Abstract In a provocative, yet scarcely discussed, argument at the end of Knowing Full Well, Ernest Sosa has attempted to determine what kind of evidence we possess in support of the belief that our cognitive capacities as human beings are reliable. According to Sosa, we can appeal to considerations of coherence to prove that such capacities are reliable (i.e., it would be epistemically self-defeating to think otherwise). However, Sosa also declares that such considerations are not “determinative, ultima facie” reasons−which is to say, they are to be regarded as defeasible. As we will try to point out, this overall strategy is ultimately incoherent. Furthermore, as we will argue, Sosa fails in attempting to provide us with an analogy between the case of doubting the reliability of the cognitive faculties of an individual and doubting such reliability in the case of the species.