Reliabilism holds that a belief is justified if and only if it is formed using a reliable method. For each belief that is formed, however, there seem to be multiple methods used to form it, and it may be that some of these methods are reliable but others are not.

Suppose, for example, that you are walking home at dusk. As you approach your house, you see a distant figure walking towards you. Recognising who it is, you form the belief that your father has come to meet you. By what method did you arrive at this belief?

On one level, the method that you used was sense-perception. On another level, the method used was sight. On yet another, it was night-vision, or night-vision at a distance. Factor in your use of memory, and it is clear that there are many different processes by which you arrived at your belief.

Some of these methods may be reliable, but others are not: perhaps night-vision is reliable, but not at a distance; perhaps vision is reliable, but not at night; or perhaps sense-perception is reliable, but vision is the least reliable of our senses.

So was the belief formed using a reliable method? Yes and no. Is the belief justified? The reliabilist answer, it seems, must be the same: “Yes and no”.

If reliabilism cannot provide a definitive answer to the question as to whether or not a belief is justified, in such a wide range of cases as those in which the above problem arises, then it fails as a theory of justification.