Abstract

In a nutshell, I construct and defend an analysis of empirical knowledge. Call this analysis A . A is very eclectic, but it is best described as a reliabilistic-defeasibility analysis of knowledge. Its virtues are these: A seems to be able to handle Gettier examples. A overcomes the 'social aspects problem' of knowing. A makes sense of this phenomenon: Sometimes we are willing to attribute knowledge to S who has evidence E for believing that P. But, at other times, we are unwilling to say that S knows that P even though he has the same evidence for believing that P as he had in the first case. And , given A , knowledge is possible despite the possibility that we are brains in vats