What I reject, then, is the very first premise of Cohen's argument - that if a fact F supports a principle P, there must be an explanation why F represents a reason to endorse P. Rejecting the initial premise is, in general, the best way to handle regress arguments of this sort. To Aquinas' argument that there must be a prime mover since, whatever moves requiring a mover, the chain cannot go back to infinity, the proper response is to object at the outset, with Galileo at our side, "But it is just not true that whatever moves requires a mover.