Discussions about epistemic inequalities have for several years highlighted the need to engage critically and reflexively with the politics of citation. Many authors have called for colleagues to correct longstanding epistemic and material injustices by proactively citing scholars and scholarship from marginalised groups, thereby producing radical knowledge that disrupts power. Analysing the epistemic-political grammar of these calls, I note that they often assume that resistance and disruption are intrinsic to corrective citation – i.e. that citing names understood as marginal will by default undermine relations of power. But is that always the case? Drawing on three sets of empirical examples, I demonstrate that citation often does not have the epistemic and material effects we predict, or hope, it will, and may reinforce some inequalities at the same time as it disrupts others. I show that the effects of citation are complex and contingent because they are shaped by unpredictable interactions between different structures of power, unexpected (dis)connections between global and local inequalities, and dynamic relationships between injustice within texts and inequalities beyond them. I argue, therefore, that we must question the more binary and reifying logics of contemporary conceptualisations of citation and attempt to think about corrective citation differently. To contribute to this rethinking, I draw on several authors to propose an approach that celebrates the potential of corrective citation, but remains attentive to its limitations, foregrounding complexity and opacity, recognising the possible failures of radical epistemic practices, and probing our affective investments in them.