Canadian Journal of Philosophy

I Introduction

Let's say that two individuals are epistemic counterparts iff they happen to be in precisely the same (non-factive) mental states. If one has a veridical experience, her counterparts will undergo a subjectively indistinguishable conscious experience. If she remembers something, her counterparts will seem to recall the same event or fact. If she knows something to be true, her counterparts will believe it to be true. Counterparts always find the same things intuitive. Any difference between those who know a great deal about the external world and their systematically deceived counterparts is a difference the deceived counterparts could never appreciate.¹ [End Page 399]

Suppose that when it comes to her mundane beliefs about her immediate surroundings, Audrey gets things right as a rule. Audrey's counterpart Cooper is systematically deceived. Audrey sees that she is holding a book of matches, knows that she wants to light her stove, and strikes a match to light it. Coop seems to see a book of matches, thinks he has struck a match, and believes he has lit his stove. Because of the demon's efforts, Coop undergoes these experiences in the dark. He strikes non-existent matches with phantom limbs. Intuition suggests that while Coop might be mistaken in nearly everything he believes about the external world, he is no less justified in believing what he does than Audrey is justified in believing what she does. To bolster the intuition and explain why we ought to regard mistaken beliefs backed by hallucinatory experiences as justified and not merely excused, Audi offers these remarks:

… given the vivid hallucination, I am in no way at fault for believing what I do, nor do I deserve criticism. Far from it. I am like a surgeon who skillfully does all that can be expected but loses the patient. There I should feel regret, but not guilt; I should explain, but need not apologize; and when we know what my evidence was, we approve of what I did; we consider it reasonable.²

In general, it seems that what we ought to say about epistemic counterparts is this:

Parity: Necessarily, if S and S' are epistemic counterparts, S is justified in believing p iff S' is justified in believing p.

Once we accept Parity as true, it is hard to see how we could reject Internalism or refuse to reject Externalism:

Internalism: Necessarily, the facts about the justification of a subject's beliefs are fixed completely by the facts about the subject's non-factive mental states.

Externalism: It is not the case that the facts about the justification of a subject's beliefs are fixed completely by the facts about the subject's non-factive mental states.³ [End Page 400]

To see why, consider two versions of Externalism. According to process reliabilism, you cannot have a justified belief about p unless the processes that led to that belief reliably lead to the truth.⁴ According to the knowledge account, you cannot have a justified belief about p unless you know p.⁵ While Audrey arrives at her beliefs by means of reliable processes and those beliefs constitute knowledge, Coop arrives at his beliefs by the most unreliable of means and knows nothing of the external world. Assuming Parity is true, we have to say that he is no less justified than she in spite of these external differences. Thus it seems justification must be fixed by the non-factive mental states that Audrey and Cooper share in common. The point seems perfectly general. If you take the justification of belief to involve some condition that does not strongly supervene on a subject's non-factive mental states, it seems you've adopted a view of justification incompatible with Parity. That your view is incompatible with Parity strongly suggests that your view is mistaken.

Most think that Parity and Externalism are incompatible.⁶ Many think that their incompatibility gives us good reason to reject Externalism.⁷ I shall argue that there is no good argument from Parity to Internalism. In §II, we will see that there are two problems with arguing from Parity to Internalism. First...