On Monday, Donald Trump posted on Truth Social that anyone reporting voter fraud, past or upcoming, “will be paid $1,000,000 upon the arrest and conviction of those that created the Fraud!” [5]. Five outlets ran it: NBC News, Forbes and The Guardian in English, France 24 and RFI in French. Four of the five carry the counterweight in some form, from NBC’s flat “There is no evidence of widespread voter fraud in U.S. elections” [1] to the Guardian’s “unfounded claims” [5]; Forbes fact-checks specific claims instead, noting that Arizona machines Trump said were “broken purposely” had, independent investigators found, “a technical issue with some ballot printers used by Maricopa County” [4]. That is the coverage at its best. Around it, though, everything a reader would need to judge the offer is arranged to be taken on trust: the motive asserted, the live number invoked, the confirmation duplicated, the one tested precedent parked last.

The why, printed as fact

In its own voice, NBC explains the offer: “Trump’s offer Monday is his latest effort to exert greater influence over the midterm elections”Unsourced motive attributionUnsourced motive attribution is a news text stating what an institution or government 'wants', 'fears' or 'seeks' as plain fact, with no official, document or source behind the intention, so the reporter's hypothesis reads as a finding. [1]. No official, document, or named strategist stands behind the intention; it is the reporter’s reading in news voice. France 24 and RFI go further: the offer is “la dernière tentative en date du républicain de semer le doute quant aux futurs résultats des élections du 3 novembre” (the Republican’s latest attempt to sow doubt about the coming results of the November 3 elections) [2], a sentence RFI republishes nearly verbatim [3]. This is unsourced motive attribution, and the inference is probably sound; the same article reports the Justice Department suing, at Trump’s request, to stop Alaska certifying a Senate result if the Republican loses [1]. But the grammar still does two things to the reader. It closes the question of how we know: an inferred motive and a reported motive arrive in identical sentences, and a reader trained to accept the first cannot reject the second when it flatters power. And it substitutes psychology for mechanics: once the story is why he is doing it, the questions that would check the act, who administers the money and under what authority, have nowhere to live. Forbes shows the alternative: “The Trump administration has taken multiple steps aligned with Trump’s repeated claims that American election systems are rife with fraud” [4]. Aligned with, not intending to. The acts are listed, the claim attributed, and the reader does the arithmetic.

Unsourced motive attribution is a news text stating what an institution or government 'wants', 'fears' or 'seeks' as plain fact, with no official, document or source behind the intention, so the reporter's hypothesis reads as a finding.

The number that arrives as two words

The administration’s live allegation, litigated by the Justice Department in California, is noncitizen voting [1][4]. Across five articles, its empirical status rests on one sentence, NBC’s: “Noncitizen voting is also rare, studies show”Proof by invocationProof by invocation cites a report, a study or unnamed experts as having shown something, without giving a single figure, name or finding from them, so the reader can only take the conclusion on trust. [1]. No study named, no figure, no finding. That is proof by invocation: the sentence borrows the authority of research and hands over none of its content, so the reader can only take it on trust. The same article shows the standard two paragraphs up, giving the general fraud claim a named tracker and a count, Heritage’s 1,600-plus incidents over 40 years, “a tiny fraction of the total votes cast over that period” [1]. The French reports reach that standard, naming Brookings, which counted 39 cases of fraud among more than 100 million ballots across 32 Pennsylvania elections over three decades, and quoting the Brennan Center [2][3]. The asymmetry is the tell: the claim with no political urgency gets numbers, and the claim the government is litigating gets two words. A reader who wants to check the allegation at the center of the lawsuits has nothing on the page to hold.

Proof by invocation cites a report, a study or unnamed experts as having shown something, without giving a single figure, name or finding from them, so the reader can only take the conclusion on trust.

One draft, two mastheads

France 24 and RFI are presented as two outlets; on this story they are one text. The leads match to the word apart from an inserted date [2][3]. The Truth Social quotation is identical. The paragraph on his 2020 defeat is the same apart from one capital letter, and the no-evidence sentence differs by a single preposition [2][3]. The research paragraph is the same sentence twice: “Brookings n’a recensé que 39 cas de fraude sur les plus de 100 millions de bulletins examinés lors de 32 élections dans l’État de Pennsylvanie sur trois décennies”ChurnalismChurnalism is a press release, wire feed, or one outlet's unsourced reporting republished under several bylines and read by the public as independent confirmation. (Brookings counted just 39 cases of fraud among more than 100 million ballots examined across 32 elections in Pennsylvania over three decades) [3], identical to the letter at France 24 [2]. This is churnalism: one draft under two mastheads, read by the public as independent confirmation. Nothing in either article says where the shared copy comes from, and from the texts alone it is impossible to tell a common wire from one newsroom’s draft republished; what is visible is the duplication. The effect is arithmetic: two brands checked, one story confirmed twice, when it was reported once. The sharpest figure in the French coverage, Brookings’ 39 of 100 million, arrives pre-duplicated, with a halo of double sourcing it did not earn.

Churnalism is a press release, wire feed, or one outlet's unsourced reporting republished under several bylines and read by the public as independent confirmation.

The accusation the quotes let stand

A teaser on France 24’s page promotes a related article by asking whether Trump will “tricher à tout va”Scare quotesScare quotes are quotation marks placed around a word an outlet won't stand behind, so it can print the word anyway. (cheat at every turn) [2]. Quotation marks plus a question mark let the page print the word cheat while standing behind nothing; that is scare quotes. It is a cross-promotion line rather than the article’s own headline, but teasers get read, and the formula scales. The same page shows the honest use of the same punctuation when the Alaska lawsuit’s justification appears as the administration’s stated risk of “confusion” between two candidates with the same name [2]: there, the quotes pin the word to the government that invoked it. Same marks, two jobs. The test for a reader: is the quoted word attached to a speaker, or floating free where the outlet can disclaim it?

Scare quotes are quotation marks placed around a word an outlet won't stand behind, so it can print the word anyway.

The last paragraph is the story

Only one of the five articles notes that this scheme has been run before. The Guardian’s final paragraph reports that in 2020 Texas’s lieutenant governor, Dan Patrick, offered $25,000 for election-fraud tips, and that “the first payout went to Eric Frank, a Democratic poll worker in Pennsylvania whose tip led to the conviction of a 72-year-old Republican for voting twice”Buried ledeA buried lede is the fact that should have organized the whole story, relegated instead to the final sentence or hidden inside someone else's quote. [5]; the two ballots were cast in the voter’s own name and in his son’s, the Washington Post reported in 2021 [5]. That is the one empirical test of a fraud bounty in this coverage, and it produced a payout, a conviction, and a picture of the fraud that actually exists: individual, small, caught. It should organize the story, because it tells the reader what a $1 million version does if it works. Instead it is a buried lede, placed after the White House comment line, in the slot readers reach least. The other four outlets omit the precedent entirely [1][2][3][4], so their readers meet the offer as a novelty, with no history and no outcome.

A buried lede is the fact that should have organized the whole story, relegated instead to the final sentence or hidden inside someone else's quote.

The question registered, then dropped

Three outlets notice the money and stop, in near-identical words: “Trump did not specify where the funds for such payments would come from” [1]; “Trump did not say how the rewards would be funded” [4]; “Trump did not disclose how he plans to pay for the rewards” [5]. France 24 and RFI never raise it [2][3]. The gap is registered as a fact about the announcement, then never pursued as a question about the act. A president proposing to pay a million dollars per conviction invites questions no article asks: under what authority, from which fund, administered by whom. NBC supplies a fragment of an answer and leaves it disconnected, reporting that the offer “comes after he unveiled a $25 million government fraud whistleblower program, telling reporters he’d decide who was worthy of the prize” [1]. In the sibling program, the president personally decides who gets paid. The article never asks whether the $1 million offer inherits that structure; it files the sentence under context.

What a reader would need instead

5 checks to run on the next story like this one.

  1. The authority and the money: which fund pays, under what statute, with what role for Congress. The only official answer so far is a referral back to the post [1]; the Guardian’s White House query had produced no reply [5]. An offer with no funding path is a press release.

    Gap in the coverage

  2. Named research on the live allegation: the study, its name, its figure, on the page, the way this coverage names Heritage’s tracker [1], Brookings’ 39 of 100 million [2], and the Brennan Center [2][3]. “Studies show” is not a source.

    Proof by invocation

  3. The precedent as the spine: Patrick’s 2020 bounty and its outcome [5] are the only data anyone has on what these schemes produce. Four of the five outlets could have opened with it.

    Buried lede

  4. A label on shared copy: if France 24 and RFI publish one draft, a wire credit would say so [2][3]; until then, cross-checking French coverage double-counts a single report.

    Churnalism

  5. The construction to demand: Forbes’s “aligned with” [4] links acts to claims without asserting intent. Hold every outlet to that, including the ones whose conclusions you like.

    Unsourced motive attribution

The offer’s one precise clause, payment upon arrest and conviction [5], is also its one checkable clause. The coverage around it asks for trust on the motive, the numbers, the confirmation, and the money. A scheme that pays on proof should not be covered on trust.