This paper argues that liberal epistemology lacks the conceptual resources to diagnose indoctrination under conditions of deep semantic and ontological divergence. Standard liberal accounts treat indoctrination as an epistemic failure attributable to agents: insufficient openness, poor critical reasoning, resistance to evidence, or exposure to inadequate alternatives. Such diagnoses presuppose that epistemic norms are framework-neutral and that agents can, through deliberation, revise the very standards by which reasons and evidence are recognised. The paper challenges these presuppositions. Drawing on philosophy of language (Wittgenstein, Sellars, Quine, Davidson), Bernard Williams’s analysis of thick concepts, and contemporary political psychology, it argues that many apparent epistemic failures are better understood as grammatical failures: breakdowns in concept transfer across incommensurable semantic frameworks. The central claim is not that agents lack agency or that frameworks are immutable, but that liberal epistemology systematically mislocates explanatory constraint. Deliberative agency operates successfully within shared semantic grammars, but it does not plausibly function as the primary mechanism by which those grammars themselves are revised. When disagreement crosses framework boundaries, deliberation contributes little explanatory power even when present. The paper develops this diagnosis through: a distinction between thin (execution-level) and robust (authorship-level) agency, an analysis of why liberal remediation strategies (critical thinking, exposure to alternatives, deliberative openness) presuppose semantic commensurability, and a variance-based explanatory model showing why agency is routinely over-credited in cross-framework disagreement. On this view, education and indoctrination differ normatively rather than structurally: both involve the formation of semantic frameworks, but are evaluated differently from within those frameworks. Liberal epistemology’s attempt to diagnose indoctrination from a neutral standpoint therefore commits a category error. The argument is diagnostic rather than prescriptive. It does not propose a new epistemic regime or educational programme, but clarifies the limits of liberal epistemology’s self-understanding. The contribution lies in showing why certain forms of disagreement are systematically misclassified as epistemic vice, and why this misclassification is not a contingent failure but a structural one.
Bry Willis (Mon,) studied this question.