In this chapter I examine debates over the relative merits of qualitative and quantitative research in education from the perspective of Michael Oakeshott's critique of rationalism in the study of human conduct. Contrary to the positivist view that causal explanation based on randomized experimentation is the highest standard of knowledge, I argue that when it comes to the study of human subjects, even statistical generalizations depend upon a prior form of qualitative understanding. The chapter concludes by considering some consequences of this perspective, which I call "transcendental pragmatism,' for the practice of inquiry in education.