Dogmatism Under Miners Skepticism

7/29/2019 Dogmatism Under Miners Skepticism

1/36

Dogmatism, Underminers and Skepticism

Matthew McGrath

1 Introduction

Consider three passages from leading epistemologists of the past half-century:

Our general principle is this:

P4 For every x, if there is a way of appearing such that (i) it is self-presenting and (ii) x

is appeared to in that way, then the following is evident for x provided it is epistemically

in the clear for him and something he considers: there is something that is appearing thatway to him.

Being thus appeared to puts one in contact, so to speak, with external reality. (Chisholm

1982, 18)

direct realism can handle the problem of perception by adopting nondoxastic primafacie reasons such as the following:

xs looking red to S is a prima facie reason for S to believe that x is red.

This means that the perceptual state itself is the reason, and not a belief about the

perceptual state. (Pollock 1986, 177)

[the modest foundationalist] idea is that there is some more direct relation between

experience and belief that is crucial here. Their idea is that, at least in the typical case,when you have a clear view of a bright red object, then your experience itself justifies thebelief that you are seeing something red. (Feldman 2003, 77)

Although the terminology is distinct, these epistemologists agree that perceptual experiences

themselves can justify beliefs about how things stand in the external world. So, if they are right,

ones justification for a belief does not always trace to ones justification for anotherbelief; the

chain of justification can terminate in an experience. And if they are right, such experiences

can justify not only beliefs about experiences but about objects in the external world.

In recent work, James Pryor articulates the core element of such views.1 Suppose we say,

as a gloss, that an experience is as of Pjust if the experience is one in which it is perceptually

1 See Pryor (2000, 2004, 2005, forthcoming).

1


2/36

presented (visually, audially, tactily, etc.) to the subject that P. Trying to spell out exactly what it

is for an experience to be an experience as of P might well be a difficult matter, about which

theorists such as Chisholm, Pollock, Feldman, and Pryor might (and it seems do) disagree.

Given the notion of an experience being as of P, the commitment shared by these epistemologists

can be expressed in the thesis that experiences as of P prima facie justify one in believing that P,

where P is a proposition about how things stand in the external world, and that the experiences

do so themselves, not only in conjunction with other justified beliefs, and do so directly, not

through a chain of justification involving other justified beliefs. As Pryor puts it, experiences as

of P prima facie immediately justify one in believing that P. He calls this thesis dogmatism

about perceptual justification, or simply dogmatism. 2

Dogmatism is attractive for both phenomenological and theoretical reasons.

Phenomenologically, as Pryor (2000, 537) points out, our perceptual beliefs ordinarily do not

seem to be based on anything other than how things look, sound, feel, etc. We might have

implicit beliefs about lighting conditions, about the reliability of our faculties, etc., but we dont

seem to base our ordinary perceptual beliefs on such beliefs. These facts about basing might be

taken to be some grounds for dogmatism, at least assuming a fairly strong connection between

justifiers and basesfor belief. More compelling are considerations of the theoretical advantages

of dogmatism. If dogmatism is true, then we can see how the edifice of our justified beliefs

about the external world is ultimately supported by foundations (Silins 2008, 136-7). Not only

this, but these foundations do not seem like ad hoc posits needed to deliver the right results.

The dogmatist idea is only a more careful rendering of the intuitively attractive idea that it is

2 Strictly speaking, dogmatism is a thesis about what is sometimes called ex antejustification: justification to have

an attitude, whether one has it or not. For most of the paper the distinction between ex ante and ex postjustification

(the having of attitudes which are justified), will not be crucial. However, these matters become important in 6.

2


3/36

reasonable to take how things look, sound, feel, etc. at face value, to trust the testimony of the

senses. So, dogmatisms theoretical advantages are rooted in a basic intuitive plausibility.

There is therefore much to recommend dogmatism. But Pryor sees an additional benefit

to dogmatism. It can assist us in what he calls the modest anti-skeptical project:

The modest anti-skeptical project is to establish to oursatisfaction that we canjustifiably believe and know such things as that there is a hand, without contradicting

obvious facts about perception. (2000, 517)

This project is not easy going, however, because there are compelling skeptical arguments to the

contrary. To successfully complete the project, one must diagnose and defuse these skeptical

arguments; the goal is not to convince the skeptic on her own terms but to put our minds at

ease regarding skepticism (517).

What, then, are the best skeptical arguments? According to Pryor and I will not

challenge this they are arguments invoking epistemic principles that require, for us to be

justified in believing P from certain experiences or grounds, that we are independently justified

in rejecting various possibilities which are bad or skeptical with respect to those

experience/grounds and P.3 Pryor focuses on one such principle, what he calls the skeptics

principle of justification

SPJ: If youre to have justification for believing p on the basis of certain experiences or

grounds E, then for every q which is bad relative to E and p, you have to have

antecedent justification for believing q to be false justification which doesnt rest on or

presuppose any E-based justification you may have for believing p. (531)

3 Following Pryor, let us say that a possibility is bad, with respect to an experience and a proposition, iff it is apossibility in which one would have the experience and ordinarily rely on it in believing the proposition but in which

one would not thereby gain perceptual knowledge of the proposition. Bad possibilities are, roughly, skeptical

possibilities, although some are not as far-fetched as the brain-in-a-vatpossibility. For instance, the possibility that

one sees a plastic lawn ornament that looks exactly like a deer is a bad possibility with respect to my visual

experience as of a deer and the proposition that the thing one sees is a deer. Although a possibility is always bad

with respect to a certain proposition and a certain piece of evidence (doxastic or nondoxastic) for it, I will often

simply leave the relevant proposition/evidence pair implicit and simply speak of bad possibilities.

3


4/36

Other skeptical principles might be weaker than this but still demand that to have justification

based on certain experiences or grounds, you need to have independent justification to reject

various bad possibilities, at least if you grasp them and understand their badness. The modest

anti-skeptical project is to diagnose and defuse arguments based on such principles.

Why think dogmatism would enable us to succeed in this anti-skeptical project? The

answer seems simple, at first blush: if dogmatism is true, all that is needed for perceptual

justification is the having of the experience; its therefore notnecessary that one have

independent justifications to believe or reject other propositions; and so it is not necessary that

one have independent justifications to reject any bad possibilities; since SPJ and its variants

demand such independent justifications, they are false. Not only are they false, we can see why

they are false if dogmatism is true: its just the experience itself thats needed for perceptual

justification, and nothing more. We will examine this line of reasoning in 3.

Dogmatism has anti-skeptical punch, in Pryors words (537), insofar as it enables us to

succeed in the modest anti-skeptical project, that is, insofar as it is a plausible and intuitive

account of perceptual justification which can help us understand how and why the best skeptical

arguments go wrong. This paper will argue that dogmatism lacks anti-skeptical punch.

2 Clarifications of key terms

Before turning the relation of dogmatism to skepticism, we need to be clearer on the key

expression immediate justification, which figures in the formulation of dogmatism, as well as

expressions antecedent or independent figuring in principles like SPJ .

Pryor gives us a definition of being immediately justified in believing something:

4


5/36

Say that you are immediately justified in believing p iff youre justified in believing

p, and this justification doesnt rest on any evidence or justification you have for

believing other propositions. (2000, 532)

Rest on is not as clear as one might like (and presuppose, other term Pryor uses, is even more

problematic). Is it being used to express a claim about necessary conditions for having the

justification or the constitution of the justification? Pryor tells us in the same paper that to claim

a justification is immediate is to say something about thesource of ones justification (532); and

this suggests an interpretation in terms of constitution. In later work (2004, 2005), he is more

explicit that constitution is the relevant relation. So, E immediately justifies P just if E justifies P

and Es justification of P is not constituted by justifications to believe other propositions (i.e.,

propositions other than P). I will take immediate justification to be explained in terms of what

constitutes the justification, rather than its necessary conditions.4

Some terminology will make talk of immediate justification more perspicuous. Let us

say that a justification to believe P is justificational ancestry of P. It consists of P, something

directly justifying P, something justifying that direct justifier (if anything does), and so on. What

is it for a justification to be immediate? Suppose we model justifications with chains, in which

the arrows represent relations of directly justifying. An immediate justification of P is one that

has no other justification as a proper part, and so its chain will consist only of one arrow, which

reaches P. Thus, any justifications modeled by

(a) X (b) E (c) X

E Q Q E

4 If we were to opt for the other interpretation of immediate justification, then there would be no doubt that if

dogmatism is true, principles such as SPJ would be false. But we would have to face the question of whether the

phenomenal and theoretical considerations invoked earlier are good reasons to think that there are no conditions on

perceptual justification requiring the having of independent justification to reject skeptical possibilities. This seems

doubtful, especially for the theoretical considerations. Even if we have good reason to posit experiences as

providing foundational justification, questions would arise about whether and how we could go beyond the claim

about whatjustifies to the conditions of having that justification. This distinction is discussed in the next section.

5


6/36

P P P

are mediate, rather than immediate, justifications of P. Let us say, then, that E immediately

justifies one in believing that P iff there is an immediate justification of P in which E is the sole

justifier, or in terms of our models, in which E is the sole anchor. Thus, if E immediately

justifies one in believing that P, the corresponding chain takes this shape:

E

P5

Next: independent justifications. In asking about whether one needs to be

independently justified in rejecting a bad possibility Q, in order to have a certain justification to

believe P, we are asking, as Pryor (2000, 525) puts it, whether one needs a justification to reject

Q which does not rest on that justification to believe P. Again, I read this as being about

constitution. As a first approximation, then, an experience E as of P and proposition P, one is

independently justified in rejecting a Q which is bad with respect to E and P iff one has a

justification of ~Q which does not contain the E=>P justification as a part.

However, I think this first approximation needs amending. Consider

Chain 1:

My experience as of a red wall

I perceive a red wall

I am not dreaming

Suppose Chain 1 represents a justification of ~Q. Is the represented justification independent of

the justification represented by the Chain 2?

Chain 2:

5 If one accepted what Wright calls unearned justifications, we could model these as follows: =>P.

6


7/36

My experience as of a red wall

There is a red wall before me

Not if we are attempting to capture the intuitive idea of independent justifications. If I claimed

the wall was red, clearly relying only on my visual experience, then if you demanded that I

needed independent justification to believe I was not dreaming, then Chain 2 would not satisfy

you. But it does not contain Chain 1 as a part.

A second proposal is that in order to be independent of E=>P justification, a justification

to reject a bad Q must not contain the justifier of P, the experience E. But this is too strong.

Consider this reasoning:

I have a very realistic experience as of a red wall, and given that dreaming experiencesare not this realistic, Im not dreaming.

If we model the justification to this reasoning by a chain, we must add a justifier for the

proposition that my experience is realistic. The justification for this proposition presumably

comes from the experience itself. The resulting chain would therefore look like this:

My experience as of a red wall Background evidence (?)

My experience as of a red wall is very realistic Dream experiences are not this realistic

Im not dreaming

This justification is independent of the experience/red wall justification in the desired sense and

yet it contains the experience.

Here I will hazard an account of my own. What appears to qualify a justification to reject

a possibility which is bad with respect to E and P as independent of the E=>P justification is

its not containing E as its sole ultimate justifier. In terms of chains: a justification counts as

independent of a putative immediate justification by E of P just in case the chain correctly

7


8/36

modeling that justification does not contain E as its sole anchor. This is how we will understand

the talk of independent justification in what follows.6

One reminder before proceeding. An important qualification of the dogmatist claim is

that experiences areprima faciejustifiers. They justify in the absence of defeaters. We should

therefore think of the E=>P chain as depicting a structure of prima facie justification which, in

the absence of defeaters, is a structure of (ultima facie) justification.

3. What it would take for dogmatism to have anti-skeptical punch

Dogmatism has anti-skeptical punch iff it provides a way to diagnose and defuse skeptical

arguments demanding independent justifications to reject certain bad possibilities. The tempting

line of reasoning mentioned earlier was that since dogmatism holds that experiences as of P are

justifiers all by themselves, the truth of dogmatism guarantees that perceptual justification does

not require independent justifications to reject bad possibilities.

6

There are several minor modifications we might wish to make to this definition. Suppose that ones justificationfor ~Q were correctly represented by a chain:

E rational intuition

P P->~Q

~Q

As Robert Howell has suggested to me, one might wish to count this as independent of the E=>P justification. Or,

suppose E* is an experience which is a part of E. Consider the chain:

E*

P

~Q

This might seem not to be independent of the E=>P justification. To capture these intuitions, we could revise our

definition accordingly. The modification needed to avoid the second issue is clear: to be independent a justification

must contain something that is distinct from any part of E. As for the first issue, we might go in one of several

ways. We might allow a justification to be independent if it includes some anchor X, distinct from E (and its parts),

which is analytic or conceptual in some favored sense. Since I doubt these modifications will be important in

anything that follows, I will ignore them and work with the simpler definition in the text.

8


9/36

However tempting it is, this reasoning ignores theprima facie clause in dogmatism.

Consider how the tempting reasoning would proceed if we take account of this clause.

Dogmatism claims that experiences as of P prima facie immediately justify believing P.

Experiences as of P prima facie immediately justify believing P only if whenever someone has

an experience as of P, and lacks defeaters, then that experience immediately justifies believing P.

So, if dogmatism is true, then in the absence of defeaters, all it would take to have perceptual

justification for P is to have an experience as of P. Notice that we cannot complete this

reasoning as follows: and one can have the experience without having the independent

justifications to reject any bad possibilities; and therefore one can have perceptual justification

without having such independent justifications. Yes, one can have the experience without having

the independent justifications; but the issue is whether this would be a case of defeat.

How could it be that whenever someone had the experience as of P and lacked the

relevant independent justifications that this was a case ofdefeat? We should not read too much

into the word defeat. The guiding idea is simply that relations of justification, even of

immediate justification, obtain only under certain conditions. Some of these conditions will

concern the lack ofcounterevidence; but others might be merely enablingconditions; and still

others might be mere necessary conditions. The dogmatist might wish to say, for instance, that

someone could have an experience as of P, lack anything we would be willing to

counterevidence, but still not be justified in believing P, because the person did not understandP

or perhaps because the person lacked the capacity to believe P. These might be counted enabling

conditions. Or, again, the dogmatist might wish to say that no one could have any justification

for anything without having justification for thinking that one exists (Silins 2008, 131-2). This

seems to be a case of a mere necessary condition. Now, suppose that it was either an enabling

9


10/36

condition or mere necessary condition of an experiences immediately justifying one in believing

P that one had independent justification to reject certain bad possibilities. Then we could see

how dogmatism could be true even though a principle demanding independent justifications was

also true. Silins (2008) has argued for precisely such a view. According to Silins, an experience

as of P can immediately justify one in believing that P, but it is a necessary condition of its doing

so that one have independent justification to reject certain skeptical possibilities.

The dogmatist must concede, then, that dogmatism does not entailthe falsity of

principles demanding independent justifications to reject bad possibilities. Still, she might insist

that nevertheless dogmatism takes us most of the way to rejecting them and to doing so on

principled grounds. For, consider the familiar non-dogmatist view that a perceptual justification

cannot consist merely in the having of an experience but must also contain independently

justified beliefs in the reliability of ones experience. On this view, there is no quick way with

skeptical principles demanding independent justifications; for, on this view independent

justifications to believe otherthings are required for perceptual justification for the simple reason

that they are constitutive of it. But once we accept that a perceptual justification can be

immediate, we overcome this serious obstacle.

All the dogmatist needs if her dogmatism is to have anti-skeptical punch, in fact, is the

plausible assumption that an experiences immediately justifying a belief does not require ones

being independently justified in rejecting anybad possibilities. Given this, the truth of

dogmatism guarantees the falsity of SPJ and its relatives. The question is whether the dogmatist

is entitled to this assumption. There is a good case to be made that she is, at least prima facie.

The assumption seems supported by the basic Humean methodological principle that, other

things being equal, we should not posit necessary connections between distinct existences.

10


11/36

Immediate justifications do not have among their parts justifications to believe other

propositions, including propositions asserting the falsehood of bad possibilities; thus, an

immediate justification of P is a distinct existence from those independent justifications. In the

absence of a special reason to think that in such cases there are necessary connections, we are

entitled to think there are not.

There are exceptions to this methodological principle. As noted above, one might need to

be justified in thinking one exists to have any justifications at all. If so, then even an immediate

justification for such an external world proposition would require that one had a justification for

the proposition that one exists, and presumably one which is independent of the immediate

justification. Or again, perhaps one could not have any justifications for a proposition about a

things being a zebra without having the concept of a zebra, and perhaps having the concept of a

zebra requires having justifications to believe propositions likezebras are animals. This would

be another instance in which even an immediate justification would require that one had other

justifications. Or again, for an example more directly related to our concerns in this paper,

suppose that one could not have a justification to believe that there is a zebra before one without

also having a justification to believe that there is something before one. So, even immediate

justification would have this necessary condition.

So, the dogmatist who hopes to defuse and diagnose skepticism should agree questions

about constitution in general are distinct and not equivalent to questions about necessary

conditions. But she can nevertheless appeal to the general Humean principle in defending her

assumption that if dogmatism is true, perceptual justification does not require that one be

independently justified in rejecting any bad possibilities.

11


12/36

I agree with this assessment of the dialectical situation, except for one detail. A minor

modification is needed to make sure the dogmatist is not making life too easy for herself. A

skeptic might not feel the need to insist that someone who had an experience as of P but who

didnt grasp any bad possibilities at all say a very young child or an animal needs to be

independently justified in rejecting those bad possibilities in order to have perceptual

justification. The skeptic will remind us that we grasp these possibilities. Moreover, the skeptic

might even insist that anyone who grasps a bad possibility must have certain independent

justifications to reject it in order to have justification from the experience; perhaps a child might

grasp the possibility that one is dreaming but not understand its badness, not understand its

skeptical potential in relation to his experience and belief. Again, the skeptic will remind us that

we do understand the badness of the dreaming possibility. So, all the skeptic needs to demand

for her argument to have purchase on us is that those who understand the badness of the

relevant bad possibilities need the independent justifications. To guarantee the falsity of such

weaker skeptical principles, the dogmatist needs to strengthen the plausible assumption

accordingly. Making this adjustment and cleaning up the language to make the assumption

precise, we arrive at what Ill call:

The Humean assumption: For any bad possibility Q (with respect to E and P), whether or

not one understands the badness of Q, its not the case that in order for E to immediately

justify one in believing P, one must be independently justified in rejecting Q.

I call it Humean because its best defense appeals to the Humean methodological principle. If,

but only if, this assumption is true, does dogmatism have anti-skeptical punch. The remainder of

this paper argues that this assumption is false.

The natural way to argue against the Humean Assumption to provide a find some E/P/Q

triple such that if one understands the badness of Q with respect to E and P, then E can

12


13/36

immediately justify P only if one is independently justified in rejecting Q. We will consider two

strategies for doing this, one focusing on bad Qs which are alternatives to P and another focusing

on bad Qs which are not alternatives to P but what Pryor (2000, 527) calls underminers.

Examples of underminers are possibilities areI am dreaming; I am in a field containing not only

deer but also lawn ornaments that look from this distance like deer. I will argue that the

underminers strategy can be used to provide grounds for rejecting the Humean Assumption.

4. First strategy: bad alternatives

Suppose you are immediately justified by E in believing P, and suppose you understand

the badness of some bad alternative Q. Pick any such Q. Given your understanding of Qs

badness, you are justified in believing that P entails ~Q. So, you are justified in believing P and

you are justified in believing that P entails ~Q. Next, apply the closure principle for

justification in believing to conclude that you must be justified in believing ~Q, that is, in

rejecting Q. Finally, argue that if the alternative Q is bad, then the only way to be justified in

rejecting it is to be independently justified in rejecting it.

If the argument strategy succeeds, it shows that if one understands the badness of a bad

alternative Q, then in order to have immediate justification to believe P from experience as of P

to have dogmatic justification one must be independently justified in rejecting Q. This

would refute the Humean Assumption.

How might the dogmatist block this strategy? It would be a significant cost to the anti-

skeptical dogmatist to have to deny closure, especially since the basis of the denial could not

plausibly be the familiar just barely problem, which is a serious problem for closure

independently of skeptical worries. If I am just barely justified in believing P and I just

13


14/36

barely know that P entails ~Q, it seems I might fail (just barely) to be justified in believing ~Q.

However, in ordinary perceptual cases one is surely much more than just barely justified in

having the perceptual belief that its a deer, its a tree, I have a hand; and ordinary adults know

quite well that the relevant entailments hold and not just barely that being a deer implies not

being a non-deer which is a lawn ornament, that being a tree implies not being a mere hologram

of a tree, and that having a hand implies not being only a brain in a vat. So, just barely

difficulties with closure will not do the dogmatist any good. The dogmatist would have to mount

a full-on attack on closure.

The better response for the dogmatist, and the one I think most dogmatists would wish to

make, is to attempt to show that one can be justified in rejecting bad alternatives non-

independently. The question is whether the dogmatist can plausibly argue that there are genuine

justifications solely anchored solely in the experience as of P and terminating in the falsity of a

bad alternative to P. Call anyone who thinks such chains can correspond to genuine

justifications aMooreanabout bad alternatives. Is such Mooreanism credible?

We now are on familiar ground. This is the territory of the recent debates over Moorean

responses to skepticism. Consider the chain:

Experience as of zebra

The thing I see is a zebra.

The thing I see is not a cleverly disguised mule.

There are familiar objections to the claim that such a chain be a genuinely justificatory. The

best-known is that the chain fails to be justificatory because it is question-begging or circular in

some way (Stroud 1984). If there is a question about whether the thing seen is a cleverly

disguised mule, an inference corresponding to this chain will beg the question. More recently, a

14


15/36

probabilistic objection has also been raised (White (2006) and Cohen (2007)). How could my

experience as of a zebra serve as the sole anchor for a justificatory chain for the conclusion that

its not a cleverly disguised mule when my experience lowers the probability of that conclusion

(and even raises the probability it is false)? On a familiar and attractive probability-raising view

of evidence, the experience would count as evidence that the conclusion is false, not that it is

true.7

I will not attempt to adjudicate the debate over Moorean responses to skepticism here. I

will only note that some headway has been made in replying to the anti-Moorean objections.

Pryor (2004) and Markie (2007) argue for a distinction between an inference failing to transfer

epistemic support and an inference not begging the question against a doubting opponent. The

above chain begs the question against an opponent who doubts whether its a zebra because she

doubts whether its a painted mule; but for Pryor and Markie, it transfers justification all the

same. In response to the probabilistic objection, one might argue that the probability-raising

account is problematic as an account of evidential support. One might in particular claim that

evidential confirmation and disconfirmation of a hypothesis should be understood in terms of

providing good reasons for or against believing it but that in general probabilistic models of

confirmation and disconfirmation cannot distinguish between lacking good reasons to believe

something and having good reasons to reject it. Pryor (manuscript) has attempted recently to

construct an alternative model of confirmation and rational updating which avoids these

problems and is compatible with the Moorean E=>P=~Q chain being justificatory.8

7 White (2006, 233) suggests that his diagnosis of the failure of these arguments fits well with the transmission

failure accounts of Wright (2004) and Davies (2000). See also Weisbergs (forthcoming) No Feedback Principle.8 See Klein (1995) for another account of how the Moorean model for bad alternatives might be genuinely

justificatory, and Weatherson (2007) for argument that dynamic Keynesian models of uncertainty, which are

compatible with dogmatism, have advantages over the standard Bayesian model.

15


16/36

Having mentioned this debate, I will set it aside. What should be noticed is that in

arguing against the Humean Assumption we need not restrict ourselves to bad alternatives. As I

will argue in the next two sections, if we expand our range of attention to bad underminers, we

can devise a better strategy.

5. Second strategy: underminers

Consider two questions:

#1: Do we need to be justified in rejecting each of the underminers whose badness we

understand in order to have dogmatic justification?

#2: If we do, can we find some underminer Q, which is bad with respect to some proposition

P and some experience as of P, such that the only way to be justified in rejecting Q is tobe independently justified in rejecting it?

If the answer to both questions isyes, then the Humean Assumption is false and so dogmatism

would lack anti-skeptical punch. In this section I will argue that ifthe answer to question #1 is

affirmative, then the answer to question #2 is affirmative as well. I argue for the antecedent in

the next section.

Return to the zoo example. Suppose I am looking at one of the animals in the zebra pen,

and it is a zebra. I have an experience as of a zebra. Consider the following underminer:

HALF FAKES: half of the zebra-looking animals in the pen before me are zebras; the

other half are not zebras but mules cleverly disguised to look exactly like zebras.

I appreciate the badness of this possibility. Recall that we are assuming that the answer to

question #1 is affirmative, i.e., that having dogmatic justification requires being justified in

rejecting the underminers whose badness one understands. Given this assumption, I need to be

justified in rejectingHALF FAKES. If anti-skeptical dogmatism is true, I would need to be non-

16


17/36

independently justified in rejecting it. I would need to have a Moorean justification to reject it,

i.e., a justification which is grounded solely in my experience as of its being a zebra.

The most obvious Moorean model here is:

Experience as of a zebra

The thing I see is a zebra

TheHALF FAKESpossibility doesnt obtain

A second model includes a proposition about the experience as an element:


The thing I see is a zebra. I have an experience as of a zebra.

I have a veridical experience as of a zebra over there.

TheHALF FAKESpossibility fails to obtain.

Compare these to the model for alternatives which we discussed in the previous section:

Experience as of zebra

The thing I see is a zebra.

The thing I see is not a cleverly disguised mule.

All three models are subject to the circularity and probabilistic objections. But the models for

underminers have an additional problem: the final links in the chains are not strong enough.

Picture, if you will, G.E. Moore, standing before the zebra pen, pointing and arguing,

That animal is a zebra, so it is not a cleverly disguised mule. This argument, though it might

seem to beg the question, does at least have the appearance of a proof. Crucially, the conclusion

follows from the premise.Now picture Moore pointing and arguing, That animal is a zebra, so

its not the case that half of zebra-looking animals in the pen are zebras and the other half are

painted mules. This doesnt have the appearance of a proof. The conclusion doesnt follow

17


18/36

from the premise, nor does it seem substantially supported by it alone. Whatever justification

provided here would therefore seem relatively weak, and not enough to justify outright rejection

of theHALF FAKESpossibility. To put it probabilistically: ZEBRA doesnt substantially

increase the probability of~HALF FAKES. Nor doesZEBRA together with the fact that one has

the experience as of a zebra. At best these providesome support, nowhere nearly enough for a

proof or even a strong case of the sort needed if one is to acquire justification good enough to

reject ~HALF FAKES.

The general point is that underminers differ from alternatives in that the falsity of an

underminer U is not necessarily strongly supported by the target proposition P, nor even the

conjunction of P and the fact that one has the experience as of P. HALF FAKESis not at all

unique in this respect. It is easy to construct further examples (e.g., take U as many of the hills

you see are much farther away than they lookand P as that hill is near, or take U assome of the

fruit-looking items in the bowl on the table are plastic and P as that is an apple). And there are

examples concerning other putative sources of justification as well (e.g., take U asJohn is quite

often insincere and P as the proposition John just presented to you as true, or take U as my

memory for dates of the Stuart kings is spotty and P as some proposition you seem to remember

concerning a Stuart king).

In fact, not only is it easy to find examples like this, I suspect it can be done for any

putative case of dogmatic justification. All we need to do is cook up appropriate half-fakes or

barn-faade country-style underminer U for a given E/P pair. One recipe might be to let U = it

could easily have been that I have E but not-P, and it could easily I have been that I have E and

18


19/36

P. Another might be to let U =I am a victim of an evil scientist who seamlessly envats me every

other day. Other recipes might be available.9

Finally, I do not wish to give the impression that the Moorean models for underminers

encounter obstacles only with the sorts of underminers Ive discussed. Even the old favorites

like dreaming differ from bad alternatives in that the target proposition by itself intuitively

provides weaker evidence against them. Imagine Moore pointing at the pen and arguing, Thats

a zebra, so Im not dreaming. This does not seem like a proof in anything like the way Thats

a zebra, so its not a painted mule does. One feels one could object, Well, you could be

dreaming while standing in front of a zebra. Remember the Duke of Devonshire! Moore would

need, and the actual historical Moore knew he needed, something stronger.10

We have considered only the two most obvious Moorean models, which include a weak

step from the target proposition to the falsity of the underminer, but one might avoid the need for

such a step by taking the experience directly to justify something that strongly supports the

denial of the underminer. Thus, consider:


I perceive that the thing over there is a zebra

I know through perception that the thing over there is a zebra

HALF FAKESdoesnt obtain.

Pointing and arguing, I perceive that the thing over there is a zebra, so I know it is; so its not in

a pen of zebra-looking animals only half of which are zebras has the appearance of a proof.9 The Humean Assumption claims that any dogmatic justification could be had without the having of independent

justifications to reject any bad possibilities Q, regardless of whether one understands Qs badness or not. To refute

this assumption it only takes one E/P/Q triple to falsify it. The dogmatist might wonder if she maintains at least

some degree of anti-skeptical punch if there are plenty of E/P/Q triples which dont falsify the principle. The

existence of these recipes for cooking up bad underminers would show that even this weaker sort of anti-skeptical

punch will not be forthcoming.10 In Four Forms of Skepticism, G.E. Moore (1959, 246) maintains that if he knows hes standing up then hes not

dreaming. He does not attempt to argue that if hes standing up hes not dreaming.

19


20/36

Still, it is difficult to see how an experience as of a thing being a zebra could immediately

justify one in believing that one perceives that the thing is a zebra in a sense of perceives that

P sufficient for perceptual knowledge that P. As we noted earlier, it is important to the

motivation of dogmatism, to its avoidance of the charge of being arbitrary, that dogmatism

amount to a fleshing-out of the idea that it is reasonable to take experiences at face value, i.e., to

accept the testimony of the senses. But once we think of an experience as immediately justifying

beliefs in propositions which are stronger than the experiences own content we lose this

intuitive motivation. If the experience presents things as P, and a philosopher claims that it

justifies one in believing a proposition Q, stronger than P, we need a plausible story about how

this could be; and it is hard to see how one would not need to appeal to something additional to

the experience as of P as a further justificatory ground.

One might hope to assuage the difficulty by revising standard assumptions about the

content of perceptual experiences. On John Searles (1983) account, for instance, the experience

in question would have the reflexive content there being a zebra over there is causing this very

experience. However, this account by itself would not do, of course, because even if my

experience has this content, and even if it justifies me in believing it, this justification could not

be transmitted to ~HALF FAKES,because there is just not enough support provided. There

would remain a problematic step in the chain. The thing causing this experience is a zebra;

therefore its not the case that there are half zebras and half fakes in the pen is no proof.

The dogmatist would hope to go beyond the Searlean view by claiming that the content

of the relevant experiences which would standardly be thought to be P is rather the reflexive

contentI am hereby perceiving that P. Taking an experience at face value would then amount to

taking it to be a case of perceptual knowledge, and its being perceptual knowledge does strongly

20


21/36

support ones not being amidst many fakes. But this proposal is a non-starter. There is

controversy about whether cases of so-called veridical hallucination are possible, i.e., cases in

which there is a match between the experiential content and the scene despite the absence of any

causal connection. But it seems wrong to think that whether ones perceptual experience as of a

zebra is veridical depends not only on such causal connections but on whether there are cleverly

disguised mules lurking around out of ones area of visual focus. Henry in fake barn country

surely cannot complain, after he is apprised of the facts, that his experience of the barn was non-

veridical, or that his experience when looking at the barn was misrepresenting the object or his

relation to it. The same goes for the zebra case.

I conclude that ifone must be justified in rejecting all underminers whose badness one

understands in order to have dogmatic justification, then the only way to retain the Humean

Assumption, and so the only way to maintain that dogmatism has anti-skeptical punch, is to

embrace one of three unpleasant claims: (i) an overly strong claim about the content of

experience, viz. that experiences have the contents that they are cases of perceptual knowledge;

(ii) an ad hoc claim that experiences can justify propositions which go well beyond their own

contents; or (iii) an implausible Mooreanism about underminers.

6. The necessity of being justified in rejecting underminers

Consider, again, the zebra case. Must I be justified in rejecting the possibility of

underminers likeHALF FAKESin order to have dogmatic justification that its a zebra? Here we

cannot appeal on a widely accepted closure principle to answer affirmatively. We need a

substantive new argument. In a nutshell, my argument is this:

Either Im justified in believing HALF FAKESobtains, justified in being agnostic about

it, or justified in rejecting it. But if I am justified in believing it obtains, my experience

21


22/36

as of zebra would not justify me in believing the thing I see is a zebra. And if I am

justified in being agnostic about it, again my experience wouldnt justify me in believing

this. So, I must be justified in rejecting it. It follows that in order to have dogmaticjustification from my experience as of a zebra to believe that the thing I see is a zebra, I

must be justified in rejectingHALF FAKES.

This reasoning can be generalized as follows:

1. (Assumption): Im in a situation in which I have dogmatic justification from anexperience (E) as of P to believe P, and U is an underminer whose badness I understand.

2. Either I am justified in believing U, in being agnostic about it, or in rejecting it.

3. I am not justified in believing U (because if I were, I wouldnt have the dogmatic

justification from E to believe P, contrary to (1)).

4. I am not justified in being agnostic about U (because if I were, I wouldnt have thedogmatic justification from E to believe P, contrary to (1)).

So, 5. I must be justified in rejecting U.

From 1-5, we conclude

6: In order to have dogmatic justification from E to believe P, I must be justified in

rejecting every underminer U whose badness I understand.

Note that in (6) there is no claim about how I must be justified in rejecting these underminers

(whether I must be independently justified in doing so) but only thatI must be justified.

The crucial steps that need defense here are (2) and (4). (1) is assumed and the cost of

denying (3) for the dogmatist would be the rejection of the existence of what John Pollock

(1986) calls undercuttingdefeaters, which I assume is out of the question.

6.1 Defending Premise (2)

(2) asserts a trilemma about U: either youre justified in believing it, being agnostic about it or

in rejecting it. Before discussing the grounds for this trilemma, I should emphasize that we are

speaking of (ex ante) justification tohave an attitude, whether you have it or not, rather than the

22


23/36

(ex post) justification of a possessed attitude. It would be psychologically implausible to assert

the trilemma for ex post justification, but not for ex ante justification. We can be justified to an

attitude toward a proposition even when we have never entertained it. For instance, a few

minutes ago, you arguably did not or at least not explicitly take any attitude toward the

proposition that Red Sox played the White Sox more than twice in the 2008 season. However, if

you knew a little about American major league baseball, you were justified in believing that it

was true. Nor does (2) impose unreasonable epistemic demands on us. Following Pryor, let us

construe what it is to be justified in having an attitude toward P as a matter of its being

epistemically appropriate to have that attitude. Epistemic appropriateness is a notion of

permission rather than of obligation. So, (2) does not hold me at epistemic fault if I have no

doxastic attitudes at all toward P. It only claims that one of the three attitudes perhaps more

than one must be appropriate for me to have toward U.

There is some obvious ways (2) could be false but which are ruled out by (1). For

instance, I might not understand the underminer U at all or might not understand its badness with

respect to the experience and target proposition. Are there any other ways for (2) to be false?

Here are two suggestions. First, one might think that there are epistemic dilemmas, i.e., cases

in which, because of competing epistemic demands, no attitude is appropriate. In a case of

epistemic dilemma, for instance, I ought to believe P and also ought to be agnostic about P. But

if I ought to believe P, I am not permitted not to believe P, and so I am not permitted to be

agnostic about P nor permitted to reject P. Similarly, if I ought to be agnostic about P, I am not

permitted to believe P nor permitted to reject P. It would follow that in such a case none of the

three attitudes is permissible to take toward P. If U presented an epistemic dilemma like this, (2)

would be false. Second, one might appeal to borderline cases in which ones evidence or

23


24/36

justification leans toward (or against) U but is on the borderline of what it takes to make

believing (rejecting) U appropriate. One might think in these cases, too, that noneof the three

attitudes is appropriate.

Against the second objection, a reasonable reply is that if the evidence or justification is

on the borderline, and if we dont want to say that only one attitude is appropriate, the best

conclusion to draw is not that noattitude is appropriate but that twoattitudes are appropriate.

Suppose the evidence is on the borderline of the belief threshold for P. Then the idea would be

that both belief and agnosticism toward P are appropriate, but that neither is required. If this is

right, then if Q presented a case of borderline evidence/justification, the trilemma (2) would still

be true.

What about epistemic dilemmas? The most plausible cases of epistemic dilemmas come

from considerations of irresponsibility in the treatment of evidence. Suppose that you have

neglected relevant evidence which would have justified agnosticism about Q. However, you

have now forgotten about it, and later you reject Q in accord with the evidence you have at that

later time. It might be claimed that you ought not reject Q at the later time, nor be agnostic about

it, nor believe it. So, you wouldnt be permitted to have any of the three attitudes. Against this

account, Feldman (2004) replies that in the sort of case described your later rejection of Q may

be irresponsible or blameworthy, but surely at the later time what it makes sense for you to think

is whatever attitude fits the evidence (reasons, etc.) that you have at that later time, and this is

rejection of Q. So, on a notion of epistemic justification tied closely with that of epistemic

rationality of what it makes sense to think at the time this is not a case of an epistemic

dilemma at all.

24


25/36

But the main problem with both arguments is that they have such limited scope. Not

every underminer whose badness one understands presents a borderline case. And if epistemic

dilemmas arise at all, they arise infrequently. We could, if need be, therefore simply restrict the

trilemma to cases that do not present these peculiarities. This would mean adding corresponding

qualifications about Q in assumption (1). And (6) would then be slightly weakened, accordingly.

But this slightly weaker version of (6) would not affect our overall argument against the Humean

Assumption.11

6.2. Defending Premise (4)

Premise (4) states that I am not justified in being agnostic about U. This is the crucial

step in the argument. It, in effect, takes being justified in agnosticism toward an underminer as

sufficient to lack dogmatic justification.

I will defend (4) by appealing to a principle about justification. The principle I will

appeal to is not specifically about dogmatic justification. It concerns perceptual justification to

believe P in ordinary cases in which we have an experience (E) as of P. Ordinary E-based

perceptual justification to believe P is what is found in epistemically appropriate instances of

simply taking things to be as ones perceptual experience (E) presents them as being.

Dogmatists will take such justification to amount to dogmatic justification from E. A non-

dogmatist might understand it in terms of justification from E together with background

knowledge or an unearned a priori warrant to think ones faculties are reliable. But dogmatists

11 Let me explain this. If we added to (1) the stipulation that the underminer Q not present an epistemic dilemma

or present a borderline case, then we would need to add that same qualification to conclusion (6). The examples

mentioned in 5 would then need to be shown not to involve epistemic dilemmas or borderline cases. I think this is

extremely plausible for the case we considered involvingHALF FAKES.

25


26/36

and a great many non-dogmatists can agree that there is ordinary E-based perceptual

justification. 12

The principle to be defended about ordinary E-based perceptual justification is the

Agnosticism Principle: If U is an underminer with respect to E and P, and one understands

Us badness, then if one is justified in being agnostic about U, one lacks ordinary E-based

perceptual justification to believe P.

This is the parallel of the more familiar and logically weaker principle about justification to

believe in underminers

Undercutters Principle: If U is an underminer with respect to E and P, and one

understands its badness, then if one is justified in believing that U obtains, one lacksordinary E-based perceptual justification justify one in believing P.

If the Agnosticism Principle is true, then (4) will be true. This is because if that principle is true,

then if I am justified in being agnostic about U, I will not have ordinary E-based perceptual

justification to believe P; and if I do not have that, then a fortiori I do not have dogmatic

justification from E to believe P.13

I offer two arguments in favor of the Agnosticism Principle. Each parallels an argument

supporting the widely accepted Undercutters Principle. The first argument is based on the claim

that the Agnosticism Principle provides a plausible account of why there is something irrational

about combiningan attitude of agnosticism about an underminer while also believing P in the

ordinary way simply on the basis of E. Suppose I am agnostic about whether the zebra-looking

12 There are prominent positions in the epistemology of perceptual belief which deny what Im calling ordinary E-

based perceptual justification, and among these are disjunctivism as well as Williamsons (2000) knowledge-firstaccount. The disjunctivist will deny the highest common factor and so presumably deny the very Es we are

discussing. Williamson claims that only knowledge justifies belief. For him, if anything justifies one in believing Pin an ordinary case of perceptual justification to believe that P, it must be part of ones knowledge, but an experience

as of P does not meet this condition.

Still, ordinary E-based perceptual justification is common ground between Pryor and Wright and many

others in the dispute over dogmatism. Thanks to Jessica Brown for discussion here.13 Im assuming here that if I did have dogmatic justification from E to believe P, I would have ordinary E-based

perceptual justification to believe P. This is unproblematic because in principle one way one could have ordinary E-

based perceptual justification is to have dogmatic justification from E to believe P.

26


27/36

animals in the pen are only half zebras. Perhaps I think its more likely than not that theyre all

zebras, but the question is open for me. I am agnostic about it.14Now, there is intuitively

something irrational about retaining this agnosticism while at the same time simply relying on

my visual experience as of a zebra in the ordinary way to conclude that the animal Im looking at

is a zebra. If I think to myself, There really might be a number of mules that look exactly like

zebras here; unlikely but possible, and then, while not abandoning this agnosticism I go on

simply to trust my experience by concluding this one is a zebra, Im being irrational. But the

faultiness in my combination of states cannot automatically be pinned on a certain component in

the combination. We cannot conclude that my agnosticism is unjustified, or that my perceptual

belief is. We need an explanation for how the combination could be faulty that doesnt

automatically pin the faultiness on the agnosticism and doesnt automatically pin it on the belief.

Its the irrationality of the combination qua combination that needs to be explained. Such

irrational combinations can be found with other justificatory sources as well. Suppose I am

agnostic about whether you are sincere in your testimony that P but without giving up my

agnosticism I go on simply to rely on your word in believing that P. In having this combination

of states, Im being irrational. But again, its the combination qua combination that is irrational.

We cannot conclude that my agnosticism must be unjustified, or that my testimonially formed

belief must be unjustified. 15

14 It is immaterial to the argument which follows whether my agnosticism here is of the decided indifference sort,as one has when one thinks that the coin has a 50% chance of coming up heads on the next flip, or of a more

unsettled sort, as when one is hard-pressed to say much that is informative about how likely it is and is in the

position of thinking one just doesnt know.15 Crispin Wright (2007) makes vivid the irrationality involved in such combinations. My discussion here draws

substantially on his. However, whereas Wright takes these problems to concern higher-level propositions

propositions about ones justification I am here considering underminers in general, which need not be higher-

level. See Pryor (forthcoming) for a critique of Wrights claims about higher-level propositions.

27


28/36

The Agnosticism Principle provides a good explanation of why these combinations

would be irrational qua combinations.16 In combining the agnosticism about the underminer with

the simple trust in experience, Im being epistemically self-defeating: Im guaranteeing, by

having this combination, that one of the component attitudes is an attitude I shouldnt have, i.e.,

is (ex post) unjustified. For suppose the agnosticism about the underminer is justified. Then I

must be (ex ante) justified in being agnostic about it, and so by the Agnosticism Principle I lack

ordinary E-based perceptual justification to believe the target proposition P. But if this is so,

then my ordinary perceptual belief that P is unjustified. Alternatively, suppose the perceptual

belief is justified. Then I must have ordinary E-based perceptual justification to believe P (if I

didnt, my ordinary perceptual belief that P wouldnt be justified). But then, by the Agnosticism

Principle, I cannot be justified in being agnostic about the underminer. But I am, and so I am

unjustifiably agnostic about it. Either way, one of the two components of the combination is

unjustified.17

There are clear parallels between the intuitive irrationality of these combinations and the

intuitive irrationality involved when one believes the underminer to obtain while relying simply

on the experience. Perhaps in the case of believing the underminer, the combination is even

more irrational, but that is presumably because when you are justified in believing an underminer

U to obtain, it would be even more unjustified form an ordinary perceptual belief that P based on

the experience than it would if you were only justified in being agnostic about U. Just as the

Agnosticism Principle explains the irrationality of the combination involving agnosticism in an

16 Notice that we cannot help ourselves to the same sort of explanation we might give to explaining why its

irrational to combine belief in P with belief in ~P. Here one might hope that the obvious logical incompatibility of

the belief contents would be a sufficient explanation. But in the case of relying on a visual experience in believing P

while also being agnostic about a bad possibility, there is no incompatibility of contents.17 Several inferential steps here rely on a principle relating ex ante to ex post justification. If my belief that P is the

uptake of a certain candidate justifier for P, then my belief is justified iff the candidate justifier is a genuine

justifier for P, and similarly for agnosticism.

28


29/36

underminer, so the Undercutters Principle explains the (greater) irrationality of the combination

involving belief in an underminer.

Moreover, there are clear precedents in philosophy for explaining irrational

combinations of states or attitudes in this way. A powerful example comes from the practical

reasons literature (cf. Broome 1999 and 2002, Dancy 2000). There is something wrong about the

combination ofintending to achieve end E, believing M is a necessary means for achieving E,

and being undecided on whether to achieve M. Im going to bring about E, yes, and I need to

do M to do that, but Im still undecided on whether to do M.18 This is practical self-defeat. But

we cant pinpoint a particular element in the combination that must be irrational. The

irrationality of the combination, however, can be explained by appealing to a principle about

practical justification:

If one is justified in being undecided on whether to M, then if one is justified in believing

that M is a necessary means to E, then one isnt justified in intending to bring about E

The irrational combination mentioned before is irrational insofar as in being in it I guarantee that

one of my states is one I shouldnt have.

The second argument I offer for the Agnosticism Principle is more indirect. Consider a

strengthened variant of the Agnosticism Principle in which E-based perceptual justification is

understood not merely topermitbelief in P but to make it the case that one epistemically ought

to believe P if one has any attitude toward P at all.19 I will argue that there is excellent reason to

accept this principle. The original principle and its strengthened variant become equivalent on

the plausible assumption that ordinary E-based perceptual justification permits belief iff it makes

18 As Broome (2002) notes, we may need to add further beliefs here, such as the belief that you will execute the

intention for the end only if you have an intention for the means.19 Cf. Feldman (2004). For him, the sense in which one ought to believe a proposition P is conditional: if you have

any doxastic attitude toward P, it should be belief.

29


30/36

belief the attitude to have concerning whether P. Plausibly, if you have dogmatic justification

from experience as of a red wall to believe the wall is red, belief isnt just permitted; its the only

reasonable cognitive attitude to have on the question of the color of the wall. Because of this

plausible equivalence, my argument for the strengthened variant is transformable into an

argument for the original Agnosticism Principle.20

Here is the argument. Suppose it is possible to be epistemically permitted to be agnostic

about an underminer whose badness one understands while also having ordinary E-based

perceptual justification making it the case that one ought to believe P. Suppose that I

epistemically ought to believe there is a red wall before me in the ordinary way on the basis of

how the wall looks to me. This is an ought. Suppose also that I am epistemically permitted in

being agnostic about whether there are hidden red lights shining on the wall. This is a

permission. Notice a consequence of these suppositions: uptake on the ought precludes

uptake the permission on pain of irrationality. By uptake or taking up I mean merely the

having of the attitude on the basis of what makes it an attitude one ought/is permitted to have.

To take up the ought is to rely on my experience as of a red well in believing that the wall is red.

To take up the permission is to be agnostic about whether there are hidden red lights shining on

the wall. To do both is to have an irrational combination of states. We should be very reluctant

to think that epistemic justification could be so epistemically unfair, and so very reluctant accept

the strengthened version of the Agnosticism Principle.

20

Even without the original Agnosticism Principle, the strengthened variant could be used in an argument for astrengthened version of (6):

In order to have dogmatic justification from E to believe P which makes it the case that I ought to believe Pif I have

any attitude toward it, I must be justified in rejecting every underminer U whose badness I understand.

This version of (6), together with the results of 5, show that any dogmatism strong enough to deliver epistemic

oughts fails to have anti-skeptical punch.

30


31/36

Similar considerations support the strengthened version of the Undercutters Principle, the

one in which E-based justification understood as grounding an ought. For if this principle is

false, there will be cases in which, merely by believing what one is permitted to believe about the

lighting conditions (that they are abnormal) on the grounds that permit it, and believing what

ought to believe on the color on the ground of ones experience, one hosts an irrational

combination of attitudes. This, again, would be a case in which uptake on an ought precludes

uptake on a permission on pain of irrationality.

The reader will notice that both the Agnosticism and the Undercutters principles in their

strengthened versions allow that in certain cases taking up an ought can introduce irrationality, at

least if one maintains certain other states. But these will be cases in which the other states are

themselves ones which is not permitted to have. If I already have an impermissible agnosticism

about the lighting conditions, then I cant take up the epistemic ought concerning the color of the

wall. But it is not implausible to think that by doing wrong in having an attitude, I can open

myself up to irrationality in adding a further attitude I ought to have. Suppose poor Bill has

fallen from a two-person canoe hes sharing with me. The sea is very rough and he will not

survive unless I help him back in. Suppose I wrongly set out to achieve a horrible end, to end

poor Bills life as soon as possible. I might also know that a necessary means of doing this is to

not help him back into the boat. Still, surely, I oughtto help him back in. But if I do what I

ought concerning whether to help him back in, while retaining my horrible end, I end up with an

irrational combination of attitudes. This seems exactly what we should say about this case.

There is nothing deeply unfair about it. To put it bluntly: thats just the luck one gets if one

does wrong!

31


32/36

There is therefore good reason to accept the Agnosticism Principle. The principle

provides an attractive explanation of the irrationality of combining agnosticism about

underminers while relying on ones ordinary E-based perceptual justification. Moreover, given

the plausible assumption that if ordinary E-based perceptual justification permits a belief in P it

also makes it the case that one ought to believe P (in Feldmans conditional sense), one can argue

for the Agnosticism Principle on the ground that its falsity would imply the implausible

consequence that there can be cases in which uptake on an epistemic ought precludes one from

taking up an epistemic permission, on pain of irrationality.

These two considerations provide support for the Agnosticism Principle, just as parallel

considerations provide support for the Undercutters Principle. With premise (4) defended, this

completes the defense of the argument that in order to have dogmatic justification one must be

justified in rejecting every underminer U whose badness one understands.

7. Conclusion

In order for dogmatism to have anti-skeptical punch, what I have called the Humean

Assumption must be true; that is, dogmatic justification of a belief by an experience must not

require being independently justified in rejecting any bad possibilities, whether the subject

understands the badness of the possibility or not. When we turn our attention to underminers,

we see that the only ways to save the Humean Assumption are to accept either (i) that we do not

need to be justified in rejecting the underminers whose badness we understand in order to have

dogmatic justification for the belief; or(ii) that if we do need to be rejected in rejecting these

underminers, we at least do not ever need to be independently justified in rejecting them. 5

argues that (ii) is unacceptable. The cost of (ii) would be embracing either an implausible theory

32


33/36

of experiential content, an ad hoc claim that experiences can justify beliefs in propositions well

beyond their contents, or an overly ambitious Mooreanism. 6 argues that (i) is unacceptable.

We need to be justified in rejecting the bad underminers whose badness we understand; because

if we were not justified in rejecting them, we would either be justified in believing them or in

being agnostic about them, and in either of those cases we would not be justified in believing the

target proposition, and so a fortiori we would not be dogmatically justified in believing it.

I want to close by making clear what I have notargued. First, I have not argued against

dogmatism. For all I have argued, Silins (2008) dogmatism may well be true. On this view, the

absence of such independent justifications would count as the absence of a necessary condition

for the dogmatic justification. It remains true on this view that experiences as of P prima facie

justify beliefs that P. This is dogmatism without anti-skeptical punch, but still with some of its

principal attractions, especially its modest foundationalism. Second, although I have argued for

the thesis that in order to have perceptual justification one needs to be independently justified in

rejecting underminers whose badness one understands, I have not argued that perceptual

justification requires having a justification to think perception is reliable which does not rely on

other perceptual justifications. Imagine yourself in the zoo again, looking at a zebra. How could

you be justified in thinking the animals in the pen are not half zebras and half cleverly disguised

mules? The answer might well be: on the basis of your background empirical inductive

grounds. The answer is not: on the basis of your present experience as of a zebra.21

Works Cited

Bergmann, Michael (2006). Justification without Awareness (Oxford: Oxford University Press).

21 For helpful comments on early drafts of this paper, I thank Jessica Brown, Clayton Littlejohn, Jeremy Fantl,

Richard Feldman, Robert Howell, Peter Markie, Andrew Moon, Nico Silins, Ernest Sosa, as well as members of the

audience at the Feldmania conference at the University of Texas at San Antonio.

33


34/36

Broome, John (1999). Normative Requirements,Ratio 12, 398419.

_____ (2002). Practical reasoning, in Jos Bermdez and Alan Millar (eds.)Reason and

Nature: Essays in the Theory of Rationality, (Oxford: Oxford University Press), 85111.

Burge, Tyler (1993). Content Preservation,Philosophical Review 102, 4: 457-88.

Chisholm, Roderick (1982). The Foundations of Knowing(Minneapolis, MN: University of

Minnesota Press)

Cohen, Stewart (2007). Why Basic Knowledge is Easy Knowledge,Philosophy andPhenomenological Research70: 2, 417430.

Davies, M. (2000). Externalism and Armchair Knowledge, in P.Boghossian, and C. Peacocke

(eds.),New Essays on the A Priori (Oxford: Oxford University Press).

______ (2004). Epistemic Entitlement, Warrant Transmission, and Easy Knowledge,

Aristotelian Society Supplementary Volume, 78: 1, 213 245.

Dancy, Jonathan (2000).Practical reality (Oxford: Oxford University Press).

Feldman, Richard (2003). Epistemology (Englewood Cliffs, NJ: Prentice-Hall).

(2004) The Ethics of Belief, in hisEvidentialism (Oxford: Oxford University Press),

166-96.

Klein, Peter (1995). Skepticism and Closure: Why the Evil Genius Argument Fails,

Philosophical Topics 23, 213236.

Markie, Peter (2005). Easy Knowledge,Philosophy and Phenomenological Research 70, 1:

406416.

Moore, G.E. (1959). Four Forms of Skepticism, inPhilosophical Papers (London: Allen and

Unwin), 196-226.

Plantinga, Alvin (1993). Warrant: The Current Debate (Oxford: Oxford University Press).

34
http://www3.interscience.wiley.com/journal/119919329/issuehttp://www3.interscience.wiley.com/journal/119919329/issuehttp://www3.interscience.wiley.com/journal/119919329/issue


35/36

Pollock, John (1986). Contemporary Theories of Knowledge (Towata, NJ: Rowman and

Littlefield).

Pryor, James (2000). The Skeptic and the Dogmatist,Nous 34: 517 49.

______(2004). Whats Wrong with Moores Argument,Philosophical Issues, 14, 349-378.

______(2005). There is Immediate Justification. in Matthias Steup and Ernest Sosa (eds.)

Contemporary Debates in Epistemology, (New York: Blackwell).

______(forthcoming) When Warrant Transmits, in ed. by Annalisa Coliva (ed.) Wittgenstein,

Epistemology and Mind: Themes from the Philosophy of Crispin Wright(Oxford:

Oxford University Press).

______(manuscript). Uncertainty and Undermining

Searle, John (1983). Intentionality (Cambridge: Cambridge University Press).

Silins, Nicholas (2008). Basic Justification and the Moorean Response to the Skeptic,

Oxford Studies in Epistemology: Volume 2, 108-140.

Stroud, Barry (1984). The Significance of Philosophical Skepticism (Oxford: Oxford University

Press).

Weatherson, Brian (2007). The Bayesian and the Dogmatist,Proceedings of the Aristotelian

Society 107: 169-85.

White, Roger (2006). Problems for Dogmatism,Philosophical Studies 131, 3: 525-

557.

Williamson, Timothy (2000). Knowledge and its Limits (Oxford: Oxford University Press).

Wright, Crispin (2004). Warrant for Nothing (and Foundations for Free),Aristotelian

Society Supplementary Volume, 78: 1, 167-212.

______(2007). The Perils of Dogmatism, in Susana Nuccetelli and Gary Seay (eds.) Themes

35


36/36

from G.E. Moore, (Oxford: Oxford University Press), 25-47.

Dogmatism Under Miners Skepticism

Documents