LW LessWrong Highlights — Bilingual Study EditionLessWrong 精选 · 中英双语精读
All 50 ↩目录 ↩
#10 How To Actually Change Your Mind 1056 words · ~5 min

The Bottom Line底线

Writing the conclusion first poisons every argument that follows — your reasoning is only as honest as the algorithm that chose what to conclude.先写结论会毒化此后的每一条论证——你的推理只与选择结论的那个算法一样诚实。

01

Concise Summary简洁概述

Yudkowsky opens with a thought experiment: a 'clever arguer' is hired to advocate for whichever box the winning bidder owns, so they write 'Box B contains the diamond' at the bottom of their page first, then fill in supporting evidence above it, ignoring contrary signs. The crucial insight is that the physical ink becomes entangled with reality at the moment the conclusion is written — not when the arguments are assembled. Whatever cherry-picked reasoning follows cannot change that correlation. By contrast, a genuinely curious inquirer who surveys all evidence first, then writes a probability at the bottom, produces handwriting that actually tracks the world. This maps onto everyday motivated reasoning: when you decide the conclusion and then look for support, the algorithm driving your beliefs is not 'find truth' but something else entirely.

Yudkowsky 以一个思想实验开场:一位「狡辩者」受雇为出价最高的投标人辩护,于是他先在纸的底部写下「B 箱里有钻石」,再在上方填入支持证据,忽略反面迹象。关键洞察在于:墨水与现实的纠缠,在写下结论的那一刻就已固定——而非在论证被整理完毕之后。此后再怎么精心挑选论据,也改变不了这种相关性。相比之下,一位真正好奇的探究者先审视全部证据,再在底部写下概率,这样的笔迹才真正追踪世界。这对应日常的动机性推理:当你先决定结论、再寻找支持时,驱动你信念的算法并非「寻找真相」,而是完全别的东西。

02

Infographic信息图

2 arguers
clever arguer vs. curious inquirer — same surface words, opposite epistemic worth
狡辩者 vs. 好奇探究者——表面文字相似,认识论价值截然相反
bottom line first
the moment the conclusion is inked, its evidential entanglement with reality is fixed
先写底线的那一刻,结论与现实的证据性纠缠便已固定
85%
the honest inquirer's estimate — a probability that actually reflects the signs
诚实探究者给出的估算——一个真正反映迹象的概率
📝

Conclusion written first

先写结论

The clever arguer inks the bottom line before gathering evidence, so the handwriting tracks who paid more, not which box holds the diamond.

狡辩者在收集证据之前就写下结论,因此那段笔迹追踪的是谁出价更高,而非哪个箱子里有钻石。

🔗

Evidential entanglement is fixed at writing

证据性纠缠在书写时固定

Across all possible worlds, the correlation between ink and reality is set at the moment of writing; subsequent cherry-picked arguments cannot alter it.

在所有可能世界中,墨水与现实的相关性在书写时便已确定;此后精心挑选的论据无法改变这一点。

🔍

The curious inquirer's algorithm

好奇探究者的算法

When all signs are surveyed first and the conclusion follows from them, the handwriting is causally downstream of the evidence and tracks reality.

当所有迹象先被检视,结论由此得出,笔迹便是证据的因果下游,能真正追踪现实。

🚗

The real bottom-line algorithm

真正的底线算法

"Never repair anything expensive" is the real rule driving the car-brakes rationalization; the arguments assembled above it are post-hoc window dressing.

「绝不修任何昂贵的东西」才是驱动刹车合理化的真实规则;其上堆砌的论证不过是事后的门面装饰。

⚠️

Not a general counterargument

并非万能反驳

Calling someone a 'clever arguer' can itself be a motivated move; the world's cleverest arguer may still be right that the sun is shining.

指控某人是「狡辩者」本身也可能是一种动机性操作;世上最狡猾的辩手说「太阳正在照耀」,也可能是对的。

The argument, step by step
论证的推进链条
1
A clever arguer is hired to advocate for box B, regardless of the evidence.
一位狡辩者被雇来为 B 箱辩护,不论证据如何。
2
They write the conclusion at the bottom of the page first, then cherry-pick supporting signs above it.
他们在纸底写下结论,再在上方挑选支持性迹象。
3
At the moment of writing, the ink's correlation with reality is permanently fixed — favoring the bidder, not the diamond.
书写的那一刻,墨水与现实的相关性便永久固定——偏向出价者,而非钻石所在。
4
A curious inquirer surveys all signs first; their probability estimate is causally entangled with the actual evidence.
好奇的探究者先检视全部迹象;他们的概率估算与实际证据存在因果纠缠。
5
The same pattern applies to personal reasoning: the real algorithm is whatever chose the conclusion, not the arguments assembled afterward.
同样的模式适用于个人推理:真正的算法是选择结论的那个,而非事后堆砌的论证。
6
Caution: this insight can itself be weaponized as a motivated counter-argument — apply it to yourself, not as a trump card against others.
警示:这一洞察本身也可能被武器化为动机性反驳——把它用于审视自己,而非作为对付他人的王牌。
03

Detailed Summary详细概述

The Clever Arguer's Sheet of Paper

Yudkowsky opens with a vivid auction scenario. Two sealed boxes, A and B, one containing a diamond. A hired arguer — paid by whichever box-owner bids higher — takes out a sheet of paper. Before examining any evidence, they write "And therefore, box B contains the diamond" at the bottom. Then they fill in the space above: blue stamp, shininess, weight — selecting only signs that favor B, ignoring anything favoring A. Finally they recite this paper to the audience as if it were reasoned analysis.

The philosophical pivot: At the moment the ink hit the paper, the evidential entanglement of that conclusion with reality was fixed. Across a collection of worlds (Everett branches or Tegmark duplicates), there is some objective frequency at which box B actually contains the diamond. That frequency does not change based on what arguments are written above the bottom line. The ink is non-erasable. The boxes won't change. The correlation between the ink-conclusion and the box-contents was determined the moment the arguer chose which owner to serve.

Yudkowsky sharpens the point with an analogy from cognitive psychology: subjects shown the word "Green" printed in red ink often say "green" when asked for the color. Being literate — knowing the word's meaning — actually confuses you. The true import of a physical object is its causal entanglement with other things, not the meaning you read off it.

The Curious Inquirer

Contrast this with a second person who is genuinely curious. They first write down all distinguishing signs of both boxes, apply their knowledge and probability theory, and then write at the bottom: "Therefore, I estimate an 85% probability that box B contains the diamond."

Of what is this handwriting evidence? Trace the chain of causation: it runs through all the signs and portents of the boxes, and in worlds with different portents, a different probability is written. This handwriting is entangled with the actual evidence and thus with the actual contents. The clever arguer's handwriting tells you about the box-owner's finances; the curious inquirer's handwriting tells you about the boxes.

The two sheets of paper may contain very similar English words. Yet they carry radically different information — and someone who reads them aloud as if they were equivalent has missed everything.

Your Effectiveness as a Rationalist

The essay then applies this to first-person reasoning. If your car makes metallic squealing when you brake, and you don't want to pay for new brakes, you can search for reasons why the car might not need fixing. But your survival rate across possible worlds is determined by the algorithm that decided which conclusion to seek arguments for — not by the arguments themselves. If the real algorithm is "Never repair anything expensive," then the real bottom line was written before you assembled a single supporting reason. Good algorithm or bad, the arguments above it won't change the outcome.

The Caveat

Yudkowsky closes with a crucial warning: this is not a Fully General Counterargument. To say "my opponent is a clever arguer" can itself be a motivated move — paying yourself to keep whatever beliefs you started with. The world's cleverest arguer might still point out, correctly, that the sun is shining. The lesson is a discipline for one's own reasoning, not a rhetorical trump card against conclusions one dislikes.

狡辩者的那张纸

Yudkowsky 以一个生动的拍卖场景开场。两个密封箱子,A 与 B,其中一个装着钻石。一位受雇的辩手——由出价更高的箱主付钱——拿出一张纸。在检视任何证据之前,他先在纸的底部写下「因此,B 箱里有钻石」。然后在上方填满内容:蓝色印章、表面光泽、重量——只挑有利于 B 的迹象,忽略一切有利于 A 的线索。最后,他将这张纸朗读给观众,仿佛这是经过深思熟虑的分析。

哲学上的关键转折:墨水落纸的那一刻,该结论与现实的证据性纠缠便已固定。 在一组可能世界(Everett 分支或 Tegmark 副本)中,B 箱实际装有钻石的客观频率是固定的。这个频率不会因为底线上方写了什么论证而改变。墨水不可擦除,箱子不会改变。墨水结论与箱中内容之间的相关性,在辩手选择为哪位箱主服务的那一刻便已确定。

Yudkowsky 借用认知心理学中的类比来强化这一点:当被试看到用红色油墨印刷的「Green」(绿)字并被问及颜色时,许多人会说「绿色」而非「红色」。会读字——知道单词的含义——实际上干扰了你。一个物理对象的真正意义在于它与其他事物的因果纠缠,而非你从中读出的含义。

好奇的探究者

与之形成对照的是另一个人,他真正出于好奇。他写下两个箱子的全部区别性迹象,运用自己的知识和概率论,然后在底部写道:「因此,我估计 B 箱有 85% 的概率装着钻石。」

这段笔迹是什么的证据?追溯因果链:它流经箱子的全部迹象与征兆,而在迹象不同的世界中,底部会写下不同的概率。这段笔迹与实际证据、进而与实际内容存在纠缠。 狡辩者的笔迹告诉你箱主的财力;好奇探究者的笔迹告诉你箱子的内容。

这两张纸也许包含非常相似的英文词句。然而它们承载着截然不同的信息——把它们当作等价物朗读出来的人,什么都没明白。

你作为理性主义者的效能

文章随后将此应用于第一人称推理。如果你的车刹车时发出金属摩擦声,而你不想花钱换刹车片,你可以去寻找「车可能不需要修」的理由。但你在可能世界中的存活率,由决定要为哪个结论寻找论据的那个算法决定——而非由论证本身决定。如果真正的算法是「绝不修任何昂贵的东西」,那么真正的底线在你整理出任何一条支持理由之前便已写就。算法好也好、坏也好,其上的论证不会改变结果。

警告

Yudkowsky 以一个至关重要的提醒收尾:这并非万能反驳。 说「我的对手是狡辩者」本身也可能是一种动机性操作——付钱给自己,让自己保留一开始就有的信念。世上最狡猾的辩手说「太阳正在照耀」,也可能是对的。这个教训是对自身推理的一种自律,而非对抗你不喜欢的结论的修辞王牌。

04

FAQ常见问答

Why does it matter when the conclusion is written?结论写下的时机为何重要?

Because the physical act of writing fixes the causal chain. Whatever the ink says is now correlated only with the factors that caused it to be written — not with subsequent arguments. Arguments assembled after the conclusion was chosen cannot retroactively change what the conclusion is tracking in the world.

因为书写这一物理行为固定了因果链。墨水写下的内容,此后只与导致其被写下的因素相关——而非与随后的论证相关。在结论已选定之后才整理的论证,无法追溯性地改变该结论在世界中追踪的是什么。

What is "evidential entanglement" exactly?「证据性纠缠」究竟是什么意思?

Yudkowsky uses it to mean: across all possible worlds, does the presence of this ink (or belief, or signal) covary with the fact it supposedly indicates? The curious inquirer's probability estimate covaries with the boxes' actual contents because it was causally produced by inspecting those contents. The clever arguer's conclusion does not — it was produced by the auction bidding.

Yudkowsky 用它来表示:在所有可能世界中,这段墨水(或信念、或信号)的存在,是否与它本应指示的事实协变?好奇探究者的概率估算与箱子的实际内容协变,因为它是通过检视这些内容因果产生的。狡辩者的结论则不然——它是由拍卖出价产生的。

How does the car-brakes example extend the box analogy?刹车的例子如何延伸了箱子的类比?

The box scenario is abstract; the car example makes it personal. When you don't want to pay for brake repairs, the real algorithm selecting your conclusion is "avoid expensive repairs" — not "figure out whether the brakes are safe." All the rationalizations assembled afterward ("it's probably just dust," "brakes make noise sometimes") are the lines above the bottom line. Your survival odds depend on the hidden algorithm, not those assembled lines.

箱子场景是抽象的;刹车例子使其贴近个人。当你不想花钱修刹车时,选择你的结论的真实算法是「避免昂贵的修理」——而非「弄清刹车是否安全」。此后堆砌的所有合理化(「可能只是灰尘」「刹车偶尔会响」)都是底线上方的那些行。你的存活概率取决于那个隐藏的算法,而非那些堆砌的文字。

Doesn't this mean you can dismiss any argument by calling it motivated reasoning?这是否意味着你可以通过指责对方「动机性推理」来驳斥任何论证?

Yudkowsky explicitly warns against this. Calling your opponent a "clever arguer" can itself be motivated — a way of paying yourself to keep existing beliefs. The world's cleverest arguer may still be correct. The test is meant as a discipline for your own reasoning process, not as a trump card to dismiss conclusions you dislike.

Yudkowsky 明确对此提出警告。指控对手是「狡辩者」本身也可能是动机性操作——一种付钱给自己以保留既有信念的方式。世上最狡猾的辩手说的话也可能是正确的。这个检验旨在作为你自己推理过程的自律,而非用来驳斥你不喜欢的结论的王牌。

What should you do differently if you suspect you're being a clever arguer?如果你怀疑自己正在扮演狡辩者,应该怎么做?

Ask: what algorithm actually generated my bottom line? If the conclusion came first and the search for reasons followed, treat the resulting argument with deep suspicion. The corrective is to genuinely consider the other box — to survey all evidence, including evidence against the conclusion you prefer — before committing to any bottom line.

问问自己:究竟是什么算法产生了我的底线?如果结论先行、寻找理由在后,就要以高度怀疑的态度对待由此产生的论证。纠正方法是真正考虑另一个箱子——在确定任何底线之前,先审视全部证据,包括对你所偏好之结论不利的证据。

Is this the same thing as confirmation bias?这与确认偏误是同一回事吗?

It overlaps with confirmation bias but is more fundamental. Confirmation bias describes selectively seeking supporting evidence after a belief forms. Yudkowsky's point is that the very origin of a conclusion — which algorithm generated it — determines its epistemic worth. Even a conclusion reached without obvious bias still needs to have been generated by truth-tracking reasoning, not by social, financial, or emotional pressures that happened to produce a belief.

两者有重叠,但更为根本。确认偏误描述的是在信念形成之后有选择地寻求支持性证据。Yudkowsky 的论点在于:一个结论的起源——是什么算法产生了它——决定了它的认识论价值。即便是没有明显偏误地得出的结论,也仍需由追踪真相的推理产生,而非由碰巧产生某一信念的社会、财务或情感压力产生。

05

In-depth Analysis · Pros & Cons深入解读 · 优缺点

"The Bottom Line" is a short, precise essay that identifies a structural — rather than merely psychological — defect in motivated reasoning: conclusions written before evidence is surveyed are physically incapable of tracking reality, regardless of how compelling the assembled arguments sound. It is one of Yudkowsky's cleanest analytical moves in the Sequences.

《底线》是一篇简短而精确的文章,它指出了动机性推理中一个结构性的——而非仅仅心理性的——缺陷:在审视证据之前写下的结论,在物理上就无法追踪现实,无论随后整理的论证听起来多么令人信服。这是 Yudkowsky 在「序列」中最干净利落的分析之一。

Strengths亮点 / 优点
  • The causal-entanglement framing is genuinely novel
    因果纠缠的框架确实新颖
    Rather than saying motivated reasoning 'feels wrong,' Yudkowsky gives a structural account: ink's evidential worth depends on what caused it to be written. This reframing survives regardless of how polished the motivated argument sounds.
    Yudkowsky 没有简单说动机性推理「感觉不对」,而是给出一个结构性说明:墨水的证据价值取决于是什么导致它被写下。无论动机性论证听起来多么精致,这一重构都成立。
  • The two-arguer contrast is vivid and memorable
    两种辩手的对比生动而易记
    Placing the clever arguer and the curious inquirer side-by-side — producing superficially similar text with radically different epistemic worth — makes the abstract point viscerally clear.
    将狡辩者与好奇探究者并置——产生表面相似却在认识论价值上截然不同的文本——使这个抽象论点具体可感。
  • Applies cleanly to first-person reasoning
    干净地适用于第一人称推理
    The car-brakes extension shows this isn't just a theory about professional arguers — the "real algorithm" behind any personal belief is what actually determines your epistemic fate.
    刹车的延伸案例表明,这不仅仅是关于职业辩手的理论——任何个人信念背后的「真实算法」才是真正决定你认识论命运的东西。
  • Closes its own loophole
    堵上了自身的漏洞
    The warning that 'my opponent is a clever arguer' can itself be motivated reasoning shows unusual intellectual honesty — the essay anticipates its own misuse.
    关于「我的对手是狡辩者」本身也可能是动机性推理的警告,体现了罕见的智识诚实——文章预见了自身可能被滥用的方式。
Limits & Critiques局限 / 批评
  • The 'fixing' metaphor may be too clean
    「固定」的隐喻可能过于干净
    In practice, the bottom line is rarely written at a single moment; beliefs form gradually, arguments and intuitions interleave, and partial updating is common. The sharp ink-at-one-moment framing doesn't map neatly onto the fuzzy, iterative nature of real belief formation.
    在实践中,底线很少在单一时刻写就;信念逐渐形成,论证与直觉交织,局部更新很常见。「在某一刻落笔」这一清晰的框架,并不整齐地对应真实信念形成的模糊、迭代性质。
  • Everett/Tegmark as background furniture
    将 Everett/Tegmark 当作背景装饰
    The essay's key argument about correlation across possible worlds rests on many-worlds or modal-realist assumptions that are themselves contested. Readers who don't accept those frameworks may find the 'entanglement is fixed' claim harder to assess.
    文章关于可能世界间相关性的核心论证,依赖于多世界诠释或模态实在论假设,而这些本身就有争议。不接受这些框架的读者,可能会发现「纠缠已固定」的说法更难评估。
  • The test is easier to apply retrospectively than in real time
    这个检验事后应用比实时应用容易得多
    Asking 'what algorithm really wrote my bottom line?' is psychologically very hard when you're inside the reasoning — motivated reasoners typically do not experience themselves as having chosen a conclusion first. The essay diagnoses the problem more than it provides a practical correction procedure.
    当你深陷推理之中时,问「究竟是什么算法写下了我的底线?」在心理上非常困难——动机性推理者通常不会感觉到自己是先选了结论。文章诊断问题的力度,远大于它提供实际纠正程序的力度。
  • Understates legitimate adversarial argumentation
    低估了正当的对抗性论证
    Some contexts — legal advocacy, debate, red-teaming — deliberately assign someone to argue for a conclusion. The essay's framing implies this always corrupts the ink's evidential value, but adversarial processes can serve epistemic goods at a system level even when each arguer has a fixed bottom line.
    某些情境——法律辩护、辩论、红队演练——刻意安排人为某个结论辩护。文章的框架暗示这总是会污染墨水的证据价值,但对抗性流程即便每位辩手都有固定底线,也可能在系统层面服务于认识论的善。
Bottom line
总评

A tightly argued, genuinely useful essay that reframes motivated reasoning as a structural rather than merely psychological failure. Its deepest contribution is the insight that it is the origin algorithm, not the assembled arguments, that determines epistemic worth — a point that survives even when the motivated argument happens to reach the correct conclusion. Its main limitation is that the elegant 'moment-of-writing' framing oversimplifies the messy, iterative reality of belief formation, and the essay offers diagnosis more than cure.

一篇论证严密、切实有用的文章,它将动机性推理重构为一种结构性而非单纯心理性的失败。其最深刻的贡献在于洞见:决定认识论价值的是起源算法,而非堆砌的论证——即便动机性论证碰巧得出正确结论,这一点仍然成立。其主要局限在于,优雅的「书写时刻」框架过度简化了信念形成的混乱、迭代现实,且文章诊断的力度远大于提供解药的力度。

06

Original Text原文

There are two sealed boxes up for auction, box A and box B. One and only one of these boxes contains a valuable diamond. There are all manner of signs and portents indicating whether a box contains a diamond; but I have no sign which I know to be perfectly reliable. There is a blue stamp on one box, for example, and I know that boxes which contain diamonds are more likely than empty boxes to show a blue stamp. Or one box has a shiny surface, and I have a suspicion—I am not sure—that no diamond-containing box is ever shiny.

Now suppose there is a clever arguer, holding a sheet of paper, and they say to the owners of box A and box B: “Bid for my services, and whoever wins my services, I shall argue that their box contains the diamond, so that the box will receive a higher price.” So the box-owners bid, and box B’s owner bids higher, winning the services of the clever arguer.

The clever arguer begins to organize their thoughts. First, they write, “And therefore, box B contains the diamond!” at the bottom of their sheet of paper. Then, at the top of the paper, the clever arguer writes, “Box B shows a blue stamp,” and beneath it, “Box A is shiny,” and then, “Box B is lighter than box A,” and so on through many signs and portents; yet the clever arguer neglects all those signs which might argue in favor of box A. And then the clever arguer comes to me and recites from their sheet of paper: “Box B shows a blue stamp, and box A is shiny,” and so on, until they reach: “and therefore, box B contains the diamond.”

But consider: At the moment when the clever arguer wrote down their conclusion, at the moment they put ink on their sheet of paper, the evidential entanglement of that physical ink with the physical boxes became fixed.

It may help to visualize a collection of worlds—Everett branches or Tegmark duplicates—within which there is some objective frequency at which box A or box B contains a diamond.^1^

The ink on paper is formed into odd shapes and curves, which look like this text: “And therefore, box B contains the diamond.” If you happened to be a literate English speaker, you might become confused, and think that this shaped ink somehow meant that box B contained the diamond. Subjects instructed to say the color of printed pictures and shown the word Green in red ink often say “green” instead of “red.” It helps to be illiterate, so that you are not confused by the shape of the ink.

To us, the true import of a thing is its entanglement with other things. Consider again the collection of worlds, Everett branches or Tegmark duplicates. At the moment when all clever arguers in all worlds put ink to the bottom line of their paper—let us suppose this is a single moment—it fixed the correlation of the ink with the boxes. The clever arguer writes in non-erasable pen; the ink will not change. The boxes will not change. Within the subset of worlds where the ink says “And therefore, box B contains the diamond,” there is already some fixed percentage of worlds where box A contains the diamond. This will not change regardless of what is written in on the blank lines above.

So the evidential entanglement of the ink is fixed, and I leave to you to decide what it might be. Perhaps box owners who believe a better case can be made for them are more liable to hire advertisers; perhaps box owners who fear their own deficiencies bid higher. If the box owners do not themselves understand the signs and portents, then the ink will be completely unentangled with the boxes’ contents, though it may tell you something about the owners’ finances and bidding habits.

Now suppose another person present is genuinely curious, and they first write down all the distinguishing signs of both boxes on a sheet of paper, and then apply their knowledge and the laws of probability and write down at the bottom: “Therefore, I estimate an 85% probability that box B contains the diamond.” Of what is this handwriting evidence? Examining the chain of cause and effect leading to this physical ink on physical paper, I find that the chain of causality wends its way through all the signs and portents of the boxes, and is dependent on these signs; for in worlds with different portents, a different probability is written at the bottom.

So the handwriting of the curious inquirer is entangled with the signs and portents and the contents of the boxes, whereas the handwriting of the clever arguer is evidence only of which owner paid the higher bid. There is a great difference in the indications of ink, though one who foolishly read aloud the ink-shapes might think the English words sounded similar.

Your effectiveness as a rationalist is determined by whichever algorithm actually writes the bottom line of your thoughts. If your car makes metallic squealing noises when you brake, and you aren’t willing to face up to the financial cost of getting your brakes replaced, you can decide to look for reasons why your car might not need fixing. But the actual percentage of you that survive in Everett branches or Tegmark worlds—which we will take to describe your effectiveness as a rationalist—is determined by the algorithm that decided which conclusion you would seek arguments for. In this case, the real algorithm is “Never repair anything expensive.” If this is a good algorithm, fine; if this is a bad algorithm, oh well. The arguments you write afterward, above the bottom line, will not change anything either way.

This is intended as a caution for your own thinking, not a Fully General Counterargument against conclusions you don’t like. For it is indeed a clever argument to say “My opponent is a clever arguer,” if you are paying yourself to retain whatever beliefs you had at the start. The world’s cleverest arguer may point out that the Sun is shining, and yet it is still probably daytime.

^1^Max Tegmark, “Parallel Universes,” in Science and Ultimate Reality: Quantum Theory, Cosmology, and Complexity, ed. John D. Barrow, Paul C. W. Davies, and Charles L. Harper Jr. (New York: Cambridge University Press, 2004), 459–491, http://arxiv.org/abs/astro-ph/0302131.

有两个密封的箱子正在拍卖,箱子 A 和箱子 B。这两个箱子中,有且只有一个装着一颗珍贵的钻石。有各种各样的迹象和征兆,可以指示一个箱子是否含有钻石;但我没有任何我知道完全可靠的迹象。比如其中一个箱子上有一枚蓝色印章,而我知道,含有钻石的箱子比空箱子更有可能出现蓝色印章。又或者一个箱子的表面很光亮,而我有一种感觉——虽然并不确定——没有任何含有钻石的箱子是光亮的。

现在,假设有一位狡辩者,手拿一张纸,对 A 箱和 B 箱的主人说:「来竞标我的服务吧,谁赢得了我的服务,我就为谁的箱子辩护,说那个箱子里装着钻石,这样那个箱子就能卖出更高的价格。」于是箱主们开始出价,B 箱的主人出价更高,赢得了狡辩者的服务。

狡辩者开始整理思路。首先,他在那张纸的底部写道:「因此,B 箱里装着钻石!」然后,在纸的顶部,狡辩者写道:「B 箱上有一枚蓝色印章」,在它下面写道:「A 箱的表面是光亮的」,再下面写道:「B 箱比 A 箱轻」,如此写下许多迹象和征兆;然而狡辩者忽略了所有可能有利于 A 箱的迹象。然后,狡辩者来到我面前,从他的纸上朗读道:「B 箱上有一枚蓝色印章,而 A 箱的表面是光亮的」,如此这般,直到他读到:「因此,B 箱里装着钻石。」

但请想一想:在狡辩者写下他的结论的那一刻,在他将墨水留在纸上的那一刻,那段实物墨水与实物箱子之间的证据性纠缠便已固定。

将一组世界——Everett 分支或 Tegmark 副本——可视化,也许会有所帮助;在这些世界中,A 箱或 B 箱装有钻石存在某种客观频率。^1^

纸上的墨水被构成了奇特的形状和曲线,看起来像这样的文字:「因此,B 箱里装着钻石。」如果你碰巧是一个识字的英语使用者,你可能会感到困惑,以为这些有形状的墨水在某种程度上意味着 B 箱装有钻石。被要求说出印刷图片颜色的被试,当被展示用红色墨水印刷的「Green」(绿色)一词时,经常会说「绿色」而非「红色」。不识字会有所帮助,这样你就不会被墨水的形状所迷惑。

对我们而言,一件事物的真正意义在于它与其他事物的纠缠。再次设想那组世界,Everett 分支或 Tegmark 副本。在所有世界中的所有狡辩者都将墨水落于他们纸张底线的那一刻——假设这是同一个时刻——墨水与箱子的相关性便已固定。狡辩者用的是不可擦除的钢笔;墨水不会改变。箱子不会改变。在那些墨水写着「因此,B 箱里装着钻石」的世界的子集中,A 箱装有钻石的世界已经占有某个固定的百分比。无论在上面的空白行上写下什么,这一点都不会改变。

因此,墨水的证据性纠缠已然固定,我留待你来判断它可能是什么。也许那些相信自己能为箱子作出更好论证的箱主,更倾向于雇用推销员;也许那些担心自身缺陷的箱主出价更高。如果箱主们自己也不了解那些迹象和征兆,那么墨水将与箱子的内容完全无关,尽管它可能告诉你一些关于箱主们的财力和出价习惯的信息。

现在假设在场的另一个人是真正好奇的,他们首先两个箱子的所有区别性迹象写在一张纸上,然后运用自己的知识和概率定律,在底部写道:「因此,我估计 B 箱有 85% 的概率装着钻石。」这段笔迹是什么的证据?检视导致这段实物墨水出现在实物纸上的因果链,我发现这条因果链蜿蜒穿过箱子的所有迹象和征兆,并依赖于这些迹象;因为在迹象不同的世界中,底部会写下不同的概率。

因此,好奇探究者的笔迹与迹象、征兆以及箱子的内容纠缠在一起,而狡辩者的笔迹只是哪位箱主出价更高的证据。墨水的指示意义大相径庭,尽管愚蠢地朗读那些墨水形状的人可能会觉得那些英文词语听起来相似。

你作为一名理性主义者的效能,由实际上写下你思维底线的那个算法决定。如果你的车在刹车时发出金属摩擦声,而你不愿意面对更换刹车片的经济代价,你可以决定寻找你的车可能不需要修理的理由。但在 Everett 分支或 Tegmark 世界中存活的你的实际百分比——我们将其视为对你作为理性主义者之效能的描述——由那个决定你将为哪个结论寻求论据的算法决定。在这种情况下,真正的算法是「绝不修理任何昂贵的东西」。如果这是一个好算法,那很好;如果这是一个坏算法,那也无奈。你事后写在底线上方的论证,无论如何都不会改变任何事情。

这是对你自己思维的一个警示,而非一个针对你不喜欢之结论的万能反驳。因为如果你在付钱给自己以保留一开始就有的信念,那么说「我的对手是狡辩者」确实也是一个狡辩的论证。世上最狡猾的辩手可能会指出太阳正在照耀,而那仍然很可能是白天。

^1^ Max Tegmark,《平行宇宙》,收录于 Science and Ultimate Reality: Quantum Theory, Cosmology, and Complexity,John D. Barrow、Paul C. W. Davies、Charles L. Harper Jr. 编(纽约:剑桥大学出版社,2004),459–491,http://arxiv.org/abs/astro-ph/0302131