Concise Summary简洁概述
Bayesian evidence, legal evidence, and scientific evidence are three nested categories. A police commissioner's private tip that Wulky is a crime boss is genuine rational evidence — it shifts your probability estimate and you'd be foolish to ignore it — but courts won't admit it because legal evidence requires extra social standards to resist corruption. Scientific evidence is even more restrictive: it requires publicly reproducible generalizations that anyone can verify without trusting an authority. Historical knowledge and tomorrow's sunrise prediction are rational beliefs but not scientific ones in this strict sense. These categories don't compete; they serve different social purposes. Whatever label we attach, information in a closed-access journal or a commissioner's quiet word still counts as Bayesian evidence.
贝叶斯证据、法律证据、科学证据是三个嵌套范畴。警察局长私下告诉你沃尔基是黑帮老大，这是真实的理性证据——它改变了你的概率估计，忽视它的人才是傻瓜——但法庭不会采纳，因为法律证据需要额外的社会标准以抵御腐败。科学证据的要求则更严格：它需要任何人都能自行验证、无需信任权威的公开可重复的规律性知识。历史知识和「明天太阳会升起」的预测是理性信念，但在这种严格意义上并非科学陈述。这些范畴并不竞争，它们服务于不同的社会目的。无论我们贴什么标签，封闭期刊里的信息或局长的私下保证，仍然算作贝叶斯证据。
Infographic信息图
Bayesian evidence is the widest net
贝叶斯证据是最宽的网
Anything that gives you a lopsided likelihood ratio — a tip, a hunch, a private assurance — is rational evidence, even if no court or journal would touch it.
任何能给出不对称似然比的东西——小道消息、直觉、私下保证——都是理性证据，即便没有法庭或期刊愿意受理它。
Legal evidence adds social safeguards
法律证据附加了社会防护
Courts demand special standards because unrestrained power attracts corruption; the rules filter out illegitimate pathways to belief, not just weak evidence.
法庭要求特殊标准，因为不受约束的权力会滋生腐败；这些规则筛除的是通往信念的非法途径，而不仅仅是弱证据。
Science requires public reproducibility
科学要求公开可重复性
Scientific knowledge consists of generalizations anyone can test themselves; it excludes unique historical events and single future predictions — even if those are perfectly rational to believe.
科学知识由任何人都能亲自检验的规律性陈述组成；它排除了独特的历史事件和单一未来预测——即便相信那些预测完全合乎理性。
Tomorrow's sunrise is rational but not scientific
明天日出是理性的但非科学的
Predicting a future event from a theory is an extrapolation; labeling it a 'scientific belief' before the experiment risks removing the motivation to run it.
从理论外推未来事件是一种推断；在实验运行之前就称之为「科学信念」，有失去做实验之理由的风险。
Definitions are pragmatic choices
定义是务实选择
No mountain has p < 0.05 carved on it; the threshold for 'scientific' is a social decision we can revise — including toward open-access-only publication.
没有哪座山上刻着 p < 0.05；「科学」的阈值是一个社会决策，我们可以修改——包括将其限定为仅开放获取的出版物。
Detailed Summary详细概述
The Opening Puzzle: Wulky and the Police Commissioner
Yudkowsky opens with a vivid scenario: your friend the police commissioner tells you in strictest confidence that Wulky Wilkinsen is the city's crime kingpin. Should you believe it? As a rationalist, yes — it is strong Bayesian evidence that shifts your probability estimate, and only a fool would then go insult Wulky.
But a court won't imprison Wulky on that basis. The statement is not legal evidence. Why not? Because if you jailed everyone accused by a police commissioner, you'd initially catch many real criminals — but unchecked power attracts corruption. Over time, you'd jail more innocents and fewer actual bosses. So the legal system deliberately imposes extra social standards. All legal evidence should ideally be rational evidence, but not the other way around.
The Scientific Layer: Reproducibility as the Key
Yudkowsky then introduces a different kind of statement: he is wearing white socks at a specific time. Is that scientific? No. You can believe it; he could testify to it in court. But you cannot run an experiment yourself to verify it. Science consists of generalizations — claims about many instances, testable by anyone, without deferring to authority. Science is the publicly reproducible knowledge of humankind.
This reveals why historical knowledge, though reasonable to hold, is not scientific. Alexander the Great existed? A rational belief, yes — but we're dependent on Plutarch; we cannot verify it independently.
Tomorrow's Sunrise and the Experiment Problem
Here the essay makes its subtlest point. The Sun will rise tomorrow — surely that is scientific? No, not quite. It is a prediction extrapolated from scientific generalizations, based on testable models. But consider: if you label this prediction a "scientific belief" before running any confirming experiment, why bother to do the experiment? The label does premature epistemic closure. Science lives in the generalizations from which predictions flow, not in any particular predicted event.
"There is a 'conventional prediction' or 'theory Q's prediction.' But if you already know the 'scientific belief' about the result, why bother to run the experiment?"
Definitions Are Pragmatic Choices
Yudkowsky is explicit that the question "Is my definition of scientific knowledge true?" is not well-formed. The special standards we impose are pragmatic. p < 0.05 is written on no mountain; many argue it should be 0.01 or 0.001. Future generations might reasonably restrict the label "scientific" to open-access-only publications — because if knowledge costs $20,000/year to access, can it truly be the knowledge of humankind, and can results be fully trusted if critics must pay to engage?
The Unified Bottom Line
Whatever definitional box we use, the private commissioner's tip and the closed-access journal both continue to count as Bayesian evidence. Rational evidence is the universal category; legal and scientific evidence are socially curated subsets, each serving a distinct institutional purpose. Confusing the categories leads to two opposite errors: dismissing genuine evidence because it isn't "official," or over-trusting official labels as if they guaranteed truth.
开篇谜题：沃尔基与警察局长
Yudkowsky 以一个生动场景开场：你的朋友——警察局长——以极度保密的方式告诉你，沃尔基·威尔金森是这座城市的黑帮头目。你该相信吗？作为一名理性主义者，答案是肯定的——这是强有力的贝叶斯证据，改变了你的概率估计，只有傻瓜才会去侮辱沃尔基。
但法庭不会凭此拘押沃尔基。这条陈述不是法律证据。为什么？因为如果你把所有被警察局长指控的人都投入监狱，最初确实会抓住许多真正的罪犯——但不受约束的权力会滋生腐败。随着时间推移，你会关押更多无辜者，而越来越少的真正罪犯。所以法律体系故意设置了额外的社会标准。所有法律证据理想上都应该是理性证据，但反过来并不成立。
科学层面：可重复性是关键
Yudkowsky 随后引入另一种陈述：他在某个具体时间穿着白色袜子。这是科学陈述吗？不是。你可以相信它；他可以在法庭上作证。但你无法自己做实验加以验证。科学由规律性陈述构成——关于众多实例的主张，任何人都可以检验，无需信任权威。科学是人类公开可重复的知识。
这揭示了为什么历史知识尽管合理，却并非科学知识。亚历山大大帝真实存在？当然是理性信念——但我们依赖普鲁塔克，无法独立验证。
明天的日出与实验问题
这里文章提出了最微妙的一点。太阳明天会升起——这难道不是科学吗？不尽然。这是从科学规律外推出来的预测，基于可检验的模型。但想想：如果你在任何确认实验之前就把这个预测贴上「科学信念」的标签，为什么还要做实验？这个标签过早地封闭了认识论。科学活在产生预测的规律性陈述中，而不在任何特定的被预测事件里。
「有『传统预测』或『理论Q的预测』。但如果你已经知道关于结果的『科学信念』，为什么还要费心做实验？」
定义是务实选择
Yudkowsky 明确指出，「我对科学知识的定义是否正确？」这个问题本身就构型不良。我们强加的特殊标准是务实的。p < 0.05 没有刻在任何山上；许多人认为应该是 0.01 或 0.001。未来的世代可能合理地把「科学」标签限定于开放获取的出版物——因为如果知识每年需要 20,000 美元才能获取，它真的能算作人类的知识吗？如果批评者必须付费才能参与，结果真的可以被充分信任吗？
统一的底线
无论我们使用什么定义框架，局长的私下线报和封闭获取期刊都继续算作贝叶斯证据。理性证据是最普遍的范畴；法律证据和科学证据是经过社会筛选的子集，各自服务于不同的制度目的。混淆这些范畴会导致两种相反的错误：因为某事不够「官方」就忽视真实证据，或过度信任官方标签，仿佛它们能保证真理。
FAQ常见问答
If a tip is Bayesian evidence, why don't courts admit it?如果线报是贝叶斯证据，法庭为什么不采纳？
Courts impose extra standards not because the tip lacks informational value, but to prevent systematic abuse. Unrestrained power attracts corruption: a commissioner who can imprison anyone by accusation eventually will. Legal evidence standards are a social defense mechanism, not an epistemological ruling that tips are worthless.
法庭设置额外标准，不是因为线报缺乏信息价值，而是为了防止系统性滥权。不受约束的权力会滋生腐败：一个可以凭指控就关人的局长，最终一定会这么做。法律证据标准是一种社会防御机制，而非「线报毫无价值」的认识论裁决。
Why isn't 'The Sun will rise tomorrow' a scientific statement?为什么「明天太阳会升起」不是科学陈述？
It is a prediction extrapolated from scientific generalizations, not a generalization itself. More critically, labeling it a "scientific belief" before the experiment implies you already know the result — which undercuts the motivation to run the experiment at all. Science lives in the testable generalizations from which predictions flow, not in the individual predicted events.
它是从科学规律外推出来的预测，本身不是规律性陈述。更关键的是，在实验之前就把它称为「科学信念」，意味着你已经知道结果——这会削弱做实验的动机。科学活在产生预测的可检验规律中，而不在各个被预测的单一事件里。
Is historical knowledge worthless since it's not scientific?历史知识既然不是科学的，是否就毫无价值？
Not at all — and Yudkowsky doesn't say so. A rationalist is licensed to believe Alexander the Great existed; the evidence, though dependent on authorities like Plutarch, is far better than maximum entropy. The point is only that "scientific" is a special designation requiring independent verifiability, not a synonym for "credible."
完全不是——Yudkowsky 也没有这样说。理性主义者完全可以相信亚历山大大帝存在；相关证据虽然依赖普鲁塔克等权威，却远优于最大熵状态。关键只在于：「科学的」是一个特殊称谓，要求独立可验证性，而非「可信的」的同义词。
What does Yudkowsky mean that the definition of science is not a 'well-formed question'?Yudkowsky 说科学的定义不是「构型良好的问题」，是什么意思？
He means asking whether a definition is true confuses the issue: definitions are not true or false, they are useful or not. The standards we set for "scientific" — like p < 0.05 — are pragmatic social choices that can be revised for better outcomes. There is no cosmic rulebook; we design the institution to achieve our epistemic goals.
他的意思是：追问一个定义是否正确混淆了问题——定义无所谓真假，只有有用或无用之分。我们为「科学」设置的标准——比如 p < 0.05——是务实的社会选择，可以为了更好的结果而修订。没有任何宇宙规则手册；我们是在设计这个制度来实现我们的认识论目标。
Why bring up open-access publishing at the end?为什么结尾提到开放获取出版？
To show that the pragmatic logic cuts in multiple directions. If science is defined as publicly reproducible knowledge of humankind, then knowledge locked behind a $20,000/year paywall is questionable on that definition's own terms: it is neither fully public nor freely criticizable. The open-access argument is a live implication of taking the definition seriously.
为了说明这种务实逻辑可以在多个方向上产生后果。如果科学被定义为人类公开可重复的知识，那么锁在每年 20,000 美元付费墙后的知识，按这个定义自身的标准就值得质疑：它既不完全公开，也不能被自由批评。开放获取的论证是认真对待这个定义的一个现实推论。
Does this mean we should distrust peer-reviewed papers in expensive journals?这是否意味着我们应该不信任昂贵期刊上的同行评审论文？
Yudkowsky stops short of that: the information in a closed-access journal "will still count as Bayesian evidence." His claim is narrower — the label "scientific" may be under-deserved in strict terms, and the social incentives may be distorted. The epistemology is not destroyed; the institutional design is questionable.
Yudkowsky 并未走到那一步：封闭获取期刊中的信息「仍然算作贝叶斯证据」。他的主张更为有限——在严格意义上，「科学」这个标签可能不完全名副其实，社会激励机制也可能出现扭曲。认识论本身并未崩溃；有问题的是制度设计。
In-depth Analysis · Pros & Cons深入解读 · 优缺点
This short essay draws a careful three-level map of evidence — Bayesian, legal, scientific — and shows that the narrower categories are socially constructed filters on the broader one, not replacements for it. The stakes are practical: confusing the categories leads either to ignoring real evidence or to fetishizing official labels.
这篇短文细致地绘制了一张三层证据地图——贝叶斯证据、法律证据、科学证据——并指出较窄的范畴是对较宽范畴的社会建构过滤，而非替代。其现实意义在于：混淆这些范畴，要么导致忽视真实证据，要么导致对官方标签的迷信。
- The three-level framework is genuinely clarifying三层框架确实具有澄清作用Most people conflate "scientific" and "credible," or assume legal inadmissibility means epistemic worthlessness. The nested-categories model dissolves both confusions cleanly.大多数人混淆「科学的」与「可信的」，或认为法律上不可采纳就意味着认识论上毫无价值。嵌套范畴模型干净利落地消解了这两种混淆。
- Concrete examples carry the argument具体例子承载了论证Wulky the crime boss, the white socks, Alexander the Great, and tomorrow's sunrise each isolate exactly one variable, making abstract distinctions vivid without requiring technical probability theory.沃尔基黑帮老大、白色袜子、亚历山大大帝、明天的日出——每个例子恰好隔离了一个变量，使抽象区分变得生动，无需借助技术性概率论。
- Anticipates the open-access movement预见了开放获取运动Written in 2007, the essay's argument that paywalled journals fail their own 'public reproducibility' standard reads as remarkably prescient; the debate it gestures at became central to science policy within a decade.写于2007年，该文关于付费期刊未能通过其自身「公开可重复性」标准的论点，读来预见性非凡；它所指向的辩论在十年内成为科学政策的核心议题。
- Honest about the pragmatic nature of definitions对定义的务实本质保持诚实By explicitly calling definitional choices pragmatic rather than true/false, Yudkowsky avoids a common philosophical trap and invites productive debate about what standards actually serve epistemic goals.通过明确将定义选择称为务实而非真/假，Yudkowsky 避免了一个常见的哲学陷阱，并邀请关于哪些标准真正服务于认识论目标的富有成效的辩论。
- The sunrise example conflates prediction with generalization日出例子混淆了预测与规律Yudkowsky says a specific sunrise prediction is not "scientific" because it's a unique future event — but physics generalizations routinely yield specific quantitative predictions that are treated as scientific claims. The line he draws is less crisp than the essay suggests.Yudkowsky 说具体的日出预测不是「科学的」，因为它是独特的未来事件——但物理学规律常规地产生具体的定量预测，这些预测被视为科学主张。他所划定的界线并不像文章所暗示的那样清晰。
- The corruption argument for legal standards is incomplete法律标准的腐败论证不够完整Yudkowsky's account of why courts restrict evidence focuses on abuse-of-power corruption, but legal evidence standards also guard against cognitive biases of jurors, hearsay unreliability, and procedural fairness — richer reasons the essay ignores.Yudkowsky 对法庭为何限制证据的解释集中在权力滥用腐败上，但法律证据标准还防范陪审团的认知偏差、传闻的不可靠性和程序公正——文章忽视了这些更丰富的理由。
- 'Publicly reproducible' is an idealization, not a description「公开可重复」是一种理想化，而非描述Much of modern science — particle physics, genomics, climate modeling — requires equipment or data sets that only a handful of institutions can replicate. The essay's definition describes an ideal that large swaths of real scientific practice don't meet.现代科学的很大一部分——粒子物理学、基因组学、气候模型——需要只有少数机构才能复制的设备或数据集。文章的定义描述了一个理想，现实科学实践的大量领域并不符合。
- The essay proves less about Bayesian evidence than it claims文章对贝叶斯证据的证明不及其声称的多Calling something 'Bayesian evidence' doesn't specify how much it should move your credence, which is the hard practical question. The commissioner's tip being non-zero evidence says little about what probability you should actually assign to Wulky's guilt.称某事为「贝叶斯证据」并未指明它应该在多大程度上改变你的可信度——这才是困难的实践问题。局长线报是非零证据，对于你实际上应该赋予沃尔基有罪多大概率，几乎没有说明。
A tightly argued, highly useful conceptual clarification that every rationalist and anyone who argues about 'scientific consensus' or 'legal proof' should internalize. Its core move — separating what moves a probability from what a social institution endorses — is genuinely important. The essay is less useful as a guide to the harder quantitative questions (how much should evidence move you?), but it does not claim to answer those.
这是一篇论证严密、极具实用价值的概念澄清，每一位理性主义者以及任何争论「科学共识」或「法律证明」的人都应该将其内化。其核心动作——将「改变概率的东西」与「社会制度所背书的东西」分离——具有真正的重要性。这篇文章对于更困难的定量问题（证据应在多大程度上改变你的信念？）帮助有限，但它也未声称要回答这些问题。
Original Text原文
Suppose that your good friend, the police commissioner, tells you in strictest confidence that the crime kingpin of your city is Wulky Wilkinsen. As a rationalist, are you licensed to believe this statement? Put it this way: if you go ahead and insult Wulky, I’d call you foolhardy. Since it is prudent to act as if Wulky has a substantially higher-than-default probability of being a crime boss, the police commissioner’s statement must have been strong Bayesian evidence.
Our legal system will not imprison Wulky on the basis of the police commissioner’s statement. It is not admissible as legal evidence. Maybe if you locked up every person accused of being a crime boss by a police commissioner, you’d initially catch a lot of crime bosses, and relatively few people the commissioner just didn’t like. But unrestrained power attracts corruption like honey attracts flies: over time, you’d catch fewer and fewer real crime bosses (who would go to greater lengths to ensure anonymity), and more and more innocent victims.
This does not mean that the police commissioner’s statement is not rational evidence. It still has a lopsided likelihood ratio, and you’d still be a fool to insult Wulky. But on a social level, in pursuit of a social goal, we deliberately define “legal evidence” to include only particular kinds of evidence, such as the police commissioner’s own observations on the night of April 4th. All legal evidence should ideally be rational evidence, but not the other way around. We impose special, strong, additional standards before we anoint rational evidence as “legal evidence.”
As I write this sentence at 8:33 p.m., Pacific time, on August 18th, 2007, I am wearing white socks. As a rationalist, are you licensed to believe the previous statement? Yes. Could I testify to it in court? Yes. Is it a scientific statement? No, because there is no experiment you can perform yourself to verify it. Science is made up of generalizations which apply to many particular instances, so that you can run new real-world experiments which test the generalization, and thereby verify for yourself that the generalization is true, without having to trust anyone’s authority. Science is the publicly reproducible knowledge of humankind.
Like a court system, science as a social process is made up of fallible humans. We want a protected pool of beliefs that are especially reliable. And we want social rules that encourage the generation of such knowledge. So we impose special, strong, additional standards before we canonize rational knowledge as “scientific knowledge,” adding it to the protected belief pool. Is a rationalist licensed to believe in the historical existence of Alexander the Great? Yes. We have a rough picture of ancient Greece, untrustworthy but better than maximum entropy. But we are dependent on authorities such as Plutarch; we cannot discard Plutarch and verify everything for ourselves. Historical knowledge is not scientific knowledge.
Is a rationalist licensed to believe that the Sun will rise on September 18th, 2007? Yes—not with absolute certainty, but that’s the way to bet.^1^ Is this statement, as I write this essay on August 18th, 2007, a scientific belief?
It may seem perverse to deny the adjective “scientific” to statements like “The Sun will rise on September 18th, 2007.” If Science could not make predictions about future events—events which have not yet happened—then it would be useless; it could make no prediction in advance of experiment. The prediction that the Sun will rise is, definitely, an extrapolation from scientific generalizations. It is based upon models of the Solar System that you could test for yourself by experiment.
But imagine that you’re constructing an experiment to verify prediction #27, in a new context, of an accepted theory Q. You may not have any concrete reason to suspect the belief is wrong; you just want to test it in a new context. It seems dangerous to say, before running the experiment, that there is a “scientific belief” about the result. There is a “conventional prediction” or “theory Q’s prediction.” But if you already know the “scientific belief” about the result, why bother to run the experiment?
You begin to see, I hope, why I identify Science with generalizations, rather than the history of any one experiment. A historical event happens once; generalizations apply over many events. History is not reproducible; scientific generalizations are.
Is my definition of “scientific knowledge” true? That is not a well-formed question. The special standards we impose upon science are pragmatic choices. Nowhere upon the stars or the mountains is it written that p < 0.05 shall be the standard for scientific publication. Many now argue that 0.05 is too weak, and that it would be useful to lower it to 0.01 or 0.001.
Perhaps future generations, acting on the theory that science is the public, reproducible knowledge of humankind, will only label as “scientific” papers published in an open-access journal. If you charge for access to the knowledge, is it part of the knowledge of humankind? Can we fully trust a result if people must pay to criticize it?
For myself, I think scientific practice would be better served by the dictum that only open, public knowledge counts. But however we choose to define “science,” information in a $20,000/year closed-access journal will still count as Bayesian evidence; and so too, the police commissioner’s private assurance that Wulky is the kingpin.
^1^ Pedants: interpret this as the Earth’s rotation and orbit remaining roughly constant relative to the Sun.
假设你的好友——警察局长——以极度保密的方式告诉你，你所在城市的黑帮头目是沃尔基·威尔金森。作为一名理性主义者，你有理由相信这个陈述吗？换个说法：如果你因此去侮辱沃尔基，我会说你太鲁莽了。既然谨慎的做法是把沃尔基作为黑帮老大的概率当成显著高于默认值来行事，那么局长的陈述必定是强有力的贝叶斯证据。
我们的法律体系不会仅凭警察局长的陈述就监禁沃尔基。它在法律证据的意义上是不可采纳的。也许如果你把每一个被警察局长指控为黑帮头目的人都关起来，你起初会抓住很多黑帮头目，而相对较少只是局长不喜欢的人。但不受约束的权力吸引腐败，就像蜂蜜吸引苍蝇：随着时间推移，你会抓到越来越少的真正黑帮头目（他们会竭力确保匿名），而越来越多的无辜受害者。
这并不意味着警察局长的陈述不是理性证据。它仍然具有不对称的似然比，而去侮辱沃尔基仍然是愚蠢之举。但在社会层面，为了追求某种社会目标，我们刻意将「法律证据」定义为只包括特定种类的证据，比如局长在4月4日夜间的亲身观察。所有法律证据理想上都应该是理性证据，但反过来并不成立。在我们将理性证据封为「法律证据」之前，我们施加了特殊的、强有力的、额外的标准。
在我写这句话的时刻——太平洋时间2007年8月18日晚上8点33分——我穿着白色袜子。作为一名理性主义者，你有理由相信这个陈述吗？有。我能在法庭上就此作证吗？能。这是一个科学陈述吗？不是，因为你无法自己做实验来验证它。科学由规律性陈述构成，这些陈述适用于许多具体实例，使你能够运行新的现实世界实验来检验该规律，从而自行验证规律是否成立，而无需信任任何人的权威。科学是人类公开可重复的知识。
如同法庭体系，科学作为一种社会过程，是由易错的人类构成的。我们希望有一个受保护的信念池，其中的信念特别可靠。我们还希望有社会规则来鼓励产生此类知识。因此，在我们将理性知识封为「科学知识」并将其纳入受保护信念池之前，我们施加了特殊的、强有力的、额外的标准。理性主义者有理由相信亚历山大大帝在历史上存在吗？有。我们对古希腊有一个粗略的图景，不可靠但优于最大熵状态。但我们依赖普鲁塔克这样的权威；我们无法抛开普鲁塔克自行验证一切。历史知识不是科学知识。
理性主义者有理由相信太阳会在2007年9月18日升起吗？有——不是百分之百的确定性，但这是合理的押注。^1^ 这个陈述，在我于2007年8月18日写这篇文章时，是一个科学信念吗？
否认「太阳会在2007年9月18日升起」这类陈述的「科学」形容词，似乎有些怪诞。如果科学无法对未来事件——尚未发生的事件——作出预测，那它就毫无用处；它就无法在实验之前作出任何预测。太阳会升起这一预测，确实是从科学规律中做出的外推。它基于你可以自己通过实验检验的太阳系模型。
但想象你正在构建一个实验，在新的背景下验证已接受理论Q的第27条预测。你可能没有任何具体理由怀疑这个信念是错的；你只是想在新的背景下检验它。在实验运行之前，称对结果存在「科学信念」，这似乎很危险。有「传统预测」或「理论Q的预测」。但如果你已经知道关于结果的「科学信念」，为什么还要费心做实验？
我希望你开始明白，为什么我把科学与规律性陈述等同，而不是与任何一个实验的历史等同。一个历史事件只发生一次；规律性陈述适用于众多事件。历史不可重复；科学规律可以。
我对「科学知识」的定义是否正确？这不是一个构型良好的问题。我们对科学施加的特殊标准是务实选择。星辰或山峦上都没有刻写 p < 0.05 应作为科学发表标准这件事。现在许多人认为0.05太宽松了，把它降低到0.01或0.001会更有用。
也许未来的世代，在「科学是人类公开、可重复的知识」这一理论的指引下，只会把发表在开放获取期刊上的论文标注为「科学的」。如果你对获取知识收费，它还是人类的知识吗？如果人们必须付费才能批评某项结果，我们能完全信任它吗？
就我个人而言，我认为科学实践会因「只有开放、公开的知识才算数」这一准则而得到更好的服务。但无论我们如何选择定义「科学」，每年两万美元封闭获取期刊中的信息仍然算作贝叶斯证据；同样，警察局长私下保证沃尔基是头目，也算。
^1^ 学究们：将此解释为地球的自转和公转相对于太阳保持大致恒定。