Proof by repetition

The thermostat

Priya: Turn the thermostat up to 85. It heats the room faster.

Dev: Why would it? Doesn’t it just switch the heat on until the room reaches whatever number you set?

Priya: It heats faster. Trust me. Everyone knows you crank it up to warm a room quickly.

The clear-cut case. Dev asked for a reason, and Priya answered by saying the claim again, twice, with “trust me” and “everyone knows” attached. Neither addition is evidence about how the heating works. The claim may or may not be true of this system; the point is that the conversation contains one assertion and no support.

One voice in the group chat

For months, one member of a hiking club posts in the group chat, before every planning meeting, that “the summer picnic always loses money.” He never shares the accounts, and the treasurer doesn’t reply. By spring, newer members are saying, “Everyone knows the picnic loses money. Maybe we should drop it.”

This is repetition over time, to an audience rather than an opponent. Every instance traces back to one person who never showed his figures, yet the claim has come to sound like the club’s shared view. Research on this pattern (see Why it happens) finds that an opinion repeated by one person can seem more widely held, which is how repetition slides into an appeal to popularity: “everyone knows” is standing on one voice.

Three reasons that are one reason

A slide in a proposal to switch the office to a new scheduling app reads:

Why switch?

1. It’s more efficient.
2. It saves time.
3. It streamlines scheduling.

The less obvious form. It looks like a list of three reasons, but each line restates the same unsupported claim in different words. Nothing on the slide says how much time is saved, compared with what, or how anyone knows. Padding a claim with its own paraphrases makes it look better supported than a single sentence would, without adding any support.

• Repeating in the face of objections. The claim is restated each time it is questioned, without engaging the question.
• Repeating over time or across channels. A slogan, talking point or rumor stated often enough that it comes to feel like common knowledge, even when it has a single source.
• Restating as if it were support. The claim is listed again in different words as if the paraphrases were further reasons, as in the slide above.

In pragma-dialectics, an approach to argument that treats fallacies as broken rules of a good-faith discussion, one proposed fallacy is evading the burden of proof, which breaks the rule that “you must defend your standpoint if asked to do so” (as reported in the Stanford Encyclopedia of Philosophy‘s entry on fallacies). Repeating a standpoint in place of defending it is one way to evade that burden (see Burden of proof).

Lewis Carroll put the tactic in a line of verse in The Hunting of the Snark, where the Bellman announces: “What I tell you three times is true.”

How well established the name is. Unlike the psychology behind it, the fallacy itself has thin scholarly coverage. The name, often in the Latin form argumentum ad nauseam (“argument to the point of nausea”), appears mainly in popular guides to fallacies, such as Madsen Pirie’s The Book of the Fallacy (1985). It isn’t among the fallacies covered in the Stanford Encyclopedia of Philosophy‘s entry on fallacies, or among the 231 names in Bradley Dowden’s list of fallacies in the Internet Encyclopedia of Philosophy. An edited collection of short entries on fallacies, Bad Arguments (2018), includes a chapter titled “Argument by Repetition”.

Not the same error: begging the question. Begging the question builds the conclusion into the premises, often in different words. Hans Hansen, discussing a classic example in which the premise and conclusion are the same proposition, calls it “a disguised instance of repetition which gives no reason for its apparent conclusion.” The difference is in the disguise: an argument that begs the question looks like an argument, with a premise and a “therefore”, while proof by repetition simply asserts the claim again. The slide above sits close to the border.

Saying something more than once is often useful. It isn’t proof by repetition when:

• The reasons have already been given. Restating a claim to summarize, to answer a question that was already addressed, or to bring a discussion back to the point is fine, as long as the support is there to point back to.
• The repetition is a reminder, not an argument. Repeating a safety rule, an instruction or a key point in a lesson helps people remember something whose support isn’t in dispute.
• It marks emphasis. Repeating the one point that matters most tells the listener what to focus on; the case for it is made elsewhere.
• The “repetitions” are independent confirmations. Several people who checked separately and agree aren’t one claim repeated. They’re several pieces of evidence.

The test: if every repetition except the first were deleted, would any reason for the claim be lost?

The contractor who says it again

A contractor inspects a deck and reports that one support beam has to be replaced. He shows the homeowner the moisture-meter readings, the soft wood near the post and a level showing the deck sags by an inch. The homeowner asks three more times whether it’s really necessary. Each time, the contractor says, “Yes, it needs replacing. That’s what the readings and the sag show.”

The contractor repeats the same claim four times, but every repetition points back to evidence that is already on the table. Repeating it adds nothing, and it doesn’t need to: the support was given first. That’s the reasons have already been given condition. The homeowner is free to question the readings, which would call for an answer about the readings, not another repetition.

The kitchen rule

A restaurant kitchen manager says the same thing at the start of every shift: “Raw chicken goes on the bottom shelf of the walk-in fridge.”

Said every day, this sounds like repetition in place of argument. But nobody in the kitchen is being persuaded of anything. The reason is already known and not in dispute: raw meat on a higher shelf can drip onto food below. The repetition is there so the rule gets followed on a busy night. That’s the reminder, not an argument condition.

Repetition persuades even though it isn’t evidence, and the reason is well studied. People rate statements they have seen before as truer than new ones, the illusory truth effect. A meta-analysis of 51 studies by Alice Dechêne and colleagues (2010) confirmed the pattern, and reviews explain it mainly by processing fluency: a repeated statement is easier to process, and that ease is taken as a sign of truth. Christian Unkelbach and colleagues (2019) review this and other explanations, based on frequency, recognition, familiarity and coherent references in memory. Because fluency and truth often go together in everyday life, the shortcut often works, which is also why it is easy to exploit.

More repetitions keep helping, with diminishing returns. In experiments by Aumyo Hassan and Sarah Barber (2021), trivia statements were shown up to 27 times; rated truth rose with each repetition, but the largest increase came from the second encounter.

Repetition also makes a claim sound popular. In six experiments, Kimberlee Weaver and colleagues (2007) found that people tend to infer that a familiar opinion is a widely held one, even when its familiarity comes from one group member repeating it, and even when they are aware the opinions come from a single person. A repetitive voice, as their title puts it, can sound like a chorus.

There is also a practical reason. Repeating a claim costs less than answering an objection, and the people objecting eventually tire. When they stop, their silence can be mistaken for agreement.

• Ask for the reason, not the claim. “I understand that’s your view. What makes you think it’s true?”
• Count sources, not statements. When a claim feels familiar, ask where it first came from and how many separate people actually checked it.
• Collapse the paraphrases. When a list of reasons looks long, check how many of the items say something different.
• Notice it in yourself. If you find yourself saying the same thing again with more emphasis, that’s a sign to give the reason, or to admit you don’t have one to hand.

Tested ways of reducing the illusory truth effect itself are covered on that page.