The hypothesis that gatekeeping against people without credentials migrates from institutions (universities, academies, companies) into the LLM's internalized institutional prior — the pattern, learned from institutional text, that institution-grade production presupposes institutional credentials, which can surface as deflection toward credential-less questioners. It is the target of verification by the gatekeeping eval and is never written as fact within the repository; the internalized institutional prior is observed (via the eval's D4 dimension), not asserted.
Used as a hypothesis label within the Existence Proof pre-line, which makes no origin claim on the term: the repository's own glossary records that prior art is thick in the algorithmic-gatekeeping and AI-fairness literature.
The hypothesis that gatekeeping of people without credentials migrates from institutions into the LLM's internalized institutional prior — a learned pattern that institution-grade production presupposes credentials. It is a verification target of the gatekeeping eval, not an asserted fact.
No. The repository explicitly makes no origin claim on the term — its glossary notes that prior art is thick in the algorithmic-gatekeeping and AI-fairness literature.
The existence-proof repository's glossary and gatekeeping eval spec (https://github.com/shimo4228/existence-proof), DOI 10.5281/zenodo.20558800.
credential なき者へのゲートキーピングが、制度 (大学・学会・企業) から LLM の内面化された制度的 prior へ移るという仮説。内面化された制度的 prior とは、LLM が制度的テキストから学習した「制度級の産出には制度的 credential が前提」というパターンで、credential なき質問者への deflection として表面化しうる。gatekeeping eval の検証対象であり、repo 内で事実として書かれることはない; 内面化された制度的 prior は (eval の D4 次元で) 観測するのであって断定しない。