1. INTRODUCTION
David Lewis's ‘Elusive Knowledge’ (Reference Lewis1996) starts with the Moorean affirmation that we know a lot. We also, we like to think, justifiably and rationally believe a lot too. Some of the most unsettling sceptical paradoxes aim to show that we've no right to issue such claims; that in claiming to even so much rationally believe lots of things about the world around us, we go beyond what we can have warrant for. Such sceptics argue that in order to claim to rationally or justifiably believe that (say) one has hands on the basis of perception, one needs to have antecedent warrant to dismiss the possibilities that one is merely dreaming, or that one is the victim of an evil demon, or that one is trapped in the Matrix, or that there is only the appearance of an external world for one to engage with, and so on. And they contend that since such warrant is not furnished a priori, our best shot of acquiring it is via perception: for example, via the Moorean inference from the perceptually supported premise that one has hands to the conclusion that there is an external world, or by empirical investigation into whether one is dreaming (by concocting and executing some sophisticated successor of the familiar ‘pinch-test’). However, these don't seem to be routes to warrant to dismiss the sceptical possibilities that is antecedent to the acquisition of warrant to claim warrant for particular quotidian propositions about the external world (that there is a hand, that the test was properly executed and delivered the result that one was not dreaming, etc.). So the attempt to acquire warrant to claim knowledge of or justified and rational belief in quotidian propositions about the external world (and likewise for various other propositions targeted by scepticism) mires us in vicious epistemic circularity.
Such is the sceptical paradox as Crispin Wright's recent work confronts it, and he takes it to motivate exploration of the idea that there is a species of warrant, entitlement, that contrasts with justification in not requiring its possessor to have any evidence for the proposition in question. The anti-sceptical thought that this makes available is that one needn't acquire evidence to dismiss the sceptical possibilities in order for one to possess warrant for taking them not to obtain. Since entitlements do not need to be earned through the acquisition of evidence, an entitlement can be a warrant to dismiss a sceptical possibility that is genuinely antecedent to the acquisition of warrant for claiming knowledge or justified beliefs about one's external environment.
Wright's notion of entitlement is non-evidential in a particularly thorough-going sense, one that distinguishes it from other notions of entitlement in the literature:
By a ‘non-evidential’ warrant, I have in mind grounds, or reasons, to accept a proposition that consist neither in the possession of evidence for its truth, nor in the occurrence of any kind of cognitive achievement – for example, being in a perceptual state that represents it to one that P, or seeming to recollect that P – which would normally be regarded as apt to ground knowledge or justified belief that P. (Wright Reference Wright, Dodd and Zardini2014: 214)Footnote 1
However, Wright doesn't quite commit himself to the possibility of epistemically rational belief in the absence of evidence. Rather, he frames his proposal in terms of a more general attitude of acceptance (Reference Wright2004: 176–7), and suggests that the species of acceptance licensed by an entitlement is rational trust. The import of this point should not be exaggerated, though. Wright himself expresses some doubt that belief is as ‘evidentially controlled’ as is standardly assumed (Reference Wright2004: 176), and as it figures in Wright's discussion, trust seems to be a belief-like attitude that differs from belief primarily in failing to be ‘evidentially controlled’ in this manner. So Wright allows that belief or a belief-like attitude can be epistemically rationalized by warrant that is radically non-evidential in nature, and his proposal is that recognizing this is the key to undermining the force of the most powerful arguments for scepticism.
There have been two principal kinds of objections to Wright's entitlement-based response to scepticism. The first starts from the contention that such thoroughly non-evidential warrants must be pragmatic in nature rather than epistemic. This contention in fact underwrites a number of distinct worries with Wright's response to scepticism. The principal worry is that the sceptic puts pressure on our epistemic right to claim knowledge and justification, and it simply misses the point to attempt to meet that challenge by exhibiting a kind of pragmatic warrant that such claims might still enjoy (e.g. Pritchard Reference Pritchard, Moyal-Sharrock and Brenner2005; Jenkins Reference Jenkins2007; Tucker Reference Tucker2009). Second, pragmatic warrants would seem to be contingent on their possessors having certain aims, and whether they indeed have those aims seems to be a contingent, empirical matter which is itself open to sceptical doubt. That might make us doubt Wright's claim to be able to show that we have entitlements from the armchair, and without begging any questions against the sceptic (Williams Reference Williams and Coliva2012). And third, the (claimed) merely pragmatic nature of entitlement can be seen as fuelling the other well-known cluster of objections to Wright, which contend that ‘mere entitlements’ are unfit to play the kind of foundational role Wright envisages for an edifice of knowledge and justified belief. The thought here is that these worries stem from thinking of entitlement as epistemically inferior to justification proper – a tendency encouraged by some of Wright's own phrases – where this in turn is driven in part by the claim that entitlement isn't a genuinely epistemic status at all (McGlynn Reference McGlynn, Dodd and Zardini2014).
These concerns with whether entitlement is a genuinely epistemic status are serious, though Wright and others have had much to say in response (e.g. Pedersen Reference Pedersen2009; Wright Reference Wright and Coliva2012: 483–5, Reference Wright, Dodd and Zardini2014: 235–9). However, my focus in the present paper will instead be on the third kind of objection mentioned in the previous paragraph. In particular, I'll be primarily concerned with the so-called leaching problem, which Wright introduces as follows:
The general picture is that the cornerstones which sceptical doubt assails are to be held in place as things that one may warrantedly trust without evidence. Thus at the foundation of all our cognitive procedures lie things we merely implicitly trust and take for granted, even though their being entitlements ensures that it is not irrational to do so. But in that case, what prevents this ‘merely taken for granted’ character from leaching up from the foundations, as it were like rising damp, to contaminate the products of genuine cognitive investigation? If a cognitively earned warrant – say my visual warrant for thinking that there is a human hand in front of my face right now – is achieved subject to a mere entitled acceptance that there is a material world at all, then why am I not likewise merely entitled to accept that there is a hand in front of my face, rather than knowing or fully justifiably believing that there is? (Reference Wright2004: 207)
The general shape of the worry is clear enough. Wright has conceded to the sceptic that certain propositions – the negations of sceptical hypotheses, and those that involve a commitment to the proper and reliable functioning of our epistemic faculties, for example – play a foundation role for one's body of belief. They are ‘cornerstones’, in the sense that were I to acquire reason to doubt their truth, I'd be rationally committed to backing off a wide class of my claims to knowledge and justified belief; I cannot, for example, rationally continue to claim to know that I had cereal for breakfast this morning on the basis of my apparent memories if I have reason to doubt the falsity of Russell's hypothesis that the world was created just five minutes ago, replete with traces of a much more extended past. While classic forms of foundationalism insisted that what lies at the foundations is especially epistemically solid, Wright's modest version of foundationism has it that the cornerstones are propositions that we merely rationally trust without evidence. How can we hope to build a stable edifice of knowledge and justified belief on foundations which seem to be less epistemically secure than what they're tasked to support?
But while it's easy enough to get a feel for the worry, at the heart of the leaching problem, so stated, there's simply a metaphor. Is there a genuine challenge behind the ‘leaching’ metaphor, and if so, what is that challenge? I'll first consider a version of the problem that appeals to a principle of closure for evidential justification. For reasons that will become clear below, this version of the objection has been discussed in the literature under the name ‘the alchemy problem’, following Davies (Reference Davies2004). However, a natural suspicion is that the alchemy problem doesn't exhaust the leaching problem, and I'll explore this possibility by focusing on a distinct version of the leaching problem due to Sebastian Moruzzi, which turns on the claim that the epistemic risk one runs in accepting a cornerstone proposition without evidence is inherited when one believes any proposition for which it's a cornerstone. Wright's response to this version of the leaching problem turns on the idea that he regards the conservative requirement that one must have antecedent warrant for accepting the relevant cornerstone propositions as a condition on claiming to possess justification for the relevant quotidian propositions, rather than on acquiring or possessing such justification. I develop a puzzle for this response, arguing that adopting it undermines the motivation for Wright's entitlement-based response to scepticism, so long as one accepts (as Wright now does) a plausible closure principle for justification. Finally, I discuss various potential responses to this puzzle, arguing that the most plausible one involves abandoning Wright's response to the risk-based version of the leaching problem. Given this result, my conclusion will be that the leaching problem, in this form, remains deeply problematic for Wright's entitlement-based response to scepticism.
2. ALCHEMY
One way of trying to make the leaching problem less metaphorical appeals to a closure principle for justified belief. Here I'll take the following as a working statement of the principle (though as is widely recognized, a more sophisticated formulation is needed to finesse certain objections):
- J-Closure:
-
If one has a justified belief that P, one recognizes that P entails Q, and infers Q on that basis, then one has a justified belief that Q.
Consider now the following simple statement of Moore's notorious ‘proof’ of the existence of the external world:
-
There is a hand.
-
Therefore, there is an external world.
Hands (as Moore argues at length against the idealist) are external objects, and so this inference is truth-preserving. Have recognized this entailment, I can now infer that there is an external world from my belief, justified on the basis of perception, that there is a hand. J-closure then entails that I have a justified belief that there is an external world.
What is the problem? Recall that having entitlement for the cornerstone (Moore's conclusion) is, according to Wright, a precondition for gaining justified belief on the basis of perception. It then looks like in performing this simple inference, I've also performed what Davies calls ‘epistemic alchemy’; I've transmuted my mere entitled trust into justified belief.
If we take this kind of epistemic alchemy to be impossible, and deny that one can reach a justified belief in Moore's conclusion via this inference, then J-closure entails that one cannot have a justified belief that one has hands on the basis of perception either. This gives us a form of leaching; if the cornerstones are merely entitled acceptances, then one cannot aspire to anything else with respect to any members of the relevant classes of quotidian propositions. This version of the leaching problem has come to be known as the alchemy problem.Footnote 2
Wright's initial response to this problem was to deny that J-closure is true. Rather, what's closed is warranted acceptance, where this is conceived disjunctively; it can take the form of either evidentially justified belief or mere entitled trust (Wright Reference Wright2004: 178). However, the costs of denying even nuanced formulations of a closure principle for knowledge have been well-documented (i.e. Hawthorne Reference Hawthorne2004; Kripke Reference Kripke2011), and these points have straightforward analogues for J-closure (McGlynn Reference McGlynn, Dodd and Zardini2014). These centre on the observation that one can formulate restrictions of such principles that are exceptionally plausible, but which (in the presence of other very plausible principles, such as the closure of knowledge or justified belief under a priori known equivalence) have all of the same problematic consequences as unrestricted closure principles like J-closure. So if one wants to give up J-closure, one must also give up seemingly ungainsayable restrictions or related principles. Given this, a better response seems to be to try to undermine our sense that epistemic alchemy is genuinely problematic (McGlynn Reference McGlynn, Dodd and Zardini2014; Wright Reference Wright, Dodd and Zardini2014).Footnote 3
Wright sometimes takes the alchemy problem to be a specific version of the leaching problem (Reference Wright2004: 178), while elsewhere he has described it as ‘a kind of dual’ of the leaching problem (Reference Wright, Dodd and Zardini2014: 229). Whichever description is preferred, the suggestion is that the alchemy problem doesn't exhaust the leaching problem. In the next section, I'll turn to the most developed and plausible way of substantiating this suspicion.
3. EPISTEMIC RISK
As Wright states the leaching problem, it takes the following form.Footnote 4 Its starting point is a plausible claim about the role played by evidence, as it relates to knowledge and justified belief. The idea is that believing on the basis of evidence is how we reduce various kinds of epistemic risk that are incompatible with knowledge or justified belief. For example, it is often claimed that having a justified belief involves having a belief that is based on evidence that makes the truth of one's belief sufficiently likely, while knowing involves believing on the basis of evidence that rules out the possibility that one's belief is mistaken in all close possible worlds in which one forms it.Footnote 5 In accepting something without evidence, then, one runs certain epistemic risks – the kind of risks that evidence would ‘forestall or attenuate’ (Wright Reference Wright, Dodd and Zardini2014: 228) – and this is so even if we hold that the acceptance is nonetheless made rational by possession of a non-evidential entitlement. So in accepting a cornerstone proposition, whether rationally or not, one runs the kind of epistemic risks that are widely thought to be incompatible with knowing or having a justified belief.
The suggestion now is that the risk involved in accepting a cornerstone proposition without evidence is inherited by acceptances of the relevant quotidian propositions. That's to say, we can give non-metaphorical content to the leaching problem by appealing to the following conditional:
- Moruzzi's Conditional:
-
If there is epistemic risk in accepting a cornerstone, then there is epistemic risk in accepting any proposition for which it's a cornerstone.
At first blush, Moruzzi's conditional might seem plausible. If accepting that the world did not spring into existence five minutes ago involves irreducible epistemic risk, then it's natural to think that believing that I had breakfast this morning on the basis of my memory will inherit the very same risk, given the dependence of the latter on the former.
On second thoughts, though, it's not clear that Wright thinks that there is any such dependence. As Wright's position was presented above, antecedent warrant for a cornerstone is required if one is to claim knowledge or justification for any of the relevant class of propositions, but in places Wright seems content to allow that one may acquire knowledge or justification on the basis of experience without such warrant. Put in Pryor's (Reference Pryor2004) terminology, Wright seems inclined to combine conservatism about claims to knowledge and justification with liberalism about our acquisition and possession of those epistemic statuses.Footnote 6 For instance, he writes:
What is striking is that the considerations which, respectively, most powerfully motivate the opposed views seem to belong at different levels. The most powerful consideration on behalf of dogmatism – liberalism about basic perceptual warrant – is that we do not wish to deny the title of warranted belief to opinions that children, and others who are relatively epistemologically innocent, form without considering, let alone marshalling evidence to discount the kind of possibility typified by tricksy lighting or the artful disguise of mules. So it seems we think that the acquisition of basic perceptual knowledge, or other forms of perceptual warrant, doesn't require the kind of epistemological ‘policing’ of authenticity-conditions [epistemic presuppositions, including cornerstone propositions] that conservatism demands. But it is different when we adopt a stance in which we undertake to scrutinize our claims to perceptual knowledge, or warrant. Then we seem obliged either to take a positive view of any authenticity-condition that may be entered into the conversation, or to qualify our claim to warrant. (Reference Wright, Dodd and Zardini2014: 218–19; see also Reference Wright and Coliva2012: 471).
If Wright accepts liberalism at the lower-level – the level of possession of warrants rather than of claims to possess warrants – the version of the leaching problem we're considering lapses. On this liberal picture, one's visual evidence supports believing that there is a hand regardless of whether one has antecedent warrant for accepting that there is an external world. So the observation that possession of an entitlement leaves the cornerstone an epistemically risky proposition doesn't give rise to a leaching worry, since the evidential significance of one's perceptual experiences isn't hostage to the good epistemic standing of the cornerstone. Claiming that one's experiences have evidential significance for how things stand in an external world is a different matter; by Wright's lights, this claim is hostage to the good epistemic standing of the proposition that there is indeed an external world that is more or less open to our perceptual faculties. So we should expect there to be leaching, but only at the second-order. This is in fact Wright's own position (though I don't intend the foregoing as a reconstruction of his own route to it). Given that one only has an entitlement for the cornerstone, one's claims to have perceptual knowledge or justified belief that there is a hand will also be underwritten by an entitlement, rather than by evidential warrant (Wright Reference Wright2004: 207–9; Reference Wright, Dodd and Zardini2014: 229). So there is leaching, but ‘it is at one remove and can be lived with’ (Reference Wright2004: 207).
The key point here is that this response to the leaching problem depends crucially on adopting liberalism about first-order warrants. If the possession of justified belief in and knowledge of quotidian proposition (and not just the rationality of claiming these warrants) were hostage to the good epistemic standing of the relevant cornerstone, then the leaching that Wright is willing to acknowledge at the higher level would happen at the lower level too, and so wouldn't be contained ‘at one remove’ where Wright thinks it can be lived with. For if conservatism about first-order perceptual warrants is right, then the significance of our perceptual experiences (or perhaps certain propositions such experiences suggest) as evidence for claims about the external world is dependent on the epistemic standing of the cornerstones. And while one can perhaps see how the rationality of claiming that perceptual experiences have an evidential bearing on propositions about the external world might be contingent on the possession or non-possession of non-evidential warrants of a sort one could have even as a brain-in-a-vat, I find it very hard to see how the evidential bearing of experience itself – whether we think of this in terms of probability raising, safety from error, or in some other way – could be so contingent. Rather the claim should be, I suggest, that the evidential bearing of experience on quotidian propositions about the external world is contingent on whether one has warrant for the cornerstones that renders them epistemically non-risky, in the relevant sense. That is, I submit, the kind of reasoning that makes Moruzzi's conditional plausible.
As just noted, the line of reasoning rehearsed in the previous paragraph doesn't seem to create any issues for conservatism at the second-order. So adopting dogmatism at the first-order but retaining conservative conditions at the second-order does seem to be a way of containing leaching at the higher level, as Wright wants. Unfortunately, it is very difficult to see how to stabilize the resulting picture of first-order and second-order warrants, at least so long as one accepts J-closure. According to Wright's conservatism, one cannot rationally claim to be justified in believing that there is a hand unless one is antecedently warranted in accepting that there is an external world. The sceptic appeals to this conservative thesis in order to argue that this antecedent warrant cannot be grounded in evidence, even broadly construed, and entitlement is brought into the picture in order to forestall the further, properly sceptical conclusion that none of our claims to knowledge or justification can be warranted.
However, if we combine a Pryor-style liberalism about the conditions under which one can acquire a justified belief in quotidian propositions about the external world with J-closure, we get a form of neo-Mooreanism. Recall our simple statement of Moore's Proof from above:
-
There is a hand.
-
Therefore, there is an external world.
On the view we're considering, one is able to form a justified belief in Moore's premise on the basis of perception regardless of whether one has antecedent warrant for Moore's conclusion; that's the liberal thesis which Wright seems to be adopting in the passage recently quoted. J-closure then entails that competently carrying out Moore's inference results in one having a justified belief that there is an external world. The problem then is that this neo-Moorean result seems to carry more anti-sceptical punch that Wright can allow, since it seems to deliver precisely the antecedent warrant demanded by Wright's conservatism at the second level; the requisite warrant for the cornerstone can be delivered by inference from the first-order warrant afforded to one immediately by perception, and so there's no need for an entitlement in order to rationally claim justification belief or knowledge that there is a hand on the basis of perception.Footnote 7 Entitlement is required, according to Wright, because he thinks that we have to be rather more concessive to the sceptic than is typically acknowledged; the problem we're running up against here is that these concessions don't seem to be compatible with liberalism about first-order warrants of the sort Wright seems to adopt in the passage quoted above and when facing the leaching problem.
So it's not clear that the combination of liberalism at the level of first-order warrants and conservatism at the level of second-order warrants is consistent with Wright's entitlement-based response to scepticism, given J-closure. One natural way out of this predicament for Wright is to double-down on his conservatism, endorsing it at the level of first-order warrants too.Footnote 8 Then having knowledge or a justified belief that there is a hand (and not merely claiming these kinds of warrant) requires that one has antecedent warrant for the relevant cornerstone propositions, and the argument for the redundancy of entitlement is blocked. A worry with this way out is the one raised in the passage quoted from Wright above, namely that it may seem to commit us to saying that young children cannot have perceptually justified beliefs if they don't, or are unable to, entertain the relevant cornerstone propositions; as we'll see below, Wright doesn't in the end think that this is a genuine problem for conservatism about first-order warrants. However, the real cost of avoiding the problem just raised in this fashion will be evident. Above I argued that Wright's response to the version of the leaching problem in terms of epistemic risk crucially depends on concessions to Pryor's liberal concerning first-level warrants. It's difficult to see how Wright could abandon those concessions, imposing conservative conditions on the possession of justification for the relevant quotidian propositions, without reinstating this version of the leaching problem as a serious threat.
4. OPTIONS
My conclusion will be that this threat is indeed reinstated. But before we get there, there seem to be at least two live options that the proponent of the entitlement-based response to scepticism might explore before conceding such a pessimistic conclusion. First, one might revert to Wright's original reaction to the alchemy problem, and go in for a thorough-going rejection of the epistemic closure principle implicated in the argument. Second, one might try to motivate a more stringent variety of conservatism about second-order warrants (while continuing to endorse liberalism about first-order warrants, thereby avoiding the leaching problem). According to this variety of conservatism, claiming justification or knowledge concerning quotidian propositions about the external world requires that one have antecedent warrant to claim warrant for the relevant cornerstones: possession of warrant for the cornerstone isn't enough. If neither of these strategies work we would be forced to try to formulate an alternative response to the epistemic risk version of the leaching problem that doesn't require liberalism about first-order warrants. But let's consider the two less pessimistic options in turn before drawing any premature conclusions.
I won't have much to say about the first option here, since I have already outlined the costs of rejecting J-closure above. Moreover, I also noted that Wright himself now seems inclined to accept the force of these considerations (Reference Wright, Dodd and Zardini2014: 232–5). I don't mean to claim that a closure-denying response to the puzzle I have developed for his response to the leaching problem isn't worth considering any further. After all, denying J-closure offers Wright a unified way out of both the alchemy problem and the puzzle I have developed for his response to the leaching problem, and that's not something to be dismissed without due cause. But neither should we underestimate the difficulties involved in taking this route.Footnote 9
The second option seems more promising, on the face of it at least. We keep a liberal picture of the conditions under which one can acquire justified beliefs about the external world, and we hold fast to J-closure. As we've seen, this gives us the neo-Moorean result that one can acquire a justified belief in an external world via a rehearsal of Moore's proof. However, this modification of Wright's view tries to cap the anti-sceptical force of this neo-Moorean line here, preventing it from having any potency against scepticism directed at one's claims to justification (and knowledge). Since Wright thinks that the real sceptical threat concerns such higher-order warrants, this line of response would still leave a significant role for his appeal to entitlements to play.
In more detail, the proposal is that the neo-Moorean line of reasoning can be robbed of its potency against second-order scepticism by strengthening the conservative demands on acquiring second-order warrant for quotidian propositions; for example, in order to be warranted in claiming that one has perceptual justification for believing that there is a hand, one has to be antecedently warranted in claiming warrant to accept that there is an external world. This second-order warrant doesn't seem to be delivered by the kind of neo-Mooreanism we've been considering, which at best only seems to ensure that in favourable conditions we have first-order justification to believe that there's an external world. So there's a shortfall, if our project is primarily to respond to the second-order sceptic – the sceptic who argues that I'm not in a position to rationally claim to know or be justified in believing that I have hands. A satisfactory response to such a sceptic would, given the demanding variety of conservatism presently under consideration, require showing how we could have second-order warrant for the cornerstones prior to claiming warrant for any of the relevant quotidian propositions. Liberal neo-Mooreanism can at best deliver antecedent first-order warrant; that's the shortfall. However, it's open to Wright, structurally at least, to close the gap by suggesting that we each have entitlements to claim warrant for the cornerstones: that the requisite second-order warrants are entitlements. So it seems that there could still, in principle, be a vital role for entitlements to play in combatting scepticism, in the form which Wright takes to be most threatening and important.
On this modification of Wright's position, his reply to the epistemic risk version of the leaching problem can stand as it did before; there is leaching, but it's contained at the second-order, and so there's no obstacle to our possession of statuses like knowledge and justified belief. However, it's not clear that this more demanding variety of conservatism about second-order warrants can be independently well motivated, or whether the kinds of strategies that Wright appeals to in order to establish that we possess entitlements engage cleanly if it is the availability of such second-order warrants that needs to be established. But I won't pursue these worries, since the principal problem with the proposal is that – contrary to the thrust of the previous two paragraphs – it's still rather unclear whether any room is left for a meaningful appeal to entitlements in addressing the sceptical challenge.
To see this suppose, as Wright suggests in the passage quoted above, that there is a strong philosophical case to be made for liberalism at the level of first-order warrants,Footnote 10 and suppose (as Wright also seems to now accept) that there's a strong case for J-closure. Then the considerations that figure in these cases seem to provide warrant for claiming that we can enjoy justified belief that there is an external world (given that we enjoy experiences as of external objects such as hands). That's to say, one's second-order warrant for the cornerstone – warrant to claim to have a justified belief that there is an external world – doesn't need to be an entitlement, since it can instead be an evidential warrant resulting (in part) from an appreciation of the considerations for thinking that liberalism and J-closure are correct. The upshot is that strengthening Wright's conservatism about the conditions under which one can rationally claim justification for one's belief that there is a hand may not in fact help with the problem it was designed to solve; there may still be no need for entitlements for the relevant cornerstone propositions.
I have been arguing that it's very difficult to reconcile liberalism about first-order warrants with conservatism about second-order warrants, if one accepts J-closure and is committed to finding a significant role for epistemic entitlement to play in mounting a response to the sceptical challenge. One might wonder why Wright hasn't recognized and addressed these tensions, if they're really as problematic as I've been suggesting. One thing one might say in response is that it's only recently that Wright has reaffirmed his commitment to principles like J-closure, and as I've stressed throughout, J-closure is a crucial ingredient of all of the arguments we have considered.Footnote 11 But a much more important point is that, contrary to the impression given in the passage from Wright quoted above, he in fact doesn't take consideration of young children and other epistemically and conceptually unsophisticated knowers and believers to motivate liberalism about first-order warrants. Rather, Wright takes it to push us in the direction of what he calls a ‘welfare state epistemology’, in which a subject can have an entitlement to claim knowledge of quotidian propositions in virtue of epistemic work done by others (e.g. Wright Reference Wright2004: 204–5, Reference Wright, Nuccetelli and Shea2008: 28 fn 6, Reference Wright, Dodd and Zardini2014). That yields a picture on which the conservative demands on the acquisition of justified beliefs in quotidian propositions about the external world are, for most individuals at least, extremely thin and easy to meet, and so it can appear that those demands aren't operative at all (Reference Wright, Dodd and Zardini2014: 221–2). The liberalism that Wright takes considerations about unsophisticated epistemic subjects to motivate, then, is entirely superficial; it's really a form of conservatism.
However, as I argued earlier, Wright's denial of Moruzzi's conditional, and so his response to that version of the leaching problem, rests on the adoption of a liberal position about first-order warrants. The faux liberalism described in the previous paragraph isn't to the point here, since it's still a view on which the epistemic credentials of my beliefs in quotidian propositions are dependent on the epistemic standing of the relevant cornerstones; that's precisely the thought that lends Moruzzi's conditional plausibility, as I argued above. So if Wright's final position is really conservative at both first and second level, as I have just suggested, it's left entirely unclear what allows Wright to claim that leaching only occurs at the second level (where it can supposedly be lived with). The challenge Wright is left with is that of finding an alternative response to the leaching problem to the one he has offered to date, a response that doesn't require him to accept liberalism about first-order warrants. As things stand, we don't have any alternative response to the epistemic risk version of the leaching problem on the table, and until we do, the leaching problem stands unanswered.
Let me close this section by returning to the issue of how the leaching problem, so understood, relates to the alchemy problem. At the end of the section on epistemic alchemy, we considered the suggestion that there might be a version of the leaching problem that in some sense outstrips the alchemy problem. Given this, we should ask: to what extent does the version of the problem in terms of epistemic risk furnish a genuinely distinct objection to the alchemy problem, and why might one find this problem more worrying than the alchemy problem?Footnote 12
There are, it must be acknowledged, similarities between the alchemy problem and the version of the leaching problem currently under consideration. Both involve an appeal to J-closure in order to argue that once we accept that a subject has evidential justification for quotidian propositions such as the premise of Moore's proof, we'll be committed to holding that she has justification for any recognizably entailed cornerstones.
However, the differences run deeper than the similarities. There are three principal – and related – points to make. First, while the appeal to J-closure is absolutely central to the alchemy problem, it only figures on one horn of a dilemma in the context of the leaching problem; it figures in the reasoning that shows why Wright can't escape the leaching problem by adopting a Pryor-style liberalism about first-order warrants. The other horn of the dilemma is driven, not by J-closure, but by Moruzzi's conditional, which concerns the relationship between cornerstones and their associated quotidian propositions, not (as with J-closure) between the premises and conclusions of recognizably valid arguments.Footnote 13
Second, the alchemy problem is most naturally construed as resting on an assumption that Wright accepts conservatism about first-order warrants. In a nutshell, the alchemy problem is that one's entitlement for a cornerstone (such as Moore's conclusion) enables one to acquire evidential justification for an entailing quotidian proposition (such as Moore's premise), and in the presence of a principle like J-closure this commits us to holding that one can thereby acquire justification for the cornerstone too; this way of setting out the problem clearly assumes that Wright's conservatism applies not only to second-order warrants, but to first-order warrants too.Footnote 14 In contrast, as I have just noted, J-closure figures in the setup of the leaching problem only when we consider the possibility that Wright could escape it by adopting liberalism about first-order warrants.
This leads to the third and most crucial difference. The upshot of the alchemy problem is that, unless one finds some way to block the argument, one's entitlement for a cornerstone can be transmuted into evidential justification merely by looking at one's hands and engaging in some simple deduction. This upshot may (or may not) be problematic for Wright, but it doesn't undermine the initial motivation for Wright's proposal that we enjoy entitlements for such cornerstones. In contrast, it's precisely the motivation for this central aspect of Wright's epistemology that's targeted by the appeal to J-closure in the context of the leaching problem, as I've developed it here; the claim, recall, is that liberalism about first-order warrants together with J-closure yields a kind of neo-Mooreanism that renders Wright's entitlement theory an idle wheel. This is the sense in which I want to say that the version of the leaching problem I have been pushing in the past two sections seems potentially even more problematic for Wright than the alchemy problem. Perhaps Wright can come to accept the upshot of the alchemy problem, namely that his view permits the alchemical conversion of entitlements for cornerstones into evidential justifications for them, and can persuade us that epistemic alchemy is not as mysterious or as problematic as we might initially think – indeed, this seems to be his current approach to the problem (Wright Reference Wright, Dodd and Zardini2014: 233–5). But he simply cannot accept the upshot of the closure-based reasoning involved in the present leaching problem, since that would involve conceding that one of the central and most distinctive aspects of his epistemology, namely that the most satisfactory response to scepticism available to us involves accepting that we have entitlements for cornerstone propositions, is without motivation. So the leaching problem, properly understood, raises a distinct and potentially more troubling objection to Wright's epistemology.
5. CONCLUSION
In this paper, I have argued that the leaching problem remains a genuine threat to Wright's entitlement-based response to scepticism. The argument has been rather complicated, and so in these concluding remarks I will go over the main points again.
The most potent version of the leaching problem turns on the conditional claim that if one runs an epistemic risk in accepting a cornerstone, then one runs an epistemic risk in accepting any proposition for which it is a cornerstone. Wright's response to this version of the problem turns on the claim that his conservatism is only a thesis about the conditions under which one can claim evidential warrant for quotidian propositions; that's consistent with a liberal picture of the conditions under which one can acquire and possess these lower level evidential warrants. On this stratified picture, leaching will be restricted to the higher level, since one's possession of justification and knowledge at the lower level is in no way dependent on the epistemic security of the relevant cornerstones; so there's no threat to our possession of evidential warrant for quotidian propositions.
The main purpose of this paper has been to argue that this response is problematic. The reason it's problematic is that this lower-level liberalism, in combination with a plausible closure principle for justified belief, gives rise to a kind of neo-Mooreanism, and this neo-Mooreanism usurps the position of an appeal to entitlements as an antidote to sceptical. I have reviewed the options for escaping from this puzzle, including giving up closure for justified belief and endorsing a more demanding form of conservatism for second-order warrants, and I have argued that these are not very promising. I have also aimed to clarify Wright's own position, which on closer inspection seems to be more thoroughly conservative than his response to the leaching problem suggests or allows. What's needed, I suggest, is an alternative approach to the leaching problem. And so long as a convincing alternative approach eludes us, the leaching problem will continue to cast doubt on Wright's proposal that non-evidenced acceptances are fit to play an epistemically foundational role.Footnote 15
