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》（2020年第25期，总第89期）“人物专栏”与大家分享本站采编人员訾姝瑶对Andrea Moro教授进行采访的访谈录。
Andrea Moro教授 ，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，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。
本期访谈中，Andrea Moro 教授首先就大脑和句法之间存在的关系提出了深刻而独到的见解，随后与我们分享了他对脑科学和认知科学近期发展状况的评价和看法，最后提出了他对这些领域未来发展的期待。
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。
访谈录
01.
訾姝瑶: 您认为大脑与句法之间的关系是怎样的？
Andrea Moro教授：
这是一个核心问题，不是三言两语就能说清楚的。至少有两个不同的方面需要考虑：
一是句法形式属性的知识和它们在大脑中的实现方式之间的关系；二则涉及一个更宽泛的问题，即人类头脑是否有能力不受任何机械或环境条件所限地表达思想，即所谓的“创造性”。由于我们的理解水平存在巨大差异，这两个问题很难被理解。至于语法的形式属性与大脑之间的关系，我们的研究可以基于至少以下一项重要结论，即人类语言的这种特定格式(这种格式规则不是基于词汇的线性序列， 而是基于递归性产生的层级结构)并不是一种文化或规约的人工产物，而是基于人类基因所呈现出来的大脑神经生物学结构的表达。
此外，由于递归程序会产生许多潜在的无限性结构，而无限性又不能逐步实现(不存在“部分无限”)，因此我们可以推翻“原型句法”的假设，这对进化观有着巨大的影响：从定义上讲，句法一定是突然出现的，它是由一个器官所表达， 尽管这个器官必须在突变的压力下逐渐进化。存在于句法和大脑之间关系的悖论是个谜，却也是一个事实。对于第二个方面，即语言创造力，在我看来，我们对于该问题的认识并没有超越十七世纪伟大的思想家们的直觉，尤其像笛卡尔这样的思想家，我们必须承认，这仍然是一个科学研究无法企及的事实，一个当下的、甚至是永远的谜团。
在设计研究项目时，人们可以通过语言学过去五十年的发展来了解句法与大脑间的关系。自布拉格学派的音系学和生成语法学派的句法学研究开始，人们就在寻求使用尽可能少的术语分解原始元素，并且使基本的组合操作实现个性化。与此同时，研究人员正试图观察大脑是否按照与发现的最小因素以同构的方式进行工作。通过测试大脑对学习递归语法和非递归语法的反应，我们发现，虽然这两种语法都是可学习的，但只有递归语法能激活神经回路，而且这些神经回路通常都是服务于语言的。这一结果揭示了句法中“可能的语言”这一概念在神经生物学上的等同性，但并没有解释其为什么不能扩展到其他领域，比如人类语言的语音和语义方面。
值得注意的是这些研究也证实了另一种来自于传统思维的观念，即所有动物都有语言，许多动物虽然拥有自己的符号，但它们并不能重组有限元素并创造出新的潜在的无限意义。换言之，只有人类才有句法。因此，研究句法才能理解什么因素使大脑成为了人脑。
02.
訾姝瑶：您如何看待近来脑科学和认知科学的发展？
Andrea Moro教授：
时至今日，这无疑是最重要的研究领域之一，但论及人类语言时，我们仍然面临着巨大的挑战。
在二十一世纪的后五十年里，诺姆·乔姆斯基基于语言形式特征的开创性研究，通过阐释若干核心公理，证明了统计学（特别是马尔可夫模型）并不能描述人类句法结构。当今计算机的运算能力是往日所不可比拟的，过去人们认为只要收集大量的数据（所谓的大数据）就能够在纯数据的基础上描述人类句法，日益提高的计算机技术重燃了人们的这一希望。
不幸的是，统计数据无法推翻公理， 大数据可能是有用的，但并不能在这个领域里取代“伽利略式的研究”。我们要认识到直接利用统计数据进行的模拟与实验科学所做出的解释无关。然而，在一个以销售技术工具为导向的世界中，模拟研究不仅更容易获得，而且还能比其他类型的研究更快地响应更多的需求。当然我不是说我们必须放弃技术和模拟，但我们必须意识到理解人类并不同于模仿人类的能力。从哲学的角度来看，这当然是正确的。事实上，纯粹的人类语言统计模型，本质上无法捕捉语言创造力并且会将其降低为机械能力，因此我们不能将人类和机器等同。从其他角度来看亦是如此，如在心理学和临床学中，融合入语法的纯统计模型所遵循的路径远远不能与儿童自发语言习得的真实过程相媲美，因此它无法为语言病理相关的治疗提供任何的帮助。
03.
訾姝瑶: 您对这些领域的未来有怎样的期待？
Andrea Moro教授:
我个人认为这取决于技术-模拟的研究视角能否为理解真实大脑功能的研究提供足够的空间和可信度，以及学术界和普遍的现代文化将多大的空间赋予平行进行但却必不可少的研究领域，旨在将语言的原始元素和基本操作减少至最小。
最后，我们应该将语言研究的目标设定为与最著名的科学符号相媲美的事物，如门捷列夫元素周期表。目前还不能说我们能够研究得出句法的周期表，但是通向这一目标的道路必然要经过对不同语言中不同结构进行比较这一过程，正如句法制图所指出的那样：它可以被选为理解人类心智特点的材料来源。
INTERVIEW
01.
Shuyao Zi: How do you think of the relationship between syntax and brain?
Prof. Andrea Moro:
This is a core question and one which cannot be dismissed shortly. There are at least two different aspects to be considered:
one is the relationship between our knowledge of the formal properties of syntax and the way they are implemented in the brain circuits;
the other pertains to a broader issue, namely the capacity of the human mind to elaborate thoughtsthat are not determined by any mechanical or environmental conditions, the so-called “creativity”. Both issues are very far from being understood but there are substantial differences with respect to our level of comprehension.
As for the relationship between the formal properties of grammars and the brain, we can safely rely on at least one major result, namely that the specific format of human languages (the fact that rules are not based on the linear sequences of words but on the hierarchical structures generated by recursive procedures) is not a cultural nor a conventional artifact but rather the expression of the neurobiological structure of the brain as expressed by the human genome. Moreover, since the recursive procedures generate potentially infinite structures and infinity cannot be realized gradually (there is no such a thing as “partial infinity”) we can eliminate the hypothesis of a protosyntax with enormous consequences for the evolutionary perspective: syntax, by definition, must have appeared all of a sudden, although it is expressed by an organ which must necessarily been evolved gradually under the pressure of mutation. This paradox concerning the relationship between syntax and a brain is a mystery but it is a fact. As for the second aspect, namely language creativity, it seems to me that we are not going any further than the intuitions of the great thinkers of the XVII century, notably Descartes, and we must consider this still an inaccessible fact to scientific research; a mystery for the present and possibly forever.
When it comes to design research projects, the first aspect concerning the relationship between syntax and the brain can only be approached by pursuing the lines of reasoning matured during the last fifty years in linguistics, that is ever since the Prague school in phonology and Generative Grammar in syntax: decomposing the primitive elements in the minimal possible terms and individuating the basic assembling operations. Simultaneously, researchers are trying to see if the brain works in an isomorphic way with respect to the minimal factors discovered. One result was accomplished by testing the brain’s reaction to learning recursive vs. non-recursive grammars: it was proved that while both types of grammars are learnable, only recursive ones involve the activation of neural circuits which are normally devote do language. This result uncovered the neurobiological equivalent of the notion of “possible language” for syntax but there is no reason as to why it cannot be extended to other domains, including phonological and semantics aspects of human language.
Ultimately, notice that these researches also confirm another intuition which is part of the traditional thought, namely that all animals have a language and many animals can generate isolated symbols but they all lack the capacity to produce new potentially infinite meanings by recombining a finite set of elements: in other words, only humans have syntax. Therefore studying syntax is an obligatory step toward the understanding of what makes a brain a human brain.
02.
Shuyao Zi：What do you think about the current development of brain science and cognitive science?
Prof. Andrea Moro:
This field of research is certainly one of the most important nowadays but there is a concrete risk especially when it comes to human language. In the last fifty years of the XXI century, Noam Chomsky’s pioneering research on the formal properties of languages proved by demonstrating several core theorems that statistics (in particular Markovian models) could not capture the structure of human syntax.
The advancements in the capacity of computation of the machine nowadays is incommensurable higher than those of the past years: this enhanced technology revived the hope that collecting a huge amount of data (the so-called big data) is enough to capture human syntax on a purely statistical base. Unfortunately, statistics cannot disproof a theorem and big data may be useful but cannot substitute the “Galilean style of research” in this domain: what is crucial is to realize that simulation performed by a brute exploitation of statistics has nothing to do with explanation performed by experimental science.
Unfortunately, in a world led by the necessity to sell technological tools simulation is not only easier to obtain but it also responds to many more needs than other types of research and in a much faster way. I am not of course saying that we must abandon technology and simulation but we must be aware that understanding humans is different from simulating their capacities. This is certainly true from a philosophical point of view: in fact, a purely statistical model of human language is intrinsically incapable to capture language creativity and reduces it to a mechanical capacity, thus assimilating humans to machines. But this is also true from other points of views, among all psychological and clinical ones: for example, a pure statistical model converging to a grammar follows a path which is not even remotely comparable to the real process of spontaneous language acquisition by children and so it can offer no hints to the cure of language pathologies.
03.
Shuyao Zi: What are your expectations for the future of these areas?
Prof. Andrea Moro:
My personal and limited feelings depend on whether the technological-simulation perspective will leave enough space and credits to the research of the understanding of thereal functioning of the brain and, of course, to how much space the academic world and more generally the modern culture will endow the parallel but essential lines of research aiming at the reduction of the linguistic primitive elements and operations to their minimal entities. Ultimately, we should aim in language studies to something which should be reasonably comparable to the most famous scientific icon, namely Mendeleev period table of elements.
Whether we could reach a periodic table of syntax cannot be said at the moment but the path toward this must necessarily go through the comparison of different linguistic structures in different languages, exactly as indicated in the cartographic project: this, if anything, can be chosen as the source of major reflections toward the understanding of this fingerprint of the human mind.
Andrea Moro教授简介
Prof. Andrea Moro
Andrea Moro 教授，1962 年生于意大利帕维亚。他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）”。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，并于 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。
Andrea Moro 教授的主要研究领域有：
形式句法；
句法语义间的关系和神经语言学。
他在研究中的主要贡献：
逆判断句的发现；
动态反对称理论；
“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。
Andrea Moro (Pavia, 1962) graduated in Classical Literature at the Università di Pavia, obtained a Ph.D. in Linguistics at the 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. 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).
His main fields of research:
theoretical syntax；
the relation between syntax and semantics and neurolinguistics.
His contributions:
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.
本文版权归“理论语言学五道口站”所有，转载请联系本平台。
文字：李蔷薇 刘婷婷 辛雨晴 张晴 訾姝瑶
排版：安镜儒
审校：陈旭 王丽媛