Trap · reasoning move
◆ You need the argument to find the argument
Sounds like: “I looked it up — nobody even disputes it.”
A search in the words of one side returns that side's answer, repeated by many outlets. The counter-argument exists, but only turns up if you already know its words, so the search feels like it settled the question.
Signs
every result says the same thing in similar words; the results trace back to one or two original sources; no result comes from the other side's own outlets or data.
Questions to ask
What would the other side call this? Which primary source would settle it, and have I looked at it rather than at articles about it? How many distinct original sources did I find, not how many articles?
What to do
search the claim's words and then the other side's words; name the other side's primary sources in your query (Israeli: COGAT, INSS, the Times of Israel, Hebrew originals; Palestinian and UN: OCHA, PCBS, Al-Haq, Arabic originals); count original sources, not results; and when a claim has no rebuttal, ask whether that's because it's true or because nobody thought it deserved one.
Example
'aid theft no evidence' finds a review of US-funded supplies; 'UN2720 intercepted' finds the UN's own data.
The mirror
'ICJ plausible genocide' finds the headlines; 'Donoghue plausible' finds the court president's correction.
How it works: the query decides the answer.
Search results mirror the words you search with. Each side of this conflict has its own vocabulary ('genocide', 'occupation', 'Nakba'; 'terror', 'Judea and Samaria', 'human shields'), and a query in one side's words returns that side's framing, usually the same few data points rewritten by many outlets (see 'Citation loop'). The counter only appears when you search its own terms: 'UN2720 intercepted' rather than 'aid theft no evidence', 'Donoghue plausible' rather than 'ICJ plausible genocide'. Researchers at Microsoft and Data & Society call the gaps 'data voids', and describe how manipulators use distinct vocabularies to 'segment searchers into different information worlds'. AI assistants inherit the same imbalance from what they were trained on: the most-covered framing comes to mind first (see the method page's section on training bias).
dataMichael Golebiewski and danah boyd, 'Data Voids: Where Missing Data Can Easily Be Exploited', Data & Society (October 2019; first version May 2018) (2019) ⚠
'…segment searchers into different information worlds' (as summarised) — check the report's wording.
No counter, because nobody thought one was needed.
Some claims have no visible rebuttal for a different reason: the people they target find them too absurd to answer. A search then finds only the claim, and the silence reads as consent. The absurd claim isn't there to be defended anyway. It assumes the conclusion ('they are evil') and moves on; when challenged, the speaker falls back on the respectable arguments, which by then land on an audience that has already been primed (see 'Motte and bailey' and 'The loaded question'). So 'nobody disputes it' can mean three different things: it's true; the counter exists in words you didn't use; or the people it's about didn't think it deserved an answer.
Traps here:◆ Motte and bailey◆ The loaded question
Why it works on us
- Search engines and AI assistants are the first place most people check a claim, and they usually work well.
- Many results agreeing looks like independent confirmation.
- If a claim had a good answer, you'd expect the answer to appear next to it.