This paper is concerned with Foucault's historical methodology. It argues that the coherence of his project lies in its development of a set of tools for unearthing the historical principles that govern thought and practice in the epochs that have shaped the present age. Foucault claimed that these principles are, at once, transcendental and historical. Accordingly, the philosophical soundness of Foucault's project depends on his having developed a satisfactory way of passage between the absolutist purism of the transcendental and the mundane contingency of the historical. The paper shows that the key to seeing how Foucault achieved this desideratum lies in a surprising and largely unexplored methodological tradition that he himself explicitly acknowledged: Husserlian phenomenology as it was taken up, modifed, and practiced in the thought of the philosopher of logic and mathematics, Jean Cavaillès—what I call the phenomenology of the concept.

The essay has four parts. The first sketches the two most prominent lines of interpretation of Foucault's methodology and argues that both are inadequate, not least because they both dismiss Foucault's phenomenological heritage. The second part lays out the rudiments of the neglected strand of the phenomenological tradition inaugurated by Cavaillès's important critique and appropriation of Husserlian method. This serves, in turn, to set the stage for the third part that examines, first, Canguilhem's and then Foucault's distinct projects for grasping the transcendental within the historical, and the historical within the transcendental—their respective continuations of Cavaillès's phenomenology of the concept. The essay concludes with a brief consideration of the pathways that this way of reading Foucault opens up for understanding the nexus of power, knowledge, and subjectivation that came to define his work.