Quine poses two problems for Peirce’s concept of truth considered as convergence of opinions. The first is the use of numerical analogy, and the second is the uniqueness of the final opinion. This paper responds to Quine’s criticism by arguing that, contrarily to Quine’s assumptions, the underlying space of inquiry is non-Hausdorff for Peirce and that mathematical analogies, which need not be numerical analogies, play a fundamental role in logical reasoning. In both cases I argue that Quine’s interpretation lacks generality. In contrast to this, I highlight the importance of generalization in Peirce’s way of thinking.