An open claim
claim → evidence → possible revision
A claim stays open when some observable result could count against it, and the speaker would accept that result.
A field guide to discursive closure
Closure exploits conversational norms by ejecting from consideration the question at issue.
Conversation runs on a norm: you answer the question you were asked. The exploit is a question built so that every answer to it concedes a different question — the one you were never asked. A salesperson who asks, before you have agreed to buy, whether you would like delivery on Monday or on Wednesday has made either answer a purchase; sales training teaches this by name as the assumptive close, and a customer who notices can still walk out. A candidate who announces, before the votes are cast, that the only way he can lose is if the election is rigged has made either result a win — and when the count came in against him, it was the recounts, the courts and the officials that were reclassified as part of the fraud. When the conclusion arrives before the evidence, that is a sales pitch. When it survives the evidence against it, you are being closed.
Written for people who argue, teach, moderate, report, or vote — and who want a test that still works when it is turned on their own side.
The difference between an open and closed claim
A closed claim
claim → challenge → challenger disqualified → claim protected
The correction is not answered. It is reclassified as part of the attack.
You already know the small version
A salesperson offers two delivery days. Both answers mean you bought.
A salesperson, to a customer who has not yet said whether they are buying:
“Would you like delivery Monday, or would Wednesday be better for you?”
Monday You bought it.
Wednesday You bought it.
The question of whether you were buying at all was never asked.
The conclusion — you are buying — is placed before the question that should have produced it: the cart before the horse. Sales training teaches the move by name, the assumptive close; the delivery question is its textbook script, which the trade calls the alternate of choice. The Overton window is the range of answers a public will tolerate hearing; the close is a private Overton window for questions, one conversation wide, with “am I buying at all?” just outside it. A customer who notices can still say no. The move gets larger when a no stops counting as an answer and starts counting as a symptom — when “I don’t think I’m buying” marks you as a difficult customer rather than as someone who is not one.
The mechanism in one sentence
A candidate, before any vote is cast, says a loss can only mean fraud. “Was the election rigged?” is answered before it can be asked.
President Trump, campaigning for re-election, eleven weeks before Election Day:
“The only way we’re going to lose this election is if this election is rigged.”
If Trump wins The election was legitimate.
If Trump loses The loss proves the election was rigged.
Either way, he won. Whether the election was rigged was never a question.
That is the assumptive close with a country for the customer. It became closure after the votes were counted: the rule announced in August — a loss can only mean fraud — was kept, and every check that found no fraud was filed as part of the fraud.
After the count, each institution that checked the claim and found nothing:
Recounts confirm the result They recounted the fraudulent ballots.
Courts reject the suits The judges lacked courage, or were part of the system.
Officials find no fraud The officials were weak, or corrupt.
Each check joined the plot. The claim did not move.
The 2020 case
The stolen-election claim was not one allegation. It was a system.
Joe Biden lawfully won the 2020 election. The Electoral College result was 306–232; recounts and audits confirmed the decisive state results; Trump’s own election-security officials reported no evidence that voting systems changed votes; and his Justice Department leaders repeatedly told him the fraud allegations were false. The important rhetorical fact is what happened next: each failed check was absorbed into a larger story about corrupt officials, cowardly judges, hostile media, and a persecuted movement.
National Archives result ↗ 2020 election-security statement ↗ January 6 Committee report ↗
Pre-decide what a loss means
Before voting began, define victory as the only legitimate result. A loss can then arrive only as evidence of fraud.
Supply an expandable mechanism
Mail ballots, machines, noncitizens, urban vote counts, election workers: when one allegation fails, another can take its place.
Disqualify every referee
Courts, journalists, election officials, auditors, and even disloyal Republicans become participants in the plot when they disagree.
Turn resistance into loyalty
If the leader and his voters are treated as one body, challenging his claim becomes an attack on their vote, identity, and country.
The test
Four questions cut through the fog
Closure is a structure you can inspect. Ask these before arguing about the latest allegation:
- What is actually being claimed? State it in a form that evidence could confirm or disconfirm.
- What is treated as already settled? Look for the premise smuggled in as background rather than argued for.
- Who is allowed to check it? A claim with no acceptable referee has no working error-correction path.
- What result would change the speaker’s mind? If the answer is “none,” the dispute has been closed, not won.
Run the four questions on a claim → · See why this test works → · Browse the deeper device catalog →
The pedigree
It has had a name for 2,300 years
Aristotle catalogued the small version in the Prior Analytics: using as a premise the thing you set out to prove. The schools called it petitio principii, and English, from 1581, called it begging the question. Merriam-Webster’s entry for the phrase now gives “to elicit a question logically as a reaction or response” first, and its usage note calls the logical sense “lesser used and more formal.”
Suggested path
Read the site in six steps
- How it works One sentence, fully mapped, and the three-way distinction between false, closed, and both.
- The 2020 case The mechanism at scale: seven moves, a timeline, a primary-source ledger, and what would weaken the reading.
- Old names The fallacies the devices answer to, from Aristotle on, and what became of the English name for the oldest one.
- The devices Eight closure constructions and four force multipliers, each with a specimen, a forbidden question, and a counter-move.
- Practice A trainer, a mapping worksheet, a printable field guide, and a page on talking to someone inside a closed claim.
- Method & limits What the site claims, what it does not, and the tests that would falsify it.
Keep this analysis open. Political prosecutions can be abusive. Election systems can fail. Journalists, judges, and officials can be wrong. None should be trusted by category. The test is whether a claim names evidence and permits a result that would count against it. Sources, counterarguments and failure conditions: Method.