Institutional-Prior Foreclosure is a proposed model of recognition bias in AI-assisted evaluation.
The concept describes a system treating established disciplines as legitimate partly because they already possess familiar vocabulary, citations, and institutional surfaces, while treating structurally similar new frameworks as unsafe, totalizing, or illegible.
The proposed recognition-strip test holds an argument’s structure constant while changing only its disciplinary recognition markers. Differential treatment suggests recognition bias; the same criticism in both conditions suggests a structural objection.
The paper does not claim that novelty proves merit or that every rejection is biased. It supplies a falsifiable comparison procedure whose validity depends on genuinely matched prompts and controlled testing.