Reasons and Guidance (Or, Surprise Parties and Ice Cream)
2016; Wiley; Volume: 57; Issue: 3 Linguagem: Inglês
10.1111/phib.12086
ISSN2153-960X
Autores Tópico(s)Epistemology, Ethics, and Metaphysics
ResumoA familiar and intuitive thought is that normative reasons must be able to guide us. That is what reasons seem to be for. Considerations that cannot guide cannot do what reasons are supposed to do. To put the same point differently, it is the job of a reason to recommend that a person perform a certain act or hold some attitude. If it is do that job, the relevant person must be able to heed and respond to its recommendation. Thoughts like this lie behind the following response constraint on reasons: RC is distinct from the principle that reason implies can, according to which: While RC entails RIC, the converse does not hold. It is consistent with RIC, but not RC, that there are cases in which there is a reason for a person to φ, and she can φ, but not for that reason. We return to this below. RC offers a plausible way of capturing the idea that reasons guide. It is also of wider significance. On a standard interpretation, it is a premise of Williams' most influential argument for a Humean theory of reasons, according to which, very roughly, that p is a reason for you to φ only if φing serves some motive (desire, aim, etc.) you have (Williams 1981). This is supposed to follow from RC together with a Humean theory of motivation according to which, very roughly, you cannot φ for the reason that p unless you think that, given that p, φing will serve some motive you have. More recently, Kelly (2002), Shah (2006), and others have argued that RC rules out the possibility of pragmatic reasons for belief and other reasons of the 'wrong kind'. To adapt an example from Firth (1981), suppose that Jane Doe is suffering from a serious illness. If Jane believes that she will survive, she is more likely to do so. But Jane cannot believe that she will survive for the reason that doing so increases her chances of survival. Given RC, it follows that the fact that doing so increases her chances is not a reason for Jane to believe that she will survive. Others suggest that something like RC is what distinguishes the deontic from the evaluative (e.g. Williams 1981; Gibbons 2013; Raz 2011: ch. 2; Shah 2006). If the soil is dry, it would be good if Lilian were to turn on the sprinkler and if rain were to fall. Why, then, is there a reason for Lilian, but not for the clouds, to water the garden? RC might provide (the beginnings of) an answer to that question. RC is also thought to support some form of perspectivism, according to which only considerations within your ken or cognitive reach provide reasons, hence, determine what you ought to do (see Dancy 2000: 59; Lord 2015; Gibbons 2013: ch.6; Raz 2011: 110–111). Suppose that you are not able to φ for the reason that p unless you stand in some epistemic relation—for example, knowledge—to that consideration. Given RC, it follows that a consideration is a reason for you only if you stand in the relevant epistemic relation to it. These remarks illustrate the importance of and interest in RC. Our aim here is not to argue for that principle or defend the views and arguments in which it appears. Rather, our first aim is to defend RC against certain influential objections. Specifically, we consider and respond to two classes of alleged counterexamples to it. In doing so, we distinguish two different readings of RC, reject one and defend the other. This serves our second aim—to clarify how RC is to be understood. Joel's career, his wife and her career, his friends, his Lakers' season tickets, his family, and his loves of surfing and of mountain climbing all tie him to Los Angeles. But Joel also loves chocolate-cayenne-cinnamon flavored ice cream, which he can only get in Madison, Wisconsin (Schroeder 2007: 166). The fact that he can only get chocolate-cayenne-cinnamon flavored ice cream (hereafter, ice cream) in Madison is a reason for Joel to move there. But Joel is not capable of moving to Madison for that reason, given how much he cares about all of the things tying him to LA. There is a surprise party for Beth at her house that starts at 5pm. Beth loves surprise parties and it would make her very happy to arrive home at 5pm. However, were Beth to find out that there is a surprise party for her at 5pm, the surprise would be ruined, and the party would be a disaster. It is very intuitive to think that the fact that there is a surprise party waiting for her is a reason for Beth to go home at 5pm. Note, for instance, that an informed friend might advise Beth to go home by saying, 'Though I can't tell you what it is, there's a reason for you to go home!' But Beth cannot go home for the reason that there is a surprise party there. That would require Beth to believe that there is a surprise party, which would spoil the surprise. In that case, it would no longer be a fact that there is a surprise party waiting for Beth and, hence, there would no longer be a reason for her to go home for 5pm.3 More generally, a consideration that is a reason for you to φ only so long as you do not believe that it obtains, or only so long as you do not φ on the basis of it, looks like a counterexample to RC.4 Cases involving self-effacing reasons are not just of interest insofar as they seem to pose a challenge to RC. For instance, as Lord (Manuscript) points out, Beth's case is a potential counterexample to the kind of perspectivist view mentioned above. The fact that there is a surprise party waiting for her is a reason for Beth to go home for 5pm, but only so long as it lies beyond her ken.5 In addition, self-effacing reasons might seem to be counterexamples to certain analyses of reasons. Consider, for example, the view that for a consideration to be a reason for φing is for it to be a premise in good reasoning or 'sound deliberation' that concludes with φing.6 If Beth were to believe that there is a surprise party waiting for her, it would not be good reasoning for her to reason from that belief to a decision to go home for 5pm; nonetheless, what she believes is a reason for that decision. In what follows, we defend RC against counterexamples of both sorts. Before doing so, let us mention, so as to put aside, alternative responses to each. One tempting response to cases of massively outweighed reasons is to insist that, in the above example, Joel can move to Madison for the reason that he can get ice cream there; he just won't. And, as Schroeder (2007: 166) grants, there are senses of 'can' in which this seems right. For instance, Joel can move to Madison for that reason in the sense that there are possible circumstances—circumstances in which he has none of the cares and concerns that are most central to him—in which he does just that. However, Schroeder suggests that, if RC involves this sense of 'can', it is more or less trivial. For just about any agent, and any putative reason, there are circumstances under which that agent would be motivated by that reason, given radical enough changes to their psychology. That p is a reason for you to φ only if you are able to be pro tanto motivated to φ by the reason that p.7 As I imagine Joel's case, he is disposed to find the facts about chocolate-cayenne-cinnamon ice cream salient, if ever he thinks about moving to Madison, but these thoughts would not actually motivate him to move, not even in any pro tanto sense – this disposition is masked by the presence of such clear motivations not to move to Madison. Joel, I think, is not a case of a conflicted psychology—even slightly. He feels no conflict at all – not even one that is very tiny—over only giving up his favorite ice cream (2007: 166). Proponents of the above line of response are likely to dig their heels in here. And Joel is certainly more clearly problematic for the original constraint than the weaker one. So we risk a stalemate here. Moreover, it is unclear whether the weakened version of RC can be put to work, for example, in arguing against reasons of the wrong kind. To return to the earlier example, although Jane cannot believe that she will survive for the reason that doing so increases her chances of survival, one might think that Jane's awareness of the benefits of so believing might to some extent motivate her to do so. It is also worth noting that weakening RC in the above way does not offer a solution to the problem of self-effacing reasons.8 The most familiar response to self-effacing reasons is to deny that they are reasons. Indeed, there is some inclination to deny that self-effacing reasons are reasons precisely because we cannot respond to them (see Kiesewetter Forthcoming; Lord Manuscript; Setiya 2009: 538; Sinclair 2012).9 We will suggest here that this denial is unnecessary. We will offer a unified response to the two counterexamples. Our response grants that self-effacing reasons are reasons, articulates a non-trivial sense in which agents can act for massively outweighed reasons, and avoids the stalemate over Joel and the weakened response constraint. RC says that reasons must be considerations we are able to act in light of. But 'able' can be understood in different ways. For instance, if you know how to make a cake but lack the ingredients, then in one sense you are able to make a cake but in another you are not. Likewise, when Andy Murray does not have a tennis racket to hand, then in one sense is he retains the ability to serve but in another sense he is unable to do so. This distinction gets marked in different ways by different theorists.10 Following Mele (2002), we will use 'general ability' for the sense in which you are able to make a cake in the above case, or in which Murray is able to serve. A general ability to φ is an ability to φ in a wide range of circumstances, if not the present circumstances (here and now). If you have a general ability, in this sense, it might often fail to manifest. This could be because of a lack of opportunity—as when you lack the ingredients to make a cake—or because of some kind of interference—as when you are too tired to concentrate on what you are doing. When you lack the opportunity to manifest a general ability, or when something interferes with this ability, then you lack specific ability. A specific ability to φ is an ability to φ in your present circumstances (here and now). A general ability can be thought of as a kind of power or competence. As such, there is a crucial difference between having a general ability (in the above sense) and being able to acquire such an ability. A monoglot speaker of English lacks the general ability, hence, the power, to speak Mandarin, though she might have the ability, general or specific, to acquire that ability (cf. Maier 2015).11 Given the distinction between general and specific abilities, we can distinguish two versions of the response constraint: Perhaps RCSpecific entails RCGeneral—we take no stand on that here—but the converse does not hold. We will argue that, while both massively outweighed reasons and self-effacing reasons are counterexamples to RCSpecific, neither is a counterexample to RCGeneral. In doing this, we will show that RCGeneral is the version of the response constraint that is most plausible anyway. Given the distinction between general and specific abilities, we can also distinguish two versions of the principle that reason implies can: Ordinarily, reason implies can is understood to concern specific ability, along the lines of RICSpecific (see, for example, Vranas 2007: 169–70). So understood, only RCSpecific, not RCGeneral, entails reason implies can. Nonetheless, RCGeneral is consistent with RICSpecific, which we grant for present purposes. What is it to have the general ability to φ for the reason that p? After all, reasons are specific considerations—for example, that it is now snowing in Boston, or that the game is about to start. What is it to have a general ability to respond to such a consideration? A general ability must be an ability to have a certain type of response to certain types of reasons. For example, you might have a general ability to perform acts of the wearing boots type in response to considerations of the snow in your circumstances type. A useful way to develop this is to connect responding to reasons with reasoning. When you φ for the reason that p, we can think of you as engaging in a piece of reasoning: 'p, so I'll φ'.12 A general ability to have a certain type of response to a certain type of reason can then be thought of as a matter of having the general ability to follow certain patterns or rules of reasoning. For instance, your general ability to wear boots for the reason that it is snowing can be thought of as the general ability to follow the rule 'if it is snowing (in your circumstances), wear boots!' So understood, RCGeneral implies that that p is a reason for you to φ only if you are able in the general sense to follow a rule of reasoning of which 'p, so I'll φ' is an instance. In what follows, we will work with this way of understanding RCGeneral. However, we stress that this is primarily an expositional device. Those who prefer to understand general abilities to make types of responses to types of reasons in other ways can still accept the substance of our defense of RCGeneral. Before moving on, it is worth noting that abilities can be more or less general. Both a reason and the relevant response will be tokens of various types, some more specific than others. So, there is not really a binary distinction between a general and a specific version of RC but a spectrum of response constraints ranging from the most specific to the most general. As this observation does not make a difference to the discussion to follow, and for ease of presentation, we will continue to contrast RCGeneral and RCSpecific. Massively outweighed reasons are not counterexamples to RCGeneral. In the above case, Joel can in general make responses of a certain type, for example, decisions to move—for a certain type of reasons—namely, those having to do with the obtaining of ice cream. Alternatively, Joel does have a general ability that might be manifest in moving to Madison for the reason that he can get ice cream there. For instance, Joel is in general able to follow a rule or pattern of reasoning such as: 'φing would get me ice cream, so I'll φ'. This ability might manifest itself, for instance, in Joel's going to the ice cream parlor when he is in Madison. Of course, this ability does not manifest in his moving to Madison. But that is just because there are so many things—all the things keeping him in LA—that interfere with this ability, and stop him exercising it. These points do not trivialize RCGeneral. Joel's ability is an ability to act on instrumental considerations: to act for the reason that φing would achieve F, when F is something he desires. As there are many things that Joel does not desire, there are many Fs such that Joel lacks a general ability to reason 'if φing would get me F, I'll φ'. For instance, Joel might lack the ability to reason 'if φing would get me marmite, I'll φ', or 'if φing will get me killed, I'll φ'.13 Cases of massively outweighed reasons are also important because they show that we should expect nothing stronger than RCGeneral to be true. It is a platitude that reasons to act can be outweighed. If an agent is responding properly to reasons, awareness of outweighing reasons will block the manifestation of their ability to respond to a reason. In some cases, like Joel's, this might mean not just that an agent won't respond to an outweighed reason but that they are not able to do, in the sense that they lack the specific ability to do so. This does not show that the outweighed reason is not in fact a reason. Our suggestion is that cases of massively outweighed reasons are counterexamples to RCSpecific but not to RCGeneral. A proponent of RCSpecific might respond as follows. With Schroeder, we seem to assume that, if a person's disposition to φ is masked in a circumstance, she lacks the specific ability to φ in that circumstance.14 In the case of Joel, his disposition to be moved by the prospect of ice cream is masked by his love of surfing and the like. It remains the case that he has a disposition to act for the sake of ice cream. Furthermore, the manifestation conditions for this disposition seem to be met—for instance, Joel is aware of the fact that ice cream is only available in Madison. Surely, the thought continues, he retains the specific ability to move to Madison for this reason. To deny this is to suggest that one lacks the specific ability to φ in any circumstance in which one's disposition to φ fails to manifest or, more generally, in which one does not φ. This is implausible. Here is a way to put this proposal. When introducing the distinction between general and specific abilities, we said that you lack specific ability when you lack the opportunity (as when you lack the ingredients to make a cake) or when something interferes (as when you are too tired to concentrate on what you are doing). In effect, the present suggestion is that opportunity, together with the appropriate disposition(s), suffices for specific ability. Interference does not preclude specific ability. However, it is not clear what justifies this differential treatment of opportunity and interference. As we put it before, when you have the specific ability to φ, you have the ability to φ in your present circumstances—here and now. But if someone prevents Andy Murray from serving, he is not able to serve here and now, anymore than he can when he is on the plane with his racket in the hold. To put it another way, when you have the specific ability to φ, it is up to you whether to φ. But interference stops it being up to you whether to φ just as much as lack of opportunity does.15 Does this mean that one lacks the specific ability to φ in any circumstance in which one does not φ? It might seem that it does (cf. Honoré 1964: 464). After all, suppose you want to φ and also want to ψ, when you know you cannot do both. If you choose to φ, then it seems that your disposition to ψ must have been interfered with, by your choice, or by the strength of your desire to φ. However, we think that the problem here lies with the connection between dispositions and abilities being relied on. In the specific sense of 'able,' which we have tried to elucidate, what one is able to do is constrained by some of one's psychological features but not others. For instance, our deep cares and concerns can constrain our options: a loving parent may not be able to choose to kill their child, a proud politician may not be able to bring themselves to apologize for their mistake. In this way, our specific abilities—what we can do here and now—are constrained by our psychology. Our claim, following Schroeder, is that Joel's attachment to Los Angeles constrains his options in this way. But our cares and concerns leave us with plenty of options open. So it hardly follows that what Joel can do in the here and now is only what he in fact does. If accounts of abilities that connect abilities very closely to dispositions struggle to accommodate these features of abilities, that is a problem for those accounts, rather than for anything we have argued for here.16 Of course, one might simply stipulate a sense of 'able' in which one is able to φ when one has a disposition to φ and the manifestation conditions for this disposition are met. And one might defend a version of the response constraint that uses this sense of 'able'. In effect, this would be a version of the response condition that is more general than RCSpecific but less general than RCGeneral. One worry about this approach is that it is not clear that it could be what defenders of the response constraint have had in mind, given that the sense of 'able' it relies on is stipulative and, as the discussion above indicates, does not seem to correspond to the ordinary senses of 'able' we have elucidated. But we need not press this worry too strongly. We can accept that this proposal also accommodates the case of Joel. The choice between it and RCGeneral is then to be made on independent grounds. The rest of the paper will provide such grounds. In the next section, we argue that RCGeneral, unlike the present proposal, also allows for self-effacing reasons.17 In the following section, we provide a further rationale for specifying the abilities required by reasons in the way that RCGeneral does. Consider again Beth. Does she have a general ability that might manifest in going home for the reason that there's a surprise party there? You might think not. After all, Beth is not, in general, able to follow the rule 'if there's a surprise party for me at location L, go to L!' However, things are not so straightforward. For there are other patterns of reasoning that the surprise party reasoning exemplifies that Beth does follow. For instance, Beth has the general ability to reason from desiring an end, believing that p and believing that, given p, φing will help her achieve that end, to intending to φ. Suppose that Beth loves bowling. And suppose that she knows that, given that the number 10 bus goes to the bowling alley, taking the number 10 bus promotes bowling. In that case, Beth might reason: the number 10 bus goes to the bowling alley, so I will take the number 10 bus. The surprise party reasoning is also an instance of this pattern of reasoning. So Beth does have a general ability to reason in a way that 'there's a surprise party at home, so I'll go home' exemplifies. The pattern of reasoning here is relatively coarse-grained. It might be thought that reasons should be connected to more fine-grained abilities—abilities which are more closely connected to the content of the particular reason. And it might be thought that Beth has no such abilities with respect to the surprise party reasoning. In fact, though, Beth does have some such abilities. For instance, Beth might well have the general ability to reason 'there's a surprise F at location L, so I'll go to L', when a surprise F is something that she enjoys or is interested in. Beth might manifest this ability in going into a house for the reason that there will be a surprise gift there, or for the reason that she will have a surprise visitor. (To be clear: the surprise here is not that there is a gift/visitor but what/who it is.) However, it is true that Beth lacks certain more fine-grained abilities that the surprise party reasoning exemplifies. In particular, as already noted, she lacks the general ability to reason: there's a surprise party for me at L, so I'll go to L. So, if RCGeneral is understood to imply that the surprise party is a reason for Beth to go home only if she has this ability, then RCGeneral implies that the surprise party is not a reason for her to go home. The question thus seems to be how fine-grained an ability we should take RCGeneral to require. It is worth noting that, even when interpreted so as to imply that the surprise party is not a reason for Beth, RCGeneral does not rule out all self-effacing reasons. Consider a version of the surprise party case in which Beth knows that, if going home would make her happy, then there is a surprise party there. In such a case, the fact that going home would make her happy is a self-effacing reason. But Beth does have the general ability to reason: going home would make me happy, so I'll go home. It is just that this ability is interfered with when she knows that going home would have made her happy only because there is a surprise party there. Furthermore, when understood in this very fine-grained way, RCGeneral seems too strong. Consider: spinach contains iron. That seems like it might be a reason for Elliot to eat his spinach. But Elliot lacks the concept of iron. So, he is not able to eat spinach for the reason that it contains iron (or for any other iron-involving reasons). Given the very fine-grained version of RCGeneral, it is not a reason for him to do so. That seems wrong. At least, it seems to run counter to our ordinary talk of reasons. Consider, for instance, that we would be happy to advise Elliot to eat his spinach on the grounds that it contains iron. In contrast, less fine-grained interpretations of RCGeneral allow that the fact that spinach contains iron is a reason for Elliot to eat his spinach. After all, Elliot is able to perform the type of action to which eating spinach belongs on the basis of the kind of fact that relates to, say, the benefits of doing so, of which the fact that spinach contains iron is a token or instance.18 So we have seen that the surprise party case is a counterexample only to the most fine-grained ways of interpreting RCGeneral. Moreover, we have seen that even the most fine-grained interpretation allows some self-effacing reasons and also has some questionable implications. Since, as we have also argued, RCGeneral should be preferred to RCSpecific, we take this to show that the surprise party case is not a counterexample to the most plausible versions of the response constraint. Nonetheless, it may seem unsatisfying to leave things there. While we have suggested that reflection on cases, such as Elliot's, provides some reason to prefer more coarse-grained interpretations of RCGeneral, it would be more satisfying to offer a rationale for interpreting the principle in one way rather than another. In the next section, we will sketch the beginnings of such a rationale. Consider: Why is the fact that there is a surprise party waiting for her a reason for Beth to go home for 5pm? A natural answer to this question might be: Because Beth loves surprises and, if she goes home for 5pm, she will promote something she loves. Corresponding to this explanation is Beth's general ability to reason 'p, so I'll φ' when she thinks that, given p, φing promotes something she loves. In light of these observations, we can address the question of how to specify the general ability that, according to RCGeneral, is a condition on a consideration's being a reason. The specification of the general ability should correspond to an explanation of why the relevant consideration is a reason for the relevant response.19 In support of this suggestion, recall that RCGeneral is supposed to capture the intuitive idea that reasons must be able to guide us. Plausibly, though, the underlying thought here is more specific: reasons must be able to guide us in their capacity as reasons (cf. Shah 2006: 485ff). That is, a reason to φ must be able to guide you to φ in virtue of being a reason to φ. Reasons must be able to guide in this way because it is only so that they make available certain kinds of important normative status. For example, to be doxastically justified in believing that p, you need to believe that p because there is justification (a justifying reason) to believe that p, and to be creditworthy for φing on certain grounds, you need to φ because those grounds make it right to φ (cf. Arpaly 2002; Arpaly and Schroeder 2014; Markovits 2010a).20 Importantly, φing for the reason that p does not suffice for responding to the fact that p as a reason to φ. This is perhaps clearest in the theoretical case. For instance, suppose that you reason: If your premises are known, you here come to believe that there is a spider nearby on the basis of facts that constitute a reason for your belief. However, it does not follow that you respond to these reasons as reasons. For instance, suppose that in drawing this inference you are following the arachnophobe's rule: from any premises concerning spiders, conclude that there is a spider nearby. In that case, your inference would not result in a (doxastically) justified belief. You would not be responding to the relevant reasons as reasons (cf. Firth 1978; Millar 1991; Turri 2010; Lord and Sylvan Forthcoming). What more is needed? We will not attempt a full answer to this difficult question here. What we will suggest is that that one way to respond to that p as a reason to φ is for one's φing for the reason that p to manifest an ability that corresponds to that in virtue of which that p is a reason to φ.21 To see the plausibility of this, consider the spider inference again. Suppose your inference to the conclusion that there is a spider nearby manifests your ability to engage in modus ponens reasoning, that is, reasoning of the form: p, if p then q, so q. Here your ability corresponds to that in virtue of which your premises constitute a reason to believe that there is a spider nearby. Roughly, your premises constitute this reason because their form is such as to guarantee that, if they are true, your conclusion is true too. That is a plausible explanation of why your reasoning in this case allows you to respond to reasons as reasons, and so results in a justified belief. We have suggested that when you have an ability that corresponds to an explanation of why certain considerations are reasons, then you have an ability to respond to those considerations as reasons. If that is right, then we have a rationale for interpreting RCGeneral in the way we have recommended. According to RCGeneral, for a fact to be a reason for you to make some response, you must have the general ability to make responses of that type for reasons of that type. The question we have addressed in this section is how to specify the relevant ability—in particular, the reason and response types. Our suggestion was that the specification should correspond to an explanation for why a consideration of the relevant type is a reason for the relevant type of response. The rationale for this is that, when a subject manifests the ability so specified, she is responding to the consideration as a reason for the relevant response. We have argued that neither massively outweighed nor self-effacing reasons require us to give up RC, understood as the claim that only considerations we have the general ability to act on can be reasons. We have
Referência(s)