Reasons: Right and Wrong

Reasons: Right and Wrong Sean Aas Draft of 10/16/14 Comments welcome at [email protected] Please do t uote o ite ithout pe issio . I. The Wrong-Kinds-of-Reasons Problem A demon threatens catastrophic destructio if ou do t alue his toe ail lippi gs. “o ou ha e e good easo to do so. Yet that does t espe t ou ake the alua le. You a t get ahead at o k u less ou oss he s a asshole, ut he a tell if ou e faki g it . That does t ake hi respe ta le. You f ie d s dep essed and could really use a win. Even if letting him win is the right thing to do, that does t ake it a show of good sportsmanship. You e got by believing in God—but that does t u h to gai a d othi g to lose ake that belief credible. In each of these cases, a normative concept which seems to be defined in terms of a certain kind of activity is t definable, simply, as doing that thing or having that attitude as you have most reason to do it. Some reasons which seem to bear on valuing/ espe ti g/spo ti g/ elie i g, do t see to ea o whether some instances of these things in question are respectable/valuable/sportsmanlike/credible. These easo s a e, as it s so eti es said, of the wrong kind to do so. This notion is important to our understanding of the role of reasons in the normative domain in at least two ways. Some, notably Pamela Hieronymi, think that distinguishing right-kind reasons from wrong-kind reasons requires, ultimately, revisiting our notion of a reason itself; wrong-kind reasons 1 clearly count in favor of the things they are reasons for, but yet, so the thought goes, are not reasons for them; thus being a reason for is not a matter of counting in favor.1 And some think that the existence of wrong-kind reasons constitutes a decisive objection to philosophical accounts that try to understand all normativity in terms of the idea of a normative reason. Some clearly normative concepts, like the concept of value, involve an unanalyzable notion fittingness; and a corresponding metaphysics of normative facts for the relevant actions and attitudes to fit.2 These views are bolstered by the inability of any existing account to distinguish wrong-kind reasons from right-kind reasons without, either, substantial modification of our account of the reason-providing relation; or, substantial appeal to normative notions not analyzable in terms of reason-provision. My goal here is to give a general account of the wrong-kind/right-kind distinction that is prima facie consistent with the view that all normative concepts can be analyzed using only non-normative concepts and the concept of a normative reason, where that is understood as a consideration counting in favor of an action or attitude. The first step is to properly understand the scope of the problem. In the next section, I argue that the WKR (for: wrong-kind-of-reason) problem arises not just for attitudes, but also more broadly for any activity where it makes sense to talk about there being a right way and a wrong way to engage in the activity—any practice, as I will say.3 In all such cases there is a question as to whether considerations which count in favor of doing something within a practice count in favor of the conclusion that doing that thing would be the right way to engage in that practice. But there is also the converse question: do 1 Cf. he e The W o g Ki d of ‘easo Journal of Philosophy 102 (2005): 437-457; The Use of Reasons in Thought (and the Use of Earmarks in Arguments), Ethics, no. 1 (2013): 114–27. 2 Wlodek Rabinowicz and Toni Rønnow-‘as usse , The Strike of the Demon: On Fitting Pro-Attitudes and Value*, Ethics 114 (2004): 391–423. 3 This idea is a ge e alizatio of the idea of ‘a ls s idea of a p a ti e, e pli ated at Joh ‘a ls, T o Co epts of ‘ules, The Philosophical Review 64, no. 1 (1955): 3–32. 2 considerations which bear on what counts as the right way to engage in a practice always bear on whether one ought to perform the moves of that practice? In the following section I argue that they do ot; a d he e that ei g a easo of the ight ki d to ϕ does not always entail being a reason, simpliciter, to ϕ. In the concluding section I sketch an account which makes sense of this feature of the wrong kinds of reasons problem: on which, very roughly, reasons are of the right kind with respect to a practice if they are appropriately responsive to the characteristic goals and normal methods of that practice. This, I argue, is at least prima facie consistent with giving reasons-as-considerations-in-favor-of a fundamental role in our normative conceptual framework; thus there is no direct argument from o g ki d of easo phe o e o to the eed fo a fu da e tal ethi ki g of ou o ept of a normative reason or of its possibly unique fundamentality to our normative conceptual framework. II. The Scope of the Problem The wrong kinds of reasons problem arose, originally, in the debate over so-called buck-passing or fitting-attitudes theories of value.4 At a first pass, these theories say that to think that something is valuable is to think that one has most reason to value it; they thus attempt to analyze one central normative concept, the concept of value, in terms of the concept of a reason – thereby supporting the general hypothesis that the fundamental normative notion is the notion of a normative reason. But so e easo s hi h ea o alui g do t see to ea o alue—demon-given reasons, say. As this problem arose, theorists immediately noticed a similarity to a problem, familiar in epistemology, concerning the relevance of broadly practical reasons to the justification of beliefs.5 Even if the fact that it ould ake e happ to elie e a pleasa t falsehood is a easo to do so, it is t, to any extent, a 4 Or anyway arose under that name; as was soon recognized, and as I discuss below, similar issues had long since been raised concerning pragmatic reasons for belief. 5 Rabinowicz and Rønnow-Rasmussen, op cit. 3 o side atio that episte i all justifies that elief, o easo s fo elief, these theo ists thought, is that the akes it edi le. What s e ot suita l the belief in question: that it ould e i e to elie e that p does t o g ith p a ti al elated to the o je t/ o te t of ake it a o e likel that p is true. And perhaps we could say the same thing about demon-given reasons for valuing: my reason to alue the de o s toe ail lippi gs, afte all, does t see to ha e a thi g much to do with the clippings themselves. As a defense of the fitting-attitudes account of value, this move has its problems.6 But following a suggestio of Ma k “ h ode s, e should suspe t i a ase that it is ot suita le as a odel fo a general distinction between doing things the right way and doing them as we have most reason to do them.7 Clearly, it applies only to intentional attitudes; and that distinction applies to many other things besides. There is a right way and a wrong way to value, to be sure— ut the e s also a ight a a d a wrong way to tie a four-in-hand, or to play a game of basketball. And in all three cases, doing the thing i uestio as e ha e ost easo to do it is t the sa e as doi g it the ight a : i ea h ase so e reasons are of the wrong kind, while some are of the right. If something like the state-given/object-given disti tio is ight fo the ase of alui g, o elie i g, the , the e s easo to thi k it s a spe ial ase of a more general account which covers non-intentional attitudes and activities as well. Schroeder suggests that the relevant general class is the class of things that can be done correctly or incorrectly. This is not implausible. We can believe correctly or incorrectly; plausibly, we can value correctly or incorrectly too. And reasons do seem to be of the right kind to ground a judgment of 6 Notably, that it seems likely to be circular: in the case of belief, the RK-reasons need to be related to the truth of the proposition believed. But e d etter not say, analogously, that RK-reasons for valuing need to be related to the value of the valued object—that would be circular in the context of the buck-passing account. Cf. here Rabinowicz and Rønnow-Rasmussen, ibid. 7 Value and the Right Kind of Reaso s Oxford Studies in Metaethics 5: 25-55, September 2010. Schroeder seems to have changed his view on these matters since then, for reasons I examine and reject below (n10). “ h oede as used he e efe s to “ h oede -ca.-2010. 4 credibility/value just when they go to the correctness of believing or valuing. Similarly, it seems like we can tie a knot or set a table correctly or incorrectly; and that this fact enforces a distinction between reasons which bear on how to engage in that activity on any particular occasion, and reasons which bear o ho to get it ight. “o “ h oede s a ou t of the s ope of the WK‘ p o le is e tai l oader, and correspondingly more plausible, than some extant alternatives. But it s ot oad e ough. To see h , o side so e p edi a e ts ou pla i g poke . You e et o e tha ou a affo d to lose o a ga e—so ight fi d ou self i hile u h that ou re family will be destitute if you lose. An opportunity arises to cheat and win—you get a minute where you think you a deal ou self a i i g a df o the otto of the de k. You e p ett su e ou a get a a it. A d the pla e s ou e pla i g ith, let s sa , a e ho i le o ste s, ho ill use thei i ith i gs i terrible ways if they win.) So you have overwhelming reason to deal from the bottom of the deck. But that does t ake it the o e t a to deal. So far, so good, for Schroeder. But compare the following case. Same card game, same stakes, same o ste s. But ou e a a ged fo a f ie d, ho s at hi g f o a oss the oo , to o the ga e a d split the winnings with you, when he observes you fold your hand in a certain way, at a certain time. As it happens, at the appointed time, (without cheating) you draw a fantastic hand—one that would po a l i , if ou did t fold it. You ha e o e would, after all—also ou ll get the ould t e heati g—it ould, fo o e i the hel i g easo to fold ou p o ised ou f ie d ou o ste s allets, too . Foldi g i this i u sta e all speaki g, ot e a i o e t o e i the ga e of poke . But it s ot the right way to play: to play poker right, you have to play to win. It seems, then, that there are two sorts of normative concepts that engender wrong-kinds of reasons problems: concepts of correct conduct in an activity narrowly construed, and something more 5 general, like the idea of doing an activity right. The poker case shows this, but we should have suspected it anyway: concepts like the concept sportsmanship and the idea of fidelity to promises seem to ge e ate WK‘ p o le s, e e though the do t see to ha e a thi g o iousl to do ith correctness. It seems clear, then, that the scope of the problem is broader still than the set of concepts that serve to judge correct or incorrect engagement in an activity: broad enough to include concepts which encode the looser norms that go with the idea of there being a right way and a wrong way to do something. Pa ela Hie o o i has e e tl a gued, agai st ie s like “ h oede s a d i e, that the o g kinds of reasons problem as such only arises for (only some) attitudes. The thought is that we only have a loose intuitive grip on the problem to start with – there seems to be some general problem involved in distinguishing pragmatic from evidential reason to believe, value-relevant from value-irrelevant reasons to value, and so forth. Schroder and I offer reasons to add things to this list; but all we have to go on in doing so, really, is felt similarity. Hieronymi maintain that if we have a clear way of drawing the disti tio i so e of these ases, that does t o e othe s, that ould e a fault i ou i itial a of posing the problem rather than a flaw in our solution. She proposes, in particular, a solution to the problem that seems quite convincing for the wrongkinds-of-reasons problem as it arises for certain attitudes; but which (like the object-given/state-given distinction, of which it is perhaps a descendant) does not apply at all to activities, like dealing a poker hand or playing basketball. The intuitive thought is that forming certain attitudes is in some sense equivalent to answering certain corresponding questions. So, to take the clearest case, to come to believe that p is to affirmatively answer the question: is p true? A right-kind reason for believing, or any other attitude, is one that bears on the relevant corresponding question. A wrong-kind reason bears 6 instead only on some other, related question – that believing in God is a good bet bears on whether is it good for me to believe in god, not on whether there is a god. This is a very plausible account of the wrong-kind/right-kind distinction as that applies to belief, at least. It offers an elegant, informative account of what distinguishes right-kind from wrong-kind reasons for belief, in terms of a plausible and illuminating account of what believing is. It is somewhat less clear, however, that it works for valuing, or for other related attitudes. There does seem to be some intimate elatio et ee o i g to elie e so ethi g a d o i g to affi it s t uth; so e e ui ale e between coming to believe that p a d o i g to a s e the uestio : is p t ue , ith a es . But it is less clear what questions will correspond analogously to attitudes like valuing, or admiring. One might be tempted to say: to value something is to determine that, yes, it is valuable; to admire someone is to determine that, yes, they are admirable. But this temptation should, probably, be resisted, as it links these attitudes too closely to the corresponding judgments. I can judge that opera is valuable without self alui g it. I a judge that i al s o k is ad i a le, ithout the e a tually admiring it. The e is othi g st a ge a out e p essi g these attitudes, eithe : Ope a is e self do t alue it ; He o k is ad i a l ead it ; eithe so st a ge as the lassi alua le i deed; ut I igo ous; ho e e , I feel othi g ut ese t e t he I It is t ue that … the at is o the at, ut I do t elie e it . Perhaps, however, the account could be extended to cover attitudes apart from belief – perhaps, e.g., as Hieronomyi suggest, these correspond to some more complex set of questions, affirmative answers to which really do constitute valuing, admiring, etc. In any case the account clearly does not extend to distinguishing wrong-kind from right-kind reasons for engaging in activities, as well. One does not deal a hand of poker, or play a game of basketball, or set a table, by answering a question. So the plausi ilit of Hie o i s solutio to the o g-kinds-of-reasons problem, where it is plausible, 7 o stitutes a halle ge to “ h oede s so e hat a d et o e e pa si e eading of the scope of that problem. Here I want, simply, to accept this challenge. My response will be to offer a solution to the problem that is oad e ough to at h ie of its s ope; a solutio hi h, I hope, e plai s h Hie o s solution is for persuasive when it is persuasive, while doing at least as good a job in the cases which are harder for her to account for. III. One Proposed Solution ‘etu i g to “ h oede : his u de sta di g of the p o le is oade tha Hie o i s; still, it is probably too narrow to cover all the cases that seem to raise an intuitive problem of distinguishing doi g so ethi g the ight a f o doi g it as e ha e ost easo to do it. This is not in itself a decisive objection. Perhaps, per Hieronymi, the appearance that all these cases raise a similar problem needs to be, not explained, but explained away, by a solution to the problem that is good enough in some cases to convince us that something else must be going on in the others. What matters in the end, then, is what Schroeder has to say in the way of solving the problem as he understands it. Schroeder thinks that WKR problems arise in basically two ways. They arise, first, in cases where so eo e has a easo hi h does t see to appl to e e od ho e gages i an activity—robbery schemes for poker players, or demon threats for valuers. Second, they arise in cases where, though e e o e ho e gages i a p a ti e has the ele a t easo s, that s o l e ause e e od pe se has the relevant reasons—say, when there are categorical moral reasons not to laugh at a joke. The ensuing 8 thought is that these easo s a e of the o g ki d ith espe t to a a ti it e ause the do t ha e enough to do with the nature of that activity itself. This is a good thought. But I less su e a out “ h oede s o e spe ifi i te p etatio of it: RKR: The right kind of reasons with respect to an activity A, are all and only those reasons which are shared by necessarily every able person engaging in A, because they are engaged in A, together with all reasons which are derivative from such reasons. 39 . The idea here is that reasons are of the right kind with respect to an activity iff they are (a) general with respect to that activity, in the sense that anyone who engages in that activity has them; but also (b) specific to engagement in that activity, in the sense that such engagement explains why people have them when they have them. It seems to me that there are problems with both conditions: The idea of the second, specificity, condition is to rule out reasons which are, as it were, too general to be reasons of the right kind with respect to an activity—paradigmatically, moral reasons which bear on conduct within an activity but not on whether that conduct is correct with respect to the standards of that activity. In those cases the reasons in question apply for whatever reason moral reason apply more ge e all ; he e, the do t appl to o du t ithi a a ti it just e ause o e is e gaged i that activity. There is, I think, reaso to e o e ed that this ules out too u h. He e s o e easo : a t moral reasons sometimes be reasons of the right kind? So, e.g., perhaps moral reasons are, at least sometimes, the wrong kind of reasons to determine whether a joke is funny. But are t the so eti es the ight ki d of easo s? It see s to e that the so eti es a e: so e jokes, a e t fu e ause they are so morally objectionable. Perhaps these problems with the specificity condition can be finessed. The problems facing the generality condition are much more serious. The purpose of this condition is to rule out reasons which are too specific or idiosyncratic to be reasons of the right kind with respect to a given activity. The idea is that, of all the reasons that might apply to the conduct of an activity, only those that apply to that 9 o du t ge e all a e of the ight ki d . This akes ei g of the ight ki d est i ti e: if R is a reason of the right kind to ϕ, it s a easo to ϕ. But it does t look like this is ight, as the following cases show:8 “uppose I sitti g ith o ste s agai , pla i g a ga e of poke . I k o the ll i , a d the ll use the it s fault that o e to do so ethi g ho i le to a f ie d of o that if I do t heat, i e. Add, if ou like, that friend is in this predicament. It seems pretty plausible here to say that whatever reasons I might have to play the game right are not just swamped or outweighed, but cancelled or defeated, the po e of o al easo s to help f ie d. But that does t make cheating the right a to pla a ga e of poke : it s just that the easo s that o t i ute to that e di t do t go to hethe I should play that way in this case or not. Or consider a more general class of cases, where terrible people have some unusually impressive qualities. These qualities, in these people, might sometimes be admirable. But make these people terrible enough, and it starts to look implausible to think that we have *any* reason to admire them—is the e a hethe easo to ad i e “tali s de isi e ess? I do t thi k so: he e should ad i e “tali a eas thi g to de ide e e t i g to figu e out e do t ha e to eigh the easo s to admire his decisiveness against the reason to deplore his murderousness—given the latter, the former just have no weight at all. Or: try another case, adapted from Schroeder himself. Suppose that I work at the White House, and, at the last i ute, fi d self alled o to set the ta le fo the P eside t. I dis o e that the se e I replacing was part of an assassination plot, and that the signal to call the whole thing off is to flout eti uette a d put the salad fo ks o the i side. “uppose also that I a t safel that the ll put the pla i otio ight a a . Plausi l , I do t ha e a 8 a a o e, o fea easo at all to put the salad A tuall the e e a le pe so ualifi atio already shows this; what follows shows it in a very different, and I think much less Schroeder-congenial, way. 10 fo k o the outside if ou dou t this, odif the ase so I espo si le so eho fo the hole situatio . But that s still the o e t a to set the ta le. I sofa as the latte otio is to e understood, as Schroeder would have us understand it, in terms of the force of the right kinds of easo s i fa o of doi g so, it follo that I e e though I do t ha e a ust ha e easo s of the ight ki d to set the ta le ight easo s si pli ite to do so. Ad ittedl , o e of these ases is a utte l de astati g ou te e a ple to “ h oede s RKR. “ h oede ould su el espo d that the el o u elia le egati e e iste tial i tuitio s a out reasons. In all these cases, he could say, there is some reason to do thi gs the ight o o e t a . It is just too weak to be worth mentioning, in light of the reasons on the other side.9 In response, note that the claim that there is always some easo to do so ethi g the ight o o e t a egi s to look less plausible when we focus our attention on activities and practices that are not just morally neutral, like table-setting or card-playing, but morally objectionable. There is, appa e tl , a ight a a d a su je t s o g a to ate oa d so eo e: to do it ight you have to force the outh ope , so that the ate pou s i to his th oat as ell as his ose. Does a p ospe ti e waterboarder always have some easo to ope his su je t s outh, efo e pou i g the ate o e ? M intuition – yours too, I hope – is that he does not always have some reason to do this. We could be o g, he e. But e ould eed so e a tual, positi e, easo to elie e that e a e o g. A d it s ot at all clear that we have it. Schroeder could say that the other strengths of RKR give us such a reason, in the absence of a compelling alternative. But of course this response depends on the absence of compelling alternatives. I will argue, below, that there is one – one that has all the strengths of “ h oede s a ou t a d othe s besides. 9 Weighti g fo a Plausi le Hu ea Theo of ‘easo s, Noûs 41, no. 1 (2007): 123–124. 11 IV. A Different Approach A asi i sight of “ h oede s a ou t is the thought that o side atio s a e easo s of the ight kind with respect to an activity when they not idiosyncratic with respect to that activity: when they are, in some sense, normal for that activity. RKR understands normality in terms of generality: reasons are normal with respect to an activity because they apply generally, to anyone who engages in that activity. But if, as I ha e suggested, ight ki d easo s a easo s, the ge e alit does t look like a e e ot only outweighed but cancelled by other plausi le i te p etatio of o alit : so e easo s which apply to the normal conduct of an activity might well fail to apply to it in every case. Still, I think the idea of understanding the right kind/wrong kind distinction in terms of normality/abnormality is on the right track. We just need a better understanding of what it is for reasons to be relevantly normal. To find one, think back to one of the clearer wrong-kind-of-reasons cases—the ase of p ag ati to elie e that I easo s fo elief. Wh does t the fa t that it d ake e feel eall good supe -good looking make this belief credible? Proponents of the state-based/object- given approach discussed above will sa that it s e ause this easo does t ha e a thi g to do ith the o te t of the elief i uestio . But of ou se it does ha e so ethi g to do ith it: afte all, it s being good-looki g that I d e happ a out. The p o le , i stead, has to e the o e that Hie o a ou t di e ts us to: that it does t ha e e ough to do ith the truth of the elief that I is eall good- looking. Belief aims at the truth: representing the truth is what beliefs are for, in some important sense. That plausibly, is wh hat e e goi g fo i elie i g so ethi g is a s e i g the uestio : it is t ue? , i the affi ati e – elie i g is the sa e as hat e e goi g fo i a s e i g uestio s a out the world; and that is, in both cases, getting the world right, saying and believing it to be as it is. A d, so the thought o ti ues, si e this is hat elief is fo , e d do ell to ha e a o ept hi h is used to evaluate how well beliefs do this. If this is the role of the concept of credibility (aka: epistemic 12 justification), then it should be no surprise that credibility is immune to non-truth-related reasons: we use judgments of epistemic-justification/credibility to evaluate beliefs precisely in light of how they help us get at true picture of the world. Part of the story, then, seems to be this: some of our concepts are used to evaluate certain kinds of action or attitudes in light of how these serve certain characteristic purposes. Reasons are of the right ki d to g ou d the appli atio of these o epts he the e related in the right way to these purposes. But this is o l pa t of the sto . What s the ight a ? Co side elief agai . “uppose that, as it happens, I am indeed really, ridiculously, good-looking. But I have lots of really good evidence to the co t a to the people a ou d e do t i di ate thei ad i atio , si e the do t a t eaut to go head . Ma e I should elie e it fo p ag ati easo s, ha i g to do ith ho happ I d e if I were really good-looking. This time the content is true. But t uth i p ag ati easo s a e t elated to its the ight a : the do t ha e a thi g to do ith the ki d of e ide e I ha e to elie e i that truth. The problem here would be that my belief, though related to truth, is not related in the way that believing faculties typically try to relate beliefs to truth: by disposing them to covary systematically with evidence. So it is important, in the case of belief, that wrong-kind reasons be related , not just to the characteristic aims of believing, but also to the characteristic methods believers use to achieve these aims. Can we generalize from these observations about belief? Wrong-kinds-of-reasons problems arise for the concept of epistemic justification because it serves to evaluate beliefs in light of a conception of hat eliefs a e supposed to do ep ese t the t uth a d ho the e o all supposed to do it co-varying systematically with evidence). But consider: they also arise for the concept of sportsmanship, 13 because we use that concept to evaluate athletic performances in light of a conception of what these ae o all supposed to do e fu , uild ha a te , et . a d ho the e supposed to do it t to i , follow the spirit of the rules in addition to the letter, etc). In both cases, we conceive an attitude or an activity under a description which takes it to serve some distinctive purpose in some distinctive way— we conceive it, as I will say, as a practice. The ole of the idea of the e ei g a ight a a d a o g a to engage in that activity is to allow us to evaluate instances of it in light of those purposes and methods. I d suggest, the , that e epla e “ h oede s ‘K‘ ith so ethi g like: RKR*: The right kind of reasons with respect to any practice, P, are the reasons that bear on performances within that practice in light of its characteristic aims and characteristic methods for achieving those aims. This account has se e al ad a tages o e “ h oede s: First, it gets the intuitive scope of the wrong kinds of reasons problem right. Recall that Schroeder thinks that the problem arises when activities have correctness-conditions. But the poker-robbery case suggest that this was too narrow. A single recognizable problem arises whenever doing something the right way can be distinguished from doing it as one has most reason to do it. RKR* explains this. Practices define particular ways of pursuing particular aims. Sometimes these methods are definable in some relatively determinate way—as with the formal rules of a game. Correct performance of an activity consists in following these methods to the letter. But sometimes—as with the kind of informal norms that often attach to games, as well—our sense of these methods is not so well-defined. The o espo di g o s a e t determinate enough to be correctness-norms, strictly speaking: they have to do, rather, with the looser idea of consistency with the spirit of the practice in question—of doing it right, rather than doing it correctly. 14 Second, it has no trouble explaining why moral reasons could sometimes be reasons of the right kind—sometimes recognizably moral goals are among the goals that characterize a practice. Maybe, for i sta e, offe si el a ist jokes a e t fu e ause joki g a ou d is a out solidif i g o ds of mutual respect and affection, by showing that these bond can survive minor transgression. If so, jokes that put so eo e ho should e pa t of the solidified o u it outside it do t fit the ai s a d ethods of joking as a practice; they are, therefore, not just unethical, but also, unfunny. Thi d, this a ou t is i a positio to i di ate ou i tuitio s that e so eti es do t ha e a reason at all to do things the right way. Basically, the idea here is that there are cases where it makes sense to understand an action or attitude as a performance within a practice, even though the norms ha a te isti of that p a ti e do t a tuall appl . “o, e.g., i the poke - heati g ase, e a t e plai h I d a t to deal f o the otto of the de k ithout eference to the norms of poker; even though, gi e the da ge to f ie d, I do t ha e a easo at all to follo those o s. “i ila l , i the table-setti g ase, the e fa t that I ha e o easo to flout eti uette a t e e plai ed ithout reference to the norms of etiquette themselves—the terrorists are looking for an incorrect setting of the ta le as a sig al to set off the o , a d it s this e fa t that a els easo to do so. Attitudi al cases raise a somewhat more complex set of issues. Still, the basic idea will be that there are cases where we need to think of a given state as an attitude of a type characterized by certain norms; even though e thi k those o s do t a tuall appl i the ase—as he e thi k that e do t ha e reason to ad i e a di tato s de isi e ess, e e if e thi k de i e ess per se is admirable. This a ou t, I a t to sa , also o t asts fa o a l ith Hie o i s. It allo s us to: a sa everything her account plausibly says about wrong-kind and right-kind reasons for belief; (b) extend the 15 truth in those things to cover many other things besides; and (c) do so without thereby changing our views about the fundamental nature of the reason-providing relation. Believing that p can be, as she says, a matter of affirmatively answering the question: is p true? And considerations relevant to that question can be right-kind reasons to believe that p; it is this question, after all, that point us to the truth-representing purpose believing is supposed to serve in the first place. This, combined with the proposal advanced here, helps to explain an important feature that wrongkind reasons have in the belief case, and in the case of some other attitudes, but lack in many cases of performance in non-attitudinal practices. Because forming a belief just is answering a question; and a wrong-kind reason is not relevant to that question, it is very difficult to consciously and reflectively believe on the basis of (what you know to be) the wrong kinds of reasons.10 Here the very nature of the attitude on which a putative reason bears constrains the description under which it can be considered; it makes it very hard to see something plainly unrelated to the aims and methods of belief as a reason to believe. There is no such problem when playing poker or setting a table or waterboarding. For the performances involved in doing this can be seen under multiple descriptions. Putting the fork on the right side of the plate is setting the table wrong; but in the White House case considered above, it is also saving many lives. Unlike in the case of belief, there is nothing about how an agent has to understand himself as he does this that constrains his normative perspective to that of polite table-setting; no constitutive question that setting a table answers, and therefore, no internal connection between setting a fork down and promoting the achievement of a certain complex pattern of dinnerware. 10 “ h oede all this oti atio as et i a late dis ussio tha op. cit.; and takes its status as an earmark of the right-kind/wrong-kind distinction as a reason to resist expanding the scope to include nonattitudes. The a gu e t i the te t sho s h he eed ot ha e do e this. “ h oede , The U i uit of “tateGi e ‘easo s Ethics 122 (2012): 457-488. 16 So, because actions that constitute performance in practice are not, like beliefs, constitutively linked to the norms that govern those practices, agents can respond much more easily to wrong-kind reasons for action than to wrong-kind reason for belief. However, pace Hieronymi, we can understand this without denying that wrong-kind reasons to believe are (at least sometimes) reasons to believe. Reasons are, just, considerations counting in favor. Further, that sometimes it is difficult or impossible to respond to a class of reasons does not mean that they are not reasons after all. Appearances to the contrary can be explained away, by reference to the resources developed here and by Hieronymi herself. Any given attitude or action can be seen under multiple descriptions; put, as it were, in more than one space of reasons. But for some attitudes in particular, the fact that forming the attitude comes to answering a question makes it very difficult for the attitude-bearer to take up any perspective antithetical to answering that question when forming that attitude. So there is a principled basis for saying, that some o side atio s that eall , ge ui el , ou t i fa o of attitudes a t see to us to do so he ea e considering whether or not to form or maintain said attitudes. Thus the fact that some considerations counting in favor of an attitude cannot be consciously used to form that attitude does not show that they are not reasons for that attitude after all; thus, the wrong-kinds of reasons problem does not require us to rethink our conception of reasons as considerations counting in favor. V. Morals So, I have proposed: to be a reason of the right kind is to be a reason that bears on an attitude or practice in light of its characteristic aims and methods. If this is what it is to be a reason of the right/wrong-kind, why do WKR-problems arise? That is to say: why are there concepts which evaluate a ti ities a o di g to the fo e of o side atio s hi h a e t all pe the o igi al p o le a do l (per the new twist pursued here) the reasons which apply to the conduct of those activities? I think 17 e e o i a positio to sa : e ha e su h o epts e ause e ha e practices—at a first pass, distinctive ways of recognizing and responding to reasons for actions and attitudes. Practices so defined bring norms with them; and we need concepts which encode consistency with those norms. Some things typically fall under the norms of a practice; the practice-given reason which apply to those things are rarely outweighed or cancelled by other sorts of reasons. But even for these things, there are many importance exceptions—demon-threat-motivated valuings, cancelled reasons of correctness, etc. These exceptions point us to the existence of the corresponding practices and their distinctive norms. In so doing they indicate the existence of normative concepts which are, though in some sense defined in terms of an activity, not definable simply as performing that activity as we have most reason to perform it. But though these concepts are not definable along the simple schema: Xing as we have most reason to X, there is no obvious reason to think that they must have some kind of normativity outside the normativity of reasons (construed, per above, as considerations counting in favor). They must be defined on some instance of the more complex schema: Xing in accord with the characteristic purposes and methods of the practice of Xing; but it s ot lea h e d eed a thi g o ati e othe tha the idea of a normative reason to fill this in. To say that something is a purpose seems to be to say that there are reasons of a certain kind to pursue it; to call something a method of pursing a purpose it to say that it is a reason because in some defined set of circumstances it is the most effective way of pursuing some purpose. If the foregoing is correct, then, there seems to be little reason to posit any irreducibly o ati e otio of fitti g ess to e plai WK‘-cases: we can do the same work with the (somewhat) clearer idea of a practice as a distinctively structured set of considerations counting in favor. This shows that the bare existence of wrong-kinds-of- easo s phe o e o is t i o siste t ith the idea that all 18 normativit is the o ati it of ou ti g i fa o of —the e s a ethod fo e plai i g those phenomena which need not obviously appeal to any other kind of normativity. 19 

Reasons: right and wrong Draft of 10/21/2015 The Wrong-Kinds-of-Reasons Problem A demon threatens catastrophic destruction if you don’t value his toenail clippings. So you have very good reason to do so. Yet that doesn’t make them valuable. You can’t get ahead at work unless you respect your boss (he’s an asshole, but he can tell if you’re faking it). That doesn’t make him respectable. Your friend’s depressed and could really use a win. Even if letting him win is the right thing to do, that doesn’t make it a show of good sportsmanship. You’ve got much to gain and nothing to lose by believing in God—but that doesn’t make that belief credible. In each of these cases, a normative concept which seems to be defined in terms of a certain kind of activity isn’t definable, simply, as doing that thing or having that attitude as you have most reason to. Some reasons which seem to bear on valuing/respecting/sporting/believing, don’t seem to bear on whether some instances of these things in question are respectable/valuable/sportsmanlike/credible. These reasons are, as it’s sometimes said, of the wrong kind to ground the application of these sorts of normative concepts. This notion is important to our understanding of the role of reasons in the normative domain in at least two ways. Some, notably Pamela Hieronymi, think that distinguishing right-kind reasons from wrong-kind reasons requires, ultimately, revisiting our notion of a reason itself; wrong-kind reasons clearly count in favor of the things they are reasons for, but yet, so the thought goes, are not reasons for them; thus being a reason for is not merely a matter of counting in favor. Cf. here “The Wrong Kind of Reason” Journal of Philosophy 102 (2005): 437-457; More modestly, but still significantly, some think that the existence of wrong-kind reasons constitutes a decisive objection to philosophical accounts which say that what makes a concept ‘normative’ is that it contains some notion of a ‘reason’, in the sense of a consideration ‘counting in favor’ of an action or attitude. Normative concepts, they say, involve an unanalyzable normative notion of fittingness; and a corresponding metaphysics of normative facts for the relevant actions and attitudes to fit. These views are bolstered by the inability of any existing account to distinguish wrong-kind reasons from right-kind reasons without, either, substantial modification of our account of the reason-providing relation; or, substantial appeal to normative notions not analyzable in terms of reason-provision. My goal here is to sketch a general account of the wrong-kind/right-kind distinction, which does not have these strong anti-‘reason'The first step is to properly understand the scope of the problem. In the next section, I argue that the WKR (for: wrong-kind-of-reason) problem arises not just for attitudes, but also more broadly for any activity where it makes sense to talk about there being a right way and a wrong way to engage in the activity—any practice, as I will say. This idea is a generalization of the idea of Rawls’s idea of a practice, explicated at In all such cases there is a question as to whether considerations which count in favor of doing something within a practice count in favor of the conclusion that doing that thing would be the right way to engage in that practice. But there is also the converse question: do considerations which bear on what counts as the right way to engage in a practice always bear on whether one ought to perform the moves of that practice? In the following section I argue that they do not; and hence that being a reason ‘of the right kind’ to ϕ does not always entail being a reason, simpliciter, to ϕ. In the concluding section I sketch an account which makes sense of this feature of the wrong kinds of reasons problem: on which, very roughly, reasons are of the right kind with respect to a practice if they are appropriately responsive to the characteristic goals and normal methods of that practice. I conclude by arguing, briefly, that this account is consistent, both, with taking reasons to be considerations-counting-in-favor; and, with the reducibility of concepts like ‘value’ to this concept of a reason. Thus, the wrong kinds of reasons problem need not spell the end of the idea that ‘favoring’ is, uniquely, fundamental to the normative domain. The Scope of the Problem The wrong kinds of reasons problem arose, originally, in the debate over so-called buck-passing or fitting-attitudes theories of value. Or anyway arose under that name; as was soon recognized, and as I discuss below, similar issues had long since been raised concerning pragmatic reasons for belief. At a first pass, these theories say that to think that something is valuable is to think that one has most reason to value it; they thus attempt to analyze one central normative concept, the concept of value, in terms of the concept of a reason – thereby supporting the general hypothesis that the fundamental normative notion is the notion of a normative reason. But some reasons which bear on valuing don’t seem to bear on value—demon-given reasons, say. As this problem arose, theorists immediately noticed a similarity to a problem, familiar in epistemology, concerning the relevance of broadly practical reasons to the justification of beliefs. Even if the fact that it would make me happy to believe a pleasant falsehood is a reason to do so, it isn’t, to any extent, a consideration that (epistemically) justifies that belief, or makes it credible. What’s wrong with practical reasons for belief, these theorists thought, is that they’re not suitably related to the object/content of the belief in question: that it would be nice to believe that p doesn’t make it any more likely that p is true. And perhaps we could say the same thing about demon-given reasons for valuing: my reason to value the demon’s toenail clippings, after all, doesn’t seem to have anything much to do with the clippings themselves. As a defense of the fitting-attitudes account of value, this move has its problems. Notably, that it seems likely to be circular: in the case of belief, the RK-reasons need to be related to the truth of the proposition believed. But we’d better not say, analogously, that RK-reasons for valuing need to be related to the value of the valued object—that would be circular in the context of the buck-passing account. Cf. here Rabinowicz and Rønnow-Rasmussen, ibid. But following a suggestion of Mark Schroeder’s, we should suspect in any case that it is not suitable as a model for a general distinction between doing things the right way and doing them as we have most reason to do them. “Value and the Right Kind of Reasons” Oxford Studies in Metaethics 5: 25-55, September 2010. Schroeder has changed his view on these matters since then, for reasons I examine and reject below (n10). ‘Schroeder’ as used in the main text refers to Schroeder-ca.-2010. Clearly, it applies only to intentional attitudes; and that distinction applies to many other things besides. There is a right way and a wrong way to value, to be sure—but there’s also a right way and a wrong way to tie a four-in-hand, or to play a game of basketball. And in all three cases, doing the thing in question as we have most reason to do it isn’t the same as doing it the right way: in each case some reasons are of the wrong kind, while some are of the right. If something like the state-given/object-given distinction is right for the case of valuing, or believing, then, there’s reason to think it’s a special case of a more general account which covers non-intentional attitudes and activities as well. Schroeder suggests that the relevant general class is the class of things that can be done correctly or incorrectly. This is not implausible. We can believe correctly or incorrectly; plausibly, we can value correctly or incorrectly too. And reasons do seem to be of the right kind to ground a judgment of credibility/value just when they go to the correctness of believing or valuing. Similarly, it seems like we can tie a knot or set a table correctly or incorrectly; and that this fact enforces a distinction between reasons which bear on how to engage in that activity on any particular occasion, and reasons which bear on how to get it right. So Schroeder’s account of the scope of the WKR problem is certainly broader, and correspondingly more plausible, than some extant alternatives. But it’s not broad enough. To see why, consider some predicaments you might find yourself in while playing poker. You’ve bet more than you can afford to lose on a game—so much that you’re family will be destitute if you lose. An opportunity arises to cheat and win—you get a minute where you think you can deal yourself a winning card from the bottom of the deck. You’re pretty sure you can get away with it. (And the players you’re playing with, let’s say, are horrible monsters, who will use their winnings in terrible ways if they win.) So you have overwhelming reason to deal from the bottom of the deck. But that doesn’t make it the correct way to deal. Now compare the following case. Same card game, same stakes, same monsters. But you’ve arranged for a friend, who’s watching from across the room, to rob the game and split the winnings with you, when he observes you fold your hand in a certain way, at a certain time. As it happens, at the appointed time, (without cheating) you draw a fantastic hand—one that would probably win, if you didn’t fold it. You have overwhelming reason to fold (you promised your friend you would, after all—also you’ll get the money in the monster’s wallets, too). Folding in this circumstance wouldn’t be cheating—it would, formally speaking, not be an ‘incorrect’ move in the game of poker. But it’s not the right way to play: to play poker right, you have to play to win. It seems, then, that there are two sorts of normative concepts that engender wrong-kinds of reasons problems: concepts of correct conduct in an activity narrowly construed, and something more general, like the idea of doing an activity right. The poker case shows this, but we should have suspected it anyway: concepts like the concept sportsmanship and the idea of fidelity to promises seem to generate WKR problems, even though they don’t seem to have anything obviously to do with correctness. It seems clear, then, that the scope of the problem is broader still than the set of concepts that serve to judge correct or incorrect engagement in an activity: broad enough to include concepts which encode the looser norms that go with the idea of there being a right way and a wrong way to do something. Pamela Hieronomyi has recently argued, against views like Schroeder’s and mine, that the wrong kinds of reasons problem as such only arises for (only some) attitudes. The thought is that we only have a loose intuitive grip on the problem to start with – there seems to be some general problem involved in distinguishing pragmatic from evidential reason to believe, value-relevant from value-irrelevant reasons to value, and so forth. Schroder and I offer reasons to add things to this list; but all we have to go on in doing so, really, is felt similarity. Hieronymi maintain that if we have a clear way of drawing the distinction in some of these cases, that doesn’t cover others, that could be a fault in our initial way of posing the problem rather than a flaw in our solution. She proposes, in particular, a solution to the problem that seems quite convincing for the wrong-kinds-of-reasons problem as it arises for certain attitudes; but which (like the object-given/state-given distinction, of which it is perhaps a descendant) does not apply at all to activities, like dealing a poker hand or playing basketball. The intuitive thought is that forming certain attitudes is in some sense equivalent to answering certain corresponding questions. So, to take the clearest case, to come to believe that p is to affirmatively answer the question: is p true? A right-kind reason for believing, or any other attitude, is one that bears on the relevant corresponding question. A wrong-kind reason bears instead only on some other, related question – that believing in God is a good bet bears on whether is it good for me to believe in god, not on whether there is a god. This is a very plausible account of the wrong-kind/right-kind distinction as that applies to belief, at least. It offers an elegant, informative account of what distinguishes right-kind from wrong-kind reasons for belief, in terms of a plausible and illuminating account of what believing is. It is somewhat less clear, however, that it works for valuing, or for other related attitudes. There does seem to be some intimate relation between coming to believe something and coming to affirm it’s truth; some equivalence between coming to believe that p and coming to answer the question: ‘is p true’, with a ‘yes’. But it is less clear what questions will correspond analogously to attitudes like valuing, or admiring. One might be tempted to say: to value something is to determine that, yes, it is valuable; to admire someone is to determine that, yes, they are admirable. But this temptation should, probably, be resisted, as it links these attitudes too closely to the corresponding judgments. I can judge that opera is valuable without myself valuing it. I can judge that my rival’s work is admirable, without thereby actually admiring it. There is nothing strange about expressing these attitudes, either: ‘Opera is very valuable indeed; but I myself don’t value it’; ‘Her work is admirably rigorous; however, I feel nothing but resentment when I read it’; neither so strange as the classic ‘(It is true that …) the cat is on the mat, but I don’t believe it’. Perhaps, however, the account could be extended to cover attitudes apart from belief – perhaps, e.g., as Hieronomyi suggest, these correspond to some more complex set of questions, affirmative answers to which really do constitute valuing, admiring, etc. In any case the account clearly does not extend to distinguishing wrong-kind from right-kind reasons for engaging in activities, as well. One does not deal a hand of poker, or play a game of basketball, or set a table, by answering a question. So the plausibility of Hieronymi’s solution to the wrong-kinds-of-reasons problem, where it is plausible, constitutes a challenge to Schroeder’s (somewhat) and my (yet) more expansive reading of the scope of that problem. Here I want, simply, to accept this challenge. My response will be to offer a solution to the problem that is broad enough to match my view of its scope; a solution which, I hope, explains why Hieronmy’s solution is for persuasive when it is persuasive, while doing at least as good a job in the cases which are harder for her to account for. One Proposed Solution Returning to Schroeder: his understanding of the problem is broader than Hieronymi’s; still, it is probably too narrow to cover all the cases that seem to raise an intuitive problem of distinguishing doing something ‘the right way’ from doing it as we have most reason to do it. Schroeder thinks that WKR problems arise in basically two ways. They arise, first, in cases where someone has a reason which doesn’t seem to apply to everybody who engages in an activity—robbery schemes for poker players, or demon threats for valuers. Second, they arise in cases where, though everyone who engages in a practice has the relevant reasons, that’s only because everybody per se has the relevant reasons—say, when there are categorical moral reasons not to laugh at a joke. The ensuing thought is that these reasons are of the wrong kind with respect to an activity because they don’t have enough to do with the nature of that activity itself. This is a good thought. But I’m less sure about Schroeder’s more specific interpretation of it: RKR: “The right kind of reasons with respect to an activity A, are all and only those reasons which are shared by necessarily every able person engaging in A, because they are engaged in A, together with all reasons which are derivative from such reasons.” (39). The idea here is that reasons are of the right kind with respect to an activity iff they are (a) general with respect to that activity, in the sense that anyone who engages in that activity has them; but also (b) specific to engagement in that activity, in the sense that such engagement explains why people have them when they have them. It seems to me that there are problems with both conditions: The idea of the second, specificity, condition is to rule out reasons which are, as it were, too general to be reasons of the right kind with respect to an activity—paradigmatically, moral reasons which bear on conduct within an activity but not on whether that conduct is correct with respect to the standards of that activity. In those cases the reasons in question apply for whatever reason moral reason apply more generally; hence, they don’t apply to conduct within an activity just because one is engaged in that activity. There is, I think, reason to be concerned that this rules out too much. Here’s one reason: can’t moral reasons sometimes be reasons of the right kind? So, e.g., perhaps moral reasons are, at least sometimes, the wrong kind of reasons to determine whether a joke is funny. But aren’t they sometimes the right kind of reasons? It seems to me that they sometimes are: some jokes aren’t funny because they are so morally objectionable. Perhaps these problems with the specificity condition can be finessed. The problems facing the generality condition are much more serious. The purpose of this condition is to rule out reasons which are too specific or idiosyncratic to be reasons of the right kind with respect to a given activity. The idea is that, of all the reasons that might apply to the conduct of an activity, only those that apply to that conduct generally are ‘of the right kind’. This makes being ‘of the right kind’ restrictive: if R is a reason of the right kind to ϕ, it’s a reason to ϕ. But it doesn’t look like this is right, as the following cases show: Actually the ‘every able person’ qualification already shows this; what follows shows it in a very different, and I think much less Schroeder-congenial, way. Suppose I’m sitting with monsters again, playing a game of poker. I know now that if I don’t cheat, they’ll win, and they’ll use the money to do something horrible to a friend of mine. Add, if you like, that it’s my fault that my friend is in this predicament. It seems pretty plausible here to say that whatever reasons I might have to play the game right are not just swamped or outweighed, but cancelled or defeated, by the power of my moral reasons to help my friend. But that doesn’t make cheating the right way to play a game of poker: it’s just that the reasons that contribute to that verdict don’t go to whether I should play that way in this case or not. Or consider a more general class of cases, where terrible people have some unusually impressive qualities. These qualities, in these people, might sometimes be admirable. But make these people terrible enough, and it starts to look implausible to think that we have *any* reason to admire them—is there any reason to admire Stalin’s decisiveness? If you don’t like those, try this one, adapted from Schroeder himself. Suppose that I work at the White House, and, at the last minute, find myself called on to set the table for the President. I discover that the server I’m replacing was part of an assassination plot, and that the signal to call the whole thing off is to flout etiquette and put the salad forks on the inside. (Suppose also that I can’t safely warn anyone, on fear that they’ll put the plan in motion right away). Plausibly, I don’t have any reason at all to put the salad fork on the outside. But that’s still the correct way to set the table. Insofar as the latter notion is to be understood, as Schroeder would have us understand it, in terms of the force of the right kinds of reasons in favor of doing so, it follow that I must have reasons ‘of the right kind’ to set the table right even though I don’t have any reasons simpliciter to do so. Admittedly, none of these cases is an utterly devastating counterexample to Schroeder’s RKR. Schroeder would surely respond that they rely on unreliable ‘negative existential intuitions’ about reasons. In all these cases, he could say, there is some reason to do things the ‘right’ or ‘correct’ way. It is just too weak to be worth mentioning, in light of the reasons on the other side. In response, note that the claim that there is always some reason to do something the ‘right’ or ‘correct’ way begins to look less plausible when we focus our attention on activities and practices that are not just morally neutral, like table-setting or card-playing, but morally objectionable. There is, apparently, a right way and a wrong way to ‘waterboard’ someone: to do it right you have to force the subject’s mouth open, so that the water pours into his throat as well as his nose. Does a prospective waterboarder always have some reason to open his subject’s mouth, before pouring the water over? My intuition – yours too, I hope – is that he does not always have some reason to do this. We could be wrong, here. But we would need some actual, positive, reason to believe that we are wrong. And it’s not at all clear that we have it. Schroeder, to be sure, can say that the other strengths of RKR give us such a reason, in the absence of a compelling alternative. this response depends on the absence of compelling alternatives. I will argue, below, that there is one – one that has all the strengths of Schroeder’s account and others besides. A Different Approach A basic insight of Schroeder’s account is the thought that considerations are reasons of the right kind with respect to an activity when they not idiosyncratic with respect to that activity: when they are, in some sense, normal. RKR understands normality in terms of (necessitated) generality: reasons are normal with respect to an activity because they apply generally, to anyone who engages in that activity. But if, as I have suggested, ‘right kind’ reasons can be not just outweighed but cancelled by other reasons, then generality doesn’t look like a very plausible interpretation of normality: some reasons which apply to the normal conduct of an activity might well fail to apply to it in every case. Still, I think the idea of understanding the right kind/wrong kind distinction in terms of normality/abnormality is on the right track. Think back to one of the clearer wrong-kind-of-reasons cases—the case of pragmatic reasons for belief. Why doesn’t the fact that it’d make me feel really good to believe that I’m super-good looking make this belief credible? Proponents of the state-based/object-given approach discussed above will say that it’s because this reason doesn’t have anything to do with the content of the belief in question. But of course it does have something to do with it: after all, it’s my being good-looking that I’d be happy about. The problem, instead, has to be the one that Hieronymi’s account directs us to: that it doesn’t have enough to do with the truth of the belief that I’m really good-looking. Belief aims at the truth: representing the truth is what beliefs are for, in some important sense. That plausibly, is why believing something is answering the question: ‘it is true?’, in the affirmative – what we’re going for in believing is the same as what we’re going for in answering questions about the world; and that is, in both cases, getting the world right, saying and believing it to be as it is. And, so the thought continues, since this is what belief is for, we’d do well to have a concept which is used to evaluate how well beliefs do this. If this is the role of the concept of credibility (aka: epistemic justification), then it should be no surprise that credibility is immune to non-truth-related reasons: we use judgments of epistemic-justification/credibility to evaluate beliefs precisely in light of how they help us get at true picture of the world. Part of the story, then, seems to be this: some of our concepts are used to evaluate certain kinds of action or attitudes in light of how these serve certain characteristic purposes. Reasons are of the right kind to ground the application of these concepts when they’re related in the right way to these purposes. But this is only part of the story. What’s ‘the right way’? Consider belief again. Suppose that, as it happens, I am indeed really good-looking. But I have lots of really good evidence to the contrary (the people around me don’t indicate their admiration, since they don’t want my beauty to go to my head). Maybe I should believe it for pragmatic reasons, having to do with how happy I’d be if I were really good-looking. This time the content is true. But my pragmatic reasons aren’t related to its truth in ‘the right way’: they don’t have anything to do with the kind of evidence I have to believe in that truth. The problem here would be that my belief, though related to truth, is not related in the way that believing faculties typically try to relate beliefs to truth: by disposing them to covary systematically with evidence. So it is important, in the case of belief, that wrong-kind reasons be related , not just to the characteristic aims of believing, but also to the characteristic methods believers use to achieve these aims. Can we generalize from these observations about belief? Wrong-kinds-of-reasons problems arise for the concept of epistemic justification because it serves to evaluate beliefs in light of a conception of what beliefs are supposed to do (represent the world) and how they’re normally supposed to do it (by co-varying systematically with evidence). But consider: they also arise for the concept of sportsmanship, because we use that concept to evaluate athletic performances in light of a conception of what these are normally supposed to do (be fun, build character, etc.) and how they’re supposed to do it (try to win, follow the spirit of the rules in addition to the letter, etc). In both cases, we conceive an attitude or an activity under a description which takes it to serve some distinctive purpose in some distinctive way—we conceive it, as I will say, as a practice. The role of the idea of there being a ‘right way’ and a ‘wrong way’ to engage in that activity is to allow us to evaluate instances of it in light of those purposes and methods. I propose, then, that we replace Schroeder’s RKR with: RKR*: The right kind of reasons with respect to any practice are the reasons that bear on performances within that practice in light of its characteristic aims and characteristic methods for achieving those aims. This account differs from Schroeder’s in several (helpful) ways: First, it gets the intuitive scope of the wrong kinds of reasons problem right. Recall that Schroeder thinks that the problem arises when activities have correctness-conditions. But the poker-robbery case suggest that this was too narrow. A single recognizable problem arises whenever doing something the right way can be distinguished from doing it as one has most reason to do it. RKR* explains this. Practices define particular ways of pursuing particular aims. Sometimes these methods are definable in some relatively determinate way—as with the formal rules of a game. Correct performance of an activity consists in following these methods to the letter. But sometimes—as with the kind of informal norms that often attach to games, as well—our sense of these methods is not so well-defined. The corresponding norms aren’t determinate enough to be correctness-norms, strictly speaking: they have to do, rather, with the looser idea of consistency with the spirit of the practice in question—of doing it right, rather than doing it correctly. Second, it has no trouble explaining why moral reasons can sometimes be reasons of the right kind—sometimes recognizably moral goals are among the goals that characterize a practice. Maybe, for instance, offensively racist jokes aren’t funny because joking around is about solidifying bonds of mutual respect and affection, by showing that these bond can survive minor transgression. If so, jokes that put someone who should be part of the solidified community outside it don’t fit the aims and methods of joking as a practice; they are, therefore, not only unethical, but also, unfunny. Third, this account is in a position to vindicate the intuitions, developed above, that we sometimes don’t have any reason at all to do things the right way. Basically, the idea here is that there are cases where it makes sense to understand an action or attitude as a performance within a practice, even though the norms characteristic of that practice don’t actually apply. So, e.g., in the poker-cheating case, we can’t explain why I’d want to deal from the bottom of the deck without reference to the norms of poker; even though, given the danger to my friend, I don’t have any reason at all to follow those norms. Similarly, in the table-setting case, the very fact that I have no reason to flout etiquette can’t be explained without reference to the norms of etiquette themselves—the terrorists are looking for an incorrect setting of the table as a signal to set off the bomb, and it’s this very fact that cancels my reason to do so. Attitudinal cases raise a somewhat more complex set of issues. Still, the basic idea will be that there are cases where we need to think of a given state as an attitude of a type characterized by certain norms; even though we think those norms don’t actually apply in the case—as when we think that we don’t have reason to admire a dictator’s decisiveness, even if we think decisiveness per se is admirable. This account, I want to say, also contrasts favorably with Hieronymi’s. It allows us to: (a) say everything her account plausibly says about wrong-kind and right-kind reasons for belief; (b) extend the truth in those things to cover many other things besides; and (c) do so without thereby changing our views about the fundamental nature of the reason-providing relation. Believing that p can be, as she says, a matter of affirmatively answering the question: is p true? And considerations relevant to that question can be right-kind reasons to believe that p; it is this question, after all, that point us to the truth-representing purpose believing is supposed to serve in the first place. This, combined with the proposal advanced here, helps to explain an important feature that wrong-kind reasons have in the belief case, and in the case of some other attitudes, but lack in many cases of performance in non-attitudinal practices. Because forming a belief just is answering a question; and a wrong-kind reason is not relevant to that question, it is very difficult to consciously and reflectively believe on the basis of (what you know to be) the wrong kinds of reasons. Schroeder call this ‘motivation asymmetry’ in a later discussion than op. cit.; and takes its status as an earmark of the right-kind/wrong-kind distinction as a reason to resist expanding the scope to include non-attitudes. The argument in the text shows why he need not have done this. Schroeder, “The Ubiquity of State-Given Reasons” Ethics 122 (2012): 457-488. Here the very nature of the attitude on which a putative reason bears constrains the description under which it can be considered; it makes it very hard to see something plainly unrelated to the aims and methods of belief as a reason to believe. There is no such problem when playing poker or setting a table or waterboarding. For the performances involved in doing this can be seen under multiple descriptions. Putting the fork on the right side of the plate is setting the table wrong; but in the White House case considered above, it is also saving many lives. Unlike in the case of belief, there is nothing about how an agent has to understand himself as he does this that constrains his normative perspective to that of polite table-setting; no constitutive question that setting a table answers, and therefore, no internal connection between setting a fork down and promoting the achievement of a certain complex pattern of dinnerware. So, because actions that constitute performance in practice are not, like beliefs, constitutively linked to the norms that govern those practices, agents can respond much more easily to wrong-kind reasons for action than to wrong-kind reason for belief. However, pace Hieronymi, we can understand this without denying that wrong-kind reasons to believe are (at least sometimes) reasons to believe. Reasons are, just, considerations counting in favor. Further, that sometimes it is difficult or impossible to respond to a class of reasons does not mean that they are not reasons after all. Appearances to the contrary can be explained away, by reference to the resources developed here and by Hieronymi herself. Any given attitude or action can be seen under multiple descriptions; put, as it were, in more than one space of reasons. But for some attitudes in particular, the fact that forming the attitude comes to answering a question makes it very difficult for the attitude-bearer to take up any perspective antithetical to answering that question when forming that attitude. So there is a principled basis for saying, that some considerations that really, genuinely, count in favor of attitudes can’t seem to us to do so when we are considering whether or not to form or maintain said attitudes. Thus the fact that some considerations counting in favor of an attitude cannot be consciously used to form that attitude does not show that they are not reasons for that attitude after all; thus, the wrong-kinds of reasons problem does not require us to rethink our conception of reasons as considerations counting in favor. Morals So, I have proposed: to be a reason of the right kind is to be a consideration that bears on an attitude or practice considered in light of its characteristic aims and methods. If this is what it is to be a reason of the right/wrong-kind, why do WKR-problems arise? That is to say: why are there concepts which evaluate activities according to the force of considerations which aren’t all (per the original problem) and only (per the new twist pursued here) the reasons which apply to the conduct of those activities? I think we’re now in a position to say: we have such concepts because we have practices—at a first pass, distinctive purposes for and methods of recognizing and responding to reasons. Practices so defined bring norms with them; and we need concepts which encode consistency with those norms. Some things typically fall under the norms of a practice; the practice-given reason which apply to those things are rarely outweighed or cancelled by other sorts of reasons. But even for these things, there are many importance exceptions—demon-threat-motivated valuings, cancelled reasons of correctness, etc. These exceptions point us to the existence of the corresponding practices and their distinctive norms. In so doing they indicate the existence of normative concepts which are, though in some sense defined in terms of an activity, not definable simply as performing that activity as we have most reason to perform it. But though these concepts are not definable along the simple schema: Xing as we have most reason to X, there is no obvious reason to think that they must have some kind of normativity outside the normativity of reasons (construed, per above, as considerations counting in favor). They must be defined on some instance of the more complex schema: Xing in accord with the characteristic purposes and methods of the practice of Xing; but it’s not clear why we’d need anything normative other than the idea of a normative reason to fill this in. To say that something is a purpose seems to be to say that there are reasons of a certain kind to pursue it; to call something a method of pursing a purpose it to say that it is a reason because in some defined set of circumstances it is the most effective way of pursuing some purpose. If the foregoing is correct, then, there seems to be little reason to posit any irreducibly normative notion of ‘fittingness’ to explain WKR-cases: we can do the same work with the (somewhat) clearer idea of a practice as a distinctively structured set of considerations counting in favor. This shows that the bare existence of wrong-kinds-of-reasons phenomenon isn’t inconsistent with the idea that all normativity is the normativity of ‘counting in favor of’. PAGE \* MERGEFORMAT 1