于其他能计划一系列行动的动物来说，它们的“语言”中没有语法。
03.
陈金玉：大脑中的语言加工在不同语言层次上进行，比如，音位、形态和句法等层次。模块主义认为，在语言加工中，这些语言层次是相互独立的，并按照一定的顺序进行。然而，互动主义认为在不同的加工层次之间存在重叠的部分。您认同哪一观点？
Andrea Moro 教授：这两个观点我都支持。一方面，我们必须意识到语法模块的观点对于语言理解是必要的。比如，句法必须独立于语义和音系，否则我们就无法理解为什么我们能说出诸如“this triangle is a circle”之类的无厘头的话或诸如“I saw a woman on a car”之类的模棱两可的话。这些现象在中世纪时期的西方文化中很常见，并不是什么新鲜事，比如把句子的意义和形式分开来考虑也很常见。此外，模块主义对于解释临床数据也是必要的：患者的语言障碍也可能只因在其中一个模块上出了错而造成，如果语法不是以模块化的方式组织的，这将是无法解释的。然而，另一方面互动主义也是必要的。它是一种非常激进的方式。以最近实验得出的结果为例，我有机会使用了所谓的外场手术技术，通过这种技术，接受手术的病人可以在脑皮层暴露的情况下进行特定的测试。这样做是为了检查皮层哪些部分不应该被外科医生割断，但直接从大脑中获取电生理信息也很有用（就是我之前提到的what-problem）。结果表明，当受试者在脑海中想到一个词或一个句子而不说出来时，大脑仍然产生电波，该电波包含与语言表达相关的语音信息。换句话说，就好像我们现在可以提取“思想的声音”，即此种情况下的言语思想。有趣的是，这种类型的音系信息呈现在Broca区，而不是声学区，这一事实表明，互动主义比之前认为的要激进得多。似乎对于任何语言的符号来说，在索绪尔的意义上，能指和所指在大脑中并不能独立存在，它们必须始终结合在一起。（该实验在之前的文章中已经解释过。）
04.
陈金玉：从脑机制和神经机制方面解释语言现象，是语言学与脑科学、神经科学相结合的结果。两者的结合是一种新的尝试。毫无疑问，脑科学的研究有助于我们进一步理解人类语言的生成机制。另一方面，语言学的进步会在哪些方面推动脑科学的发展？
Andrea Moro 教授：当科学家研究除大脑以外的所有其他器官时，比如肝脏或心脏，可以采取一个相对简单的方法，即根据输入的定量测量来获得输出的定量测量。但是对大脑进行研究时情况则完全不同：除非由非量化实体进行调节，否则定量的测量实际上是无关紧要的。以单词和句子之间的区别为例：这种区别不是定量的而是定性的,它建立在非常抽象的概念的基础上，如“真理”和“真值”(关于词汇和句子之间区别的详细说明，见Moro, A. 2010. A brief history of the verb “to be”, MIT press.)。这只是一个简单的例子，事实上，任何一个与语言有关的概念都不是纯粹的定量。这意味着，当涉及到语言时，脑科学同音乐和数学一样，必须依赖语言学。除非这些研究要停留在问题表面或者等待从亚里士多德时代就为人们所熟知的语言学概念被重新发现。这种风险在很多情况下都是真实存在的，尤其是涉及到语法的时候。
参考文献：
[1] Friederici et al. 2018. “Language, mind and brain”, Nature Human Behaviour, vol 1, pp. 713-722.
[2] Moro, A. 2010. A brief history of the verb “to be”, MIT press.
[3] Moro, A. 2000. Dynamic Antisymmetry, MIT Press.
[4] Moro, A. 2017. Impossible Languages, MIT Press.
[5] Moro, A. 1988. Per una teoria unificata delle frasi copulari, Rivista di Grammatica Generativa, 13, pp. 81-110.
[6] Moro, A. 2015. The Boundaries of Babel II edition, MIT press.
[7] Moro, A. 2013. The Equilibrium of Human Syntax. Symmetries in the brain, Leading Linguists Series, Routledge, New York.
[8] Pollock, J.-Y. 1989. “Verb Movement, UG, and the Structure of IP,” Linguistic Inquiry, 20, 365-424.
English Version
01.
Jinyu Chen: What is the difference between biolinguistics and neurolinguistics and what is the connection between them? How to study language recursion from the perspective of biolinguistics?
Prof. Andrea Moro: The two terms are very commonly used as synonyms but I do think that they should properly identify two separate domains. It’s easier to define “neurolinguistics”, in a sense: it is the domain where the regularities pertaining to grammar are correlated with specific neuronal activities. These measures constitute the proper domain of neurolinguistics. The term “biolinguistics” is much more general and, at least to me, sometimes used in a metaphorical sense: on the one hand, it coincides with the general presupposition that human languages are the product of biological instructions and in the sense it simply coincides with the program inaugurated by generative linguistics some 70 years ago; on the other, it is a way to consider linguistics as immersed within the domain of biological quest, including prototypically, language acquisition, clinical studies and evolutionary issues. In this sense, biolinguistics is a way to avoid formal solutions that contrast with the restrictions imposed by the biological instructions. That said, recursion, which is a fundamental property of human syntax can and, in fact, it must be studied in both neurolinguistics and biolinguistics terms: to get the neurological correlates of recursion one must first have a biologically plausible formal model. Actually, the biological nature of recursion, as opposed to a cultural and arbitrary conventions, in Lenneberg’s own words, was possibly the first neurolinguistics piece of evidence to be obtained with respect to syntax. This was done by testing the brain reaction to non-recursive, i.e. impossible languages, and proved that the brain does not consider as linguistics instructions those rules based on linear order. This was done by observing that learning an impossible language does not involve the typical neuronal network which is involved in linguistic computation. Nowadays, research is shifting from this type of evidence, which we may call the “where-problem” as it aims at telling us where in the brain a certain network is activated, to what we may call the “what-problem” which aims at decoding the electrophysiological information the neurons exchange to elaborate linguistic information. This is a way more ambitious program and we are just moving the first steps. Recognising recursion in terms of electrophysiological information is arguably one of the fundamental steps of the near future. [for a general discussion see my The Boundaries of Babel II edition (MIT Press) or an abridged version in Impossible Languages (MIT Press).]
02.
Jinyu Chen: Generative linguistics holds that human language has biological attributes, and the creation of biolinguism marks the transformation of linguistics research to the biological paradigm. From the point of view of biological attributes, in the long evolution process, why were only human beings that acquired language ability?
Prof. Andrea Moro: As for the first part of this important question I think one really has to make a clear distinction between neurolinguistics and biolinguistics in the way I suggested in the answer to the first question because formal linguistics is in a sense methodologically far from the “biological paradigm”, for one thing it is almost never quantitative. As for the second part of the question one must understand that the word “why” in fact has two meanings: “why” in the sense as to what mechanisms make it possible in humans, and “why” as to what reason lead evolution to produce such a singularity. I do not think we have answers for neither questions but there are differences. There are in fact proposal concerning what makes the difference in the architecture of the human brain such as in the works conducted by Angela Friederici but as for the evolutionary reasons it seems that at this stage of the research one should rather try to exclude irrational answers as a preliminary step. For example, it seems clear to me that the structure of human languages have not been “designed” to enhance communication much in the same sense as one cannot think that hands have been “designed” to write. Moreover there have been attempts to reduce the rule that govern sequences of words (grammar) to the same rule that govern sequences of actions sometimes linked to Mirror Neurons: this similarity can only be metaphorical and in a sense this is a much better situation than the opposite since otherwise we should explain why other animals who can plan sequences of actions do not have syntax in their languages.
03.
Jinyu Chen: Language processing in brain is carried out in different language levels, such as phoneme, morphology, syntax etc. Modularism believes that, in language processing, these levels are independent with each other and are carried out in a certain order. However, interactionism suggests that there are overlapping parts between different processing levels. Which opinion do you support and why?
Prof. Andrea Moro: I support both and I think there is no way out. On the one hand, the idea that we must identify modules of grammar is necessary to understand language. For example, syntax must be consider as independent from semantics and phonology, otherwise we could not understand why we can say something nonsensical like “this triangle is a circle” or ambiguous like “I saw a woman on a car”. This is not a novelty: these considerations such as those separating meaning to wellformedness of sentences were also common in the Medieval time in western culture. Moreover, modularism is also necessary to explain clinical data: patients may show language impairment by manifesting errors only with respect to one module and this would not be interpretable if grammar were not organised in a modular fashion. On the other hand, though, interactionism is necessary and it is so in a very radical way if, for example, we take the results of some recent experiment I had the opportunity to make with the so-called away surgery technique. With this technique, patients who undergo surgery may perform specific tests while their brain cortex is exposed: this is done to check what portion of the cortex should not be severed by the surgeon but it is also useful to get electrophysiological information directly from the brain (the what problem I was mentioning before). The result we obtained was that when subject think of a word or a sentence in their mind without uttering it, still the brain produces an electric wave containing phonological information associated to the linguistic expression; in other words, it is as if we can now extract the “sound of thought”, verbal thought in this case. What is interesting is that this type of phonological information is presented in Broca’s area which is not an acoustic area: this fact suggest that interactionism is much more radical than previously thought. As if for any sign of language, in the sense of Saussure, the signifier and the signified do not have autonomous life in the brain: they must always be combined. [The experiment is fully described in the texts indicated before.]
04.
Jinyu Chen: Explaining language phenomena from the aspects of brain mechanism and neural mechanism is the result of combining linguistics with brain science and neural science. The combination of the two is a new attempt. The research on brain science is helpful for people to further understand the generating mechanism of human language. On the other hand, in what ways will the progress of linguistics promote the development of brain science?
Prof. Andrea Moro: When scientists study all other organs but the brain, say the liver or the heart, they can rely on a relatively simple methodological schema: you can get a quantitative measure of the output based on the quantitative measure of the input. When it comes to the brain the situation is radically different: the quantitative measure is practically irrelevant unless mediated by non-quantitative entities. Let me explain. Take the distinction between words and sentences: this distinction is not quantitative but qualitative and it is based on very abstract notions such as “truth” and “truth values” (for a detailed illustration of the distinction between words and sentences see Moro, A. 2010. A brief history of the verb “to be”, MIT press). This is only a simple example but practically every single notion pertaining to language is never purely quantitative. This implies that brain science, when it comes to language, but similarly so for music and mathematics, simply must rely on linguistics unless it decides either to remain stuck at the surface of the problem or to wait until the same notion which are familiar in linguistics ever since Aristotle’s own work, are re-discovered. A risk which is not too far from being real in many cases, especially when syntax is involved.
References:
[1] Friederici et al. 2018. “Language, mind and brain”, Nature Human Behaviour, vol 1, pp. 713-722.
[2] Moro, A. 2010. A brief history of the verb “to be”, MIT press.
[3] Moro, A. 2000. Dynamic Antisymmetry, MIT Press.
[4] Moro, A. 2017. Impossible Languages, MIT Press.
[5] Moro, A. 1988. Per una teoria unificata delle frasi copulari, Rivista di Grammatica Generativa, 13, pp. 81-110.
[6] Moro, A. 2015. The Boundaries of Babel II edition, MIT press.
[7] Moro, A. 2013. The Equilibrium of Human Syntax. Symmetries in the brain, Leading Linguists Series, Routledge, New York.
[8] Pollock, J.-Y. 1989. “Verb Movement, UG, and the Structure of IP,” Linguistic Inquiry , 20, 365-424.
往期推荐
李富强：汉语“有+VP”结构中“有”的句法制图研究
Mark Baker & Julie Legate | Case
乔姆斯基: 关于人性和人类发展
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：马晓彤 闫玉萌 陈金玉 王平
排版：马晓彤 闫玉萌 高洁
审校：王丽媛 李芳芳