A lot of what we believe is dependent upon the testimony of others. Moreover, it is hard to see how we could verify for ourselves much of what we are told via testimony since such verification would itself involve making appeal to further testimony-based beliefs that we hold, and so would simply be circular. One response to this problem is reductionism, which claims that we need to be able to offer non-testimonial support for our testimony-based beliefs if they are to be rightly held. The problem is, however, that for a large number of our testimony-based beliefs this is practically impossible, and so reductionism seems to entail that we know very little of what we usually think we know. Another alternative is credulism. This view maintains that we can rightly hold a testimony-based belief even if we are unable to offer independent support for it, at least provided there are no special reasons for doubt. The chief worry about credulism is that it might be thought simply to license gullibility. One way around this is to offer an epistemic externalist rendering of credulism. We consider the epistemology of memory, and find that it raises many of the same issues that testimony does.