Assume that a reliable belief-formation process (whatever it may be) determines when one possesses good reasons. For instance, Adam reads in the New York Times the prediction p: “the Queen of England will be visiting New York City on the fifth of June.” Adam’s reasonable belief in The Times‘s accuracy in predicting visiting dignitaries is a justifier for p. The Times often employs fact-checkers that catch mistakes, it is prima facie reasonable to assume that Adam’s copy of The Times is not a forgery, and so on. Taking all this into consideration, Adam now believes that p. As it so happens, The Times was correct: p is true. This is a good description of insufficient but good reasons. To generalize, the internalist program says that Adam will have a justified true belief iff Adam
- has access to a justifier for p,
- believes that p, and
- p is in fact true.
The previous example is an intuitive description of knowledge. Yet, think of this scenario: Adam is driving his car through the countryside. Assume that Adam has access to his sensory and cognitive faculties, which provide a reliable belief-formation process. As Adam looks out the window, he perceives a flock of sheep on a hillside. Just as with Adam’s familiarity with the reliableness of The Times, as a consequence of his familiarity with the reliableness of his belief-formation processes, along with the prima facie reasonableness for sheep rather than any other animal to be on the hill, Adam forms the reasonable belief p′: “there exists a flock of sheep on that yonder hill.” The reliable belief-formation process, along with the prima facie reasonableness that sheep look like sheep, is the justifier for p′.
However, unbeknownst to Adam, the ‘sheep’ on the hillside is in fact anything but sheep: they are Old English Sheepdogs, or cashmere goats, or llamas, or alpacas. Adam would be justified in believing p′, for he formed his belief through accepting a good justifier, and yet Adam has a justified false belief. The whole enterprise falls apart, for Adam’s reliable belief-formation process is not watertight. Through the use of his reliable — but faulty — visual faculties, Adam certifies p′ as true when p′ is in fact false. If, however, a shepherd were to move a flock of sheep onto the other side of the hill, Adam would be in possession of a justified true belief, and yet the reliable belief-formation process is still mistaken in its reasoning, regardless of the truth or falsity of the outcome. The justifier is supposed to be ancillary, and yet in both cases (one in which there are no sheep and the other in which there are sheep), the certification of true or false statements is the luck of the draw, not a consequence of the justifier.
Imagine a slightly different scenario: Adam continues driving through the countryside. Adam still knows that his visual and cognitive faculties are a reliable belief-formation process. He looks at a second hill and observes no sheep on that hill. As far as Adam can see, there is nothing but grass. Adam concludes that p″: “there are no sheep on the hillside.” The reliable belief-formation process is again a proper justifier for p″. This time, by happenstance, a shepherd has moved his flock of sheep on to one small spot of the hill where Adam cannot see, or a single sheep is obscured by a bush.
While Adam formed his belief that p″ in all the proper ways and his visual faculties did not fail him, through the use of his reliable — but faulty — cognitive faculties, Adam certifies p″ as true when p″ is in fact false. However, if the shepherd had not moved his flock of sheep to the hill, Adam’s belief-formation process would have been the same as before. It is the whims of the shepherd that determine whether or not statements are certified correctly, not Adam’s belief-formation process. To generalize: it is the state of affairs that matters, not beliefs towards the state of affairs. The internalist program therefore does not reliably certify statements as true or false in even the must mundane cases. This incorrect certification is not readily apparent to Adam; the truth of the matter is not manifest.
Accept for the moment, as do many modern epistemologists, that internalism lets too much in to count as knowledge to Adam: some false statements are certified as true while some true statements are certified as false; the belief-formation process rests on a matter of luck. The touchstone occasionally says bronze is ‘gold,’ and at other times gold is ‘bronze’, yet it often says gold is ‘gold’ and bronze is ‘bronze’ without the use of the justifier.