Outwitting Old Nick: Some Remarks on Modality, Skepticism, and the Evil Demon

• Charles W. Johnson
• Auburn University
• 20 February 2003, Memphis, Tennessee

1. Mylan Engel offers an admirable response to a hoary menagerie of skeptical arguments based on the possibility of thoroughgoing error. Descartes’ brilliant illustration of the worry in the Meditations has tagged them as Evil Demon arguments, but the same strategy is employed when, for example, Chuang Tzu wonders whether he knows he is not a butterfly dreaming he is a man, or when we modern science fiction readers wonder whether we aren’t all just brains in vats. (¶ 1)

The skeptic asks us to ponder a possibility: what if—the whole apparent world is nothing but smoke and mirrors (except without the smoke or the mirrors)? What if—it is all a web of deceit, crafted by some monstrous intelligence of incomparable power and malice? It makes our heads spin, and we feel an epistemological dizziness. We thought the ground was firm, but maybe our ordinary beliefs are the work of the Deceiver. For all we know, the demon-world could be the one we inhabit. How could we know whether or not any external objects exist, when Old Nick could have been bamboozling us all along? (¶ 2)

2. Engel is exactly right when he points out a confusion between two different sorts of possibility in such arguments. Alan White makes a similar point in his under-appreciated book Modal Thinking, where he distinguishes what he calls possible actualities from actual possibilities. Actual possibilities are metaphysical possibilia, that we use to speak of potentialities and counterfactuals. We refer to them when we say that it is possible for X to be Y. Possible actualities, on the other hand, are more or less epistemic possibilities: they are what might be real, and we refer to them when we say that it is possible that X is Y. (Consider: I’m wondering about the weather. If someone says, It’s possible that it’s raining outside, I may take an umbrella. But if someone says, It’s possible for it to rain outside, I am no better off for this bit of modal enlightenment.) (¶ 3)

Engel aptly diagnoses skeptical worries in terms of the confusion between these two distinct modal concepts. Because the skeptic confuses metaphysical and epistemic possibility, he starts to think that any damn thing he can imagine might be true—and since it might be true, we can’t say that we know it is false. But this is just a muddle. Just as with the rain, the bare metaphysical possibility of the Evil Demon does not necessarily have any bearing on my actual epistemic predicament. (¶ 4)

3. I am also convinced by Engel’s argument that an epistemic version of the argument cannot get off the ground without question-begging. His argument is, I think, interestingly connected to Moore’s two-fisted Proof of an External World. As with Moore’s response to skeptical arguments, Engel observes that although EA is logically valid, simply trotting it out will not prove the skeptic’s conclusion. A valid deductive argument shows only what follows if certain premises are true. EA shows that the e-possibility of the Demon and knowledge of an external world cannot both be accepted. But consider which of these the skeptic would have us reject—that I know that Here is one hand, and here is another. If the Demon’s epistemic possibility excludes that, then why should I grant it in the first place? Unless the skeptic can come up with some reason to see her murky philosophical intuitions about diabolical beings as more plausible than the hand in front of my face, her argument merely gives a reason to deny one of its premises—and the only response that the Demon deserves is a certain gesture of the hands. (¶ 5)

4. For Engel, though, the skeptic begs the question only in the fallibilistic version of EA. He thinks that EAi is not only valid, but sound. It’s not that the Evil Demon argument doesn’t work—but rather, it only works for infallibilistic knowledge, which is not the only, nor even the most interesting, sort of knowledge out there. (¶ 6)

I want to prod Engel on his distinction between infallibilistic and fallibilistic knowledge. It’s not that I think his distinction is mistaken, exactly. Rather, I’m not sure whether it does what he wants it to do, in terms of describing external world knowledge and diagnosing the skeptic. (¶ 7)

This distinction between fallible and infallible epistemic possibility is itself a modal distinction. Engel characterizes it in terms of the logical possibility of falsity: we have fallibilistic knowledge where we have fallible justification; we have fallible justification if there is a logically possible world in which the belief in question has the justification in question, but is still false. We have infallible justification, on the other hand, if the evidence for a belief entails its truth—if no possible world counts as one where the belief is false but justified. (¶ 8)

Thus, cogito statements count as infallible because, for example, my justification for believing that I am in pain is that I feel pain. But there is nothing that counts as feeling pain without being in pain. I cannot have the justification unless the belief is true. But when it comes to the external world, Engel says: (¶ 9)

Presumably, your current visual and tactile experiences are what justify you in believing that there is a piece of paper before you, and it’s obvious that those experiences do not entail that there is a piece of paper before you, for you can have phenomenologically indistinguishable experiences […] when no paper is there. (¶ 10)

The picture Engel has in mind here seems to be parallel to the picture for a case like this: I see that the person at the lectern is flipping something hidden behind it as he reads an essay. In the past, I have often observed this happening when someone has paper behind the lectern, and rarely when there is no paper behind it, so I infer that there is a paper on the lectern. Now I can’t see whether there is or not, so it could be that, in spite of all the evidence, the essay is actually impressed on clay tablets. But as long as it does turn out there’s paper on the lectern, it seems the evidence was good enough for our ordinary affairs. It would be perverse to say I didn’t really know there was paper there. (¶ 11)

This may be the right story to tell when it comes to inferring the existence of paper hidden behind a lectern. But is it the right story to tell about seeing paper in front of me? Hume called any such picture into serious question when he pointed out that A single perception can never produce the idea of a double existence, and that as a result, we can never observe [a conjunction] between perceptions and objects (Treatise I.IV.ii). If we are supposed to infer the existence of paper from paper-like appearances, it is unclear how we could ever realize that paper-like appearances count as evidence of something else in the first place—because it is unclear how I ever come to know that there is something else for them to be evidence of. (¶ 12)

Hume thought that this meant we have no clear idea of an external world. But we may take it as a sign that we need a new picture of human perception, such that when I see paper, that piece of paper—not the visual and tactile sensations—is the immediate object of my perception. Paper-like visual sensations do not count as evidence at all except insofar as they count as seeing the paper, and they can count as seeing it only if the paper is there. The evidence for the paper is simply that I see it—but that does entail that there is a piece of paper there, and so is infallible, according to Engel’s characterization of the term. (¶ 13)

5. On this infallibilistic view of perception, it could turn out that all my inner states are exactly as they would be when I am justified, but I fail to be justified because the world is not lined up in the right way. This may seem queer. But if Hume was right, no report on a person’s sensuous states could count as justification for belief in external objects, and the fallibilistic view of perceptual knowledge is unstable. In order to avoid skepticism, we must reconsider the epistemological Pelagianism that views justification as merely fulfilling all our epistemic duties. We may instead hold that good works alone are not enough—for the believer to be justified, she also needs grace. (¶ 14)

6. The infallibilistic view of perception would also require us to alter the diagnosis that Engel offers for the pull of Evil Demon arguments. If I infallibly know that there is paper in front of me, then that just is infallible knowledge that the Demon hypothesis is false, and EAi will fall to charges of question-begging just like EAf. (¶ 15)

How ought we to diagnose the skeptic? I think that Engel is right to draw a line here between cogito knowledge and external world knowledge, and he is right to do so in terms of fallibility. But instead of the possibility of falsity, perhaps we should think in terms of the possibility of failure—not failure of the truth of the belief, but rather its justification. (¶ 16)

Suppose that the Deceiver makes me think I see paper. It would seem to me as if the visual sensations I experience are justifications of my belief—but in fact, because I do not see the paper, they are not. By contrast, consider Descartes’ famous cogito ergo sum. There is no such thing as a deception about justification here: if the Deceiver makes it seem to me as if I am justified in believing that I exist, then I am justified in believing I exist. (¶ 17)

Now, as long as there really is a piece of paper I am seeing, I actually am justified, and actually do know that there is paper there. This is truth-infallible knowledge, so there is neither a fallibilistic nor an infallibilistic epistemically possible world without the paper there. However, it is fallible in the justification-failure sense, giving the justification the interesting feature that there is a logically possible world in which I seem to be justified, but am not. (¶ 18)

I wonder, then, whether Engel hasn’t given the skeptic more than he should. I wonder whether the pull of the Evil Demon hypothesis comes not from a sound but irrelevant version of EA, but rather because the skeptic has taken note of a complicated aspect of the metaphysically possible demon-worlds (that is, the failure of apparent justifications), and then confused metaphysical possibility with epistemic possibility all over again. (¶ 19)

7. What I have said here amounts mostly to certain incidental questions. I do not mean that they are trivial—far from it. But even if it turns out that that they reveal some misstep on Engel’s part, I certainly do not think that they undermine his main insights concerning the confusions that are involved in Evil Demon arguments. I am, quite simply, convinced of the truth of his central insights. Think of my comments as trying to fine tune the picture, rather than trying to change the channel. (¶ 20)

What I’ve tried to offer is some quick sketches of where I am not entirely sure of Engel’s picture. Consequently, I have been somewhat sketchy on details and argument. But I hope in further discussion we may return to these questions, and the others that arise from Mylan Engel’s fine essay, in more depth. (¶ 21)

Let’s begin. (¶ 22)