The system

An ecology, not a list

You can debunk a list of tricks one at a time. You cannot debunk a system that processes debunking as an attack. This page documents the interlock, and what happens to corrections that enter it.

The self-sealing loop

External correction arrives: an indictment, an audit, a fact-check, a lost lawsuit. D1 reclassifies it as an attack. D2 has pre-disqualified the correctors. D12 makes the attack an attack on the audience. Loyalty deepens; the frame’s prediction (“they will come after us”) is confirmed; and the next correction lands on harder ground.

1 · correction
Indictment, audit, fact-check, adverse ruling.
2 · reclassification
D1: this is persecution. D2: the referees are enemies.
3 · transfer
D12: they’re coming after you.
4 · confirmation ↻
The frame predicted this. Loyalty deepens. Return to 1, on harder ground.

The anti-fragility point matters most: the optimal adversarial strategy — more prosecution, more fact-checking — feeds the loop. This is why the system is stable under attack.

The interaction matrix

Rows act on columns. Hover or focus a filled cell for the one-sentence relation. Built from the same data files as the dossiers, so it cannot drift from them.

Directional relations across the full catalog — eight closure devices and four force multipliers. 30 directed relations documented so far; blank cells are undocumented, not ruled out.
D1D2D3D4D5D6D7D8D9D10D11D12
D1 ·consumesfedpowered
D2 mutually·pre-clearsprerequisite
D3 protected·laundersretail
D4 depends·upgraded
D5 retail·gradient
D6 feedsrequiressupplied·
D7 exitcomplements·
D8 upgradesupgrades·licenses
D9 end-stagerequires·
D10 force-multipliesforce-multiplies·ambient
D11 enablesambient·
D12 load-bearingtarget·

The loop model

This is a cartoon, not an estimate. Its purpose is to make one property visible: in a system where attacks confirm the frame, increasing attack intensity can increase cohesion. The parameters are yours to move; the point survives any of them. [illustrative model]

━ in-group cohesion · ━ belief updating toward correction. Model: cohesion(t+1) = cohesion(t) + frame×attack×(1−cohesion) − decay; updating ∝ (1−frame)×trust. When frame strength is high, raising attack intensity raises cohesion; when it is weak, the same attacks work as corrections. Move the frame slider through ~40% to see the inversion.

Empirical anchors

Go deeper — what the polling and the courts show, and what they don’t

Polling from multiple independent series (e.g., Monmouth, CNN, PRRI) tracks the persistence of beliefs in 2020 election fraud among Republican respondents across years of adverse adjudication. [verification pending — exact trendlines to be cited from at least two independent series]

The post-2020 election litigation record — roughly sixty losses, including before judges appointed by the plaintiff’s own side — was absorbed without frame revision. [verification pending — case-tracker citation]

Honesty note: persistence-of-belief is consistent with the loop but does not by itself prove the mechanism. Correlation and mechanism are kept separate on this site as a matter of method — see Method.