Extraordinary Claims Require Extraordinary Evidence

The principle

“Extraordinary claims require extraordinary evidence.” This principle, most famously associated with Carl Sagan, is one of the most powerful tools in the skeptic’s toolkit. Its meaning is intuitive: the more unusual, improbable, or far-reaching a claim is, the stronger the evidence must be before we are justified in accepting it.

If a friend tells you they had coffee this morning, you believe them without hesitation. The claim is ordinary, consistent with your background knowledge, and requires no special evidence. If the same friend tells you they were abducted by aliens last night, you would (and should) require considerably more evidence before accepting the claim. Not because you distrust your friend, but because the claim contradicts a vast body of established knowledge about the world.

This principle is not arbitrary skepticism. It is a formalization of how rational people already evaluate claims in every other domain of life. The question is why so many people suspend this standard when it comes to religious claims.

Historical origins

Although Sagan popularized the phrase, the underlying idea is much older. The Scottish Enlightenment philosopher David Hume articulated it in his 1748 essay “Of Miracles”: “No testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous than the fact, which it endeavours to establish.”

Hume’s insight was mathematical in spirit even before formal probability theory existed. When evaluating a claim, you should compare the probability of the claim being true against the probability of the evidence being misleading. For a miracle — defined as a violation of natural law — the evidence must be stronger than our accumulated experience of natural law operating without exception. This is an extraordinarily high bar, and Hume argued that testimony alone could never meet it.

The French mathematician Pierre-Simon Laplace formalized a similar principle: “The weight of evidence for an extraordinary claim must be proportioned to its strangeness.” In modern terms, this is a consequence of Bayesian reasoning — the prior probability of an extraordinary claim is low, so it requires proportionally stronger evidence to shift our confidence.

Applying the standard to religion

Consider the central claims of major religions. Christianity asserts that a man was born of a virgin, performed miracles including walking on water and raising the dead, was himself resurrected from the dead three days after crucifixion, and ascended bodily into heaven. Islam claims that an angel dictated the Quran to Muhammad. Hinduism includes claims about reincarnation and the literal activities of deities on earth. Mormonism asserts that golden plates containing ancient scripture were discovered in upstate New York in the 1820s.

Each of these claims is extraordinary by any reasonable standard. They assert events that contradict everything we know about biology, physics, and the natural world. They are the kind of claims that, in any other context, we would demand overwhelming evidence to accept. If someone today claimed to have walked on water, we would not believe them without extraordinary proof. The question is why claims made thousands of years ago, recorded in ancient texts by anonymous authors, should be held to a lower standard.

The problem of testimony

Most religious claims are supported primarily by testimony — the accounts of witnesses, often recorded decades or centuries after the alleged events. Testimony is the weakest form of evidence, and this is not a matter of philosophical preference but of well-documented empirical fact.

Psychological research has shown that human memory is reconstructive, not reproductive. We do not record events like cameras; we rebuild them each time we remember, incorporating errors, suggestions, and biases. Eyewitness testimony is notoriously unreliable even for ordinary events observed directly — the Innocence Project has documented hundreds of wrongful convictions based on sincere but mistaken eyewitness identification.

For extraordinary events, the problems multiply. People sincerely report miraculous experiences in every religious tradition. Christians see the Virgin Mary. Hindus see Ganesh. Muslims experience divine revelation. These testimonies cannot all be veridical, since they point toward mutually exclusive theological conclusions. The most parsimonious explanation is that human beings are capable of having experiences they sincerely interpret as supernatural regardless of which (if any) supernatural claims are actually true.

Sagan’s Baloney Detection Kit

In The Demon-Haunted World(1995), Sagan offered a set of tools for evaluating claims, which he called the “baloney detection kit.” These tools are as applicable to religious claims as they are to any others:

Seek independent confirmation of the facts. If a miracle is reported, is there corroborating evidence from independent sources? Encourage debate. Arguments from authority carry little weight on their own. Consider multiple hypotheses. If prayer seems to work, consider alternative explanations: coincidence, selective memory, natural recovery. Try not to get attached to a hypothesis just because it is yours.This is perhaps the hardest of all — our beliefs become part of our identity, and questioning them feels like questioning ourselves.

Quantify wherever possible.Vague claims (“prayer works”) should be converted to testable predictions (“prayed-for patients recover at higher rates”) and then tested. When this has been done — notably in the STEP study (Study of the Therapeutic Effects of Intercessory Prayer, 2006) — the results have consistently shown no effect of prayer on medical outcomes.

Common objections

“You can’t apply scientific standards to matters of faith.” This objection concedes the point. If a claim cannot be evaluated by ordinary standards of evidence, it is not clear why it should be believed. Faith — belief without or contrary to evidence — is precisely what the extraordinary-claims principle challenges. If faith is a virtue, it must explain why believing without evidence is better than believing with it.

“The resurrection is well-attested by historical standards.” Even granting (for argument’s sake) that the resurrection accounts are as reliable as advocates claim, Hume’s principle still applies. The question is not whether someevidence exists but whether the evidence is proportional to the extraordinariness of the claim. A dead man coming back to life is the most extraordinary claim imaginable. The evidence — anonymous accounts written decades later, in a pre-scientific culture saturated with miracle stories — is not proportional.

“The standard is biased against religion.” The standard is not biased against anything. It applies equally to all extraordinary claims: UFO abductions, psychic powers, conspiracy theories, and religious miracles alike. The principle does not target religion; religion simply happens to make some of the most extraordinary claims humans have ever proposed.

A tool for honest inquiry

The principle that extraordinary claims require extraordinary evidence is not a weapon. It is a compass. It does not tell you what to believe; it tells you how to evaluate what others ask you to believe. For anyone in the process of questioning their beliefs, this principle provides a framework: ask not whether a claim is comforting, traditional, or widely held, but whether the evidence is proportional to what is being claimed. If you apply this standard consistently — to your own beliefs as well as to others’ — you will find that some beliefs survive and others do not. The ones that survive are the ones worth keeping.