智力的本质，第2集：语言与思维的关系

毫无疑问，无法接触语言会产生毁灭性的影响，但似乎并不意味着你根本无法学习某些复杂的知识。

It is absolutely the case that not having access to language has devastating effects, but it doesn't seem to be the case that you fundamentally cannot learn certain kinds of complex things.

来自圣塔菲研究所，这里是复杂性。

From the Santa Fe Institute, this is Complexity.

我是梅兰妮·米切尔。

I'm Melanie Mitchell.

我是阿巴·埃利·福布。

And I'm Aba Eli Fobu.

想想你现在正在收听的这个播客。

Think about this podcast that you're listening to right now.

你很可能只是通过听我们说话就在学习。

You're hopefully learning by just listening to us talk to you.

你能通过这种本质上是复杂声音的方式吸收新信息，这确实令人惊叹。

And the fact that you can take in new information this way, through what basically comes down to sophisticated vocal sounds, is pretty astonishing.

在上一期节目中，我们讨论了人类学习的主要方式之一是通过身处世界并与之互动。

In our last episode, we talked about how one of the major ways humans learn is by being in the world and interacting with it.

但我们还使用语言来相互分享信息和想法，而无需亲身经历。

But we also use language to share information and ideas with each other without needing first hand experience.

语言是人类文化的支柱。

Language is the backbone of human culture.

很难想象如果没有语言，我们会变成什么样。

It's hard to imagine where we'd be without it.

如果你曾经去过一个你不会说当地语言的国家，你就知道被切断基本交流是多么令人迷失方向。

If you've ever visited a country where you don't speak the language, you know how disorienting it is to be cut off from basic communication.

因此，在今天的节目中，我们将探讨语言在智力中的作用，你们将听到的声音是在不同国家、城市和工作场所远程录制的。

So in today's episode, we're going to look at the role language plays in intelligence, and the voices you'll hear were recorded remotely across different countries, cities, and workspaces.

人类是因为拥有语言才聪明，还是因为我们聪明才拥有语言？

Are humans intelligent because we have language, or do we have language because we're intelligent?

语言和思维如何相互作用？它们中的一种能否独立于另一种而存在？

How do language and thinking interact and can one exist without the other?

第一部分。

Part one.

为什么人类会有语言？

Why do humans have language?

在动物界中，没有其他物种的交流方式能与人类语言相提并论。

Across the animal kingdom, there are no other species that communicate with anything like human language.

这并不是说动物没有以复杂的方式交流，而且很多这种复杂性都未被注意到。

This isn't to say that animals aren't communicating in sophisticated ways, and a lot of that sophistication goes unnoticed.

但人类通过长时间对话和复杂句法进行交谈的方式是完全独特的，这也是我们进化的一部分。

But the way humans talk, with our long conversations and complex syntax, is completely unique, and it's part of how we evolved.

几十年来，关于人类语言的一个主流理论被称为生成语言学或生成语法。

For several decades, a dominant theory of human language was something called generative linguistics or generative grammar.

语言学家诺姆·乔姆斯基让这个观点广为人知，其基本观点如下。

The linguist Noam Chomsky made this idea popular, and it basically goes like this.

所有语言都遵循一种内在的底层规则结构。

There's an inherent underlying structure of rules that all languages follow.

从出生起，我们就天生倾向于语言，而非其他形式的交流。

And from birth, we have a hardwired bias toward language as opposed to other forms of communication.

我们天生就倾向于语言和这些句法规则。

We're biologically predisposed to language and these syntactic rules.

这就是为什么乔姆斯基认为，人类语言是我们物种独有的，并且在不同文化中具有普遍性。

This is why human language is, according to Chomsky, unique to our species and universal across different cultures.

这一理论影响深远，但事实证明它似乎并不正确。

The theory has been incredibly influential, but it turns out it doesn't seem to be right.

我一直不太认同生成语言学，以及乔姆斯基关于普遍语法或先天语法知识必要性的核心论点。

So I've never been a fan of generative linguistics, Chomsky's kind of core arguments about universal grammar or the need for innate grammatical knowledge.

这是加里·卢平。

This is Gary Lupien.

你好，我是加里·卢平，威斯康星大学麦迪逊分校的心理学教授。

Hi, I'm Gary Lupien, professor of psychology at University of Wisconsin Madison.

我是一名认知科学家。

I'm a cognitive scientist.

我研究语言的演化，以及语言对认知和感知的影响。

I study the evolution of language, the effects of language on cognition, on perception.

在过去几年里，为了理解像大语言模型这样的东西，我和很多人一样。

And, over the last few years, to make sense of large language models like lots of other people.

近年来，大语言模型的发展加强了加里对生成语法的反感。

In recent years, the development of large language models has bolstered Gary's dislike of generative grammar.

过去人们认为，要想很好地使用语言，就必须天生具备这些语言规则的知识。

The old thinking was that in order to use language well, you needed to be biologically wired to know these language rules from the start.

但大语言模型并没有内置任何语法规则，却能生成极其连贯的文本。

But LLMs aren't programmed with any grammatical rules baked into them, And yet they spit out incredibly coherent writing.

甚至在这些大语言模型出现之前，就已经有很多观点反对这种看法。

And so even before these large language models, there were plenty of arguments against that view.

我认为这些就是最后一根钉子，彻底钉死了那种观点。

I think these are the last nails in the coffin.

所以我认为，这些模型能够生成语法复杂、甚至可以说语义连贯的语言，即使按照现代标准，它们的训练量也并不算巨大。

So I think producing grammatically sophisticated, even, you know, I'd argue semantically coherent language, these models can do all that, even without, you know, by modern standards, huge amounts of training.

这表明，原则上，我们根本不需要任何这种与生俱来的语法规则知识。

It shows that, in principle, one does not need any of this type of innate grammatical knowledge.

那么这里到底发生了什么？

So what's going on here?

史蒂夫·皮安塔多西是加州大学伯克利分校的心理学和神经科学教授，研究儿童如何学习语言和数学。

Steve Piantadosi is a psychology and neuroscience professor at UC Berkeley, studying how children learn language and math.

他说，语言确实有规则，但这些规则是涌现出来的。

He says that language does have rules, but those rules are emergent.

它们并不是一开始就存在的。

They're not there from the start.

是的。

Yeah.

所以我认为关键区别在于，乔姆斯基以及主流语言学倾向于在高度抽象的层面上陈述他们的理论。

So I I think the the key difference is that Chomsky and maybe mainstream linguistics tends to state its theories already at the high level of abstraction.

因此，他们会说：这些是我认为系统所遵循的规则。

So they say, here are the rules that I think the system is following.

而在构建大型语言模型时，你并不会告诉它关于语言运作的高层规则。

Whereas in a large language model, when you go to build one, you don't tell it the high level rules about how language works.

对吧？

Right?

你告诉它的是关于如何学习和构建其内部结构的底层规则。

You tell it the low level rules about how to learn and how to construct its own internal configurations.

你告诉它，应该以能够良好预测语言的方式来做到这一点。

You tell it that it should do that in a way that predicts language well.

当你这样做时，它会以某种方式自我配置。

And when you do that, it kind of configures itself in some way.

什么是高层规则的一个例子？

What's an example of a high level rule?

是的。

Yeah.

例如，英语中的一个高层规则是：如果你有一个句子，可以用‘that’这个词把它嵌入到另一个句子中。

So for example, a high level rule in English is if you have a sentence, you can put it inside of another sentence with the word that.

所以我可以说：我今天喝了咖啡。

So I could say, I drank coffee today.

这是一个完整的句子。

That's a whole sentence.

我可以说，约翰相信我今天喝了咖啡。

And I could say, John believed that I drank coffee today.

对吧？

Right?

因为这条规则是关于如何用一个句子构造另一个句子的，所以你可以重复使用它。

And because that that rule is about how to make a sentence out of another sentence, you can actually do it again.

所以我说，玛丽怀疑约翰相信我今天喝了咖啡。

So I can say, you know, Mary doubted that John believed that I drank coffee today.

因此，如果你要坐下来写一本英语语法书，试图描述哪些英语句子是合乎语法的，哪些是不合语法的，你就必须有一个这样的规则。

And so if you were gonna sit down and and write a grammar of English, if you're gonna try to describe what the grammatical and ungrammatical sentences of English were, you'd have to have some kind of rule that said that.

对吧？

Right?

因为任何英语母语者都会告诉你，是的，约翰说‘我今天喝了咖啡’是一个可以接受的英语句子，而‘我今天喝了咖啡’也是一个可以接受的英语句子。

Because any English speaker you ask is gonna tell you that, yeah, you know, John said that I drank coffee today is a is an acceptable English sentence, and also I drank coffee today is an acceptable English sentence.

大型语言模型在构建时，并不知道类似这样的规则。

Large language models, when they're built, they don't know anything like that rule.

对吧？

Right?

它们只是由一堆参数、权重和连接组成的混乱体，必须通过接触足够的英语来发现这条规则。

They're just a mess of kind of parameters and and weights and connections and they have to be exposed to enough English in order to figure out that rule.

我敢肯定ChatGPT知道这条规则，因为它能构造出那种包含嵌套句子的句子。

And I'm pretty sure ChatGPT knows that rule, right, because it can form sentences like that that have an embedded sentence in that way.

所以当你创建ChatGPT时，并不会从一开始就告诉它这条规则。

So when you make ChatGPT, you don't tell it that rule from the start.

它必须自己构建并发现这条规则。

It has to construct it and discover it.

我觉得有趣的是，构建像ChatGPT这样能发现这条规则的系统，并不否定这条规则在英语使用者头脑中的存在。

And I think what's kind of interesting, right, is that building a system like ChatGPT that can discover that rule doesn't negate the existence of that rule in English speakers' minds.

对吧？

Right?

所以，在ChatGPT内部的某个地方，必须存在对这条规则或类似规则的某种体现。

So, like, internally in ChatGPT somewhere, there has to be some kind of realization of that rule or or something like it.

因此，我认为其他这些关于语言的理论或基本观察的希望在于，它们最终会在这些模型所形成的内部配置中以某种方式体现出来。

And so the hope for these other theories I think or at least these other kind of basic observations about language is that they will be realized in some way inside the the internal configurations that these models arrive at.

我认为事情没那么简单，因为大型语言模型的表现远超我们的理论。

Think it's not quite that simple because the large language models are much better than our theories.

所以我们目前没有任何基于规则的解释能接近它们所能实现的能力，但它们必须具备类似的东西，因为它们确实表现出这种行为。

So we don't have any kind of rule based account of anything that comes close to what they can do, but they have to have something like that because they exhibit that behavior.

我们应该说明，我们所讨论的这些规则与你在学校学到的所谓规则并不相同，比如老师告诉你如何使用介词或不要拆分不定式。

And we should say these rules we're talking about are not the same as the quote, unquote rules you learn in school, like when your teacher tells you how to use prepositions or don't split an infinitive.

是的。

Yeah.

抱歉。

So sorry.

让我澄清一下。

Let let me just clarify.

在语言学或认知科学中，当人们谈论这类规则时，并不是指像‘不要拆分不定式’这样的规则。

In linguistics or in in cognitive science, when people talk about rules like this, they don't mean the rules like don't split infinitives.

基本上，你从英语老师那里听到的那些规则，在认知科学和语言学中完全应该忽略。

Like, basically, you heard from an English teacher, you should just completely ignore in cognitive science and linguistics.

这些规则纯粹是人为编造的。

It's it's just made up.

我的意思是，这些规则常常就是为了强化阶级差异之类的目的而凭空捏造出来的。

I mean, it it's literally made up often just to reinforce class distinctions and things.

语言学和认知科学所关注的规则是描述性的，也就是描述人们实际如何说话的规则。

The kinds of rules that linguistics and cognitive science are interested in are are ones which are descriptive, right, that talk about how people actually do speak.

人们确实会拆分不定式，确实会以介词结尾句子，而且几乎你从英语老师那里听过的任何规则，他们之所以要告诉你，是因为这些规则都违背了你自然的说话方式。

And people do split infinitives, right, and they do end sentences with prepositions and, you know, pretty much, like, any rule you've ever heard from an English teacher, they had to tell you because it's going against how you naturally speak.

所以我认为，这只不过是一种奇怪的阶级现象。

So that's just some weird class thing, I think, that's that that that's going on.

而我们感兴趣的是描述性规则，也就是这个系统在自然界中实际运作的方式。

And what we're interested in are the kind of descriptive rules of, you know, how the system is kind of actually functioning in in nature.

在这种情况下，大多数人甚至根本不知道这些规则。

And in that case, most people are just not even aware of the rules.

向所有英语老师致歉。

Apologies to all the English teachers out there.

但总结一下，语言确实有与生俱来的规则，就像史蒂夫所描述的那条规则。

But to recap, language does have innate rules, like the that rule that Steve described.

但我们并不是生来就带着这些规则，它们早已被硬编码在我们的大脑中。

But we're not born with these rules already hardwired into our brains.

语言学家迄今为止所记录的规则，还不足以完整和精确地反映实际存在的规则。

And the rules that linguists have documented so far aren't as complete and precise as the actual rules that exist.

那些ChatGPT在训练过程中可能已经发现并编码的统计模式。

The statistical patterns that CHAT GPT has probably figured out and encoded at some point during its training period.

然而，这些都无法解释为什么人类使用复杂的语言，而其他动物却没有。

Yet none of this explains why we humans are using complex language, but other animals aren't.

我问了加里，他对这个问题有什么看法。

I asked Gary what he thought about this.

关于语言在智力中所起的作用，存在着很多争论。

So there's a lot of debate about the role language plays in intelligence.

你知道吗？语言是人类在某些认知能力上超越其他动物的原因，还是结果？

You know, is language a cause of or a result of humans superiority over other animals in certain kinds of cognitive capacities?

我认为语言是人类智力如此发展的主要原因之一。

I think language is one of the major reasons why human intelligence is what it is.

所以，语言更像是原因，而不是结果。

So more the cause than the result.

在我们的进化历程中，显然有什么东西让我们天生倾向于语言。

There is something obviously in our lineage that makes us predisposed to language.

我认为，这种倾向更多地源于分享信息、社交的驱动力，而不是与语言或语法本身直接相关的东西。

I happen to think that what that is has much more to do with the kind of drive to share information, to socialize than anything language specific or grammar specific.

你在婴儿身上就能看到这一点：婴儿渴望互动，渴望分享信息，而不仅仅是把语言当作工具使用。

And you see that in infants, infants want to engage, they want to share information, not just use language in an instrumental way.

因此，它让我们获得了原本无法获得的信息。

So it gives us access to information that we otherwise wouldn't have access to.

而语言是一种极其强大的协作工具。

And then it's a hugely powerful tool for collaboration.

你可以制定计划，互相请求帮助，以更有效的方式分配任务。

So you can make plans, you can ask one another to help, you can divide tasks in much more effective ways.

因此，如果没有语言，即使你考虑像人类这样非常社会化的协作物种，一旦拿走语言，也就拿走了创造和传递文化的主要工具。

And so without language, even if you take very social collaborative species like humans, you take away language and you take away the major tool for creating culture and for transmitting culture.

为了跟进一下，黑猩猩和倭黑猩猩是非常社会化的物种，它们群体内部也有大量交流。

Just to follow-up, chimps and bonobos are very social species and have a lot of communication within their groups.

那为什么它们没有发展出你所说的这种语言驱动力呢？

Why didn't they develop this drive you're talking about for language?

为什么是我们发展出了这种能力，而不是它们？

Why did we develop it and not them?

这种能力只对某种特定的物种、某种特定的生态位才有用。

It's only useful to a particular kind of species, a particular type of niche.

所以它的启动成本非常高。

So it has a really big startup cost.

所以孩子们必须学习这些东西。

So kids have to learn this stuff.

在花上多年时间学会语言之前，语言对他们来说几乎是无用的。

Language is kind of useless to them before they put in the years that it takes to learn it.

而且，许多人都写过这一点，语言也非常容易被用来撒谎。

It's also, and many have written on this, language is also very easy to lie with.

所以这是一种不可靠的系统，言语是廉价的。

So it's an unreliable system, words are cheap.

因此，依赖语言只有在社会已经具备基本信任水平的情况下才有意义。

And so reliance on language, sort of only makes sense in a society that already has kind of base level of trust.

所以我认为，理解语言出现的关键在于理解这种亲社会性的出现——语言随后会反过来促进并加速这种特性，但这种特性必须先存在。

And so I think the key to understanding the emergence of language is understanding the emergence of that type of prosociality, that language then feeds back on and helps accelerate, but it needs to be there.

如果你观察其他灵长类社会，会发现它们只在亲属群体内部有合作，没有大规模的合作，常常存在攻击性，也不分享。

And so if you look at other primate societies, there is cooperation within kin groups, there is not broad scale cooperation, there is often aggression, there's not sharing.

所以语言根本就毫无意义。

So language just doesn't make sense.

正如加里所提到的，学习语言需要巨大的初始投入。

As Gary mentioned, there's a huge startup cost for learning language.

人类的童年期比其他物种要长得多。

Humans have much longer childhoods than other species.

从我们出生起，就开始关注所接收到的各种输入中的规律，包括语言输入。

Ever since we're born, we start paying attention to all sorts of regularities in the inputs we get, including in linguistic inputs.

这是埃夫·费德连科。

This is Ev Federenko.

埃夫是麻省理工学院的神经科学家，过去二十年来她一直研究语言。

Ev's a neuroscientist at MIT, and she's been studying language for the past two decades.

她提到，我们从第一天起就开始学习语言。

And she mentioned we start learning language from day one.

这种学习包括内化语言学家曾认为是与生俱来的结构和模式。

That learning includes internalizing the structure and patterns that linguists used to assume were innate.

我们首先关注声音如何组合形成规律性的模式，比如音节以及各种或多或少常见的过渡方式。

We start by paying attention to how sounds may go together to form kind of regular patterns like in, you know, syllables and various, various transitions that are maybe more or less common.

注意这一点。

Pay attention to that.

然后，一旦我们发现输入中的某些部分对应着特定含义，比如我常举的例子：每次妈妈说‘cat’的时候，旁边总会有一只毛茸茸的东西。

Then once we figure out that some parts of that input correspond to meanings, right, like, you know, the example I often say is, like, every time mama says cat, there's this fuzzy thing around.

也许这并不是随机的。

Maybe it's not random.

对吧？

Right?

于是你开始将语言输入的某些部分与现实世界中的事物联系起来。

And you kinda start linking parts of the linguistic input to, parts of the world.

接着，当然，你会学到如何将词语组合起来，以表达更复杂的想法。

And then, of course, you learn what are the rules for how you put words together to express more complex ideas.

所以，所有这些知识似乎都存储在我所谓的语言系统中。

So all of that knowledge seems to be stored in this, what I what I call the language system.

这些表征在我们理解他人所说的话时会被调用，因为我必须进行映射。

And those representations are accessed both when I understand what somebody else is saying to me, because I have to map.

我必须使用这种形式到意义的映射系统来解码你的信息。

I have to use this form to meaning mapping system to decode your messages.

当我脑海中有一些抽象的想法，想用共享的代码——也就是英语——向别人表达时。

And when I have some abstract thing in my mind, an idea, and I'm trying to express it for someone else using the shared code, which in this case is English.

对吧？

Right?

而且，我们通常通过与周围环境互动来学习这种共享代码。

And often, we learn the shared code by interacting with our surroundings.

就像我描述的那样，当你身边有一只猫时，你就学会了关于猫的知识。

Like, as I've described, learning about a cat if there's a cat in the room with you.

但你也可以在没有实际接触猫的情况下了解猫。

But you could also learn about cats without being able to interact with one.

有人可以告诉你关于猫的事，于是你开始形成对这种叫作猫的生物的概念，即使你从未见过它，但你知道它有尖耳朵、毛茸茸的，而且在满足时会发出低沉的呼噜声。

Someone could tell you about a cat and you could start to create an idea for this thing called cat, which you've never seen, but you know that it has pointy ears, it's furry, and it makes a low rumbling sound when it's content.

这就是语言的力量。

That's the power of language.

这是加里又出现了。

Here's Gary again.

我们所学的很多东西都很难量化，很难用数字来表示我们通过与他人交谈或阅读所学到的知识所占的百分比。

So much of what we learn, and it's very difficult to quantify, to put a number on what percent of we know we've learned from talking to others, from reading.

正规教育大多承担了这一角色，对吧？

Most of formal education takes that role, right?

如果没有语言，这显然是不可能的，甚至如果没有书面语言也是如此。

Like it would not be possible in the absence of certainly without language, but even without written language.

如果你接受了足够的语言训练，你就可以直接将语言映射到视觉世界。

If you have enough language training, you can just kind of map onto the visual world.

在我的实验室里，我们做了一些研究，将这一点与之前收集的先天性盲人数据联系起来，他们出人意料地学到了许多关于视觉世界的知识——这些知识本以为只有通过直接经验才能获得，这表明，正常视力的人可能通过直接经验学习，但大量这类信息其实都蕴含在语言的结构中。

And we've done in my lab some work connecting it to previously collected data from people who are born congenitally blind and the various things that they surprisingly learn about the visual world one would think is only learnable through direct experience, showing that, well, normally sighted people might be learning it through direct experience, but a lot of that information is embedded in the structure of language.

当我们通过语言学习时，我们学习的不仅仅是物理对象。

And when we learn through language, we're not just learning about physical objects.

语言还赋予我们命名抽象概念和类别的能力。

Language gives us the ability to name abstract concepts and categories too.

例如，当你思考‘鞋’这个词的含义时，它指的是一种物体，而不是某个特定的东西。

For instance, if you think about what the word shoe means, it refers to a type of object, but not one specific thing.

我们写过一篇关于这个的论文，举了用茄子皮做的鞋子的例子。

We wrote a paper about this and gave the example of shoes that were made out of eggplant skins.

好的。

Okay.

你可以想象一下，把茄子皮晾干，然后把两边缝起来，加上鞋带，再套在脚上，等等。

And, like, you could imagine doing that, like drying out an eggplant skin and kind of sewing up the sides and adding laces and kind of fitting it around your feet and and whatever.

你可能从未见过用茄子做的鞋子，但我们所有人都同意这是可能发生的。

And you've probably never encountered shoes made out of eggplants before, but we all just agreed that that could happen.

对吧？

Right?

你可能会找到这样的鞋子。

That you could find them.

所以这表明，定义这个概念的并不是具体的物理物体。

And so that tells you that it's not the physical object exactly that's defining what the concept means.

对吧？

Right?

因为我刚刚给你介绍了一个新的实物。

Because I just gave you a new physical object.

定义一个事物的，必须是某种更抽象的东西，更多地关乎它的关系和用途。

It has to be something more abstract, more about the relationships and the use of it that defines what the thing is.

我认为，认为语言在某种程度上是特殊的，并不那么疯狂。

I don't think it's so crazy to think that language is special in some way.

当然，我们通过语言习得了很多东西。

There's certainly lots of things that we acquire through language.

对吧？

Right?

我认为，这一点尤其明显，当你和孩子交谈时，他们不断问‘为什么’，而你用语言解释一些抽象、无法直接展示的概念，他们却能对从未接触过的系统形成相当好的理解。

This is, I think, especially salient if you talk to a kid and they're asking why questions and, you know, you explain things that are abstract and that you can't show them just in language, and they can come to pretty good understandings of systems that they've never encountered before.

你知道的，比如他们问云是怎么形成的，或者月亮在做什么，之类的，对吧？

You know, if they ask how clouds form or what the moon is doing or or whatever, right?

所有这些都通过语言系统被我们习得。

All of those are things that we learn about through a linguistic system.

一个更准确的图景可能是：记忆容量出现了微小的连续或定量变化，从而促成了语言的产生。

The right picture might be one where there's a small kind of continuous or quantitative change in memory capacity that enables language.

但一旦拥有了语言，就开启了巨大的学习潜力，使我们能够通过文化传承来学习观念，并从父母和社区其他成员那里学习复杂的事物。

But then once you have language, that opens up this kind of huge learning potential for cultural transmission of ideas and and learning complicated kinds of things from your parents and from other people in your community.

所以，阿巴，我们在本集开始时提出了一个问题：为什么人类拥有语言。

So, Abba, we asked at the beginning of the episode why humans have language.

到目前为止，我们从加里、史蒂夫和埃夫那里听到的观点是，语言很可能是人类渴望社交与合作的结果。

And what we've heard from Gary, Steve, and Ev so far is that language probably emerged as a result of humans' drive to socialize and to collaborate.

这些社交驱动力与语言本身之间存在一种反馈效应。

And there's a feedback effect between these social drives and language itself.

因此，语言是一种极佳的协作工具，而协作推动了我们的智力发展。

So language is an incredible tool for collaboration, and collaboration drives our intelligence.

例如，加里认为，语言是人类智力达到如今水平的主要原因。

Gary, for example, thinks that language is a major cause of human intelligence being what it is.

是的

Right.

对

Right.

史蒂夫还指出，语言开启了一种全新的学习方式和文化演化方式，这很有趣。

It was interesting how Steve also pointed out that language enables a whole new way of learning and of cultural evolution.

语言让我们能够快速从周围的人那里学习新东西，比如我们的父母、朋友以及其他我们互动的人。

Language allows us to quickly learn new things, you know, from the people around us, say our parents, our friends, and other people we interact with.

它还让我们无需亲身经历就能学到东西。

It also lets us learn without having to experience something ourselves.

比如，当我们小时候和父母一起走路时，他们说：别跑到车前面去。

Say, for example, when we are walking with our parent when we were little and they said, you know, don't jump out in front of the car.

我们通常会相信他们，而不需要自己去亲身体验，而这正是语言带来的作用。

We tend to trust them and not have to experience it ourselves, and this is enabled because of language.

对吧？

Right?

是的

Yeah.

我们确实应该更加感激我们的父母。

We should definitely appreciate our parents more.

但另一方面，加里也指出，语言使得撒谎和欺骗他人变得很容易。

But on on the downside, Gary also pointed out that language makes it easy to lie and to trick people.

因此，只有当社会具备基本的信任水平时，仅依赖语言才有意义。

So relying on language only makes sense when society has a basic level of trust.

这太真实了。

That is so true.

我的意思是，如果我们彼此不信任，社会就很难正常运转，但信任的代价也非常高。

I mean, if we don't trust each other, it's hard to function as a society, but trust comes at such a high cost too.

语言的另一个缺点是需要漫长的学习过程，因为我们不可能一夜之间学会一门语言。

And the other downside of language requires a long learning period because we can't learn a language overnight.

我们生来并不是就会说某种语言的。

We're not born necessarily speaking a language.

我们的童年如此漫长，这也是另一个高昂的代价。

Our childhood is so prolonged, and that's another high cost.

是的。

Yeah.

因此，在进化过程中，语言的优势一定超过了这些缺点。

So the advantages of language must have outweighed those downsides in evolution.

对。

Yes.

另一个刚刚提到的有趣观点是，当今的大语言模型表明某些语言学理论是错误的。

Another interesting point that just came up is that today's large language models have shown that certain linguistic theories are just wrong.

史蒂夫声称，大语言模型已经否定了诺姆·乔姆斯基关于大脑中存在先天普遍语法的观点。

Steve claims that LLMs have disproven Noam Chomsky's notion of an innate universal grammar in the brain.

对吧？

Right?

是的。

Yeah.

人们对语言在大脑中如何运作的看法已经发生了巨大变化。

People have really changed their thinking about how language works in the brain.

在第二部分中，我们将探讨脑成像技术能告诉我们关于语言的哪些信息，以及当人们丧失语言能力时会发生什么。

In part two, we'll look at what brain imaging can tell us about language and what happens when people lose their language abilities.

第二部分。

Part two.

语言和思维在大脑中是分开的吗？

Are language and thought separate in the brain?

埃夫的标志性方法之一是使用fMRI脑扫描，观察我们在使用语言时大脑中哪些系统会被激活。

One of Ev's signature methods is using fMRI brain scans to examine which systems in the brain light up when we use language.

她和她的合作者设计了实验，以研究语言与其他形式认知之间的关系。

She and her collaborators have developed experiments to investigate the relationship between language and other forms of cognition.

这很简单。

It's very simple.

我的意思是，我们研究语言与思维关系的实验逻辑基本都是一样的，只是用了不同类型的思维任务。

I mean, the logic where, of the experiments where we've looked at the relationship between language and thought is all pretty much the same, just using different kinds of thought.

但这个想法是，你选取一些个体，让他们进入MRI扫描仪，并让他们完成一项能可靠激活语言区域的任务。

But the idea is you take, individuals, put them in an MRI scanner, and you have them do a task that you know reliably engages your language regions.

例如，可以在大脑被扫描时，让他们阅读和聆听连贯的句子。

This could be, for example, reading and listening to coherent sentences while your brain is being scanned.

然后，将这个映射图与听到一系列随机词语和声音时大脑激活的区域进行比较，这些声音虽然听起来像语言，但完全无意义。

Then that map would be compared to the regions that light up when you hear sequences of random words and sounds that sound speech like but are completely nonsensical.

如果你们来麻省理工学院，我可以给你们做扫描，并打印出你们的语言系统图谱。

And if you guys visit MIT, I can scan you and print your map of your language system.

找到它大约只需要五分钟。

It takes about five minutes to find.

非常可靠。

Very reliable.

而且，如果我今天扫描你，或者十年后再扫描你——我已经对一些人做过这种十年间隔的扫描——它始终出现在完全相同的位置。

And, again, if I scan you today or ten years later, I've done this on some people or ten years apart, It's in exactly the same place.

它在个体内部非常非常稳定。

It's very, very reliable within people.

所以我们找到了这些语言区域。

So we find those language regions.

然后我们基本上会问：好吧，让你进行某种形式的思考。

And then we basically ask, okay, let's have you engage in some form of thinking.

让我们让你解决一些数学题，或者做一些类似模式识别的任务。

Let's maybe have you solve some math problems or do something like some kind of pattern recognition task.

我们基本上想知道：当你处理语言时激活的神经回路，是否与你在进行数学推理时——比如做加法题之类的——活跃的神经回路有重叠？

And we basically ask, do the circuits that light up when you process language overlap with the circuits that are active when you process, for example, when you engage in mathematical reasoning, like doing addition problems or whatnot?

在我们研究过的众多思维领域中，我们几乎一致地发现这一点。

And, we basically very consistently find across many domains of thought pretty much everything we've looked at so far.

我们发现，语言区域几乎完全没有活跃。

We find that the language regions are not really active hardly at all.

而另一个与语言区域不重叠的系统正在全力运转。

And some other system non overlapping with the language regions is working really hard.

因此，我们并不是通过调动语言机制来解决这些其他问题的。

So it's not the case that we engage the language mechanisms to solve these other problems.

我知道关于解读fMRI结果的难度一直存在一些争议。

I know there's been some controversy about how easy it is to interpret the results of fMRI.

你能告诉我们，这很难做到吗？

What can you tell us about, like, is that a hard thing to do?

这很容易做到吗？

Is it an easy thing to do?

我认为解读fMRI数据并没有比其他数据特别大的挑战。

I don't think there's any particular challenges in interpreting fMRI data than any other data.

我的意思是，你希望进行稳健而严谨的研究。

I mean, you want to do robust and rigorous research.

比如，在基于任何发现做出强有力主张之前，你希望确保你的发现确实反映了你所认为的内容。

Like, you want to make sure before you make a strong claim based on whatever findings, you want to make sure that your findings tell you what you think they are.

但这是任何研究都会面临的挑战。

But that's kind of a challenge for any research.

我认为这与你所采用的具体测量方式无关。

I don't think it's related to particular measurements you're taking.

我的意思是，fMRI 确实有局限性，其中之一是我们无法观察信息处理的快速时间尺度。

I mean, there are certainly limitations of fMRI, and one of them is that we can't look at fast timescales of information processing.

我们根本无法获取毫秒、几十毫秒甚至几百毫秒时间尺度上发生的情况，而对某些问题来说，这并不重要。

We just don't have access to what's happening on a millisecond or tens of milliseconds or even hundreds of milliseconds timescale, which for some questions, it doesn't matter.

但对某些问题来说，这确实很重要。

But for some questions, it really does.

因此，对于那些关键的时间尺度问题，fMRI 并不适用。

And so that makes the fMRI not well suited for those questions where it matters.

但总的来说，fMRI 得出的良好、稳健的发现是非常可重复的。

But in general, good robust findings from fMRI are, you know, very robustly replicable.

我实际上被尤夫的论点深深说服了。

I've been actually very convinced by Ev's arguments in particular.

又是史蒂夫·皮安塔多西。

Steve Piantadosi again.

你可以找到一些在某个领域非常专业的人，比如数学专家、国际象棋特级大师之类的，但他们却丧失了语言能力。

You can find people who are experts in some domain, right, like mathematics experts or chess grandmasters or whatever who have lost linguistic abilities.

这是一种非常不错的自然实验，它表明语言能力并不是这些领域推理的底层基础，因为你即使失去了语言能力，仍然可以保留推理能力。

And that is a very nice type of, you know, natural experiment that shows you that the linguistic abilities aren't the kind of substrate for reasoning in those domains because you can lose the linguistic abilities and still have the reasoning abilities.

可能仍存在一个学习过程，比如，没有语言的话，学习下棋或数学可能会非常困难。

There might still be a a learning story, like it would probably be very hard to learn chess, right, or or learn mathematics without having language.

但我认为，一旦你学会了，或者学得足够好成为专家，似乎就会出现某种其他类型的系统或非语言的处理方式。

But I think that once you learn it or learn it well enough to become an expert, it seems like there's some other kind of system or some other kind of processing that happens nonlinguistically.

非语言的。

Nonlinguistically.

这表明，原则上，你可以擅长语言，却缺乏那种似乎构成人类思维特征的序列性多步推理能力。

What it shows you is that you can be, in principle, good at language without having the ability to do the kind of sequential multistep reasoning that seems to characterize human thinking.

我认为这令人惊讶。

And that I think is surprising.

对吧？

Right?

它本不必是这样的。

It didn't have to be like that.

有可能语言是我们用于一切事物的底层基础，或者语言是一个如此困难的问题，以至于只要你解决了语言问题，就必然需要具备人类所拥有的所有底层推理机制。

It could have been that language was the substrate that we use for everything or that language was such a difficult problem that if you solved language, you would necessarily have to have all of the underlying reasoning machinery that people have.

但看起来这并不正确。

But it seems that that's not right.

对吧？

Right?

你可以在没有多少推理能力的情况下，依然在语言方面做不少事情。

That you you can do quite a bit in language without having much reasoning.

另一方面，你也可以在没有语言的情况下进行大量推理。

And on the flip side, you can do a lot of reasoning without language.

正如埃夫之前提到的，她和她的合作者在大脑中发现了语言系统，这些系统在fMRI扫描中非常稳定地出现。

As Ev mentioned before, she and her collaborators have identified language systems in the brain that show up very reliably in fMRI scans.

这些语言系统主要位于左半球。

These language systems are mostly in the left hemisphere.

那么，如果一个人完全失去了这些系统，会发生什么？

So what happens if someone loses these systems completely?

这种fMRI方法很好地补充了对患有严重语言障碍的患者的研究，对吧？

This fMRI approach is very nicely complemented by investigations of patients with severe language problems, right?

另一种方法，比fMRI要早得多，就是研究那些语言系统遭受严重损伤的个体。

So another approach, this one we've had around for much longer than fMRI, is to take individuals who have sustained severe damage to the language system.

有时，左侧大脑半球的中风范围很大，几乎完全摧毁了整个语言系统。

And sometimes, left hemisphere strokes are large, they they pretty much wipe out that whole system.

这些被称为全失语症患者。

So these are so called individuals with global aphasia.

他们无法理解，如果你给他们一句话，他们根本无法从中推断出任何意义。

They can't like, if you give them a sense, they just cannot infer any meaning from this.

我们知道这并不是低级的缺陷，因为你可以证明这种缺陷跨越了多种模态，比如书面语和口语等。

And we know it's not a low level deficit because you can establish that it's across modalities like written and spoken and so on.

因此，似乎他们一生所学习的那套语言与意义之间的映射关系已经丧失，确实被彻底摧毁了。

So it seems like the linguistic representations, that set of four meaning mappings that they've spent their lifetime learning is lost, is is really, you know, destroyed.

然后你可以去研究这些个体的认知能力。

And then you can, ask about the cognitive capacities in these individuals.

他们还能进行复杂的思考吗？

Like, can they still think complex thoughts?

对吧？

Right?

你该怎么测试这一点？

And how do you test this?

嗯，你可以给他们做一些行为任务。

Well, you you give them behavioral tasks.

对于其中一些人，你必须是个非常聪明的实验者，因为你不能再用语言来解释了。

And for some of them, of course, you have to be a very clever experimentalist because you can no longer explain things verbally.

但你知道，人们还是想出了各种方法来传递和接收指令。

But, you know, people come up with ways to give and get instructions across.

他们能理解比个大拇指向上或向下的判断。

They understand kind of thumbs up, thumbs down judgments.

对吧？

Right?

所以你会给他们一些结构良好或结构不良的数学表达式、音乐模式，或者类似的东西。

So you keep them give them, like, well formed or ill formed mathematical expressions or, musical patterns or something like that.

你发现有一些人严重缺乏语言能力。

And what you find is that there are some individuals who are severely linguistically impaired.

比如，语言系统已经消失了，至少根据我们现有的工具所能测试的结果来看是这样。

Like the language system is gone for, you know, as best as we can test it with whatever tools we have.

但他们认知能力仍然正常。

And yet they're okay cognitively.

他们只是失去了那种将内心复杂思维转化为这种共享表征形式的编码能力。

They just lost that code to take the sophistication of their inner minds and translate it into this shared representational format.

许多这样的人会严重抑郁，因为他们被误认为智力低下，对吧？因为我们常常通过一个人的说话方式来评判他们。

And a lot of these individuals are severely depressed because they're taken to be mentally challenged, right, because that's how we often judge people is by the way they talk.

这也是为什么外国人常常以这种方式遭受不公。

That's why foreigners often suffer in this way too.

人们会对他们智力水平等各方面做出判断。

Judgments are made about their intellectual capacities and otherwise and so on.

无论如何，但确实，许多这类个体似乎保留了相当完整的思考能力，这表明至少在成人大脑中，一旦你掌握了这套知识模块，就可以移除语言系统。

Anyway, but, yeah, a lot of these individuals seem to have the ability to think quite preserved, which suggests that at least in the adult brain, you can take that language system out once you've acquired that set of knowledge bits.

对吧？

Right?

你可以移除它，而它似乎并不会影响我们迄今为止测试过的任何思考能力。

You can take it out, and it doesn't seem to affect any of the thinking capacities that we've tested so far.

所以这里有个非常天真的问题。

So here's a extremely naive question.

如果语言和思维在成年人中是分离的，那为什么当我思考时，总觉得我实际上是在用词语和语言思考呢？

So if language and thought are dissociated, at least in adults, why does it feel like when I'm thinking that I'm actually thinking in words and in language?

这是个很好的问题。

It's a great question.

我的意思是，这个问题经常被提出来。

I mean, so that's a question that comes up quite often.

一点也不天真。

Not naive at all.

这是一个非常好且深刻的问题。

It's a very good deep question.

这是一个关于内心声音的问题。

It's a question about the inner voice.

对吧？

Right?

很多人有一种感知，认为他们的脑海中有一个声音在说话。

A lot of people have this percept that there's a voice in their heads talking.

这是一个很好的问题，但作为这个领域，我们目前还没有清晰的答案来解释它的作用以及它依赖的机制。

It's a it's a good question to which I don't think we as a field have very clear answers yet about what it does, what mechanisms it relies on.

我们知道的是，这并不是一种普遍现象，这一点已经暗示了它不可能是复杂思维的关键要素，因为许多声称自己没有这种内心声音的人，比如麻省理工学院的教授，他们会说：你在说什么？

What we do know is that it's not a universal phenomenon, which already kind of tells you that it cannot be a critical ingredient of complex thought because certainly a lot of people who say that they don't have it in a voice, some of them are like MIT professors, and they're like, what are you talking about?

你脑子里有声音吗？

You have a voice in your head?

那可不太好。

That's not good.

你看过医生吗？

Have you seen a doctor?

就是你知道的那种？

Like you know?

这现在是一个非常活跃的研究领域。

And it's a very active area of research right now.

很多人对这个产生了兴趣。

A lot of people got interested in this.

你可能听说过，大约十年前，关于“心盲症”——即一些人无法在脑海中形成视觉图像——也曾引起过广泛关注。

You may have heard, like, about ten years ago, there was a similar splash about aphantasia, this inability of some people to visually image things.

类似地，当你说你脑子里有声音时，有些人根本不知道你在说什么。

So similar, like, how some people don't know what you mean when you say you have an inner voice.

有些人无法在脑海中形成图像。

Some people cannot form mental images.

比如，让你想象你小时候住过的房子，他们却说：什么也想不出来。

Like, say, imagine the house you lived in when you were a child, and they're like, got nothing there.

你知道，这就像是空白一样。

You know, it's like blank.

我就没法形成那种画面，我也说不清楚。

Like, I just can't form that I can describe it.

我知道关于它的事实，但就是无法在脑海中形成那个图像。

I know facts about it, but I can't form that, mental image.

像内心独白、心理图像这类东西，用我们目前可用的方法很难研究。

And these kinds of things like inner voice, mental imagery, those are very hard things to study with the methods that we currently have available.

是的。

Yeah.

我记得曾和一个人交谈，他告诉我他没有内心独白，他只有一种感觉，但无法准确描述那种感觉。

I think I was talking to someone who actually told me they don't have an inner voice, and they actually are left with a feeling, but they can't necessarily describe the feeling.

所以当他有想法时，他不知道该如何用语言表达出来。

And so they don't know how to put it into language when they have a thought.

是的。

Yeah.

而且这确实是个很好的观点，因为我丈夫没有内在语言，他经常用这一点来论证。

Well and that's that's a very good point because my husband who doesn't have an inner voice, he often uses this as an argument.

他说，如果我们用语言思考，为什么有时候解释自己在想什么会这么难？

He's like, if we were thinking in language, why is it sometimes so hard to explain what you think?

比如，你心里明明很清楚地有这个想法，却就是难以表达出来。

Like, you know you have this idea very clearly for yourself, you just have trouble formulating it.

而且，没错，这确实是个好观点。

And, yeah, it's a that's a good point.

但加里对语言和思维之间的关系有着不同的看法。

But Gary sees the relationship between language and thought a bit differently.

他认为两者并不能如此清晰地分开。

He doesn't think that they can be separated so neatly.

我认为伊芙和她的团队正在做出色的工作，我们在很多方面都持相同观点。

I think Ev and her lab are doing fabulous work, and we agree on many things.

但这一点是我们意见不一致的地方。

This is one thing we don't agree on.