---
name: fighting-a-lie-that-gets-repeated-into-truth
description: |
  Use when a confidently-stated falsehood is being repeated so often that it is on track to become
  conventional wisdom — in politics, in a market, in a company, in a relationship. Helps the user
  recognize the mechanism (the "sleeper effect"), understand why well-intentioned fact-checking
  often fails, and structure a response that can actually hold the narrative.
  Triggers on phrases like "this lie keeps spreading no matter how many times we correct it",
  "how do I fight a false narrative that's becoming conventional wisdom", "my opponent's confident
  assertion is winning the argument", "they're defining me before I can define myself", "the facts
  aren't landing", "how do I counter someone who just repeats untruths with confidence".
  Do not use for: genuine factual disputes where reasonable people disagree; ordinary marketing
  spin; situations where the user actually holds the weak factual position.
---

# Fighting a Lie That Gets Repeated Into Truth

> *What this skill is about, in one sentence:*
> How to recognize when confident repetition is turning a false claim into accepted fact — and how
> to counter it in a way that actually shifts the narrative rather than just annotating the record.

## Where this comes from

This skill distills Chapter 8 ("Rewriting History Through the Sleeper Effect") of Sonnenfeld & Tian's
*Trump's Ten Commandments* (2025), with supplementary material from Chapter 2 (projection) and
Chapter 9 (nickname branding — covered in more depth in skill #7).

Academic foundations:

- **Carl Hovland and Walter Weiss**, "The Influence of Source Credibility on Communication
  Effectiveness," *Public Opinion Quarterly* (1951). The Yale mass-communications research that
  discovered the **sleeper effect**: people remember messages long after they forget the source,
  and unreliable sources become more persuasive over time with repetition.
- **Chris Argyris** (Harvard/Yale), cited by Sonnenfeld as a disapproving student of Hovland.
  Argyris's observation: "You tell sophisticated people both sides of the story, and they think
  you've been fair. For unsophisticated people, you just hammer them with one side of the story,
  as they do not want the nuances."
- **George Orwell**, *1984* (1949). The Ministry of Truth and the Party's slogan, "Who controls
  the past controls the future." Frames the stakes of narrative control.
- **Alan Brinkley**, *Voices of Protest: Huey Long, Father Coughlin & The Great Depression*
  (Pulitzer Prize, 1983). American historical precedents — populist demagogues who promoted false
  narratives with great effectiveness.
- **Kellyanne Conway's** "alternative facts" phrasing (January 2017) — an inelegant label for the
  same phenomenon Hovland-Weiss had described six decades earlier.

Primary source: `false`. The book applies the Hovland-Weiss framework to Trump's specific
rhetorical pattern, but the underlying psychology is not about any specific person.

## North Star

> *Does this operator change what the user actually says, to whom, with what frequency?*

Merely knowing the sleeper effect exists does not help a user fight it. The operators below each
have to shift the user's response tactic. If an operator is only descriptive of the attacker's
behavior with no implied counter-move, cut it.

## Opening Gate: is this a sleeper-effect campaign?

Three tests:

- **Repetition with certainty.** Is the claim being repeated often, with absolute confidence, by
  the same source or aligned sources?
- **Counter-evidence is available and ineffective.** Factual corrections exist but don't seem to
  be shifting the narrative.
- **The claim is becoming load-bearing.** Decisions, votes, or allocations are starting to
  depend on the claim being true.

If all three test positive, apply this skill. If only the first is true (repetition) without the
second two, you may be dealing with ordinary spin rather than a sleeper-effect campaign — use
conventional media-relations or fact-checking advice.

## The Operators

### 1. `relentless-repetition-beats-accuracy`

**Plain English:** A confident, unhedged falsehood, repeated often enough, defeats a true but
hedged statement. The messenger's certainty is more memorable than the message's accuracy.

**Source:** Sonnenfeld & Tian, Ch. 8, citing Hovland & Weiss (1951) and Argyris. Book case: the
*Washington Post* documented 30,573 false or misleading claims during Trump's first term. In July
2025, Sonnenfeld's team ran twenty of Trump's most-repeated economic, social, and diplomatic
pronouncements through major AI models — ChatGPT, Claude, Perplexity, Gemini, Grok — and all
five found most untrue. Yet the claims continued to shape public belief.

**Detect:**
- Same claim, phrased nearly identically, repeated across many venues over an extended period.
- Speaker uses absolute-certainty language ("everyone knows," "the facts are," "there is no
  question"), never hedges.
- Counter-evidence exists in the record but has not displaced the claim.
- Sub-signal: **confident false beats hedged true.** The more technically careful the correction,
  the less it lands. Careful corrections ("it's complicated, the data shows X under some
  conditions…") are systematically worse at displacing a confident false than equally confident
  true claims.
- Sub-signal: **single message for mass audience.** The attacker's message is one clean sentence.
  If the defender's correction is a paragraph with caveats, the correction loses.

**Intent behind it:** Implant the claim as impression regardless of its factual accuracy.

**Counter-move:**
- **Match the frequency.** A single strong correction is not enough. The correction needs to be
  repeated at roughly the same cadence as the original claim for the same duration.
- **Match the confidence.** Hedged corrections lose. "This is false. Here are the numbers." beats
  "Well, it's complicated, under some interpretations…"
- **Pick one-sentence corrections, not paragraphs.** The counter needs to be as memorable as the
  original. Compress your strongest factual point into one line.

**Do not use when:** the claim is genuinely contested or the "true" alternative is itself hedged
— in which case treating the dispute as a sleeper-effect campaign overstates your own certainty.

### 2. `define-the-counterparty-before-they-do`

**Plain English:** A brand, a label, a one-line characterization sticks once it's affixed.
Whoever gets the branding in first — especially if the brand captures a recognizable truth about
the target — controls the frame for every subsequent conversation.

**Source:** Sonnenfeld & Tian, Ch. 8 and Ch. 9. Cases: "Crooked Hillary," "Sleepy Joe," "Lil
Marco," "Lyin' Ted," "Low-Energy Jeb," "Crazy Bernie," "Pocahontas Warren." Boundary case: Trump
couldn't find a sticky nickname for Kamala Harris ("Laffin' Kamala," "Comrade Kamala" — both fell
flat) or for Obama ("Cheatin' Obama" — didn't land).

**Detect:**
- Your opponent is deploying a short, memorable label for you or your side.
- Media is starting to repeat the label uncritically.
- You haven't defined yourself with a comparable label.

**Intent behind it:** Reduce a complex person/position to a caricature that captures one
weakness. Once affixed, the label overshadows messaging.

**Counter-move:**
- **Define yourself first, with a label at least as memorable as the one coming at you.** "I am
  the candidate of ___" in one clause, repeated relentlessly.
- **If you've already been labeled and the label is sticky:** don't engage the label directly.
  Engaging ("I'm not crooked, here's why…") reinforces it via negation. Instead, define your
  opponent with a competing label that displaces attention.
- **The sticky label captures a real weakness, even if unfairly.** If the opposition's label
  works, there is something true about it. Address the underlying weakness, not just the label.

**Do not use when:** the label is clearly failing to stick (like the Harris/Obama cases) — in
which case engaging with it amplifies rather than counters.

### 3. `control-flow-of-records`

**Plain English:** The attacker doesn't just make false claims — they also work to prevent the
true record from becoming public. Sealed transcripts, lawsuits against reporting, intimidated
record-holders. The asymmetry is part of the pattern: they insist on *your* records while
concealing *their own*.

**Source:** Sonnenfeld & Tian, Ch. 8. Cases: Trump's demand for Obama's academic records while a
Trump staffer intimidated educational officials to keep Trump's own records sealed; Trump's $5B
lawsuit against Timothy O'Brien (dismissed after years but costing O'Brien millions and pushing
him from the *NYT*).

**Detect:**
- Your opponent demands full transparency from you while resisting disclosure of their own
  records.
- Journalists or researchers who publish truthful information face lawsuits, intimidation, or
  professional pressure.
- The public record on the contested claim is one-sided — only one party's documents have been
  released.

**Intent behind it:** Create a narrative vacuum the confident repetition can fill. Without access
to the true record, the public has only the repeated claim to go on.

**Counter-move:**
- **Proactive disclosure.** Publish your own record before being asked. This forecloses the
  "they're hiding something" frame.
- **Demand reciprocity, specifically.** "I'll release my records when you release yours" — with
  specifics about what records, not just a general challenge.
- **Support the institutions that preserve records.** Journalism, courts, archives. Weakened
  record-keeping institutions make sleeper-effect campaigns more effective.

**Do not use when:** your opponent has a legitimate reason to withhold records (privacy,
national security, trade secrets) that you would accept in principle.

### 4. `echo-a-useful-external-narrative`

**Plain English:** The attacker doesn't always originate the false claim. Sometimes they adopt
and amplify an existing false narrative that serves their purposes — a conspiracy theory, a
foreign state's framing, a fringe belief. The adoption legitimizes the narrative for mainstream
audiences.

**Source:** Sonnenfeld & Tian, Ch. 8. Cases: Trump echoing Putin's account of who started the
Ukraine conflict; Trump's claim to have seen Muslims celebrating in New Jersey after 9/11;
recycling of far-right election-fraud framings; promotion of birtherism.

**Detect:**
- A claim appears in the attacker's rhetoric that has been present in fringe spaces.
- The attacker's endorsement gives the claim access to mainstream attention.
- The attacker can plausibly disown the claim ("I'm just repeating what people are saying").

**Intent behind it:** Use a narrative whose credibility is not directly tied to the attacker's
own, while benefiting from its effect.

**Counter-move:**
- **Name the origin.** "This claim originated in [specific place]" makes the laundering explicit
  and breaks the plausible-deniability frame.
- **Refuse to debate the claim on its own terms.** The debate itself legitimizes the claim as
  "contested." Treat it as the laundering move it is.
- **If the amplification is from a foreign source:** explicit attribution ("Russian state media
  has promoted this narrative for the following reasons…") is usually more effective than
  point-by-point rebuttal.

**Do not use when:** the claim, although originating in fringe spaces, has genuine factual
merit that the attacker happens to have surfaced. Not every fringe claim is false; origin
doesn't determine truth.

### 5. `asymmetric-evidence-acceptance`

**Plain English:** Every poll that favors the attacker is real science; every poll that doesn't
is fake. Every court that rules their way is fair; every court that rules against them is rigged.
The selective acceptance of evidence is systematic and visible.

**Source:** Sonnenfeld & Tian, Ch. 8. The book notes Trump has never acknowledged a negative poll
as accurate publicly, nor dismissed a positive one as flawed.

**Detect:**
- Your opponent accepts evidence favoring their position and rejects evidence contradicting it,
  consistently and publicly.
- The criteria they offer for the rejections are ad hoc (fake, rigged, biased, deep state) rather
  than methodological.
- They never reverse a position in response to new evidence.

**Intent behind it:** Maintain an internally consistent alternative-reality narrative regardless
of external data. Over time, followers absorb the pattern: "real" evidence is what supports the
attacker.

**Counter-move:**
- **Document the asymmetry explicitly.** Side-by-side: here are five polls favoring them
  (accepted); here are five polls disfavoring them (called fake). The pattern itself is
  evidence of motivated reasoning.
- **Tie your own credibility to reversals.** If you publicly acknowledge a position you got
  wrong, you earn standing to call out asymmetric acceptance. If you never reverse, your own
  consistency is suspicious for the same reason.
- **Do not try to litigate every specific rejection.** The asymmetry is the point, not any
  individual dispute.

**Do not use when:** the "asymmetry" is actually a legitimate quality difference — one set of
evidence is methodologically better than the other, and the acceptance/rejection tracks that.

### 6. `plant-seeds-and-water-them`

**Plain English:** Some false claims are planted years before they bear fruit. A once-ludicrous
assertion ("Trump should be on Mt. Rushmore," "Trump should get a Nobel Prize") becomes more
credible through sustained repetition over time, even without new evidence. Recognition of the
plant early is the only viable counter.

**Source:** Sonnenfeld & Tian, Ch. 8. The Mt. Rushmore / Nobel Peace Prize claims went from
laughable to earnestly campaigned-for in under a decade. The "rags to riches" origin myth
similarly seeded early and defended through litigation for decades.

**Detect:**
- A claim appears in the attacker's rhetoric that is currently risible but is being repeated
  consistently.
- The claim is being aligned with culturally high-status reference points (great presidents,
  prestigious prizes, historical figures).
- The claim is accompanied by operational moves (lobbying, nominations, institutional pressure)
  that would only make sense if the claim were already half-accepted.

**Intent behind it:** Normalize the claim through repetition over a timescale longer than a single
news cycle or election. Each individual repetition seems ridiculous; the aggregate shifts public
baselines.

**Counter-move:**
- **Kill seeds early, while they are still ridiculous.** A planted claim is easier to
  de-legitimize in month one than year five.
- **Mock the claim at the level of the alignment, not the detail.** Pointing out that the attacker
  has also claimed to be greater than Washington is more effective than debating whether any
  specific accomplishment is true.
- **Do not help the watering.** Even "sincere" debate over whether the attacker deserves the
  planted honor reinforces the plant. Ignoring is sometimes the right move.

**Do not use when:** the claim is actually within normal boasting territory. Every politician
claims they'll be a great president; not every such claim is a strategic seed plant.

### 7. `projection-claim-aggrieved-victim` (the preemptive-projection operator)

**Plain English:** The attacker accuses others of doing exactly what they themselves are doing.
Accusations of "fake news" from an attacker prolifically generating false claims; accusations of
"lawfare" from an attacker prolifically suing opponents; accusations of "witch hunts" from an
attacker prolifically investigating critics. This scrambles the evidentiary waters and forces
opponents onto the defensive before they can make their own case.

**Source:** Sonnenfeld & Tian, Ch. 2 passage titled "Projecting His Own Attributes Preemptively
onto Others," and Ch. 8.

**Detect:**
- The attacker's rhetoric reliably attributes to opponents the tactic the attacker is most
  clearly using.
- Claims of victimhood accompany offensive actions.
- The accusation arrives *before* any equivalent move by the opponent.

**Intent behind it:** Reframe the attacker as victim, force defense rather than offense on the
other side, and muddy the evidentiary record so both sides look comparably culpable.

**Counter-move:**
- **Name the specific projection, not the general tactic.** "This is called projection" sounds
  accusatory and weak. "When X accuses Y of [specific tactic], here are the times X has done
  [the same tactic]" is concrete and effective.
- **Refuse the defensive posture.** If the accusation pulls you into defense, the projection
  has worked. Stay on your own message.
- **Document contemporaneously.** Keep a running record of the pattern — specific accusation +
  specific example of attacker doing the same thing + date. This record becomes your counter
  when the pattern continues.

**Do not use when:** the accusation is actually true — that is, both sides are doing the thing.
In that case, the projection frame doesn't apply; both sides' conduct is the issue.

### 8. `challenge-false-claims-with-equal-vigor` (the Conclusion's defender rule)

**Plain English:** Unless you are willing to submit to the attacker's version of reality, false
information has to be challenged, corrected, and repeated back with equal vigor. Half-efforts
lose; silence is defeat.

**Source:** Sonnenfeld & Tian, Conclusion. Amplified by the observation in Ch. 8 that Democrats
surrendered the inflation and economy narratives to Trump during 2024 because "they were reluctant
to celebrate the Biden-Harris administration's economic triumphs" despite direct urging from
their own economic advisors.

**Detect:** Not a detection operator — a meta-rule about sustained response.

**Counter-move:**
- **Assign narrative defense as an ongoing task, not an event.** If the attacker repeats the claim
  weekly, your correction must too.
- **Your own spokespeople must be willing to make the correction with confidence.** Timid
  corrections by officials who visibly don't want to engage have near-zero effect.
- **Celebrate your side's actual wins.** The book's specific observation on the 2024 campaign:
  administration officials who did not celebrate their own accomplishments ceded the narrative.
  This is a specific, replicable error to avoid.

**Do not use when:** the "equal vigor" would require asserting things you don't actually
believe. The counter is sustained honest confidence, not matching dishonesty.

## Final-answer structure

### Judgment
- Is this a sleeper-effect campaign? Which gates flagged?
- Which operators are load-bearing here?

### What Would Change My Mind
- What would suggest the claim is actually true or genuinely contested rather than a planted
  narrative?
- What would suggest the user should step away from narrative warfare (e.g., the battle is lost
  and engaging further only amplifies)?

### Next Action
- Concrete: what you say, in one sentence, repeated at what cadence, for how long. Vague
  "push back harder" is not a plan.

## References

- `references/source-notes.md` — Hovland-Weiss, Argyris, Orwell, Brinkley lineage
- `references/rejected-candidates.md`
- `references/cases.md` — 2024 campaign cases (immigration, economy, infrastructure, debt);
  origin-myth defense; O'Brien lawsuit; nickname successes and failures