Concise Summary简洁概述
Rationality prizes closeness-to-the-issue: the nearer your evidence is to the actual question, the more it screens off weaker, more distant arguments. The Wright Brothers versus Lord Kelvin illustrates this perfectly — Kelvin's authority collapses the moment you can check the calculations, and the calculations become irrelevant the moment you watch the plane fly. A theorem of causal graphs formalises the intuition: a closer node can never carry less information than a more distant one that it screens off. So whenever you can, stop arguing about credentials, rationality, or bias accusations, and get as close to the raw question as possible — hug the query.
理性推崇贴近问题：你的证据越靠近真正的问题，它能屏蔽的那些更弱、更遥远的论证就越多。莱特兄弟对阵开尔文勋爵完美地说明了这一点——一旦你能核查计算，权威就土崩瓦解；一旦你亲眼看见飞机飞起来，计算本身也不再重要。因果图的一条定理将这一直觉形式化：较近的节点所携带的信息，永远不少于被其屏蔽的更远节点。所以，每当可能，就停止争论资历、理性程度或偏见指控，尽量贴近原始问题——紧贴问题本身。
Infographic信息图
Closeness-to-the-issue
贴近问题
The discipline of seeking evidence as near to the original question as possible, so it screens off weaker, more indirect arguments.
尽量寻找贴近原始问题的证据，从而屏蔽那些更弱、更间接的论证。
Wright Brothers vs. Lord Kelvin
莱特兄弟 vs. 开尔文勋爵
Authority loses to calculations; calculations lose to watching the plane fly. Each step closer to the question trumps the previous layer.
权威输给计算；计算输给亲眼看见飞机飞起来。每靠近问题一步，就胜过上一层证据。
The Great Web of Causality
因果之网
A theorem of causal graphs: you can never get more information from distant nodes than from strictly closer nodes that screen them off.
因果图的定理：你从遥远节点获得的信息，永远不会多于屏蔽它的较近节点。
Bias accusations as a trap
偏见指控的陷阱
Cataloguing someone's biases is seductive but can't settle a factual question — it is just more indirect argument drifting away from the query.
列举对方的偏见很诱人，但无法解决事实问题——它只是更多间接论证，正在远离真正的问题。
Arguing rationality is still indirect
争论理性本身仍是间接
Even asking "who is more rational?" is a step removed from the object level. The virtue of a rationalist cannot directly cause a plane to fly.
即便问「谁更理性」，也是偏离对象层的一步。理性主义者的美德，直接上无法让飞机飞起来。
Detailed Summary详细概述
The Discipline of Closeness-to-the-Issue
Yudkowsky opens by naming a principle that operates throughout rationalist practice: closeness-to-the-issue. The discipline is simple to state — try to observe evidence as near to the original question as possible, so it screens off as many other arguments as possible — but surprisingly easy to violate.
The Wright Brothers Illustration
The essay's central example unfolds in three stages:
- Authority: The Wright Brothers are bicycle mechanics who happen to be excellent amateur physicists. Lord Kelvin is the far greater authority. If you compare authority, Kelvin wins.
- Calculations: If you demand to see the Wright Brothers' calculations, and you can follow them, and you ask for Kelvin's (he probably has none beyond his own incredulity), then authority becomes much less relevant.
- Direct observation: If you actually watch the plane fly, the calculations themselves become moot for many purposes, and Kelvin's authority is not even worth considering.
Each step closer to the question trumps the previous layer entirely.
The Causal-Graph Theorem
Yudkowsky anchors the intuition in a formal principle: the more directly your arguments bear on a question — the closer the observed nodes are to the queried node in the Great Web of Causality — the more powerful the evidence. It is a theorem of causal graphs that you can never get more information from distant nodes than from strictly closer nodes that screen off the distant ones.
Jerry Cleaver's coaching aphorism captures the same point: "What does you in is not failure to apply some high-level, intricate, complicated technique. It's overlooking the basics. Not keeping your eye on the ball."
Arguing Physics Beats Arguing Rationality
An important corollary: just as it is superior to argue physics than credentials, it is also superior to argue physics than rationality. Who was more rational, the Wright Brothers or Lord Kelvin? If we can check their calculations, we don't have to care. The virtue of a rationalist cannot directly cause a plane to fly.
The Bias-Catalogue Trap
Learning about cognitive biases is useful — until it isn't. It becomes a trap when knowing fifty biases leads you to accuse opponents of Bias #182 rather than engaging the actual evidence. You can't settle a factual issue with generic accusations of irrationality. And crucially: if there are biased reasons to say the Sun is shining, that doesn't make it dark outside. The bias charge is a more distant node; the sky itself is a closer one.
Honest Acknowledgement of Limits
Yudkowsky acknowledges that you can't always experiment, can't always check the calculations, sometimes lack background material, and sometimes face private information or time pressure. There are, he admits, a sadly large number of times when judging the speaker's rationality is the best available move. But you should always do it with a hollow feeling — a sense that something is missing — because something genuinely is.
The Prescription
The closing injunction: whenever you can, dance as near to the original question as possible — press yourself up against it — get close enough to hug the query.
贴近问题的功课
Yudkowsky 开篇就点名一个贯穿理性实践的原则：贴近问题。这个功课说来简单——尽量寻找贴近原始问题的证据，使其能屏蔽尽可能多的其他论证——但出乎意料地容易被违反。
莱特兄弟的例证
文章的核心例子分三个层次展开：
- 权威：莱特兄弟是自行车修理工，碰巧也是出色的业余物理学家。开尔文勋爵的权威高出一大截。如果你比较权威，开尔文获胜。
- 计算：如果你要求看莱特兄弟的计算，并且能看懂，同时要求看开尔文的计算（他大概除了自己的难以置信之外什么都没有），那么权威的相关性就大为降低。
- 直接观察：如果你真的亲眼看见飞机飞了起来，计算在很多意义上已经无关紧要，开尔文的权威甚至不值一提。
每靠近问题一步，就完全胜过上一层。
因果图定理
Yudkowsky 将这一直觉锚定在一个形式化原则上：你的论证越直接地指向问题——因果之网中观察节点越靠近被查询节点——证据越有力。因果图有一条定理：从遥远节点获得的信息，永远不会多于屏蔽它的较近节点。
Jerry Cleaver 的写作指导格言抓住了同一要点：「让你失手的，不是没能运用某种高层次、错综复杂的精妙技巧，而是忽视了基本功。没有盯着球。」
论物理胜过论理性
一个重要推论：正如论物理胜过论资历，论物理也胜过论理性。莱特兄弟和开尔文勋爵，谁更理性？如果我们能核查他们的计算，就根本不需要在乎！理性主义者的美德，直接上无法让飞机飞起来。
偏见清单的陷阱
学习认知偏见很有用——直到它不再有用。它变成陷阱，是当你知道了五十种偏见之后，开始指控对手犯了「你偏见库里第182号偏见」，而不是去回应实际的证据。泛泛的非理性指控无法解决事实问题。关键一点：如果存在有偏见的理由说太阳正在照耀，这并不让外面变成黑夜。偏见指控是一个更远的节点；天空本身才是更近的那个。
诚实地承认局限
Yudkowsky 承认，你不能总是做实验，不能总是核查计算，有时缺乏背景知识，有时面临私有信息或时间压力。他坦承，确实有相当多的时候，评判说话者的理性是当下最好的选择。但你应当始终带着一种空洞的感觉去这样做——一种「有什么东西缺失了」的感觉——因为确实缺失了什么。
处方
结尾的叮嘱：每当可能，就尽量靠近原始问题起舞——把自己贴上去——靠得足以紧贴问题本身。
FAQ常见问答
What exactly is "closeness-to-the-issue" and why does it matter?「贴近问题」究竟是什么，为什么它重要？
It is the discipline of seeking evidence that bears as directly as possible on the question being asked. It matters because causal-graph theory guarantees that a nearer node is never less informative than a more distant one it screens off — so indirect evidence is always a downgrade from direct evidence when direct evidence is available.
这是一门功课：寻找尽可能直接指向所提问题的证据。它之所以重要，是因为因果图理论保证了较近节点的信息量永远不少于被其屏蔽的更远节点——所以当直接证据可得时，间接证据永远是降级版。
Why does watching the plane fly trump the calculations, and the calculations trump authority?为什么亲眼看见飞机飞起来胜过计算，计算又胜过权威？
Each layer screens off the previous one. If you can see the plane in flight, you no longer need to infer whether it will fly from intermediate evidence like calculations or the reputation of the people who designed it — the flight itself is the most direct possible evidence on whether the plane flies.
每一层都屏蔽了上一层。如果你能看见飞机在飞，你就不再需要通过计算或设计者的声望这些中间证据来推断飞机是否能飞——飞行本身就是关于飞机能否飞起来的最直接可能证据。
How can learning about biases hurt you?学习偏见知识怎么会伤害你？
It hurts when it becomes a substitute for engaging with evidence. Accusing someone of exhibiting Bias #182 is an argument about the speaker, not about the question at hand — it's a more distant node. If you replace direct evidence with bias-cataloguing, you've drifted away from the query, not toward it.
当它变成回避证据的替代品时就会伤害你。指控某人展现了第182号偏见，是关于说话者的论证，而非关于手头问题的论证——它是一个更远的节点。如果你用偏见清单来替代直接证据，你正在偏离问题，而非靠近它。
Is the essay saying we should never judge someone's reasoning process?这篇文章是说我们永远不该评判他人的推理过程吗？
No. Yudkowsky explicitly acknowledges there are many situations where judging the speaker's rationality is the best available move — when you lack background knowledge, face private information, or have no time. His point is that you should recognise this as a second-best, not a triumph, and carry a "hollow feeling" that direct evidence is still missing.
不是。Yudkowsky 明确承认，有很多情况下评判说话者的理性是当下最佳选择——当你缺乏背景知识、面对私有信息或没有时间时。他的意思是：你应当认识到这是一种次优选择，而非胜利，并保持一种「直接证据仍然缺失」的空洞感。
What does "hug the query" actually mean in practice?「紧贴问题」在实践中究竟意味着什么？
It means: resist the pull toward indirect arguments (authority, rationality, bias accusations) and keep pushing toward the most direct evidence you can access. In a debate, ask yourself what evidence would most directly settle this specific question — not what makes the other person look wrong in general — and pursue that.
意思是：抵制被间接论证（权威、理性、偏见指控）吸引的冲动，不断推进到你能获取的最直接证据。在辩论中，问自己：什么证据能最直接地解决这个具体问题——而不是什么能让对方笼统地显得错误——然后去追那个。
Does the causal-graph theorem guarantee that direct observation is always available?因果图定理是否保证直接观察永远可得？
No, and the essay doesn't claim this. The theorem says that when closer nodes exist and are accessible, they dominate. Yudkowsky honestly notes that sometimes you can't experiment, can't check the calculations, or face other obstacles. The principle tells you where to aim, not that you'll always reach the target.
不，文章也没有这样宣称。定理说的是：当较近节点存在且可获取时，它们占主导。Yudkowsky 诚实地指出，有时你无法做实验，无法核查计算，或面临其他障碍。这个原则告诉你该往哪里瞄准，而非你总能到达目标。
In-depth Analysis · Pros & Cons深入解读 · 优缺点
This short essay distills a single, highly applicable principle from epistemology and causal-graph theory into a memorable prescription: get as close to the actual question as you can before settling for indirect evidence. Its brevity (436 words) belies its density — each paragraph carries real argumentative work.
这篇短文将认识论与因果图理论中的一个高度实用的原则提炼成一条令人难忘的处方：在退而求其次接受间接证据之前，尽量靠近真正的问题。它的简短（436词）掩盖了其密度——每段话都承载着真实的论证工作。
- Concrete, graduated example具体而层级分明的例子The Wright Brothers/Kelvin trio (authority → calculations → watching the flight) makes the abstract hierarchy of evidence viscerally clear and immediately memorable.莱特兄弟/开尔文三部曲（权威 → 计算 → 亲眼看见飞行）让抽象的证据层级变得直观清晰，令人过目难忘。
- Formal backing形式化的支撑The appeal to causal-graph theory elevates the principle beyond heuristic: it is a theorem, not just a tip, that closer nodes dominate screened-off distant ones.援引因果图理论将这一原则提升到超出经验法则的层面：较近节点主导被屏蔽的远节点，这是一条定理，不只是一条建议。
- Non-obvious corollary about biases关于偏见的非显然推论The insight that learning more biases can hurt you by distracting from direct evidence is surprising and important, cutting against the usual advice to catalogue biases.「学习更多偏见可能会伤害你」这一洞见——因为它分散了你对直接证据的注意力——令人意外且重要，与通常「多积累偏见清单」的建议背道而驰。
- Emotionally resonant closing情感共鸣的结尾"Hug the query" is a vivid, physical image that turns an epistemological norm into an affectionate, embodied practice rather than a cold rule.「紧贴问题」是一个生动、肉身化的意象，把一条认识论规范变成了一种亲切、具身的实践，而非一条冰冷的规则。
- The theorem is cited but not stated precisely定理被援引但未被精确陈述Yudkowsky references a "theorem of causal graphs" without naming it (it is roughly the d-separation / screening-off theorem in Bayesian networks). A reader cannot verify or apply it without filling in background the essay leaves implicit.Yudkowsky 援引了「因果图的一条定理」却未点名（大致是贝叶斯网络中的 d-分隔/屏蔽定理）。读者无法验证或应用它，除非自己补充文章隐含的背景知识。
- Ambiguity about when indirect evidence is legitimate对间接证据何时合理的模糊处理The essay says to fall back on judging rationality only with a "hollow feeling," but gives little guidance on how to weight indirect evidence when direct evidence is impossible — a common real-world situation that needs more than a disclaimer.文章说退而评判理性时应保持「空洞感」，但对于当直接证据不可能时如何权衡间接证据，几乎没有给出指引——这是一种常见的现实情境，需要的不只是一句免责声明。
- Conflates two distinct tasks混淆了两个不同任务"Hug the query" is good advice for settling factual questions, but in practical argumentation you often also need to assess source reliability (which is a form of indirect evidence). The essay treats the latter as purely a fallback rather than a legitimate parallel track.「紧贴问题」对解决事实问题是好建议，但在实际论证中你往往同时需要评估信源可靠性（这是间接证据的一种）。文章把后者视为纯粹的退路，而非合理的平行轨道。
- Short length limits scope篇幅短限制了范围At 436 words the essay can only sketch the principle. Edge cases — testimony as legitimate near-query evidence, cases where observation is itself theory-laden, the costs of "hugging" a query that is malformed — go unaddressed.仅 436 词，文章只能勾勒原则。边界情况——证词作为合理的近查询证据、观察本身负载理论的情形、「紧贴」一个构造不良问题的代价——均未涉及。
A crisp, memorable distillation of a genuinely important principle: when direct evidence is available, seize it rather than retreating to authority, biography, or meta-level rationality debates. The Wright Brothers example alone is worth the read. Its main limitation is the one inherent in its brevity: it tells you where to aim without fully equipping you to navigate the many real-world cases where aiming there is impossible.
对一个真正重要原则的精炼、令人难忘的提炼：当直接证据可得时，抓住它，而不是退守权威、传记或元层次的理性争论。光是莱特兄弟的例子就值得一读。它的主要局限是其简短固有的那种：它告诉你该往哪里瞄准，却没有充分装备你去应对众多现实中无法瞄到那里的情况。
Original Text原文
In the art of rationality there is a discipline of closeness-to-the-issue—trying to observe evidence that is as near to the original question as possible, so that it screens off as many other arguments as possible.
The Wright Brothers say, “My plane will fly.” If you look at their authority (bicycle mechanics who happen to be excellent amateur physicists) then you will compare their authority to, say, Lord Kelvin, and you will find that Lord Kelvin is the greater authority.
If you demand to see the Wright Brothers’ calculations, and you can follow them, and you demand to see Lord Kelvin’s calculations (he probably doesn’t have any apart from his own incredulity), then authority becomes much less relevant.
If you actually watch the plane fly, the calculations themselves become moot for many purposes, and Kelvin’s authority not even worth considering.
The more directly your arguments bear on a question, without intermediate inferences—the closer the observed nodes are to the queried node, in the Great Web of Causality—the more powerful the evidence. It’s a theorem of these causal graphs that you can never get more information from distant nodes, than from strictly closer nodes that screen off the distant ones.
Jerry Cleaver said: “What does you in is not failure to apply some high-level, intricate, complicated technique. It’s overlooking the basics. Not keeping your eye on the ball.”^1^
Just as it is superior to argue physics than credentials, it is also superior to argue physics than rationality. Who was more rational, the Wright Brothers or Lord Kelvin? If we can check their calculations, we don’t have to care! The virtue of a rationalist cannot directly cause a plane to fly.
If you forget this principle, learning about more biases will hurt you, because it will distract you from more direct arguments. It’s all too easy to argue that someone is exhibiting Bias #182 in your repertoire of fully generic accusations, but you can’t settle a factual issue without closer evidence. If there are biased reasons to say the Sun is shining, that doesn’t make it dark out.
Just as you can’t always experiment today, you can’t always check the calculations today.^2^ Sometimes you don’t know enough background material, sometimes there’s private information, sometimes there just isn’t time. There’s a sadly large number of times when it’s worthwhile to judge the speaker’s rationality. You should always do it with a hollow feeling in your heart, though, a sense that something’s missing.
Whenever you can, dance as near to the original question as possible—press yourself up against it—get close enough to hug the query!
^1^Jerry Cleaver, Immediate Fiction: A Complete Writing Course (Macmillan, 2004).
^2^See also “Is Molecular Nanotechnology ’Scientific’?” http://lesswrong.com/lw/io/is\_molecular\_nanotechnology_scientific.
在理性的艺术中，有一门贴近问题的功课——尽量寻找贴近原始问题的证据，使其能屏蔽尽可能多的其他论证。
莱特兄弟说：「我的飞机能飞。」如果你看的是他们的权威（两个碰巧是出色业余物理学家的自行车修理工），你就会把他们的权威与开尔文勋爵相比，会发现开尔文勋爵才是更大的权威。
如果你要求看莱特兄弟的计算，并且能看懂，同时要求看开尔文勋爵的计算（他大概除了自己的难以置信之外什么都没有），那么权威的相关性就大为降低。
如果你真的亲眼看见飞机飞了起来，计算在很多意义上已经无关紧要，开尔文的权威甚至不值得考虑。
你的论证越直接地指向问题，而不经过中间推断——因果之网中观察节点越靠近被查询节点——证据就越有力。这是因果图的一条定理：从遥远节点获得的信息，永远不会多于屏蔽它的严格意义上更近节点所携带的信息。
Jerry Cleaver 说过：「让你失手的，不是没能运用某种高层次、错综复杂的精妙技巧，而是忽视了基本功。没有盯着球。」^1^
正如论物理胜过论资历，论物理也胜过论理性。莱特兄弟和开尔文勋爵，谁更理性？如果我们能核查他们的计算，就根本不需要在乎！理性主义者的美德，直接上无法让飞机飞起来。
如果你忘记这个原则，学习更多偏见只会伤害你，因为它会让你分心，偏离更直接的论证。指控某人展现了你偏见库里第182号偏见，是再容易不过的事，但你无法仅凭这种指控就解决一个事实问题。如果说太阳正在照耀有偏见的理由，这并不让外面变成黑夜。
正如你不能总是今天就做实验，你也不能总是今天就核查计算。^2^ 有时你缺乏足够的背景知识，有时存在私有信息，有时就是没有时间。可悲地有相当多的时候，评判说话者的理性是值得做的事。你应当始终带着一种空洞的感觉去这样做，带着一种「有什么东西缺失了」的感觉。
每当可能，就尽量靠近原始问题起舞——把自己贴上去——靠得足以紧贴问题本身！
^1^Jerry Cleaver，《即时小说：完整写作课程》(Immediate Fiction: A Complete Writing Course)（麦克米伦出版社，2004）。
^2^另见《分子纳米技术是「科学的」吗？》http://lesswrong.com/lw/io/is\_molecular\_nanotechnology_scientific。