2. SEEING IS NOT (JUST) VERIDICAL PERCEIVING

ED has it that paradigm cases of perceptual knowledge that P are grounded in the state of seeing that P. But what does it take to see that P? According to Pritchard, a subject can see that P only if he occupies a situation that is “objectively epistemically good.” In other words, not only must P be true but also the subject's cognitive faculties must be functioning properly and the environmental conditions must be such that the agent's cognitive success is not “merely a matter of luck” (Reference Pritchard2012a: 32).

To some, it might seem that whether one sees that P is solely a matter of whether one has a veridical visual experience that P. For example, one might think I see that there is a barn before me whenever I have an accurate perceptual experience as of a barn and there is in fact a barn before me and my perceptual experience bears the appropriate causal connection to the barn before me. But that is not the way Pritchard understands it. For him, veridically perceiving that P is not sufficient for seeing that P (Reference Pritchard2012a: 29).

Consider, for instance, the classic fake barn case. Here a subject is unknowingly traveling through “fake barn country” – a region populated with very convincing barn façades – when he happens to look up and observe one of the region's few real barns. While we might say the subject in some sense “observes a barn,” he does not see that there is a barn before him according to ED. Seeing that P on this view not only entails that P is true but it also entails that the subject occupies an environment where he could not have easily been mistaken about P.

This robust understanding of what's involved in seeing that P might be a bit unusual but it is not without motivation. One of the core components of ED is the idea that seeing that P grounds paradigm cases of perceptual knowledge that P. If we think seeing that P can ground perceptual knowledge that P and we respect the widespread view that the subject in the classic fake barn case does not know that there is a barn before him, then we had better not say that he sees that there is a barn before him. Instead, we had better understand seeing that P in a way that excludes cases like that.

3. KNOWING YOU ARE NOT IN FAKE BARN COUNTRY

Now we are in a position to understand a problem that arises for ED. Suppose I occupy the sort of situation Pritchard calls a “good+” case. I am on an enjoyable ride through the countryside when I observe a barn before me. All of my perceptual faculties are functioning properly and the environmental conditions are such that I could not have been easily mistaken – I am not in fake barn country. My situation is objectively epistemically good. And thus I see that the thing before me is a barn. Furthermore, suppose I possess no reason to think my conditions are not objectively epistemically good. I have no reason to think I might be hallucinating or that there are trick barn façades in the area or anything of that sort. In Pritchard's terminology, my conditions are also “subjectively epistemically good” (Reference Pritchard2012a: 29). This makes for a paradigm case of perceptual knowledge. On the basis of my seeing that the thing before me is a barn, I believe and know that the thing before me is a barn.

Suppose now that I proceed through the following line of reasoning:

The 123 Argument

1. 1. I see that the thing before me is a barn.

2. 2. If I see that the thing before me is a barn, then I am not in fake barn country.

3. 3. Therefore, I am not in fake barn country.

Notice that the conclusion of this argument is not the claim that the barn in front of me is not a fake. The conclusion instead concerns the surrounding area. And it says this general area is not one populated with fake barns. Does this argument provide me with a way to know that its conclusion is true?

According to ED, it does. Recall, seeing that P is a state that is reflectively accessible. Provided, as we are assuming, that I occupy good+ conditions, premise 1 is knowable by a priori reasoning and introspection. Being a matter of the definition of “seeing that P” (at least as Pritchard understands it), the second premise is knowable in the same way. If I put these two pieces of knowledge together, I can infer and thereby come to know that I am not in fake barn country. Thus ED entails that I can look at this one solitary barn, reflect upon my experience, do a bit of a priori reasoning and thereby come to know that the surrounding area is not populated with fake barns. This, in essence, is ED's fake barn problem.

The conclusion of the 123 argument is a bit ambiguous. What exactly does it mean to say I am not in “fake barn country”? Is it that there are no fake barns in the surrounding area or that there are not a great many of them? Or is it something in between? At the very least, if I see that P and if seeing that P requires the environmental conditions to be objectively epistemically good in Pritchard's sense, then it cannot be that most of the barns in my general vicinity are fake.Footnote ² That is one way to understand the point of the 123 Argument. ED entails that, by looking at the lone barn before me and doing a bit of introspection and a priori reasoning, I can come to know that most of the barns in this general vicinity are not fake. And this is so regardless of how many barns there happen to be around here. And whether or not I see that the thing before me is a barn is not just an issue of numbers. The location of any fake barns matters too. Suppose there were only one fake barn in the area, outside my field of vision but right next door to the barn I happened to view as I rode past. In that case, had I looked up a second or so earlier – and I might very well have – I would have been duped. Or suppose most of the barns in the county are real but 95% of the ones on this particular street are fakes. I happen to have looked over at a real one but there are very convincing façades just up the road and a bit further back. These would all qualify as cases where my environment is not “conducive for the formation of the salient range of true beliefs” (Pritchard Reference Pritchard2012a: 29). So if I see that the thing before me is a barn, it follows that nothing like that is going on.

I will leave the exact meaning of “fake barn country” open to different interpretations. But the point of the 123 argument is clear. Provided I occupy good + conditions and see that the thing before me is a barn, I can, according to ED, know that I am not in any of the bad situations. I can know that most of the barns in this general area are not fake. I can also know that there are not fakes next door or just up the road. And I can do that without looking in any of those places or asking a local. All I need to do is reflect upon my experience of the barn right there in front of me and reason a priori from there.

This consequence already strikes me as deeply problematic – even if it comes about only when I occupy “good+” conditions. The problem is not that ED allows me to know that I am not in fake barn country. It is the way ED allows me to do that. There are many ways I might come to know there are not fake barns in my surrounding area. But looking at the solitary barn before me and engaging in some introspection and a priori reasoning, I would have thought, is obviously not one of them.

Although he does not directly address what I have called the “123 argument,” Pritchard explicitly endorses the idea that a subject in a good+ situation can know that he does not occupy any of the “bad” situations. And he defines bad situations to include not only cases where I am looking at a fake barn but also cases where I am looking at an actual barn when there are fakes in the vicinity (Reference Pritchard2012a: 32, 91). So it would appear that this consequence does not seem so shocking to everybody. But anyone unbothered by the 123 argument should stay tuned. In what follows, I shall further explain why the consequence makes trouble and why various efforts to deflect the problem will not succeed.

4. THAT IS NOT A FAKE BARN VERSUS THIS IS NOT FAKE BARN COUNTRY

To begin, it helps to contrast the sort of reasoning that takes place in the 123 argument with more familiar “Neo-Moorean” anti-skeptical strategies. G.E. Moore famously sought to refute the idealist by holding up his hand and announcing “Here is a hand.” Neo-Mooreans think we can know that various skeptical hypotheses do not obtain by inferring their falsehood from some item of commonsense knowledge. For example, I can know that I am not a brain in a vat by virtue of the fact that I know that I have hands; I can know that this is not a fake barn in virtue of the fact that I know it is a barn. Although some find this sort of view question begging or otherwise problematic, there are philosophers who defend it. In fact, one of the main virtues of ED is supposed to be that it can be used to ground just this sort of solution to the skeptical problem (Pritchard Reference Pritchard2012a: 6). So here are some questions. Why should anyone who thinks a Neo-Moorean solution to skepticism is defensible be bothered by the consequence highlighted by the 123 argument? And why should the fact that ED has that consequence present any problem over and above the usual gripes about Neo-Mooreanism?

The first thing to notice is that the 123 argument goes well beyond your garden variety G.E. Moore-style response to skepticism. As already noted, the conclusion of the 123 argument is not that the object before me is not a fake barn; it is about the absence of fake barns in the area around that one. The typical Neo-Moorean will emphasize that some item of ordinary knowledge such as my knowledge that the thing before me is a barn logically entails that it is not a fake barn. And, furthermore, I know about this logical entailment. So, the reasoning goes, if I know it is a barn, how could I fail to be in a position to know it is not a fake barn? But there is no logical entailment from “that's a barn” to things like “there are not a dozen or so fake barns just up the road and a little ways back.” So there is a big difference between the 123 argument and more familiar Neo-Moorean anti-skeptical arguments.

A similar point holds when we compare the problem raised by the 123 argument to the problem of easy knowledge.Footnote ³ Suppose I know that the table in front of me is red. It is necessarily true that if the table in front of me is red, then it is not white and illuminated by red lights. So it would seem that if I know that the table in front of me is red, then I ought to be able to deduce and thereby know that it is not white and illuminated by red lights. But that looks too easy. The issue here seems to arise for any non-skeptical theory of knowledge that maintains commitment to epistemic closure. Whether you see the issue as a problem will of course depend on how sympathetic you are to Neo-Moorean responses to skepticism. In any case, one might wonder whether the problem raised by the 123 argument is just the problem of easy knowledge all over again.

The problem highlighted by the 123 argument is not equivalent to the problem of easy knowledge. The problem of easy knowledge arises because there is a logical entailment from the proposition that the table in front of me is red to the proposition that it is not white and illuminated by red lights. A committed Neo-Moorean might think I can know that the table in front of me is not white and illuminated by red lights by inferring this from my knowledge that it is red. But there is no logical entailment from the proposition that the table in front of me is red to the proposition that there are not white tables illuminated by red lights elsewhere in the room. Even if we are committed Neo-Mooreans who think the sort of easy knowledge philosophers usually discuss is not a bad thing, the sort of knowledge ED lets us get from the 123 argument still looks way too easy.

Pritchard takes great care throughout the course of his book to respond to some problems faced by ED. Although he does not directly address the 123 argument, various points made in response to other problems for ED could be used for that purpose. In the next two sections, I will show how Pritchard’ s responses to what he calls the distinguishability and access problems could be applied in an attempt to solve the problem raised here. But I will also argue that these efforts do not succeed.

5. THE DISTINGUISHABILITY PROBLEM

First we shall consider what Pritchard calls “the distinguishability problem.” To begin, let us consider another example. Instead of a good+ type of case, where I see and know that there is a barn before me, imagine that I occupy much less favorable epistemic conditions. Suppose that, although I possess no reason to think anything like this is going on, I am not observing a barn but rather I am undergoing an ultra-vivid hallucination of one. This is an example of what Pritchard calls a “bad++” case (Reference Pritchard2012a: 29). In this case, I neither know nor see that the thing before me is a barn. But, Pritchard stresses, it is an “undeniable truth” (Reference Pritchard2012a: 91) that this situation is introspectively indistinguishable from the good+ case. The trouble is that, according to ED, when I am in the good+ case, I can be introspectively aware of the fact that I see that the thing before me is a barn. But if ED is true then “why doesn't it follow that one can introspectively distinguish the good+ case and the corresponding bad case after all, contrary to intuition?” (Reference Pritchard2012a: 91). In other words, if seeing is introspectively recognizable, why wouldn't I be able to distinguish seeing a barn from vividly hallucinating one just as easily as I can distinguish a pleasant sensation from painful one? This is the distinguishability problem. This sort of problem appears to be a large part of what drives worries over easy knowledge and opposition to Neo-Moorean responses to skepticism. If a real barn observation is introspectively indistinguishable from a barn hallucination how can the former ever position me to rule out the latter?

Pritchard's answer to the distinguishability problem rests on two distinctions. The first is a distinction between favoring and discriminating epistemic support (2012a: 77–100). My perceptual capacities do not enable me to discriminate between barns and realistic fake barns or ultra-vivid barn hallucinations. But, the idea is, I can still possess evidence that would favor the first over the others. Evidence favors one possibility over another when it provides positive epistemic support for the proposition that that possibility obtains. Evidence can favor even if the degree of support it provides falls short of discriminating between that possibility and all of its possible alternatives.

The second distinction is between an error-possibility that is “epistemically motivated” and one that is “merely raised” (Reference Pritchard2012a: 124). The possibility that I am looking at a fake barn is epistemically motivated when I am provided with some good reason to think it actually obtains. If the possibility is suggested without any supporting reason, for example if someone just out of the blue says “Well maybe it's a fake barn!,” then that error-possibility is merely raised.

These distinctions help solve the distinguishability problem in the following way. Even if we do not possess a capacity to discriminate barns from lifelike fakes or vivid hallucinations, we can still know that we are faced with one rather than the other provided certain other conditions obtain. According to Pritchard, if an error-possibility is merely raised rather than epistemically motivated, then one need not possess discriminating evidence in order to know it does not obtain; favoring support will do just fine. Now, if an error-possibility is epistemically motivated, discriminating evidence is called for. But since the good+ case is defined as one where I see that the thing before me is a barn and I possess no reason to think any error possibilities obtain, no error-possibility will be epistemically motivated in that sort of situation. If I see that the thing before me is a barn, then I have grounds that favor its being a barn over its being a lifelike fake even though I do not have an ability to perceptually distinguish the former from the latter. But, unless the possibility that it is a lifelike fake or a mere hallucination has been epistemically motivated, I do not need discriminating support. I only need support that favors. Thus, provided I occupy good+ conditions, my evidential basis, i.e., my seeing that the thing before me is a barn, will enable me to know that there is a barn before me, and, for the same reason, I can also know that the thing before me is not a fake barn or a hallucination. This is so even if my evidential basis does not provide discriminating support.

Here is how these thoughts might ameliorate the consequence highlighted by the 123 argument. It might be said that if it strikes us as absurd to think this argument could give me good reason to believe its conclusion, this is because we are thinking in terms of discriminating and not favoring epistemic reasons. Of course, the epistemological disjunctivist can say, seeing that the object before me is a barn does not enable me to discriminate between fake barn country and good+ country. The situation where I am looking at a real barn in a region littered with fakes is introspectively indistinguishable from the situation where I am looking a real barn in more normal circumstances. But seeing that the object before me is a barn still favors the proposition that there are not fakes in my general vicinity. And, provided I am in good+ conditions, favoring support will be all that I need because those conditions are defined as being such that the possibility of fakes, if raised at all, cannot be more than merely raised. Thus, provided I occupy the right sort of conditions, I can know that the conclusion of the 123 argument is true without having discriminating grounds.

As already noted this is, more or less, how an epistemological disjunctivist would deflect many of the common complaints surrounding Neo-Moorean solutions to skepticism. Seeing that the object before me is a barn does not enable me to discriminate between barns and barn hallucinations or convincing fakes. But seeing favors the former and, sometimes, favoring is all I need to know. But, even if we accept this sort of distinction, we are owed some sort of explanation for why the grounds given suffice for the subject to know (even though they do not discriminate). It is not enough to introduce the distinction between favoring and discriminating epistemic support. We must also show that the former kind of support is in fact present. Furthermore, even if a body of evidence favors P to a certain degree, it might not favor P enough for me to know that P. Favoring support must be not just present but present to a degree sufficient for knowledge.

Now consider again the case where I occupy good+ conditions and I believe that the object before me is not a fake barn. Why should we think that seeing that the object before me is a barn would even favor that proposition to a degree sufficient to know it? Pritchard answers this question by appeal to the factivity of seeing. If I see that the thing before me is a barn then what I see will “directly exclude the error-possibility” that the thing before me is not a barn but a lifelike fake. In other words, seeing that P provides favoring evidence against Not-P error-possibilities because seeing is factive and those possibilities are logically inconsistent with P (Reference Pritchard2012a: 125–6).

Let us assume that this explanation succeeds at providing an account of how, in virtue of seeing that P, I possess evidence that favors certain propositions to the effect that various error possibilities inconsistent with P do not obtain and this explains how I can know that those possibilities do not obtain. Seeing that there is a barn before me favors the proposition that the thing before me is not a fake and this can enable me to know that the thing before me is not a fake. Even so, we have not addressed the problem raised by the 123 argument. The trouble is that the proposition that the thing before me is a barn is not inconsistent with the proposition that I am in fake barn country. In fact, it is a bit misleading to call the case where I am in fake barn country looking at one of the few real barns an “error-possibility.” If I believe that there is a barn before me in that situation, then my belief is not in error. The classic fake barn case is a truth-possibility. The factivity of the state of seeing that the thing before me is a barn does not explain how that state provides evidence that favors the proposition that I am not in fake barn country. The relevant logical entailment does not exist here. Even if ED can answer the question of how my seeing that the thing before me is a barn provides favoring epistemic support for a belief that it is not a convincing fake, we have no answer to the question of how my seeing that it is a barn provides favoring support for a belief that there are not numerous fakes in the neighborhood. If it does do that, it is not by the same mechanism.

The sense that there is something absurd in the idea that I could observe a solitary barn before me, do some introspection and reasoning, and thereby come to know that I am not in fake barn country does not stem from ignorance or confusion over the distinction between favoring and discriminating epistemic support. It stems from a genuine concern that seeing that the thing before me is a barn does not even favor the proposition that there are not fakes in the vicinity. Or, if it does favor that proposition to some degree, it does not do so to a degree sufficient for me to know it. Thus, even if we agree that Pritchard's pair of distinctions can overturn common complaints about Neo-Moorean responses to skepticism and easy knowledge, it does not resolve the worry raised by the 123 argument.

6. THE ACCESS PROBLEM

If I see that the thing before me is a barn, the presence of fakes in the vicinity is consistent with what I see in the sense that the proposition that the thing before me is a barn does not entail the proposition that there are not fakes in the vicinity. But, one might say, the presence of fakes in the vicinity is not consistent with the fact that I see that the thing before me is a barn, at least as “seeing that P” is understood by Pritchard. Since, by definition, I see that P only if circumstances are epistemically objectively good, the fact that I see that the thing before me is a barn guarantees that there are not fake barns in the vicinity. So why not respond to the complaint raised by the 123 argument by saying that it is not what I see that provides favoring support but the fact that I see it and the fact that I am introspectively aware of the fact that I see it?

This move does not make things better for epistemological disjunctivism but worse. To see why, let us turn to another problem faced by ED and Pritchard's response to it. The “access problem” (Reference Pritchard2012a: 20) for ED arises as follows. Given that seeing is both factive and reflectively accessible, it seems that I could, via reflection, know that I see that P and, via a bit more reflection upon the factivity of seeing, infer that P is true. This is so even if P is a contingent proposition about the external world. Does this mean we have reflective access to contingent external world facts? If the answer to this question is yes, then many will think we have very good reason to reject ED. But Pritchard argues that the answer is no.

To appreciate Pritchard's reasons for thinking ED does not give us reflective knowledge of substantive external world facts, return again to the good+ case where I both see and know that the thing before me is a barn. Here I can know, by reflection alone, both that I see that the thing before me is a barn and that if I see that P then P. But it is a mistake, Pritchard argues, to think I could put these two items of knowledge together, make a modus ponens inference and arrive at reflective knowledge that the thing before me is a barn. It is part of the good+ story that I already believe and know that the thing before me is a barn. And my basis for that belief is that I see that the thing before me is a barn. In other words, reasoning from the premise that I see that the thing before me is a barn to the conclusion that the thing before me is a barn does not generate any new knowledge. In good+ conditions, I already know that the thing before me is a barn. And this knowledge is grounded in my seeing that the thing before me is a barn. Thus it is empirical and not reflective knowledge. The most one can say here is that I can be reflectively aware that my basis for believing that the thing before me is a barn is the factive reason that I see it. But this is not a problem for ED so much as a statement of the position (2012a: 51).

One could also use the 123 argument to raise the access problem. And, if we do, this response to the problem is not applicable. It is not part of the good+ situation that I believe there are not fake barns in my general vicinity. Thus should I choose to make the 123 inference, we cannot say that its conclusion was something I already knew beforehand. Furthermore, my belief that the thing before me is a barn is non-inferentially grounded in my seeing that the thing before me is a barn. But, should I employ the 123 argument, my belief in 3 will be inferentially grounded in my belief in 1 and my belief in 2. Since both of these beliefs are supported by reflection alone, we cannot contend that the conclusion arrived at is empirical. If I reason 123-ly, I arrive at a brand-new belief that I am not in fake barn country. And this belief is supported via an inference from two other beliefs. And these are both supported by reflection alone.

Pritchard may have solved the access problem as it arises for my belief that the thing before me is a barn but it shows up again when we consider the 123 argument. That still looks like a route to reflective knowledge of a contingent external world fact. If we take the fact that I see that the thing before me is a barn to entail that there are no fakes in the vicinity, then, since my awareness of this fact is introspective and my basis for thinking the entailment holds is a priori, the resulting knowledge that results, if it is knowledge, will be purely reflective. And this will mean that we can have reflective access to substantive facts about the external world. This helps us further understand why the 123 argument presents a serious problem for ED.

7. DISJUNCTIVISM EPISTEMOLOGICAL AND METAPHYSICAL

Ram Neta (Reference Neta, MacPherson and Haddock2008) defends what is sometimes called metaphysical disjunctivism (MD) against a criticism made by Mark Johnston (Reference Johnston2004). Perhaps Neta's defense of MD could be adapted as a response to the present criticism of ED. But first a word should be said about the difference between the two views. Whereas ED is a theory about the nature of rational support for experientially grounded beliefs, MD is a view about the nature of experience itself. On some philosophical theories of perception, a veridical experience of a barn and a subjectively indistinguishable vivid hallucination of a barn are fundamentally the same kind of mental state. This is what MD denies. According to MD, an experience as of a barn can be one of two things: either the state of perceiving a barn or the state of hallucinating one. And although these are subjectively indistinguishable, they are nonetheless different states of mind. While MD is a metaphysical view about the nature of experience, ED is an epistemological theory about the kind of support enjoyed by perceptually grounded empirical knowledge. And as Pritchard points out, ED does not logically entail MD. But the two do seem to make a good team (Pritchard Reference Pritchard2012a: 24).

Johnston's challenge to MD is similar to what Pritchard calls “the distinguishability problem” for ED. He imagines a case where a patient is wide awake but made to hallucinate lights above him by a surgeon stimulating his brain. At some point during the hallucination and unbeknownst to the patient, identical looking lights are turned on in exactly the place where the patient is hallucinating them. Then, without announcing it, the surgeon stops fiddling with the patient's brain but the lights remain on. Thus the patient observes lights in exactly the spot where he had been hallucinating some lights earlier. Throughout all this, the subject notices no difference at all but yet, according to MD, there is one. At the end of the experiment, the patient is not in the same mental state that he was when the experiment began. Now he perceives lights above him, before he was hallucinating lights above him. If there is a metaphysical difference between hallucination and perception, why can't the patient recognize it? Johnston thinks the proponent of MD has no good answer.

Neta offers two replies on behalf of MD. First, a proponent of MD can happily grant that sometimes we know we are perceiving and thus not hallucinating. But this does not mean the patient in Johnston's example will know that. At the end of the experiment, the patient is perceiving the lights above him. But his environment, being one where a neurosurgeon has been tinkering with his brain and causing him to hallucinate, is one that makes him prone to mistakes. And, as already pointed out, many epistemologists think if your environment makes you prone to mistakes, it also prevents you from knowing. There is no reason why a proponent of MD could not endorse that idea and offer it as an answer to Johnston's challenge. The patient cannot know he is not hallucinating the lights because he cannot know he is perceiving the lights. But this is not because there is no metaphysical difference between the two. He cannot know that he is perceiving the lights because his environment is one where he could have easily been wrong in that sort of belief.

Neta's second reply to Johnston's challenge is to appeal to attributor contextualism. On this view, raising certain kinds of skeptical possibilities can affect the semantics of our attributions of knowledge – i.e., sentences of the form “S knows that P” – in such a way that these sentences do not express true propositions even though they could if they were uttered in a context where such possibilities are not under consideration. Although one does not often see them put together, Neta argues MD and attributor contextualism are compatible. The metaphysical disjunctivist can grant that the sentence “At the end of the experiment, the patient in Johnston's example knows he is perceiving and not hallucinating” is false. But, he can add, this is not because the patient's mental state at the end of the experiment is the same as his mental state at its beginning. If the metaphysical disjunctivist is an attributor contextualist he can say it is because our consideration of the example puts certain kinds of skeptical possibilities into play for us and this makes it impossible for us to truthfully issue that sort of knowledge attribution. Since we cannot truthfully utter the sentence “At the end of the experiment, the patient knows he is perceiving lights above him,” we cannot truthfully utter the sentence “The patient knows he is not hallucinating.” But that does not require us to give up on MD.

Even if they succeed for their intended purpose, neither of Neta's replies to Johnston on behalf of MD will be of help for ED against the challenge raised by the 123 argument. The analog of Neta's first reply would be to deny that I know the conclusion of the 123 argument (I am not in fake barn country) because I do not know its first premise (I see that the thing before me is a barn) and I do not know that premise because my environment is one where I am too prone to error. But it was stipulated that the 123 argument is employed in what Pritchard calls “good+” conditions. By definition, these are conditions where I veridically perceive a barn before me and there are no fake barns in the area. According to ED, this is exactly the sort of situation in which I will be able to know that I see that the thing before me is a barn.

An attributor contextualist likely would say that considering the possibility of fake barn country raises the standards for knowledge and thus makes it much harder for us to truthfully attribute knowledge of the proposition that the thing before me is a barn. Thus the attributor contextualist will probably think the 123 argument involves some sort of context shift when we move from the first premise to the conclusion. But this will not be of much help to the proponent of ED. Recall that one of the main selling points of ED was that it promises a solution to the skeptical problem. And this solution is a Neo-Moorean one. In the right sorts of conditions, I can know that what is before me is not a fake barn because I know it is a barn. And I know that it is a barn because I see that it is a barn. The anti-Moorean intuition that I cannot know I am not looking at a fake barn in this easy way is addressed by the distinctions between discriminating versus favoring epistemic support and epistemically motivated versus merely raised possibilities. These are all essential parts of ED and this is an altogether different approach to the skeptical problem than that advocated by the attributor contextualist. If we were to extract these parts of ED and replace them with a contextualist theory, it is difficult to see what would be left of ED or of the motivation for adopting it in the first place. The epistemological disjunctivist should not take up with the attributor contextualist even if his metaphysical cousin is ready and willing.

It is worth pointing out that Johnston's example is easily addressed with resources already found in Pritchard's defense of ED. And echoes of Neta's first reply to Johnston can be heard here. Even at the end of the experiment where he veridically perceives the lights, the patient still occupies one of those “objectively epistemically bad” situations. If faced with Johnston's example, Pritchard will simply say that the lights above the patient are on but, because the environment is one where the surgeon's prodding has made him prone to error, the patient does not see that the lights above him are on.

While ED already has the resources to handle the sort of issue raised by Johnston's patient, things are not so easy with the issue raised by the 123 argument. Or are they?

8. AN EASY WAY OUT?

One way ED could avoid the trouble raised here is simply to rescind the fake barn intuition. There are some epistemologists who argue that I can know that there is a barn before me even if there are numerous fakes in the vicinity.Footnote ⁴ If the epistemological disjunctivist follows suit, he can say the same about seeing that there is a barn before me. Seeing that P, on this modified version of ED, would just be a matter of veridically perceiving that P. This move would avoid the problem raised by the 123 argument because the second premise would no longer be counted as an a priori truth. And the proponent of ED could still maintain all the other elements of the theory including its Neo-Moorean response to skepticism. Although Pritchard thinks that I do not see that the thing before me is a barn unless I am not in fake barn country, this is not an essential part of ED.

As simple as this solution is, it will be a hard pill for Pritchard to swallow. Much of his other work on epistemic luck (Reference Pritchard2007) and virtue epistemology (Reference Pritchard2012b) is built around the idea that environmental conditions of the sort described in the classic fake barn case prevent us from knowing. Giving up the fake barn intuition will come at heavy cost for him and many others.

On the other hand, certain elements of ED seem to tell against the fake barn intuition in the first place. One way to understand why people find it plausible to think that I cannot know that the object before me is a barn if I happen to be in an area where fakes are present is that I cannot discriminate between real barns and lifelike fakes.Footnote ⁵ But why not say that I do not need to? Why not say that my observation of a barn favors the proposition that it is barn even if there are fakes around? One might answer by saying that knowledge that P always requires that one could not have been easily wrong about P. But anyone who rejects the fake barn intuition will likely reject this “safety” condition on knowledge.Footnote ⁶ Therefore, if, aside from the problem raised by the 123 argument, we are convinced that ED provides an independently plausible account of perceptual knowledge, perhaps the discussion above is not so much an argument against ED as it is an argument against the fake barn intuition and its resultant requirements on knowing.Footnote ⁷