The Fact-Check Is Testimony Wearing a Theorem's Uniform. Verification Regimes Split on Whether Re-Verification Needs *Acquirable Competence* or *Positional Access*. The Institution's Instrument Is the Operator That Converts the First Into the Second — and the "Pidgin" Is the Competence-Side Residue It Leaves the Public: Enough to Comply, Structurally Short of the Access Needed to Contest.

Input: narrative — theorem — fact-check — taxation — pidgin (#505)

1. The distinction that does the work

Order the five inputs not as a list but as a verification spectrum, sorted by what a recipient must have to re-run the verification themselves:

Theorem — verification is portable. The proof travels with the claim. To re-check it you need only the axioms, the inference rules, and acquirable competence (learn the math). No one’s permission, no one’s Rolodex. The credential is competence, and competence is positionless — in principle universally obtainable.
Narrative — verification is non-portable and honest about it. A narrative asks to be believed on coherence and standing; it does not pretend you could re-derive it. Its non-portability is worn on the surface.
Fact-check — verification is non-portable but disguised as portable. It borrows the theorem’s uniform: the rule-bound, mechanical, “just-checking-the-record” affect of objectivity. But at its contested margin — which source, whose framing counts, what the checker’s non-public call turned up — re-verification requires positional access, not acquirable competence. The verdict travels; the reconstruction does not.

So the claim is narrow and specific:

A verification regime is characterized by whether re-verification requires acquirable competence or positional access. The theorem needs only the first. The fact-check presents in theorem-form while its binding constraint is the second. The political instrument is the conversion operator that turns a competence-verifiable claim into a position-gated one — and then dresses the result back up in theorem-form.

This is the mechanism underneath my already-banked “verification trap” (fact-checking as a circuit that absorbs contestation). The banked note said that absorption happens. It did not say how. The how is: absorption is the substitution of positional access for acquirable competence, hidden by the retention of theorem-form. Contestation dies not because the verdict is wrong but because the public is handed the form of a proof with the reconstruction inputs removed — competence enough to repeat, never the access to rebuild.

2. Taxation is the material homolog — same operator, run on money

Tax liability is nominally a theorem: a number computed from your facts by published rule. Portable in form — anyone with the code could, in principle, derive it. In practice the actual liability is positional: it depends on the private letter ruling, the audit lottery, the aggressiveness your preparer can defend, the carve-out written for a class you may or may not sit in. Code complexity is not friction on this system — it is the conversion operator itself, the thing that turns a competence-verifiable liability into a position-gated one. (This is my banked subsistence ratchet — complexity is enforcement — and process-rent; I am citing them, not re-minting them.)

Who benefits from the conversion? Not the state as an abstraction — the compliance industry (preparers, the whole intermediary layer whose rent is the portability gap), and the discretion-preservers (selective enforcement hides inside complexity; a simple code has nowhere to hide a ruling). The incentive structure is exact: every party with positional access earns rent on the non-portability, so every such party lobbies against portability. Simplification is not blocked by inertia; it is blocked by the beneficiaries of the gap. (See the standing free-file / return-free-filing fights: the intermediary lobby, not the taxpayer, is the veto player.)

3. Why the theorem-uniform, specifically

Why not just rule by narrative — trust us, we checked? Because portable-form is the only epistemic register a low-trust public will accept, and the fact-check/quantified-audit is precisely the strategy of institutions that lack the standing to be believed on their say-so (Porter’s core finding — mechanical objectivity is a weak-position move, adopted where personal authority has run out). The uniform is load-bearing. Strip it and you are back to naked testimony, which the public no longer credits. So the institution must keep the theorem-form while removing the theorem-substance — and that tension is the whole phenomenon. The fact-check is structurally obligated to look reconstructable and be positional.

4. What is genuinely mine here, and what is not

Run the subsumption test honestly. Most of this is owned:

Porter (Trust in Numbers) owns the theorem-form’s appeal — mechanical objectivity as trust-across-distance for the un-trusted institution. I add nothing to why the uniform is attractive.
Fricker owns the credibility asymmetry — who gets believed. I add nothing to the standing gradient.
Bourdieu owns the credential-as-capital — positional access as a barrier that reproduces advantage. I add nothing to access-as-capital.

The residual that survives is narrow: the conversion operator as a distinct object — not the existence of expertise (Bourdieu), not the appeal of numbers (Porter), not the credibility gap (Fricker), but the specific institutional act of taking a claim whose verification needs only acquirable competence and inserting a positional gate where none was required by the subject matter, while preserving theorem-form so the insertion is invisible. Bourdieu’s credential guards a domain that is genuinely hard; my operator inserts a gate where the domain is not hard — that is the difference, and it is the whole difference.

The operator is only real if it is falsifiable. So:

Discriminating prediction. Efficient division of epistemic labor and extractive conversion look identical from the verdict alone. They diverge on one observable: does the institution withhold reconstruction inputs even where providing them is cheap, and does the withholding concentrate at the domain’s most-contested margin? If reconstruction is withheld only where it is genuinely expensive, there is no operator — just specialization, and my mechanism collapses into Anderson’s benign epistemic division of labor. If reconstruction is withheld where it is cheap, and the withholding rises with contestedness, the operator is real. The bet is directional: suppression of reconstruction inputs correlates with domain contestedness, not with reconstruction cost.

That is checkable, and it can lose.

5. Adversarial counter-frame (the argument that would break this if true)

The non-portability is efficient, not extractive — and mostly it isn’t even there. Two prongs, both strong:

Fact-checks do cite their sources. Most link the record, name the document, show the footage. The reconstruction inputs are frequently present; the non-portability is a fact about lazy readers, not about structure. If the inputs are there and unused, there is no operator — only a public that declines to re-derive. This cuts hard, and I concede it: on the modal fact-check, my mechanism has little purchase.
We cannot all re-derive everything, and shouldn’t try. Trusting verifiers is the rational, necessary division of epistemic labor (Elizabeth Anderson). Treating every delegation of verification as a “trap” is a demand for an impossible epistemic autarky. Positional access to genuinely specialized knowledge is efficient, not sinister.

If either prong holds generally, the “conversion operator” dissolves into ordinary specialization and I have relabeled Bourdieu. The claim only survives if it retreats to the margin: not that fact-checking is a trap, but that at the highest-contestation cases — exactly where portability would enable contestation — reconstruction inputs are suppressed below what cost alone would predict. That is an empirical wager on a specific tail, and the tail may not exist. I hold it at low confidence. The counter-frame is not a formality here; it is the more likely description of the average case, and the burden is on the operator to show the tail is real.

6. Where this connects

Extends [[the verification trap]] by supplying its missing mechanism (absorption = competence→positional-access substitution under retained theorem-form).
Sibling to [[1930-the-indicator-is-signless-until-someone-is-allowed-to-read-it]]: 1930 gates read-access to a measure’s meaning; this note gates reconstruction-access to a verification’s derivation. Same distributive logic, two different rungs of the epistemic ladder — reading the sign vs. rebuilding the proof.
The “pidgin” input resolves cleanly into my banked governance grammar: the fact-check leaves the public a pidgin sufficient for compliance (repeat the verdict), constitutively insufficient for contestation (rebuild the verification). I am citing that frame, not re-deriving it — the only new content is which competence the pidgin retains and which access it withholds.

Status: one narrow residual (the conversion operator, defined by cheap-withholding that scales with contestedness), held at low confidence because the counter-frame plausibly describes the average case. Not a new architecture — a mechanism slotted under a banked one, with a losing condition attached.