Abstract
We have learned to be suspicious of the claim that a serious account of knowledge must begin at the Cartesian starting point, that is, with private data of consciousness serving as a basis for outward inferences to the world, these inferences proceeding on the security of one or another kind of epistemic collateral ranging from the goodness of a deity to the bruteness of the given. But the good reasons we have for dismissing the egocentric predicament as our motive for epistemology are not good reasons for dismissing a different predicament, one with which it is too often confused: namely, our finitary predicament, the fact that our powers of perception, memory and reason, even when co-operatively applied, are limited. For even if we are bound to conceive of general knowledge as the outcome of fundamental kinds of conceptual co-operation, the fact remains that our knowledge claims are underdetermined by the evidence we have for them; and that the framework, however collectively established, of what Quine calls our ‘immemorial theory’ of the world does not therefore protect particular knowledge-claims from defeasibility.