视频库 / NO.032
--:--
点击播放，这里会跟随视频显示当前句的中英字幕。
第 32 期 · 回应 Ⅰ·08「什么是「我」？」
The illusion of consciousness | Dan Dennett
节目发布 2007-05-03章节 · 点击跳转视频
0:04
哲学家的困境：人人自认意识专家
2:44
演讲目标：动摇你对内心的自信
3:53
百万亿无意识细胞如何产生意识
5:14
「真正的魔法」悖论与意识
7:06
哲学家如何解释锯人魔术
8:34
演示一：方块变色与视觉贫乏
10:25
贝洛托的桥：大脑接受暗示
13:06
心理旋转：你不知道自己怎么做的
14:22
内克尔立方体与填充争论
16:56
变化盲视实验：闪烁范式
20:01
结论：第三人称科学胜过内省
本期小问 · 档案清单
3:53
什么是「我」？
8:34
我们真的在理解世界吗？
5:14
意识可以被解释成一堆「把戏」吗？
20:01
我们对自己的意识拥有权威吗？
本期讲者
丹尼尔·丹尼特美国哲学家、认知科学家（1942–2024），塔夫茨大学教授，著有《意识的解释》《达尔文的危险思想》等，主张意识可用自然主义方法解释，反对「笛卡尔剧场」式的心智观。
01哲学家的困境：人人自认意识专家
0:04
So I'm going to speak about a problem that I have and that's that I'm a philosopher. (Laughter) When I go to a party and people ask me what do I do and I say, "I'm a professor," their eyes glaze over. When I go to an academic cocktail party and there are all the professors around, they ask me what field I'm in and I say, "philosophy" -- their eyes glaze over. (Laughter) When I go to a philosopher's party (Laughter) and they ask me what I work on and I say, "consciousness," their eyes don't glaze over -- their lips curl into a snarl.
所以我要讲一个我自己的问题，那就是——我是个哲学家。（笑声）我去参加派对，别人问我是做什么的，我说“我是教授”，他们的眼神就开始发直了。我去参加学术界的鸡尾酒会，周围全是教授，他们问我是哪个领域的，我说“哲学”——他们的眼神也开始发直。（笑声）等我去参加哲学家的聚会（笑声），他们问我研究什么，我说“意识”，这下他们的眼神不发直了——他们的嘴角开始撇起来，露出鄙夷的表情。
0:46
(Laughter) And I get hoots of derision and cackles and growls because they think, "That's impossible! You can't explain consciousness." The very chutzpah of somebody thinking that you could explain consciousness is just out of the question. My late, lamented friend Bob Nozick, a fine philosopher, in one of his books, "Philosophical Explanations," is commenting on the ethos of philosophy -- the way philosophers go about their business. And he says, you know, "Philosophers love rational argument." And he says, "It seems as if the ideal argument for most philosophers is you give your audience the premises and then you give them the inferences and the conclusion, and if they don't accept the conclusion, they die.
（笑声）我会收到一片嘘声、怪笑和低吼，因为他们心想：“那不可能！意识是解释不了的。”居然有人觉得自己能解释意识，这种厚脸皮简直是荒唐透顶。我已故的、令人怀念的朋友鲍勃·诺齐克，一位出色的哲学家，在他的一本书《哲学解释》中，评论了哲学这门学科的风气——哲学家们做事的方式。他说，你知道，“哲学家热爱理性论证。”他还说：“对大多数哲学家来说，理想的论证似乎是这样的：你先给听众你的前提，再给他们推理过程和结论，如果他们不接受这个结论，他们就会死。
1:40
Their heads explode." The idea is to have an argument that is so powerful that it knocks out your opponents. But in fact that doesn't change people's minds at all. It's very hard to change people's minds about something like consciousness, and I finally figured out the reason for that. The reason for that is that everybody's an expert on consciousness. We heard the other day that everybody's got a strong opinion about video games. They all have an idea for a video game, even if they're not experts.
他们的脑袋会爆炸。”这个想法就是要有一个强大到能把对手击倒的论证。但事实上，这根本改变不了人们的想法。要改变人们对意识这类问题的看法非常难，而我终于弄明白了原因。原因就是：人人都是意识方面的专家。我们前几天听说，每个人对电子游戏都有很强的看法。他们都有做游戏的点子，哪怕他们不是行家。
2:09
But they don't consider themselves experts on video games; they've just got strong opinions. I'm sure that people here who work on, say, climate change and global warming, or on the future of the Internet, encounter people who have very strong opinions about what's going to happen next. But they probably don't think of these opinions as expertise. They're just strongly held opinions. But with regard to consciousness, people seem to think, each of us seems to think, "I am an expert. Simply by being conscious, I know all about this."
但他们不会认为自己是电子游戏专家；他们只是有很强的看法而已。我相信，在座从事气候变化和全球变暖研究，或者研究互联网未来的人，一定遇到过对接下来会发生什么抱有强烈看法的人。但那些人大概不会把这些看法当成专业知识。那只是他们坚定持有的意见罢了。可一说到意识，人们似乎都会觉得，我们每个人似乎都会觉得：“我就是专家。光凭我有意识这一点，我就对这件事了如指掌。”
02演讲目标：动摇你对内心的自信
2:44
And so, you tell them your theory and they say, "No, no, that's not the way consciousness is! No, you've got it all wrong." And they say this with an amazing confidence. And so what I'm going to try to do today is to shake your confidence. Because I know the feeling -- I can feel it myself. I want to shake your confidence that you know your own innermost minds -- that you are, yourselves, authoritative about your own consciousness. That's the order of the day here. Now, this nice picture shows a thought-balloon, a thought-bubble.
所以，你把你的理论讲给他们听，他们会说：“不不不，意识才不是那样的！不对，你全搞错了。”而且他们说这话时带着一种惊人的自信。所以我今天要做的，就是动摇你们的这份自信。因为我懂那种感觉——我自己也能感受到。我想动摇你们的一个信念：你们并不真的了解自己内心最深处——你们对自己的意识并不具有权威。这就是今天要谈的主题。这张漂亮的图上有一个思想气泡、一个想法气球。
3:17
I think everybody understands what that means. That's supposed to exhibit the stream of consciousness. This is my favorite picture of consciousness that's ever been done. It's a Saul Steinberg of course -- it was a New Yorker cover. And this fellow here is looking at the painting by Braque. That reminds him of the word baroque, barrack, bark, poodle, Suzanne R. -- he's off to the races. There's a wonderful stream of consciousness here and if you follow it along, you learn a lot about this man. What I particularly like about this picture, too, is that Steinberg has rendered the guy in this sort of pointillist style.
我想大家都明白它是什么意思。它是用来表现意识流的。这是我见过的关于意识的最喜欢的一幅画。当然是索尔·斯坦伯格画的——它曾是《纽约客》的封面。画里这个人正在看布拉克的一幅画。这让他想到了baroque（巴洛克）这个词，然后是barrack（兵营）、bark（狗叫）、poodle（贵宾犬）、苏珊·R——他的思绪就这么一路飞奔出去了。这里有一段精彩的意识流，如果你顺着它读下去，就会了解到关于这个人的很多事。我尤其喜欢这幅画的另一点是，斯坦伯格是用点彩派那样的笔法画出这个人的。
03百万亿无意识细胞如何产生意识
3:53
Which reminds us, as Rod Brooks was saying yesterday: what we are, what each of us is -- what you are, what I am -- is approximately 100 trillion little cellular robots. That's what we're made of. No other ingredients at all. We're just made of cells, about 100 trillion of them. Not a single one of those cells is conscious; not a single one of those cells knows who you are, or cares. Somehow, we have to explain how when you put together teams, armies, battalions of hundreds of millions of little robotic unconscious cells -- not so different really from a bacterium, each one of them -- the result is this. I mean, just look at it.
这提醒我们，就像罗德·布鲁克斯昨天说的：我们是什么，我们每个人是什么——你是什么，我是什么——大约就是一百万亿个细胞级别的小机器人。这就是我们的组成。没有别的成分。我们就是由细胞构成的，大约一百万亿个。这些细胞里没有一个是有意识的；没有一个细胞知道你是谁，也不在乎。我们必须以某种方式解释：当你把上亿个无意识的小机器人细胞组成团队、军队、营队——而它们每一个其实和一个细菌没什么两样——结果却是这样。你看看这个。
4:37
The content -- there's color, there's ideas, there's memories, there's history. And somehow all that content of consciousness is accomplished by the busy activity of those hoards of neurons. How is that possible? Many people just think it isn't possible at all. They think, "No, there can't be any sort of naturalistic explanation of consciousness." This is a lovely book by a friend of mine named Lee Siegel, who's a professor of religion, actually, at the University of Hawaii, and he's an expert magician, and an expert on the street magic of India, which is what this book is about, "Net of Magic."
这些内容——有颜色，有想法，有记忆，有历史。而所有这些意识的内容，竟然是由那成群神经元的忙碌活动完成的。这怎么可能？很多人干脆认为这根本不可能。他们想：“不，意识不可能有任何自然主义的解释。”这是我一位朋友李·西格尔写的一本很棒的书，他其实是夏威夷大学的宗教学教授，同时他是个魔术高手，也是研究印度街头魔术的专家，这本书讲的就是这个，书名叫《魔法之网》。
04「真正的魔法」悖论与意识
5:14
And there's a passage in it which I would love to share with you. It speaks so eloquently to the problem. "'I'm writing a book on magic,' I explain, and I'm asked, 'Real magic?' By 'real magic,' people mean miracles, thaumaturgical acts, and supernatural powers. 'No,' I answer. 'Conjuring tricks, not real magic.' 'Real magic,' in other words, refers to the magic that is not real; while the magic that is real, that can actually be done, is not real magic." (Laughter) Now, that's the way a lot of people feel about consciousness.
书里有一段话，我很想和大家分享。它把这个问题说得太精彩了。“‘我在写一本关于魔法的书，’我解释道，然后有人问：‘真正的魔法吗？’人们说‘真正的魔法’，指的是奇迹、法术神通和超自然力量。‘不是，’我回答，‘是变戏法，不是真正的魔法。’换句话说，‘真正的魔法’指的是并不存在的魔法；而真实存在、真的能变出来的魔法，却不是真正的魔法。”（笑声）很多人对意识就是这种感觉。
5:53
(Laughter) Real consciousness is not a bag of tricks. If you're going to explain this as a bag of tricks, then it's not real consciousness, whatever it is. And, as Marvin said, and as other people have said, "Consciousness is a bag of tricks." This means that a lot of people are just left completely dissatisfied and incredulous when I attempt to explain consciousness. So this is the problem. So I have to do a little bit of the sort of work that a lot of you won't like, for the same reason that you don't like to see a magic trick explained to you.
（笑声）真正的意识不是一堆把戏。如果你要把它解释成一堆把戏，那它就不是真正的意识了，不管它到底是什么。可正如马文说过的，也正如别人说过的：“意识就是一堆把戏。”这意味着，当我试图解释意识时，很多人只会觉得完全不满足、完全不肯相信。问题就在这儿。所以我得做一些你们当中很多人不会喜欢的工作，原因和你们不喜欢别人把魔术的秘密讲给你听是一样的。
6:32
How many of you here, if somebody -- some smart aleck -- starts telling you how a particular magic trick is done, you sort of want to block your ears and say, "No, no, I don't want to know! Don't take the thrill of it away. I'd rather be mystified. Don't tell me the answer." A lot of people feel that way about consciousness, I've discovered. And I'm sorry if I impose some clarity, some understanding on you. You'd better leave now if you don't want to know some of these tricks. But I'm not going to explain it all to you.
在座有多少人，如果有个自作聪明的家伙开始告诉你某个魔术是怎么变的，你会想捂住耳朵说：“别别别，我不想知道！别把那份惊奇夺走。我宁愿一直摸不着头脑。别告诉我答案。”我发现，很多人对意识也是这种感觉。如果我强行给你们带来一些清晰和理解，那我很抱歉。如果你不想知道其中一些把戏，最好现在就离场。不过我也不会把一切都讲给你们听。
05哲学家如何解释锯人魔术
7:06
I'm going to do what philosophers do. Here's how a philosopher explains the sawing-the-lady-in-half trick. You know the sawing-the-lady-in-half trick? The philosopher says, "I'm going to explain to you how that's done. You see, the magician doesn't really saw the lady in half." (Laughter) "He merely makes you think that he does." And you say, "Yes, and how does he do that?" He says, "Oh, that's not my department, I'm sorry." (Laughter) So now I'm going to illustrate how philosophers explain consciousness.
我要做哲学家做的事。哲学家是这样解释“把女士锯成两半”这个魔术的。你们知道把女士锯成两半的魔术吧？哲学家说：“我来给你解释这是怎么做到的。你看，魔术师其实并没有真的把那位女士锯成两半。”（笑声）“他只是让你以为他锯了。”你说：“对啊，那他是怎么做到的呢？”他说：“哦，那不归我管，抱歉。”（笑声）所以现在我要给大家演示一下哲学家是怎么解释意识的。
7:43
But I'm going to try to also show you that consciousness isn't quite as marvelous -- your own consciousness isn't quite as wonderful -- as you may have thought it is. This is something, by the way, that Lee Siegel talks about in his book. He marvels at how he'll do a magic show, and afterwards people will swear they saw him do X, Y, and Z. He never did those things. He didn't even try to do those things. People's memories inflate what they think they saw. And the same is true of consciousness. Now, let's see if this will work. All right. Let's just watch this.
但我也想让你们看到，意识并没有那么神奇——你自己的意识并没有那么了不起——没有你以为的那么了不起。顺便说一句，这也是李·西格尔在他书里谈到的一点。他很惊讶：他表演完一场魔术之后，人们会发誓说亲眼看见他做了X、Y、Z。可他根本没做过那些事。他连试都没试过。人们的记忆会夸大他们自以为看到的东西。意识也是一样。好，看看这个能不能放出来。好的。我们一起看这个。
8:21
Watch it carefully.
仔细看。
06演示一：方块变色与视觉贫乏
8:34
I'm working with a young computer-animator documentarian named Nick Deamer, and this is a little demo that he's done for me, part of a larger project some of you may be interested in. We're looking for a backer. It's a feature-length documentary on consciousness. OK, now, you all saw what changed, right?
我在和一位年轻的电脑动画纪录片作者尼克·迪默合作，这是他为我做的一个小样片，属于一个更大项目的一部分，你们中有些人可能会感兴趣。我们正在找投资人。那是一部关于意识的长篇纪录片。好，大家都看到有什么变了，对吧？
8:58
How many of you noticed that every one of those squares changed color? Every one. I'll just show you by running it again.
有多少人注意到，那些方块每一个都变了颜色？每一个。我再放一遍给你们看。
9:12
Even when you know that they're all going to change color, it's very hard to notice. You have to really concentrate to pick up any of the changes at all. Now, this is an example -- one of many -- of a phenomenon that's now being studied quite a bit. It's one that I predicted in the last page or two of my 1991 book, "Consciousness Explained," where I said if you did experiments of this sort, you'd find that people were unable to pick up really large changes. If there's time at the end, I'll show you the much more dramatic case.
即使你知道它们全都会变色，还是很难注意到。你得非常专注，才能捕捉到哪怕一点点变化。这是一个例子——众多例子之一——这种现象现在被研究得相当多。我在1991年那本《意识的解释》最后一两页里就预言过这一点，我说，如果你做这类实验，就会发现人们没法察觉到真正很大的变化。如果最后还有时间，我会给你们看一个更戏剧性的例子。
9:48
Now, how can it be that there are all those changes going on, and that we're not aware of them? Well, earlier today, Jeff Hawkins mentioned the way your eye saccades, the way your eye moves around three or four times a second. He didn't mention the speed. Your eye is constantly in motion, moving around, looking at eyes, noses, elbows, looking at interesting things in the world. And where your eye isn't looking, you're remarkably impoverished in your vision. That's because the foveal part of your eye, which is the high-resolution part, is only about the size of your thumbnail held at arms length.
那么，怎么会有这么多变化正在发生，而我们却察觉不到呢？今天早些时候，杰夫·霍金斯提到了眼睛的扫视方式，你的眼睛每秒会移动三四次。他没说速度。你的眼睛一直在动，四处移动，看眼睛、看鼻子、看胳膊肘，看世界上有意思的东西。而在你的眼睛没看的地方，你的视觉贫乏得惊人。这是因为眼睛的中央凹部分，也就是高分辨率的那部分，大小只相当于你把手臂伸直时看到的自己的一片指甲盖。
07贝洛托的桥：大脑接受暗示
10:25
That's the detail part. It doesn't seem that way, does it? It doesn't seem that way, but that's the way it is. You're getting in a lot less information than you think. Here's a completely different effect. This is a painting by Bellotto. It's in the museum in North Carolina. Bellotto was a student of Canaletto's. And I love paintings like that -- the painting is actually about as big as it is right here. And I love Canalettos, because Canaletto has this fantastic detail, and you can get right up and see all the details on the painting.
那才是能看清细节的部分。感觉上不像是这样，对吧？感觉上不像，但事实就是如此。你获取的信息比你以为的少得多。这是一个完全不同的效应。这是贝洛托的一幅画。它收藏在北卡罗来纳州的博物馆里。贝洛托是卡纳莱托的学生。我很喜欢这样的画——这幅画的实际尺寸差不多就和现在屏幕上一样大。我喜欢卡纳莱托的画，因为卡纳莱托有那种极其精妙的细节，你可以凑近去看画上所有的细节。
11:01
And I started across the hall in North Carolina, because I thought it was probably a Canaletto, and would have all that in detail. And I noticed that on the bridge there, there's a lot of people -- you can just barely see them walking across the bridge. And I thought as I got closer I would be able to see all the detail of most people, see their clothes, and so forth. And as I got closer and closer, I actually screamed. I yelled out because when I got closer, I found the detail wasn't there at all.
我在北卡罗来纳州那个展厅里朝它走过去，因为我以为那大概是卡纳莱托的作品，会有那样丰富的细节。我注意到那边桥上有很多人——你勉强能看到他们正走过桥。我想，等我走近一点，应该能看清大多数人的所有细节，看清他们的衣服，等等。可当我越走越近时，我真的叫了出来。我喊出了声，因为走近之后，我发现那些细节根本不存在。
11:32
There were just little artfully placed blobs of paint. And as I walked towards the picture, I was expecting detail that wasn't there. The artist had very cleverly suggested people and clothes and wagons and all sorts of things, and my brain had taken the suggestion. You're familiar with a more recent technology, which is -- There, you can get a better view of the blobs. See, when you get close they're really just blobs of paint. You will have seen something like this -- this is the reverse effect.
那儿只有一些巧妙地点上去的小色块。我朝这幅画走过去的时候，一直期待着那些其实并不存在的细节。画家非常巧妙地暗示出了人、衣服、马车和各种各样的东西，而我的大脑接受了这种暗示。你们对一种更晚近的技术并不陌生，就是——喏，这样你能更清楚地看到那些色块。看，凑近了它们真的就只是一块块颜料。你大概见过类似这样的东西——这是反过来的效果。
12:22
I'll just give that to you one more time. Now, what does your brain do when it takes the suggestion? When an artful blob of paint or two, by an artist, suggests a person -- say, one of Marvin Minsky's little society of mind -- do they send little painters out to fill in all the details in your brain somewhere? I don't think so. Not a chance. But then, how on Earth is it done? Well, remember the philosopher's explanation of the lady? It's the same thing. The brain just makes you think that it's got the detail there.
我再给你们看一遍。那么，当你的大脑接受了暗示时，它做了什么？当画家用一两块巧妙的颜料暗示出一个人——比如马文·明斯基那个“心智社会”里的某个小成员——它们会不会派出小画匠，去你大脑里的某个地方把所有细节补齐？我可不这么认为。绝无可能。可那到底是怎么做到的呢？嗯，还记得哲学家对那位女士的解释吗？是同一回事。大脑只是让你以为细节就在那里。
08心理旋转：你不知道自己怎么做的
13:06
You think the detail's there, but it isn't there. The brain isn't actually putting the detail in your head at all. It's just making you expect the detail. Let's just do this experiment very quickly. Is the shape on the left the same as the shape on the right, rotated? Yes. How many of you did it by rotating the one on the left in your mind's eye, to see if it matched up with the one on the right? How many of you rotated the one on the right? OK. How do you know that's what you did? (Laughter) There's in fact been a very interesting debate raging for over 20 years in cognitive science -- various experiments started by Roger Shepherd, who measured the angular velocity of rotation of mental images.
你以为细节在那儿，其实并不在。大脑其实根本没有把细节放进你的脑袋里。它只是让你预期会有那些细节。我们来快速做个实验。左边的形状和右边的形状旋转之后是一样的吗？是一样的。你们当中有多少人是在脑海里把左边那个转过来，看它是否和右边那个对得上？又有多少人转的是右边那个？好。你怎么知道自己是这么做的？（笑声）事实上，认知科学界有一场非常有意思的争论，已经持续了二十多年——最早是Roger Shepherd做的一系列实验，他测量了心理意象旋转的角速度。
13:51
Yes, it's possible to do that. But the details of the process are still in significant controversy. And if you read that literature, one of the things that you really have to come to terms with is even when you're the subject in the experiment, you don't know. You don't know how you do it. You just know that you have certain beliefs. And they come in a certain order, at a certain time. And what explains the fact that that's what you think? Well, that's where you have to go backstage and ask the magician.
没错，这是可以测量的。但这个过程的具体细节仍然存在很大争议。如果你读那些文献，有一件事你不得不接受：哪怕你就是实验里的被试，你也不知道。你不知道自己是怎么做到的。你只知道自己有某些想法。而且它们以某种顺序、在某个时刻出现。那么，是什么解释了你会这么想这件事呢？这就得到后台去问问魔术师了。
09内克尔立方体与填充争论
14:22
This is a figure that I love: Bradley, Petrie, and Dumais. You may think that I've cheated, that I've put a little whiter-than-white boundary there. How many of you see that sort of boundary, with the Necker cube floating in front of the circles? Can you see it? Well, you know, in effect, the boundary's really there, in a certain sense. Your brain is actually computing that boundary, the boundary that goes right there. But now, notice there are two ways of seeing the cube, right? It's a Necker cube.
这是一张我很喜欢的图：Bradley、Petrie和Dumais做的。你可能觉得我作弊了，觉得我在那里加了一道比白还白的边界。你们当中有多少人看到那样一条边界，还有一个内克尔立方体浮在那些圆的前面？看得到吗？嗯，你知道，实际上从某种意义上说，那个边界真的存在。你的大脑其实正在计算那个边界，就是那条正好经过那里的边界。但现在，注意看这个立方体有两种看法，对吧？这是内克尔立方体。
14:57
Everybody can see the two ways of seeing the cube? OK. Can you see the four ways of seeing the cube? Because there's another way of seeing it. If you're seeing it as a cube floating in front of some circles, some black circles, there's another way of seeing it. As a cube, on a black background, as seen through a piece of Swiss cheese. (Laughter) Can you get it? How many of you can't get it? That'll help. (Laughter) Now you can get it. These are two very different phenomena. When you see the cube one way, behind the screen, those boundaries go away.
大家都能看出这个立方体的两种看法吗？好。那你们能看出这个立方体的四种看法吗？因为还有另一种看法。如果你把它看成一个漂浮在一些圆圈前面的立方体，一些黑色的圆圈，那还有另一种看法。把它看成一个在黑色背景上的立方体，透过一片瑞士奶酪看过去。（笑声）你们看出来了吗？有多少人看不出来？举手会有帮助。（笑声）现在你看出来了。这是两种非常不同的现象。当你用第一种方式看这个立方体，也就是在屏幕后面时，那些边界就消失了。
15:41
But there's still a sort of filling in, as we can tell if we look at this. We don't have any trouble seeing the cube, but where does the color change? Does your brain have to send little painters in there? The purple-painters and the green-painters fight over who's going to paint that bit behind the curtain? No. Your brain just lets it go. The brain doesn't need to fill that in.
但仍然有某种填充效应，我们看这个就能看出来。我们毫不费力就能看到这个立方体，但颜色是在哪里变化的呢？难道你的大脑得派些小画匠进去吗？紫色画匠和绿色画匠要争夺谁来画帘子后面那一小块吗？不是的。你的大脑就直接放着不管了。大脑不需要把那部分填充上。
16:07
When I first started talking about the Bradley, Petrie, Dumais example that you just saw -- I'll go back to it, this one -- I said that there was no filling-in behind there. And I supposed that that was just a flat truth, always true. But Rob Van Lier has recently shown that it isn't. Now, if you think you see some pale yellow -- I'll run this a few more times. Look in the gray areas, and see if you seem to see something sort of shadowy moving in there -- yeah, it's amazing. There's nothing there. It's no trick.
当我最初开始讲你们刚才看到的布拉德利、佩特里和杜梅的那个例子时——我回到那张图，就是这个——我说过那后面并没有填充。而我当时以为那就是个板上钉钉的事实，永远成立。但罗布·范利尔最近证明了并非如此。现在，如果你觉得看到了一点淡黄色——我再放几遍。看那些灰色区域，看看你是不是好像看到有某种影影绰绰的东西在那里移动——是的，太神奇了。那里什么都没有。这不是什么把戏。
10变化盲视实验：闪烁范式
16:56
["Failure to Detect Changes in Scenes" slide] This is Ron Rensink's work, which was in some degree inspired by that suggestion right at the end of the book. Let me just pause this for a second if I can. This is change-blindness. What you're going to see is two pictures, one of which is slightly different from the other. You see here the red roof and the gray roof, and in between them there will be a mask, which is just a blank screen, for about a quarter of a second. So you'll see the first picture, then a mask, then the second picture, then a mask.
["未能察觉场景变化"幻灯片] 这是 Ron Rensink 的研究，某种程度上是受了那本书结尾处那个建议的启发。让我先把这个暂停一下。这就是"变化盲视"。你们接下来会看到两张图片，其中一张和另一张略有不同。你们看这里有红色的屋顶和灰色的屋顶，在这两张图之间会有一个遮蔽画面，也就是一块空白屏幕，大约四分之一秒。所以你会先看到第一张图片，然后是空白，然后是第二张图片，再是空白。
17:29
And this will just continue, and your job as the subject is to press the button when you see the change. So, show the original picture for 240 milliseconds. Blank. Show the next picture for 240 milliseconds. Blank. And keep going, until the subject presses the button, saying, "I see the change." So now we're going to be subjects in the experiment. We're going to start easy. Some examples. No trouble there. Can everybody see? All right. Indeed, Rensink's subjects took only a little bit more than a second to press the button.
就这样一直循环下去，而你作为被试的任务，就是在看到变化时按下按钮。那么，先呈现原始图片 240 毫秒。空白。再呈现下一张图片 240 毫秒。空白。如此往复，直到被试按下按钮，说："我看到变化了。"现在我们就来当这个实验的被试。我们先从简单的开始。几个例子。这个没什么难度。大家都看到了吗？好的。事实上，Rensink 的被试只用了一秒多一点就按下了按钮。
18:24
Can you see that one?
那个你们能看出来吗？
18:33
2.9 seconds.
2.9 秒。
18:42
How many don't see it still? What's on the roof of that barn? (Laughter)
还有多少人没看出来？那个谷仓的屋顶上有什么？（笑声）
18:58
It's easy.
很简单嘛。
19:24
Is it a bridge or a dock?
那是一座桥还是一个码头？
19:30
There are a few more really dramatic ones, and then I'll close. I want you to see a few that are particularly striking. This one because it's so large and yet it's pretty hard to see. Can you see it? Audience: Yes. Dan Dennett: See the shadows going back and forth? Pretty big.
还有几个特别夸张的，然后我就结束。我想让你们看几个特别有意思的例子。这一个，是因为它面积那么大，但还是相当难发现。你们看出来了吗？观众：看到了。Dan Dennett：看到那些来回移动的阴影了吗？挺大一片。
11结论：第三人称科学胜过内省
20:01
So 15.5 seconds is the median time for subjects in his experiment there. I love this one. I'll end with this one, just because it's such an obvious and important thing. How many still don't see it? How many still don't see it? How many engines on the wing of that Boeing? (Laughter) Right in the middle of the picture! Thanks very much for your attention. What I wanted to show you is that scientists, using their from-the-outside, third-person methods, can tell you things about your own consciousness that you would never dream of, and that, in fact, you're not the authority on your own consciousness that you think you are.
他那个实验里，被试的中位反应时间是 15.5 秒。我特别喜欢这一个。我就用这个来收尾，因为它实在是又明显又重要。还有多少人没看出来？还有多少人没看出来？那架波音飞机的机翼上有几个发动机？（笑声）就在画面正中间啊！非常感谢大家的关注。我想向你们说明的是，科学家用他们那套从外部出发的、第三人称的方法，能够告诉你关于你自己意识的一些事情，那是你做梦都想不到的，而且事实上，在你自己的意识这件事上，你并不像你以为的那样是权威。
20:49
And we're really making a lot of progress on coming up with a theory of mind. Jeff Hawkins, this morning, was describing his attempt to get theory, and a good, big theory, into the neuroscience. And he's right. This is a problem. Harvard Medical School once -- I was at a talk -- director of the lab said, "In our lab, we have a saying. If you work on one neuron, that's neuroscience. If you work on two neurons, that's psychology." (Laughter) We have to have more theory, and it can come as much from the top down.
我们在建立一套心智理论方面确实取得了很大进展。今天早上，Jeff Hawkins 讲到他试图把理论，一套好的、宏大的理论，引入神经科学。他说得对。这确实是个问题。哈佛医学院曾经——我去听过一场报告——那个实验室的主任说："在我们实验室有句话。如果你研究一个神经元，那是神经科学。如果你研究两个神经元，那就是心理学了。"（笑声）我们必须要有更多的理论，而且它完全可以是自上而下产生的。
21:28
Thank you very much. (Applause)
非常感谢。（掌声）
视频总结 · 一句话概括与核心要点
一句话概括
丹尼特借助魔术、绘画错觉与"变化盲视"实验论证：你对自己意识的"内省权威"是一种幻觉——大脑并没有把细节装进你的头脑，只是让你"以为"细节在那里，而第三人称的科学方法能揭示你自己永远察觉不到的意识真相。
核心要点
- 意识难以讨论的根源在于"人人自认专家"：人们对气候变化或互联网未来有强烈观点却不自视为专家，唯独对意识，每个人都觉得"我有意识，所以我懂"。因此哲学论证（Nozick 所谓"不接受结论就脑袋爆炸"式的论证）根本改变不了任何人的想法；丹尼特的目标不是说服，而是动摇听众对自己内心的自信。
- 我们只是约 100 万亿个无意识的"细胞机器人"：没有任何一个细胞有意识、知道你是谁或在乎你，但把成千上万亿个与细菌差不多的细胞组合起来，却产生了颜色、观念、记忆、历史。多数人认为这不可能有自然主义解释。
- "真正的魔术"悖论同样适用于意识：Lee Siegel 在《Net of Magic》中指出，人们说的"真魔术"指不存在的奇迹，而真能做到的把戏则被认为"不是真魔术"。同理，一旦把意识解释为"一袋把戏"（Minsky 的说法），人们就说那不是"真正的意识"，于是拒绝解释——就像不想听魔术揭秘。
- 哲学家的解释方式即"锯人魔术"式解释：魔术师并没真的把女士锯成两半，"只是让你以为他锯了"；至于怎么做到的，"那不归我管"。丹尼特坦承自己对意识的解释也是这一层级：大脑并未真的呈现细节，只是让你以为有细节。
- 变化盲视实验证明视觉输入远比感觉贫乏：动画中每一个方块都变了颜色，几乎无人察觉，即便预先知道也需极度专注才能捕捉。原因是眼睛每秒扫视三四次，而高分辨率的中央凹区域只有臂展外一个拇指指甲大小，其余视野极度贫乏——"你获得的信息比你以为的少得多"。这一现象正是他 1991 年《意识的解释》末尾所预言的。
- Bellotto 画作说明大脑接受"暗示"而非填充细节：他以为桥上人群走近后会像 Canaletto 那样细节丰富，结果只是几块巧妙放置的颜料斑点，他当场惊叫。大脑并未派"小画家"去补全衣服和马车，只是让你"预期"细节存在。
- 心理旋转实验揭示"当事人也不知道自己怎么做的"：判断两个图形是否为旋转关系时，问听众"你旋转的是左边还是右边？你怎么知道自己是这么做的？"——Shepard 起自的 20 余年争论显示，即使作为被试，你也只知道自己有某些按顺序出现的信念，却不知道过程，必须"去后台问魔术师"。
- Necker 立方体有四种看法，且"填充"并非一律：Bradley-Petrie-Dumais 图形中大脑真的计算出了"比白更白"的主观轮廓；换成"透过瑞士奶酪看黑底上的立方体"时轮廓消失。紫色与绿色交界处大脑也不需要"派画家争夺"，直接放着不管。但 Rob van Lier 最近证明灰色区域确实存在某种影子式的填充运动，修正了丹尼特原先"完全没有填充"的断言。
- Rensink 的变化盲视实验给出量化数据：两幅图之间插入约 1/4 秒空白屏，各显示 240 毫秒交替，被试按键报告变化。简单案例仅需 1 秒多；其他案例 2.9 秒；某个巨大而醒目的阴影变化，中位反应时间达 15.5 秒；最后一例——波音机翼上引擎数量的改变，就在画面正中，大批听众仍看不出来。
结论与值得注意的细节
- 核心结论：科学家用"从外部"的第三人称方法能告诉你关于你自己意识的、你做梦都想不到的事实；你并不是自己以为的那个"意识权威"。
- 记忆膨胀效应：Siegel 做完魔术表演后，观众会发誓看见他做了他根本没做、甚至没尝试过的动作——意识也一样，我们的记忆会夸大自己"看到"的东西。
- 丹尼特正与动画师 Nick Deamer 合作制作一部关于意识的长篇纪录片，演讲中的方块变色演示即为其样片，当时正在寻找投资人。
- 对神经科学的呼吁：引用哈佛医学院实验室的自嘲——"研究一个神经元是神经科学，研究两个神经元就是心理学"——强调该领域亟需更多理论，而理论可以自上而下（如 Jeff Hawkins 的尝试）与自下而上同等重要。
- 演讲设有一个"离场警告"：如果你不想知道意识的把戏（就像不想听魔术揭秘），最好现在离开——这既是幽默，也揭示了人们对意识解释的情感抵触。
核心句型 · 9
1. When I … and they ask me …, I say, "…" – their eyes glaze over.
“When I go to an academic cocktail party … they ask me what field I'm in and I say, "philosophy" -- their eyes glaze over.”
三段递进的排比式自嘲：相同句式重复，最后一次打破预期（lips curl into a snarl）。适合开场制造笑点，仿写时保持结构一致、结尾反转。
2. The very X of somebody doing Y is just out of the question.
“The very chutzpah of somebody thinking that you could explain consciousness is just out of the question.”
「the very + 名词」强调某种态度本身，后接 out of the question 表示荒谬。用于转述他人的强烈反对立场。
3. I finally figured out the reason for that. The reason for that is that …
“I finally figured out the reason for that. The reason for that is that everybody's an expert on consciousness.”
先宣告「找到了原因」再顶真式重复「The reason for that is」，制造悬念并突出答案。演讲中揭示核心观点的常用节奏。
4. 'Real X,' in other words, refers to the X that is not real; while the X that is real … is not real X.
“'Real magic,' in other words, refers to the magic that is not real; while the magic that is real, that can actually be done, is not real magic.”
交叉倒装（chiasmus）结构，用词序对调揭示概念中的悖论。可用于批评某种「定义排除现实」的思维方式。
5. X isn't quite as … as you may have thought it is.
“Consciousness isn't quite as marvelous -- your own consciousness isn't quite as wonderful -- as you may have thought it is.”
「not quite as … as you may have thought」温和地降低听众期待，比直接否定更委婉。适合提出反直觉结论前的铺垫。
6. Even when you know that …, it's very hard to …
“Even when you know that they're all going to change color, it's very hard to notice.”
让步结构强调现象的顽固性——知情也无法克服。用于描述认知偏差、错觉等「知道也没用」的情形。
7. It doesn't seem that way, but that's the way it is.
“It doesn't seem that way, does it? It doesn't seem that way, but that's the way it is.”
先用反义疑问句拉近听众，再用 but 断言事实。简短有力，适合指出「感觉 vs 事实」的落差。
8. How do you know that's what you did?
“How many of you rotated the one on the right? OK. How do you know that's what you did?”
先让听众作答，再反问其依据，暴露内省的不可靠。互动式演讲中引发反思的经典追问句式。
9. X can tell you things about Y that you would never dream of.
“Scientists … can tell you things about your own consciousness that you would never dream of”
「that you would never dream of」作定语从句，强调信息出人意料。用于总结外部方法/工具带来的意外发现。
生词精讲 · 72 · 按出现顺序
glaze over
phr. v.
0:04
（眼神）变得呆滞、失去兴趣
snarl
/snɑːrl/
n.
0:04
龇牙咧嘴的表情；咆哮
hoots of derision
phr.
0:46
一片嘲笑声、嘘声
cackles
/ˈkækəlz/
n.
0:46
咯咯的怪笑
chutzpah
/ˈhʊtspə/
n.
0:46
（源自意第绪语）厚脸皮，肆无忌惮的胆量
out of the question
phr.
0:46
不可能，根本不必考虑
late, lamented
phr.
0:46
已故的、令人怀念的（固定悼念用语）
ethos
/ˈiːθɑːs/
n.
0:46
（某群体的）精神气质、风气
premises
/ˈpremɪsɪz/
n.
0:46
（逻辑）前提
inferences
/ˈɪnfərənsɪz/
n.
0:46
推理、推论
knocks out
phr. v.
1:40
击倒，击败
with regard to
phr.
2:09
关于，就……而言
expertise
/ˌekspɜːrˈtiːz/
n.
2:09
专业知识、专长
innermost
/ˈɪnərmoʊst/
adj.
2:44
最深处的，内心最隐秘的
authoritative
/əˈθɔːrəteɪtɪv/
adj.
2:44
有权威的；权威性的
the order of the day
phr.
2:44
当下的主题/议程；当务之急
stream of consciousness
n. phr.
3:17
意识流
off to the races
idiom
3:17
（思绪、事情）一发不可收拾地开始了
rendered
/ˈrendərd/
v.
3:17
（以某种风格）描绘、表现
pointillist
/ˈpɔɪntəlɪst/
adj.
3:17
点彩派的
battalions
/bəˈtæljənz/
n.
3:53
营（军事单位）；大批
bacterium
/bækˈtɪriəm/
n.
3:53
细菌（单数；复数 bacteria）
hoards
/hɔːrdz/
n.
4:37
大量、成群（此处为 hordes 的误拼）
naturalistic
/ˌnætʃərəˈlɪstɪk/
adj.
4:37
自然主义的（不诉诸超自然的）
eloquently
/ˈeləkwəntli/
adv.
5:14
雄辩地，有说服力地
thaumaturgical
/ˌθɔːməˈtɜːrdʒɪkəl/
adj.
5:14
施行奇迹的，神通的
Conjuring tricks
n. phr.
5:14
魔术戏法，变戏法
a bag of tricks
idiom
5:53
一整套手法/把戏
incredulous
/ɪnˈkredʒələs/
adj.
5:53
不肯相信的，怀疑的
smart aleck
/ˈsmɑːrt ˌælɪk/
n.
6:32
自作聪明的人
mystified
/ˈmɪstɪfaɪd/
adj.
6:32
困惑不解的，被迷住的
impose
/ɪmˈpoʊz/
v.
6:32
强加
merely
/ˈmɪrli/
adv.
7:06
仅仅，只不过
not my department
idiom
7:06
不归我管，不是我的职责范围
marvels at
phr. v.
7:43
对……感到惊叹
swear
/swer/
v.
7:43
发誓；信誓旦旦地说
inflate
/ɪnˈfleɪt/
v.
7:43
夸大，膨胀
backer
/ˈbækər/
n.
8:34
资助人，投资者
feature-length
/ˈfiːtʃər leŋθ/
adj.
8:34
（影片）长片规格的
pick up
phr. v.
9:12
察觉，捕捉到
saccades
/səˈkɑːdz/
v./n.
9:48
（眼球）快速跳动；眼跳
impoverished
/ɪmˈpɑːvərɪʃt/
adj.
9:48
贫乏的，匮乏的
foveal
/ˈfoʊviəl/
adj.
9:48
（视网膜）中央凹的
at arms length
phr.
9:48
一臂之距（伸直手臂的距离）
barely
/ˈberli/
adv.
11:01
勉强，几乎不
artfully
/ˈɑːrtfəli/
adv.
11:32
巧妙地，有技巧地
blobs
/blɑːbz/
n.
11:32
一团团、一块块（无定形的）
taken the suggestion
phr.
11:32
接受了暗示
how on Earth
phr.
12:22
到底怎么（加强语气）
Not a chance
phr.
12:22
绝无可能
mind's eye
n. phr.
13:06
脑海中的想象，心眼
raging
/ˈreɪdʒɪŋ/
adj.
13:06
激烈进行的（争论）
angular velocity
n. phr.
13:06
角速度
come to terms with
idiom
13:51
接受（不愉快的事实）
backstage
/ˌbækˈsteɪdʒ/
adv.
13:51
到后台
whiter-than-white
adj.
14:22
比白还白的（主观轮廓区域显得更亮）
Necker cube
n.
14:22
内克尔立方体（双稳态视错觉图形）
in effect
phr.
14:22
实际上
Swiss cheese
n.
14:57
瑞士奶酪（带孔洞，喻有孔的屏幕）
filling in
n.
15:41
（知觉）填充
lets it go
phr. v.
15:41
放着不管，不去处理
flat truth
n. phr.
16:07
板上钉钉的事实
shadowy
/ˈʃædoʊi/
adj.
16:07
影影绰绰的，模糊的
change-blindness
n.
16:56
变化盲视
mask
/mæsk/
n.
16:56
（实验中的）遮蔽画面
subject
/ˈsʌbdʒɪkt/
n.
17:29
（实验）被试
striking
/ˈstraɪkɪŋ/
adj.
19:30
引人注目的，显著的
median
/ˈmiːdiən/
adj./n.
20:01
中位数（的）
third-person
adj.
20:01
第三人称的（外部观察视角）
would never dream of
phr.
20:01
做梦都想不到
theory of mind
n. phr.
20:49
心智理论
from the top down
phr.
20:49
自上而下地
理解自测 · 11 题 · 是真懂了，还是以为自己懂
1. Dennett在开场说，当他告诉不同圈子的人自己的职业或研究领域时，对方分别有什么反应？
他描述了三层递进：普通派对上说「我是教授」，对方眼神发直；学术酒会上说「哲学」，对方也眼神发直；而在哲学家聚会上说「意识」时，对方不再是无趣，而是嘴角撇起露出鄙夷，并发出嘘声和怪笑。这个开场（第0–1段）的要点是：即便在哲学界内部，「解释意识」也被视为不可能的、甚至厚脸皮的企图，为全篇要挑战的偏见定下基调。
2. Lee Siegel书中「真正的魔法」那段话说了什么？Dennett用它类比什么？
Siegel说自己在写关于魔法的书，别人问是不是「真正的魔法」——即奇迹、神通、超自然力量；他回答是「戏法」。于是出现悖论：「真正的魔法」指并不存在的魔法，而真实能做到的魔法却不算「真正的魔法」（第8段）。Dennett以此类比人们对意识的态度（第9段）：一旦意识被解释成「一堆把戏」，人们就说那不是「真正的意识」，从而任何可解释的意识都被预先排除在外。
3. Rensink的变化盲视实验是怎样设计的？给出关键参数和结果数据。
采用闪烁范式（第28–29段）：两张仅有一处不同的图片交替呈现，每张240毫秒，中间插入约四分之一秒的空白遮蔽画面，循环直到被试按键报告看到变化。简单条件下被试只需一秒多；较难的例子用了2.9秒；在阴影来回变化的例子中，被试的中位数反应时间为15.5秒（第36段）。最后的波音机翼发动机数量变化位于画面正中，仍有很多观众长时间看不出。
4. 关于中央凹，Dennett给出了怎样的数据？它说明什么？
第17段引用Jeff Hawkins的说法：眼睛每秒跳动三四次；而视网膜上高分辨率的中央凹部分，大小只相当于把手臂伸直时看到的一片拇指指甲。也就是说，眼睛没有注视的地方，视觉信息极其贫乏。第18段的结论是：「你获取的信息比你以为的少得多」，但主观上视野却处处清晰。这直接支撑了全篇论点——你对自己的视觉体验的判断并不可靠。
5. Dennett为什么认为纯逻辑论证无法改变人们对意识的看法？他的诊断是什么？
他引用Nozick对哲学风气的批评：哲学家理想中的论证是让对手「脑袋爆炸」的强制性论证，但这从不真正改变人心（第1–2段）。他的诊断是：关于意识，每个人都自认为专家（第2–3段）。与气候变化、互联网未来不同——人们对那些也有强烈意见，但不会当作专业知识——对意识，人们觉得「光凭我有意识，我就了如指掌」。由于这种自信来自第一人称直觉而非论证，所以只能靠演示来动摇，而不能靠推理。
6. 「锯人魔术」的笑话在Dennett的论证中承担什么功能？
第11段中哲学家的解释是：「魔术师并没有真的把人锯开，只是让你以为他锯了」，而具体怎么做到「不归我管」。表面上是自嘲哲学的无用，实际上它交代了Dennett的方法论：哲学家的工作是消除错误的问题框架——大脑并没有真的在内部「画出」细节或「填充」画面，它只是让你以为如此（第21–22段重申「It's the same thing」）。至于神经机制如何实现这种「以为」，则交给认知科学，这也呼应了结尾对理论与神经科学分工的讨论。
7. 贝洛托的画与斯坦伯格点彩风格的人像，分别从哪个方向说明了同一个道理？
两者都说明大脑会把稀疏的信息「当作」完整内容。贝洛托的桥（第18–20段）是从远到近：远看以为桥上行人有衣着细节，走近发现只是巧妙的色块——大脑「接受了暗示」，产生了并不存在的细节预期。点彩人像与第6段的「百万亿细胞机器人」则是从微观到宏观：单个点、单个细胞都无意识、无内容，但整体上却呈现出颜色、想法、记忆。前者揭示体验的内容被高估，后者揭示体验由无意识部件构成——都指向「意识没有你以为的那么神奇」。
8. 心理旋转实验中，Dennett问「你怎么知道自己转的是哪一个」，这一追问的推理链是什么？
第22–23段：Shepard的实验测出心理意象旋转的角速度，说明确有某种可测过程。但Dennett让听众各自报告「转了左边」或「转了右边」，随即追问其依据。推理链是：（1）被试只能报告自己「相信」做了什么；（2）这种信念以某种顺序、在某时刻出现；（3）产生信念的过程本身不向内省开放；（4）所以哪怕你就是被试，你也不知道自己怎么做的；（5）要解释过程，只能「到后台问魔术师」，即依赖第三人称实验。这直接支撑结尾「你不是自己意识的权威」的结论。
9. Dennett在内克尔立方体与填充问题上做了怎样的自我修正？这体现了他什么立场？
第24–26段中他用Bradley–Petrie–Dumais图形说明：主观轮廓是大脑计算出来的，但换一种看法（透过瑞士奶酪看）轮廓就消失；而帘子后面的颜色变化大脑「直接放着不管」，不需要派「小画匠」去填充。第27段他承认，自己过去把「无填充」当成永远成立的事实，但Rob van Lier最近的实验证明某些条件下确实存在填充。这一修正体现了他的自然主义立场：关于意识的哲学论断必须服从经验研究，而不是靠直觉或论证一锤定音。
10. 如果有人反驳说：「变化盲视只说明注意力有限，不能说明我对当下体验的了解是错的」，Dennett会如何回应？
他可能会承认注意力有限是机制层面的解释，但指出这恰恰是他的论点：人们主观上「确信」自己看到了完整、连续、细节丰富的场景（第18段「感觉上不像是这样」；第12段记忆会「膨胀」），而实验表明实际获取的内容极少。错的不是「你看到了什么」，而是「你以为自己看到了多少」——即你对自己体验之丰富程度的判断。变化盲视、中央凹数据、贝洛托的桥都在证明这种判断系统性地高估。因此「注意力有限」并不削弱论点，反而是对「你并非自己意识权威」的具体解释。
11. 把Dennett「第三人称方法能揭示内省无法发现的事实」这一主张迁移到今天关于AI是否有意识的争论，会得出什么启示？
Dennett的主张（第36段）是：对意识的可靠认识来自外部、可重复的实验，而非第一人称报告。迁移到AI争论：（1）AI的自我报告（「我感到……」）与人类的内省报告一样，都只是需要解释的数据，而非真相本身；（2）反过来，人类因为「我有意识所以我懂意识」而对AI意识下的直觉判断，同样不具权威；（3）「真正的魔法」悖论提醒我们警惕一种论证：凡能被机制解释的都「不算真正的意识」——这会使问题永远无法回答。Dennett式的结论是：应该像研究人类那样，用第三人称的行为与机制研究来讨论AI，而不是靠双方的直觉互相否定。
下载便签
手机：长按图片保存
← 上一期 · NO.031Sam Altman - How to Succeed with a Startup
下一期 · NO.033 →Roger Penrose - What is Consciousness?
苏菲拉底
THE SOPHIE LAB · ASK THE BEST MINDS THE BIG QUESTIONS
内容仅供学习 · thesophielab.com