The claim
The owner, at U24 of the exchange of 2026-10-20, quoting A23: ”> The deeper thesis is that the organization of recognition and the organization of practical authority can become coupled, self-maintaining processes. / this seems to me reminiscent of themes in feminist philosophy, e.g. perpetuation of privilege”. A24: “feminist epistemology supplies one of the clearest existing articulations of the phenomenon we’ve been reconstructing”, and “Our contribution would be to connect those accounts … not to rediscover epistemic injustice under new terminology.”
The loop is held. CLAIM-277: one chain is good policy, then endorsement, then greater influence, then more chances to show competence; the other is early advantage, then greater influence, then more visibility, then greater endorsement. “This is how a system might develop both well-calibrated expertise and undeserved authority.” Its note of 2026-10-18 runs it from the side that loses: “Low exposure → Poor competence estimates → Low endorsement → Even lower exposure”. CLAIM-330 (Active, demonstrated) is its linear core: interpretation shaping authority and authority shaping interpretation destabilize symmetry exactly when βγ > (1 − α)(1 − δ).
The case is held. THEORY-327, from Fricker 2017 following Anderson: “a list of expert witnesses that no one has refreshed for a good while, so that (through no one’s fault, let us add) the same old white male witnesses tend to be drawn on”. The others “are effectively silenced”. “Their views are not sought, and this is owing to a structural prejudice (the un-refreshed list).”
The claim. The list is the loop’s state. Who is on it decides who is consulted, and being consulted produces the credentials, citations and visibility by which the list would be refreshed. A24 §3 gave the chain, “Prior institutional standing → Opportunities to contribute and be recognized → Visible indicators of expertise → Future institutional standing”, and the point that makes it self-concealing: “a measure of demonstrated competence can partly reflect the history of access to opportunities for demonstrating competence.” So the arrangement can keep an exclusion going “without requiring a continuously acting discriminator” (A24), and the institution’s own evidence appears to justify it. A26 §3: “Institutional recognition can depend on evidence of prior recognition.”
What the loop cannot do. The same dynamics entrench an earned lead and an unearned one. A24: “Our mathematical instability condition in CLAIM-330 doesn’t determine whether an emergent authority difference represents deserved expertise, legitimate specialization, or the perpetuation of privilege.” THEORY-327’s condition is the missing premise: “An omission from a list for reasons that track no misjudgement of the omitted is not covered.” That is why the claim is interpretive, and why A24 called its own version “a proposed general organizational mechanism, not a newly demonstrated theorem”. CLAIM-330’s demonstration is not inherited.
Where it bears
- QUESTION-087, whether a self-reproducing authority structure can be pathological by a standard internal to the organization. A26: “The mechanism itself is not inherently unjust. A normative conclusion requires establishing relevant prejudice, wrongful exclusion, or another justified standard of unfairness.”
- CLAIM-276: a collective interpretation’s stability “is no evidence of its truth”.
- THEORY-145 (Argyris): A24 §4’s “the mechanisms determining what it learns are themselves insulated from certain kinds of correction” is single-loop learning, with the insulation sourced in social position rather than defensive routine.
- THEORY-151 (Longino): the correction is uptake and tempered equality of intellectual authority, which CLAIM-403 separates from access.
- QUESTION-072, which asks the same of the record’s own contributors. A24: “our terminology registry and filing procedures are themselves legibility and authority-allocation mechanisms.” Until this claim, THEORY-317, THEORY-326 and THEORY-327 were cited only from that question, and CLAIM-277’s pointer to it was the one bridge to the distributed-agency line.
- Collins 1986 (LIT-1312, NOTE-1042, read in an authorized Portuguese translation): the mechanism, with the prejudice condition built in, predates this claim’s sources in that tradition. Denying a group literacy and then citing its lack of facts as proof of poor judgement is, in the translation, “atribuir a um grupo um status inferior e depois usar esse status inferior como prova da inferioridade do grupo” (assign a group an inferior status and then use it as proof of the group’s inferiority; the gloss is the record’s). The status is assigned by domination, so the prejudice the loop lacks is there from the start.
- Espeland and Sauder (NOTE-1050): standing that brings resources and reputation is produced by having had them, with no continuing prejudiced act (“It’s a self-fulfilling nightmare”, of the third tier; THEORY-854). They call the initial differences “often largely a product of measurement noise”, on cited evidence: the loop entrenching what was neither earned nor prejudiced, which is the claim’s point that the dynamics cannot tell the cases apart.
- Kay, Kasirzadeh and Mohamed (NOTE-1049) assert the loop for generative models, “a feedback loop where dominant narratives are amplified and marginalized voices are further silenced”, with no condition (THEORY-864, Proposed). New in it: the loop can run through credibility given to a system, with no hearer judging the silenced speaker. The prejudice is carried by training data, so the claim’s limit stands.
What it does not say
- That any actual roster, field or institution perpetuates privilege. It names a possible mechanism. A24: “This is a possible structural mechanism, not proof that any particular merit-based selection process is unfair”, and “not every inherited expert roster is unjust.”
- That every self-reproducing authority is privilege. That is QUESTION-087, open.
- That Fricker’s account of the wrong is right. It uses structural, pre-emptive testimonial injustice as THEORY-327 holds Fricker to state it, prejudice condition included. Whether that account is right about the wrong is QUESTION-322’s, and nothing here needs it.
- That the dynamics supply the normative criterion. A24: “Feminist social epistemology provides normative and historical distinctions that the dynamical model alone lacks.”
- That the record’s own filing is such a loop. QUESTION-072 is open, and needs a prejudice, not only a grammar (NOTE-853).
- Anything about hermeneutical injustice beyond THEORY-326, or about Pohlhaus’s willful hermeneutical ignorance and Dotson’s epistemic oppression, which A24 drew from them. Both works were sought on 2026-10-21 and are held Unavailable (LIT-1311, LIT-1329). Kay et al. gloss Pohlhaus’s term secondhand (“the continued misunderstanding or misinterpretation of marginalized experiences despite their articulation”), which is not a reading of Pohlhaus.