<p>Belief ascription in Western contexts is typically guided by the congruency principle: believing something entails, <i>ceteris paribus</i>, acting and reasoning in accordance with that belief. Some people with delusions appear to violate this, leading some to deny delusions a doxastic status. However, this contradicts both scientific characterizations of delusions and experimental evidence that ‘the folk’ consistently interpret delusions as beliefs, incongruencies notwithstanding. The issue isn’t merely philosophical but has repercussions on the agential status of people with delusions. This presents a challenge: can the intuitive appeal of doxasticism about delusions be reconciled with the congruency principle? We argue that regulative approaches to folk psychology, characterized by the ideas that belief ascription is (a) primarily geared towards normative regulation and (b) thoroughly context-sensitive, provide conceptual tools to address this challenge. This perspective builds on and further develops previous attempts to reframe the congruency principle as a <i>regulative ideal</i> that is differentially emphasized across contexts, rather than a fixed factual benchmark. This explains interpreters’ doxastic tendencies as an expression of lower congruency demands in cases of delusions compared to other interpretation contexts. Drawing on empirical literature, we propose possible contextual modulators that could explain lower congruency expectations, including the valence of belief, its ingroup-outgroup status, and the availability of alternative folk-psychological explanations.</p>