The misconception still circulates that Collingwood’s doctrine of re-enactment is a concept of empathy. This claim typically arises from the belief that his philosophy of history shares affinities with the nineteenth-century tradition of Romantic hermeneutics. It supposes that re-enactment consists in a unidirectional recapturing of past mental contents, in which are said to reside the pristine meanings of past texts as intended by their authors. By emphasising the dialectical character of re-enactment, this article makes plain that re-enactment entails no such one-sided transferal. It is right to conceive of Collingwood hermeneutically, but not in the nineteenth-century, empathy-dependent tradition. Rather, as Gadamer illuminated in acknowledging the service that Collingwood’s theories provided in the development of his hermeneutics, Collingwood is better understood as proposing a Hegelian-style integration of past and present thought. He reacted against the individualising psychologism of the anti-Hegelian German historicists and emphasised instead the shared nature of language and thought. A proper account of the context that historical investigation ought to recover involves shifting attention from a methodologically inadequate epistemological conception of re-enactment and empathy to a metaphysical one concerned with exposing the foundations of discourse upon which past agents believed, thought and acted. The myth that re-enactment belongs to a discredited hermeneutics of recovery is set against Collingwood’s attempt to depsychologise historical thinking and within his project to reconcile history and philosophy, epistemology and metaphysics.