What would a satisfying neo-Kantian and neo-Quinean account of logic be like? How should we account for its role as an instrument for acquiring knowledge, and as an instrument for the criticism of theories? How should we account for its special status in the epistemic scheme of things? I intend below to re-apply the distinctions between analytic and synthetic and betweena prioriand aposteriori (paceQuine) to answer these questions in what I hope is a novel and interesting way. My answers will put me at odds with both Kantians and Quineans; for I do not think that avoiding disagreements with either camp puts one wholly in the other.