University of Southern California Dissertations and Theses

Extending credit

(USC Thesis Other)

Extending credit

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Extending Credit by Shane Ward A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree (DOCTOR OF PHILOSOPHY) (PHILOSOPHY) August 2023 ii Acknowledgements A dissertation on credit would not be complete if I did not extend credit to those who made it possible for me to write this dissertation and succeed in grad school. I have been blessed with wonderful peers, teachers, and mentors. Even as an undergraduate, I received tremendous personal and academic support from USC faculty. I am particularly indebted to the following faculty for fostering my enthusiasm for philosophy and providing me with encouragement: Robin Jeshion, Jake Ross, Scott Soames, Gabriel Uzquiano, and Gary Watson. Some of them even guided me through difficult personal trials that I faced during this time. The quality of support I received did not diminish in grad school. I do not think I would be where I am today without the support I received from Ralph Wedgwood, Steve Finlay, Zoë Johnson King, and Mark Schroeder. It was during Zoë’s job talk that I first got interested in the philosophy of achievement and what it takes to deserve credit for a successful performance. Working with Zoë would prove to be as productive as listening to her excellent talks. She always provides sharp but fair criticism, and she also has a knack for identifying ways in which you’ve sold your own argument short. Ralph was the first person who encouraged me to delve deeply into virtue epistemology and the literature on deviant causal chains. So it’s fair to say that he got the ball rolling on half of my dissertation. His comments on drafts of my chapters were also crucial to filling in their details. At times, I’d get stuck on certain points, and Ralph would see the way forward. Steve was the first person to introduce me to the contemporary literature on moral worth, through a course I took with him in my first semester of grad school. I immediately admired his approach to philosophy. Steve is really good at doing more with less. He finds simple solutions where more complex ones are tempting. For this reason, talking to him about my work has been really useful for seeing possibilities I have neglected. There is one person I am indebted to more than any other for getting to this point: Mark Schroeder. I think it’s fair to say he changed my life. I first really began to dream of going to grad school after taking a metaethics course with him as an undergrad. The course was a blast, and I iii felt myself improving at philosophy with every assignment and reading. But I was still really hesitant to seek out guidance from faculty on the application process. I reached out to Mark (very last minute) and he guided me every step of the way, even as he was busy taking care of a newborn. In graduate school, his guidance was no less crucial. I came into graduate school with a bunch of mistaken ideas about how one gets to the point of doing philosophy professionally. As a result, in my third year, I didn’t know how I was going to write a dissertation. Mark’s guidance was ultimately what enabled me to find a topic and write productively on it. He taught me the importance of starting small and building up to larger ideas. He also made sure I didn’t lose sight of what got me into philosophy in the first place: deep curiosity about certain subjects and questions. And he of course also provided indispensable feedback both on my research and teaching. Other faculty that deserve to be acknowledged include Ed McCann, Jim Van Cleave, Jon Quong, Alexis Wellwood, Jeff Russell, David Wallace, and Kadri Vihvelin. Either through their excellent teaching or through their mentorship, each of them played an important role in my getting to this point. Special thanks are also owed to USC’s office administration staff for their constant assistance. I don’t think I could have kept track of deadlines and successfully printed out handouts for my discussion sections without Natalie Schaad, J.N. Nikolai, Michele Root, Amanda Velasco, and Donna Lugo. Natalie deserves special independent mention for calming my nerves during Hurricane Irma. In maybe my second week teaching, right before I had to lead my discussion section, I learned that Irma would be a category 5 and that my family was not going to leave South Florida. The combined stress felt like it was too much to bear, but Natalie walked me through it and got me to breathe. I got to witness firsthand what an excellent therapist she will make. My time in grad school has been a delight thanks to the wonderful graduate students who have surrounded me. I’d like to start by thanking my cohort: Simon Blessenohl, Jen Foster, Bixin Guo, Quyen Pham, and Vishnu Sridharan. I am so grateful that I had a cohort that chose to build each other up rather than tear each other down. I’m also thankful to Paul Garofalo, Philip Li, Laura Nicoară, Dan Pallies, Christa Peterson, and Douglas Wadle. Each of them provided me with emotional support, great philosophical advice, and friendship. Finally, I should thank Rebecca Carlson, David Clark, Alex Dietz, Noah Gordon, Jasmine Gunkel, Tanya Kostochka, Andrew iv Stewart, Stephanie Van Fossen, and Matthew Wiseman. Each of you has in some way made an impact on either my work or my social life during this time. Finally, I’d like to thank mom, George, Ryann, Lewis, Harrison, Sara, Richard, and Shelby. None of this would be possible without your support, and it has been such a joy and privilege to share a life with you. v Table of Contents Acknowledgments…………………………………………………………………………………i Abstract…………………………………………………………………………………………..vi Introduction……………………………………………………………………………………….1 Chapter 1: Extending Credit………………………………………………………………………5 Chapter 2: Credit for Dummies………………………………………………………………….27 Chapter 3: An Attempt at a General Solution to the Problem of Deviant Causal Chains……….49 Chapter 4: Overcoming Difficulties with Virtue………………………………………………...75 References………………………………………………………………………………………. 92 vi Abstract I try to make progress on certain difficult questions in ethics and epistemology by thinking about the nature of creditworthy performances. 1 Introduction We don’t deserve credit for all of our successes. Equivalently, not all of our successful performances are creditworthy. If I close my eyes, shoot my bow in a random direction, and manage to hit a bullseye, though my shot may have been successful, I don’t deserve credit for the success. To deserve credit for the success, my success must satisfy the non-accidentality condition that is required for a performance to be eligible for praise. That’s because, in the perhaps slightly technical way that philosophers use the term “creditworthy,” being creditworthy for a performance just is a matter of satisfying the non-accidentality condition required for one’s success to be eligible for praise. Unfortunately, my hitting the bullseye is too lucky to be eligible for praise. Therefore, it is not creditworthy. Creditworthiness is important in two different areas of philosophy. On a widely accepted understanding of moral worth, philosophers interested in what it takes for an act to have moral worth are interested in what it takes for a right action to be creditworthy. 1 They want to understand why the right actions of agents like Kant’s shopkeeper, who gives correct change only in order to maintain his profits, are too accidentally successful to be eligible for praise. The notion is also important in epistemology. There are striking parallels between knowledge and creditworthy performances. Just as my successful shot was not creditworthy when I shot in a random direction, my true belief is not knowledge if I form it on the basis of a random guess. Just as my successful shot is not creditworthy if an unfortunate wind derails it, but it is then redirected on target by a fortunate wind, my belief that someone at the bank owns a Ford is not creditworthy if the person I took to own a Ford has, unbeknownst to me, just sold it, but someone else in the bank, unbeknownst to me, owns one. And just as we value creditworthy successes over and above mere success, we value knowledge over and above mere true belief. These parallels have led certain philosophers to propose that knowledge just is creditworthy true belief. It is a core commitment of one of the most popular theories of knowledge: virtue epistemology. 2 Despite the fact that philosophers in both these areas are interested in creditworthiness, creditworthiness has received very little independent theorizing. Philosophers in the moral worth 1 This understanding of moral worth is endorsed by all of the following philosophers: Lord (2018), Way(2017), Singh (2020), Rozeboom (2017), and Sliwa (2015) 2 Ernest Sosa famously motivates virtue epistemology by observing these parallels in Sosa (2003). 2 literature have been narrowly focused on the case of creditworthy right action. And, though virtue epistemologists have been less narrowly focused, their theorizing often overcentralizes the case of knowledge. My aim in this dissertation is to help try to rectify this. On one way of reading this dissertation, the dissertation is an extended argument for a virtue-theoretic account of creditworthy performance, which holds that a successful performance is creditworthy just in case it manifests the virtues associated with the relevant activity. On this reading, the first two chapters argue against the most prominent alternative approaches to creditworthiness, which come from the moral worth literature. The first chapter argues that, whereas these alternatives do not extend to other kinds of creditworthy moral actions, a virtue- theoretic approach can. The second chapter argues that, whereas a virtue-theoretic account does not entail implausible general cognitive conditions on creditworthiness, these alternatives do. The final two chapters then develop the virtue theoretic alternative: the third chapter gives an account of the manifestation relation, and the fourth chapter gives an account of virtue. I do not think this is the best way to read this dissertation. Although I do believe that a virtue-theoretic approach is the right approach to take to creditworthiness, I am also keenly aware that philosophy is hard, and that it will take much more work to adequately defend such a position, work that I hope to continue in the future. On my preferred reading of this dissertation, it is an extended argument for a different claim: that we can make progress on philosophical questions related to creditworthiness by thinking about creditworthiness and related phenomena in their full generality. On this reading, the aim of the first chapter is to show that we gain evidence for views that are more ecumenical to the two leading accounts of moral worth by thinking about creditworthiness for other kinds of moral actions. Philosophers in the moral worth literature have gone back and forth on whether one needs to be motivated by the right-making features of acts or their rightness itself to deserve credit for doing the right thing. 3 I argue that, when we think of credit for acts with other moral properties, the case for the acceptability of either motive becomes much more compelling. 3 Arpaly (2002) and Markovits (2010) famously defend the former. Sliwa (2015) and Johnson King (2018) defend the latter. 3 The aim of the second chapter is to get clear on whether there is a general cognitive condition on creditworthy performance, a condition that agents be aware in some way that what they were doing would produce success. Such conditions have been proposed by various philosophers in order to rule out certain positions in the moral worth literature. Similar cognitive conditions have also been defended in other areas, in ways that parallel the arguments given in the moral worth literature. I argue that, when we consider the full gamut of creditworthy performances, it becomes clear that there is no general cognitive condition on creditworthy performance. Consequently, these conditions cannot be used to rule out certain positions in the moral worth literature. Additionally, this calls into question the validity of the parallel arguments for cognitive conditions in other areas. The third chapter is perhaps where it becomes clearest that this second reading is a better reading of my dissertation than the first. Virtue epistemologists have famously motivated their view by pointing out that their view subsumes the Gettier problem under the more widespread deviant causal chain problem. Once this is done, they have reasoned, they no longer face any problem: the deviant causal chain problem is a problem for metaphysicians, not epistemologists. 4 If the aim of this dissertation were simply to defend a virtue-theoretic approach to creditworthiness, I would make a similar argument: the deviant causal chain problem is a problem for metaphysicians, not philosophers working on creditworthiness, so, by invoking the metaphysical manifestation relation, I need not address it. But this is not the approach I take. Rather than punt on having an answer to the deviant causal chain problem for creditworthy performance, in chapter 3, I try to provide a solution to the deviant causal chain problem as it appears in all areas of philosophy. I then argue that the solution is more promising in all of these areas precisely because I have thought about deviance as a general phenomenon. Providing a complete defense of a general solution to one of philosophy’s hardest problems is of course not something I can do in a single chapter of a dissertation; indeed, I probably could have written a dissertation on deviant causal chains alone. But, as I have said, this is not the best way to read this dissertation. While I do hope to continue to defend this solution in the future, my aim has been instead to show that a much more promising solution arises when we think more 4 Turri (2011) makes this point. Sosa (2015) endorses it. 4 generally. We make progress on a philosophical question related to creditworthiness by thinking about the related phenomenon in its full generality. In the final chapter, I argue that, when we consider the broader class of creditworthy performances, it becomes clear that the traditional virtue epistemological account of virtue is inadequate. The virtues related to a given activity are not mere reliable dispositions to succeed at that activity. I then argue that the account of virtue that emerges from consideration of this broader class of creditworthy performances provides virtue epistemologists with a satisfying solution to the fake barn problem that has been so troubling to virtue epistemological approaches. A further potential bonus of this account is that it may allow virtue epistemologists to extend their account of creditworthy performance to the case of moral worth. 5 Chapter 1: Extending Credit Introduction We can earn credit for a much wider range of valuable actions than merely right actions. Agents who donate to charity, for instance, sometimes deserve credit for doing something supererogatory. And agents who comfort grieving strangers sometimes deserve credit for doing something kind. Philosophers have developed accounts of creditworthiness for doing the right thing, or, as they are sometimes called, accounts of moral worth. But very little has been said about creditworthiness for acts with these other kinds of normative properties. 5 Even less has been said about whether these other properties can be subsumed under a unified account of creditworthy action. In this paper, I will argue that the dominant accounts of creditworthiness for right action can’t be extended to these other domains of normativity. In particular, the two most prevalent kinds of conative requirements for creditworthiness that philosophers have endorsed are unsuited to these other domains. Imposing such requirements on these domains raises both new problems and more serious instances of problems familiar from the literature on creditworthy right action. And the most natural ways of modifying these requirements to generalize to cover the full range of cases don’t succeed. To cover the full range of cases, we will need a pluralist view that allows agents to get credit when they are motivated by either kind of desire. To this end, in section 4, I sketch a view that allows us to get credit for the wider range of positive normative properties; that is, I offer us a way to extend credit. The account I offer isn’t the only possible pluralist view. Indeed, such views have become more popular recently. 6 But it does have an attractive feature: it avoids the pitfalls that often doom pluralist views. In particular, the account is non-disjunctive, and so plausibly gives us insight into the nature of moral worth. 7 And, as I will show, the account avoids troubling sufficiency objections, which notoriously plague more relaxed theories. 5 Smith (1993), Arpaly (2002), Markovits (2010), Johnson King (2018), and Sliwa (2015) all are focused on rightness, for instance. A notable exception is Massoud (2016), who tries to extend Markovits account to explain supererogatory action. 6 See, for instance, Portmore (forthcoming) and Isserow (2020) 7 In contrast with, for instance, Isserow (2020) 6 Section 1: Two Views on Moral Worth To get credit for doing the right thing, it’s not enough to merely do the right thing. Kant’s famously showed this using a (less modern) version of the following case. 8 Shopkeeper: Tim is a shopkeeper. He regularly gives his customers the amount of change they’re owed when they make purchases. But he does so because he wants to ensure he maintains good Yelp reviews, which he knows is crucial for keeping profits up. In Shopkeeper, Tim does the right thing. Nonetheless, Tim doesn’t deserve credit for doing the right thing. His doing the right thing is too accidental to be creditworthy: he wouldn’t have done the right thing if it weren’t profitable. Short reflection on Shopkeeper yields a very natural and compelling diagnosis of what goes wrong with Tim: To be creditworthy for doing the right thing, you must do the right thing for the right reason. Since protecting one’s profits isn’t the right reason, Tim isn’t creditworthy. The reason for which Tim does the right thing, his motivating reason, is not the motivating reason required for creditworthiness. 9 You might be wondering: what motivating reason do we need to substitute in for Tim’s actual motivating reason to make him creditworthy? Kant suggested that acting with moral worth requires being motivated by the fact that the act is right. Following Kant, we might propose that Tim needs to act from a desire to do what’s right in order to get credit for doing the right thing. Tim isn’t creditworthy because he doesn’t act from this desire. This is an intuitive view to adopt in response to Shopkeeper. And it’s also philosophically motivated. As discussed earlier, part of what goes wrong with Tim is that Tim’s motive is too accidentally connected to doing the right thing. In general, you don’t deserve credit for something you do accidentally. 10 So we should expect the right motivating reason to secure a non-accidental connection to doing the right thing. But doing the right thing because it’s the right thing is about as non-accidental a connection as you 8 See Kant (1785). 9 See Alvarez (2016) for an overview on motivating reasons. 10 Kant first suggested a connection between moral worth and non-accidentality. Talking about agents who took joy in doing good, he wrote: “But I assert that in such a case the action, however it may conform to duty and however amiable it is, nevertheless has no true moral worth, but is on the same footing as other inclinations, e.g., the inclination to honor, which, when it fortunately encounters something that in fact serves the common good and is in conformity with duty, and is thus worthy of honor, deserves praise and encouragement, but not esteem” Sliwa (2015), Markovits (2010), and Johnson King (2018) stress the importance of non-accidentality. 7 can get, which makes this view philosophically appealing. I will call this the Kantian view of what motivation is required for creditworthiness. 11 Kantian View: An agent A deserves credit for doing the right thing only if the motivating reason for which the agent does the right thing is the fact that it’s the right thing to do. As obvious as the Kantian view may seem, this view has come under attack in the last thirty years or so. Michael Smith, Nomy Arpaly, Julia Markovits, and others have offered important objections to the view. 12 They have proposed, instead, that, to deserve credit for a right action, agents must be motivated to do the act by the facts that make the act right. I will call this the right- making features view (RMF). RMF View: An agent A deserves credit for doing the right thing only if the motivating reasons for which the agent does the right thing are the facts that make it right. These are the two dominant views of what conative state is required for creditworthiness in the contemporary literature. 13 In addition to a particular conative condition, some philosophers impose very particular cognitive conditions as well. 14 Paulina Sliwa, for instance, thinks that you must know that your act is right to get credit for doing the right thing. But for our purposes, what is important is that most theorists have accepted either the Kantian view or RMF view. My aim in the next section is to argue that these views cannot generalize to cover other areas of ethics. Section 2: The Objections Ethics is about more than right and wrong. We may often be concerned with doing what’s right and avoiding what’s wrong, but there are lots of other ethical properties that interest us. We also sometimes care about doing good acts, courageous acts, merciful acts, supererogatory acts, among others. And we even give people credit for doing courageous, merciful, supererogatory, and good things. Furthermore, there are cases, not unlike Shopkeeper, in which people don’t 11 Kant seems to accept an account like this for moral worth when he suggests acting from joy in doing good lacks moral worth because “the maxim lacks moral content, namely of doing such actions not from inclination but from duty.” 12 See Arpaly (2002), Smith (1994), Markovits (2010). Weatherson (2014) is also critical of this view. 13 Though Ross (1930), for instance, defends an alternative view. Johnson-King (2019) has also come to accept a form of pluralism, which contrasts with her earlier criticisms of RMF motivation. 14 See Hills (2009) and Sliwa (2015) for defenses of particular cognitive components. 8 deserve credit for performing acts like these. If you do something courageous for fame, for instance, you don’t deserve credit for doing the courageous thing. The same goes for if you do the supererogatory to secure a legacy or do the merciful thing out of fear for future retribution. The constraints on creditworthy right action seem to parallel the constraints on creditworthiness for these other actions. Our responses to creditworthy instances of these act types also mirror our responses to creditworthy right action. We often praise agents for their creditworthy supererogatory actions just as we often praise agents for their creditworthy right actions. 15 And we cease to praise agents for their supererogatory actions when we no longer believe they are creditworthy, just as we cease to praise agents for their right actions when we no longer believe they are creditworthy. These two parallels between creditworthy right action and creditworthiness for acts with these other normative properties make it look like creditworthy right action is one instance of a more general kind: creditworthy moral action. Since creditworthy right action is part of this more general phenomenon, we should expect an adequate theory of creditworthiness to be able to cover creditworthiness over a larger range of cases than creditworthy right action. So, if one of the Kantian or RMF view is an adequate theory of creditworthiness, we should expect them to be able to cover these cases. To evaluate whether these theories are adequate more generally, we must first understand what generalized versions of these theories would look like. Plausibly, the way you would generalize the Kantian view is as follows: for each normative property, creditworthiness for performing an act with that property requires that you be motivated to perform the act by its having that property. So, for instance, to be creditworthy for doing something courageous, you must be motivated to do the act by a desire to do something courageous. 16 And generalizing the RMF view works as follows: for each normative property, creditworthiness for performing an act with that property requires being motivated by the facts that make the act have that property. For instance, 15 Creditworthiness and praiseworthiness can importantly come apart: if Tim did the right thing for the right reasons, he would have deserved credit but not praise. 16 Johnson-King endorses something close to this general view in Johnson-King (forthcoming a). She holds that an act’s having a feature has to be part of your motive to be creditworthy for performing an act with that feature. The cases I will present undermine her position. 9 creditworthiness for doing the courageous thing depends on a desire to perform acts that have the properties which make that act courageous. The problem is that these analogues to the Kantian and RMF views for rightness are implausible with respect to other normative properties. The analogues face a variety of counterexamples. I will first argue that the Kantian view fails to generalize. Then, I will argue that the RMF view also fails. Against Generalizing the Kantian View Generalizing the Kantian view looks particularly implausible when we think about cases of the supererogatory. Consider the following two cases: Johnny Offers a Hand: Johnny is a starting investment banker working really tough hours. As a child, Johnny was violently abused. As a result, Johnny decided that if he ever got to a position where he could help abused children, he would spend a good amount of his time helping abused children. Now, as an adult, Johnny feels like he is in such a position. So he spends what free time he has visiting and assisting at shelters for abused children. As a result, he has no free time for himself. But the children are greatly benefited by his service. Kelly Overworks Herself: It’s the 1980s, and Kelly has volunteered at a makeshift hospital for victims of the AIDS epidemic. Kelly, seeing the horrors of the epidemic, feels an urgent need to do as much as she can, so she resolves to work herself to exhaustion. She regularly works 18-hour shifts. Her peers are stunned by her work. They can’t believe how far beyond the call of duty she’s gone. But Kelly thinks nothing of it; she thinks she’s just doing what’s required of her. She hasn’t had time to step back and reflect that she’s going far beyond what’s required. Johnny in Johnny Offers a Hand doesn’t seem to be motivated by the fact that his act is supererogatory. It’s implausible that Johnny is motivated by this fact because Johnny is in a position to do a number of different supererogatory acts. But of these various acts Johnny is in a position to do, Johnny reliably performs acts that are in some way relevant to his personal life. This is the sort of thing that needs explaining, and there’s a very obvious explanation forthcoming: Johnny is motivated to perform these particular acts by his special concern for the abused. In this 10 explanation, a desire to do the supererogatory is inoperative. But then, on the Kantian view, Johnny lacks the proper motivation for being creditworthy. This is the wrong result. Johnny is creditworthy for his supererogatory acts. The problem in Kelly Overworks Herself is much more straightforward. People often go beyond the call of duty without realizing that they are doing so. Kelly, in this case, doesn’t believe she’s going beyond the call of duty. Equivalently, she doesn’t believe what she’s doing is supererogatory. Thus, if belief that P is required for P being your motivating reason, that it’s a supererogatory thing to do can’t be what’s motivating Kelly. And so, the Kantian view predicts Kelly doesn’t deserve credit for doing something supererogatory. But Kelly does deserve credit. This is a problem for generalizing the Kantian view. Indeed, it is the same kind of problem that the Huck Finn case raised for Kantian accounts of creditworthy right action. 17 But, because cases of ignorant creditworthy supererogatory actions are far more familiar than Huck Finn cases, it is much less compelling to deny and explain away the intuitive verdict in these cases. 18 We can raise analogous problems for the attempt to generalize the Kantian view to cover courageousness. Doing so for Kelly Overworks Herself is straightforward: just add details to Kelly’s case that make her act courageous. As in the case of supererogation, she might be so directly moved to act that she doesn’t realize her act is courageous, but she’ll still be creditworthy. It’s not as easy for Johnny Offers a Hand, so consider: Soldier’s Selective Courage: Josh is an American soldier in WW2. Occasionally, some of his fellow officers go overboard in their reprimands of lower ranking subordinates. But Josh, not wanting the wrath of his fellow officers redirected to him, never intervenes. One day, Josh is out on a mission with his subordinates when a grenade is thrown at them. Josh jumps on the grenade, sacrificing his life to save his subordinates. I take it that Josh deserves credit for doing something courageous when he jumps the grenade. But Josh doesn’t seem to be motivated by a desire to do the courageous. If Josh were 17 This problem is from Arpaly (2002). 18 Johnson-King and Sliwa have adopted such strategies for Huck Finn cases. Johnson-King (2018) suggests that the verdict depends on confusing creditworthiness for character traits and creditworthiness for right actions. We feel that Huck deserves credit for something, and this inclines some to conclude he’s creditworthy for right action when he’s creditworthy for a good character trait. Sliwa (2015) also suggests that only Huck’s character is creditworthy. This strategy is unpromising in my cases: we praise people for the kinds of actions I have described regularly in ordinary life, not just their good character. 11 motivated by the fact that it’s a courageous thing to do, we would expect him to stand up to his fellow officers, since that is also a courageous thing to do and the cost to himself is much lower. But Josh, surprisingly, does the courageous thing when the costs are much higher to himself. This suggests that distinct motives with different strengths are at play in the two cases. The best explanation seems to be this. In the first case, Josh’s motive to prevent his subordinates from being berated is outweighed by his motive to avoid the criticism of his peers, presumably because he is not very strongly motivated to prevent his subordinates from being berated. But, in the second situation, a different motivation is at play: his motivation to keep his subordinates alive. His motive to keep his subordinates alive is very strong, so strong that it can even outweigh his own desire to live. Soldier’s Selective Courage is an instance of a well-known phenomenon: extraordinary acts of courage don’t always guarantee that a person will perform more ordinary acts of courage. Even more generally, courage in one area does not always guarantee courage in another area. I am suggesting that sometimes the reason this occurs is because people’s acts of courage are realized by their own special concerns, things that they care about in particular. Furthermore, even when this is the case, these people deserve credit for doing the courageous thing. The Kantian view cannot explain this. I’d like to draw a general lesson from these cases. The problem facing the Kantian view is that the standard way people go about performing creditworthy supererogatory and courageous acts differs from that which you might expect if their desires were as predicted by the Kantian view. People aren’t on the look-out trying to make sure they perform acts that go beyond the call of duty in their daily lives. They don’t have an intrinsic desire to go beyond the call of duty. Instead, they are confronted with opportunities to produce some good in the world. Sometimes, they fail to produce the good. But other times, the opportunity resonates with them in some unique way because of their unique concerns and position in life. The opportunity to help victims of abuse resonates with Johnny because of his past. Kelly feels obligated to help AIDS victims because she sees how AIDS is ravaging her community and recognizes that she is in a position to help. In these situations, people then spontaneously perform the relevant act. It’s difficult for the Kantian view to capture this phenomenon. 12 The supererogatory and courageous are structurally similar: for an act to have either property, the act must impose a great cost on the acting agent. This might tempt you to think that what I have shown is that the Kantian view struggles to accommodate creditworthiness for acts with properties like these. In fact, however, nothing turns on the structure of these normative properties. Other, structurally different normative properties raise distinctive problems of their own. Consider: Jeffrey Gives an Extension: Jeffrey is a landlord who is renting his property out to Jacob. Jacob was fired last month and therefore missed a month of rent. The time to pay rent has come around again, so Jeffrey goes to collect this month’s rent from Jacob. Once again, Jacob doesn’t have this month’s rent; he still hasn’t found a job. The lack of this revenue stream is really putting a financial strain on Jeffrey, so it would make sense for Jeffrey to evict Jacob. But Jeffrey sees how much an eviction would harm Jacob. So he gives him another extension. Jeffrey deserves credit for doing something merciful. But in Jeffrey Gives an Extension, Jeffrey is directly motivated to not evict Jacob by a concern for preventing Jacob from suffering some harm. He is directly motivated to not evict Jacob by his compassion for Jacob. It’s not the act’s status as a merciful act that drives him to do it; it’s his felt need to make sure Jacob doesn’t suffer. But this means that Jeffrey doesn’t have the right motivation to get credit for doing the merciful thing according to the Kantian view. But Jeffrey certainly is creditworthy. Thus, we have a counterexample to the Kantian view. The problem with the Kantian view is that it doesn’t allow us to get credit for being motivated to perform acts of mercy by a desire to prevent harm. But this is one of the ordinary reasons for which people perform acts of mercy. They are inspired by the thought that, by showing mercy, they can help another person avoid suffering. So the Kantian view is not in a position to explain the creditworthiness of ordinary acts of mercy. A General Problem So far, I’ve presented particular problem cases for the Kantian view that arise from the special nature of various normative properties. In closing this part of the discussion, I’d like to present a more widespread, general problem. The problem is this. It’s possible for acts to instantiate multiple normative properties. For instance, an act can be both courageous and morally required. 13 And an act can be supererogatory and merciful. And it seems possible to get credit both for doing the supererogatory and merciful in cases like these. On the Kantian view, though, to get credit for performing an act with the one normative property and also get credit for performing an act with the other normative property, we must have had the following two motivating reasons: that the act had the one normative property and that it had the other. This is certainly possible; our performing the act can be overdetermined by a variety of different motives. But this isn’t always the case, and we still sometimes deserve credit for both. Suppose Sean is a prison guard who is watching over a wrongly imprisoned and cruelly mistreated prisoner. It seems possible for Sean to free the prisoner in full recognition of the fact that what he is doing is courageous (since he risks being imprisoned himself for doing this) while being solely motivated by a desire to get justice for the prisoner. And yet this wouldn’t undermine his creditworthiness for doing a courageous thing and a just thing. There are also cases like this that involve two thick properties. Consider: A Mother Forgives: A mother of a murder victim forgives the murderer. She decides to do this because she wants there to be more forgiveness in the world. Her motivating reason to perform the act is that it’s an act of forgiveness. She decides to meet the murderer face to face and forgive him directly. This is both a courageous act and an act of forgiveness. The mother deserves credit for performing both a courageous and a forgiving act. But, plausibly, she’s only motivated to perform it by its status as an act of forgiveness. It’s noteworthy that this act is also supererogatory, and that the mother deserves credit for performing a supererogatory act even though she doesn’t perform the act because it’s supererogatory. The Kantian view can’t explain how we get credit for all the different properties. This completes my presentation of the objections to the Kantian view. Of course, they are only pressing if the Kantian can’t reply to them. So, in order to evaluate their strength, we must consider how the Kantian can respond to these worries. I now turn to this task. 14 Possible Kantian Replies I assumed that the way the Kantian would generalize her view is by constructing analogues to her view of rightness for each of the different normative properties. But perhaps there is a different way the Kantian can generalize her view. Instead of suggesting that analogous claims hold for all normative properties, the Kantian could hold that all creditworthiness depends on being motivated by certain thin normative properties. In the case of rightness, it might be rightness itself, but it could be different for other normative properties. For each normative property, creditworthiness for performing acts that have that normative property might depend on being motivated by a desire to perform acts with some thin normative property. It’s not hard to see what that normative property might be for some of my cases. A common feature of many of the supererogatory and courageous acts I have considered is that there is weighty moral reason to perform them. And, although I appeal directly to preventing harm when it comes to mercifulness, one might think that, since there’s always weighty moral reason to prevent harm, we can require that this be the motive even in cases of the merciful to perform merciful actions. This kind of move would help rule out ignorance cases. Kelly, in Kelly Overworks Herself, confesses to performing the act because she felt it was required. If performing an act because it’s required is a matter of performing it because there’s decisive moral reason to do it, then perhaps we can interpret Kelly as performing the act because there’s weighty moral reason to do it and thus as being creditworthy. It’s pretty clear, however, that this view won’t get us far enough. Cases of partial concern, such as Johnny Offers a Hand and Soldier’s Selective Courage, are not covered by this new view, since, in those cases, the agents were most plausibly interpreted as having special concern for certain right-making features. This view also doesn’t explain cases where a person performs an act with multiple normative properties but is only motivated to perform the act by its instantiation of one of the properties. The mother who forgives the murderer of her son in order to do something forgiving deserves credit for both performing a forgiving act and performing a courageous act. She is not motivated by the fact that there is weighty moral reason to do it. She’s motivated by the fact 15 that it’s the forgiving act. It’s hard to see how the view explains these cases. So I don’t think this view will work. Perhaps a way of reviving this view involves appealing to a different kind of desire to do things there’s weighty moral reason to do. Instead of appealing to a direct desire to perform acts of this kind, we could instead appeal to a regulatory desire: a desire to perform acts only if there’s weighty moral reason to perform them. Indeed, a move like this has been popularized by Kantians like Philip Stratton-Lake and Paul Benson. 19 In the cases of partial concern, the person is not moved to act directly by the act’s being supported by a weighty moral reason, but perhaps the performance of the act is regulated by a concern of the agent with making sure that he or she only performs acts that are supported by weighty moral reasons. The problem with this move is that this regulatory desire needn’t be present in some of our cases for the agents to be creditworthy. Take Johnny from Johnny Offers a Hand. It seems possible that Johnny would perform the act even if the reasons to perform it weren’t weighty. Johnny might be so strongly motivated to help that he would still assist even if his labor bore no fruit. We can imagine that, in counterfactual circumstances where his banking job became so arduous that he was not able to do anything ultimately at all effective at the shelter, Johnny would still volunteer at the shelter. Maybe his desire to assist is so strong, due to his past experiences, that it would survive even the realization that his efforts weren’t making a difference. The idea of not at least trying would be repulsive to Johnny. If this is how Johnny’s desire to help at the shelter works, then his desires aren’t regulated by a regulatory desire to only perform acts if there’s weighty moral reason to perform them. Given the actual strength of his desire to help, in counterfactual circumstances where his effectiveness is undermined and the weight of his reasons thus negated, he would still help. But this doesn’t undermine Johnny’s actual creditworthiness in the case we presented, in which he’s effective. 20 19 Stratton-Lake (2005) holds that acting with moral worth requires having the desire to perform the act only insofar as it accords with duty. Benson (1987) holds that one’s concern for duty must inhibit action when it conflicts with duty and allow action when it aligns with duty. Marcia Baron also defends a similar view with respect to the moral worth of one’s overall life conduct in Baron (1995). The former two views are subject to problems similar to the one I develop in the next paragraph and in footnote 16. Baron’s view is not applicable to the question of concern here, which is about the moral worth of individual acts. 20 Of course, if he’d work at the shelter even if it were wrong to do so, he wouldn’t deserve credit. But he shouldn’t lose credit because he’d still work at the shelter if working at the shelter were merely permissible. Note, though, that a regulatory desire to do what’s permissible is also inadequate. Conative conditions on moral worth are supposed to 16 Against the RMF View We’ve now seen that the Kantian view faces serious problems when we try to extend the theory to the wider range of normative properties. Since the RMF view was introduced because it seemed to avoid problems facing the Kantian view, one might hope the RMF view extends more seamlessly. Unfortunately, this isn’t the case. 21 There are a few different kinds of problems that the attempt to generalize the RMF view faces. One problem arises from the structure of some of these properties. Properties like courageousness, supererogatoriness, and mercifulness seem to break down into two components. What makes an act have one of these properties is that it has two distinct properties. For instance, plausibly, what makes an act courageous is that there’s sufficient moral reason to do it and doing it is scary. And what makes an act merciful is that it will prevent harm and that you are within your rights to cause the harm. 22 So, if we generalize the RMF view, to get credit for these acts, we’ll have to be motivated by the conjunction of the two components that make the act have the relevant normative property. But this is implausible. The fact that the act is scary doesn’t have to be any part of your motive for you to deserve credit for doing something courageous. And similar points hold for the other properties. Another problem with the RMF view is that, at least for normative properties beyond rightness, we sometimes deserve credit for performing acts with these properties even when we have the Kantian motivation. Consider the following case: Jessie Gives Thanks: It’s Thanksgiving, and the holiday spirit gets Jessie in the mood to do something good. So Jessie decides to donate to a charity. To decide which charity to donate to, she goes on a website that lists charities it’s good to donate to. She uses a random letter generator to generate a letter, and she chooses the first organization she sees whose first letter of its name is that letter. She then donates a large portion of her savings to that charity. illuminate what goes wrong with agents like Kant’s shopkeeper. This conative condition cannot. If Kant’s shopkeeper gives correct change only insofar as it’s permissible, but his primary motive is to get good yelp reviews, he is not creditworthy for doing the right thing. Certainly someone who does something merciful for fame, but only insofar as it’s permissible does not deserve credit for doing something merciful. 21 For different problems the RMF view faces in accommodating supererogatory action in particular, see Gert (2012) and Massoud (2016). Massoud offers a solution to the problem she discusses. 22 The accounts I offer here may not be fully accurate, but they are sufficiently accurate to illustrate the problem. 17 Intuitively, Jessie deserves credit for doing something good. But she isn’t motivated by the facts that make it good: she doesn’t know what those facts are, since she picks a good charity at random. So the RMF view can’t explain why she is creditworthy. This is of course a case of moral testimony. Some have thought there is something credit- undermining about relying on testimony when performing right actions. My case suggests this doesn’t generalize to good acts. At least some of the time, relying on moral testimony to perform good actions does nothing to undermine one’s creditworthiness. You might think we can avoid this problem by denying that Jessie is motivated by a desire to do good and claiming, instead, that she is motivated by a desire to do something charitable or generous. But we can then just regenerate the counterexample using generosity or charitability as our normative properties: Jessie is creditworthy for doing something generous or charitable even though she is motivated by the act’s being generous or charitable. Going beyond particular cases, denying the possibility of being creditworthy when motivated as the Kantian view requires runs the RMF view into a problem with respect to its account of creditworthiness for right action. A big part of what makes this view appealing is that it allows us to get credit for doing the right thing when we are motivated to do the right thing by the desire to do things that are kind, fair, honest, generous, courageous, compassionate, forgiving, and so on. 23 But if we deny that we can get credit for doing the courageous, fair, compassionate, honest, forgiving, etc. when we are motivated by a direct desire to do these things, then it turns out that, when we do the right thing for these reasons, we can’t simultaneously be creditworthy for doing the right thing and the thing with one of these properties. But this seems implausible: I deserve credit for doing the right thing and the fair thing when I do the right thing because it is fair. Indeed, the RMF supporter has no compelling justification for ruling out getting credit for doing the fair because it is fair, the honest because it is honest, etc. The main complaint RMF 23 This quote from Smith (1994) encapsulates this: “good people care non-derivatively about honesty, the weal and woe of their children and friends, the well-being of their fellows, people getting what they deserve, justice, equality, and the like, not just one thing: doing what they believe to be right, where this is read de dicto and not de re.” 18 supporters have had with doing the right thing because it is right is a complaint with having an intrinsic desire to do the right thing in particular. They have argued that there is something fetishistic about having such a desire, and so argue that good people will lack this desire. 24 But they accept that there is nothing fetishistic about an intrinsic desire to do what’s fair, honest, compassionate, etc. That is central to their view. So they don’t have any justification for denying that agents can be creditworthy for doing something fair when they do the fair because it’s fair. These are my core objections to the RMF view. Once again, we must consider replies to the objections I have presented. As we will see, the RMF theorist can reply to the first problem I raised by accepting a particular way of fleshing out the view. Unfortunately, this reply doesn’t help with the other problems I raised. Possible Replies in Defense of the RMF View Recall that the first problem the RMF view faced arises because the RMF view seems committed to holding that we must be partly motivated by properties like scariness, costliness, and our being within our rights to harm someone to get credit for courageous, supererogatory, and merciful acts. The problem is that these properties don’t seem like they are part of our motivating reason for performing these acts when we deserve credit for them. We can revise the RMF view to allow motivating reasons that only partially explain why an act has the relevant normative feature to ground creditworthiness. In Jeffrey Gives an Extension, Jeffrey is motivated by the desire to prevent Jacob from suffering some harm. This is part of what makes the act a merciful act. And so, if the RMF view allows you to get credit for having motivating reasons that partially explain an act’s having the feature it does, Jeffrey will get credit for doing something merciful. One obvious problem with this is that part of the explanation of why, for instance, courageous acts are courageous is that they are scary. But if you are motivated to perform an act that is courageous by the fact that it’s scary, you don’t deserve credit for doing the courageous thing. 25 Similar points apply to the other normative properties. 24 Smith (1994) first raises this objection. Arpaly (2015) registers her support for the fetishism objection. See Lilliehammer (1997) and Olson (2002) for important replies. 25 Of course, strictly speaking, this isn’t an objection to RMF, which states a necessary condition. But defenders of RMF typically take RMF to capture all that is needed conatively to be creditworthy. This objection shows that more is needed conatively. 19 But there’s an easy way around this problem: we adopt Markovits’ Coincident Reasons Thesis. 26 This thesis holds that you deserve credit for an act when the reasons for which you perform the act are the reasons that justify the act. But we must be careful about how we understand justification. We can’t hold that reasons justify an act if they make the act required, since supererogatory, merciful, and even some courageous acts aren’t morally required. We had better instead understand justification as follows: reasons justify an act if they make that act required or worthwhile. Since courageous, supererogatory, and merciful acts are worthwhile because these acts result in things like preventing harm, saving lives, and helping others, one will deserve credit for these acts when these kinds of facts are one’s motivating reasons, so we avoid giving people credit for being motivated by things like scariness and costliness. But we now face a new problem: for some kinds of acts, there can be multiple reasons that justify those acts, but acting for some of those reasons does not earn you credit for acts of those kinds. Showing mercy to Jacob might be a good way for Jeffrey to fulfill a promise to a friend of Jacob, but, if this is the sole reason for which Jeffrey does the merciful act, he doesn’t deserve credit for doing something merciful. This does not mean, of course, that Jeffrey isn’t creditworthy at all; he certainly deserves credit for his act of honesty. He just doesn’t deserve credit for doing something merciful. The way around this problem is to require both that the reasons for which the agent act justify the act and partially explain why it has the relevant moral property. In the last case I presented, the agent’s motivating reason does not partially explain why the act is merciful, so, on this new account, the agent doesn’t deserve credit for doing something merciful. That this objection pushes us towards this view is an interesting result. Some ways philosophers have developed RMF state the necessary condition in terms of motivation by facts that justify the act, and other ways state it in terms of motivation by facts that explain the act’s having the relevant property. Adequately responding to my objection requires conjoining these approaches. Unfortunately, this view doesn’t help us with the other problem I raised: there are cases where you get credit for the Kantian motivation. I can deserve credit for doing the fair because it is fair, the honest because it is honest, etc. And I can deserve credit for doing the right thing and 26 Markovits (2010) 20 the fair thing when I do the right thing because it is fair. What’s more, the RMF supporter has no compelling reason to deny this beyond saving their view. Far from it: RMF supporters have celebrated agents who are intrinsically concerned about things like fairness, compassion, honesty, etc. So why not praise them for doing acts with these properties when these acts issue from these praiseworthy concerns? Section 3: How to Extend Credit The views we’ve looked at faced problems when we tried to generalize them beyond rightness. This shouldn’t surprise us, since these views were constructed only to explain creditworthiness for right acts. If you construct your theory by considering only the more particular case, you are susceptible to problems from the wider range of cases. So I suggest, in coming up with a new view, that we look to the wider range of cases to construct our view. What I have in mind are the instances of courage, mercy and other virtuous acts that we’ve observed. We’ve said that there’s a connection between deserving credit and non-accidentality. The proposed accounts of non-accidentality were mistaken, but certainly non-accidentality is important. So we should look for an account of non-accidentality in these cases that explains why the relevant agents are creditworthy. I think it’s helpful to look at Soldier’s Selective Courage as a start. It’s pretty obvious why Josh’s courageous action is non-accidental: it’s non-accidental because he performs the courageous act because of his courage. Moreover, this seems like an independently plausible account of when people deserve credit for courageous actions: people deserve credit for courageous actions when their courageous actions manifest their courage. And the same goes for the other virtues. This suggests the following account of creditworthiness for acts that instantiate a virtue: Creditworthiness Virtue: An agent A deserves credit for performing an act that has virtue V iff the agent’s doing the V act manifests her possession of some degree of the virtue associated with acts that are V. In this paper, I will be officially agnostic about what the manifestation relation is, relying on intuitions about when something does and does not manifest the relevant virtue. This should 21 not be controversial. As John Turri has pointed out, the manifestation relation tracks an intuitive metaphysical distinction that we can leave it to metaphysicians to analyze. 27 I will also be agnostic on what the association relation is. 28 Once again, I will merely rely on intuitions about what the relevant associated virtues are and what their content is. And, though Creditworthiness Virtue implies that there are unique virtues for each act type, this is an assumption rather than a commitment. Though this assumption and appeal to intuitions are shortcomings, I hope they do not undermine my central aim, which is simply to illustrate how this kind of view offers a promising way to extend credit. 29 Still, I ought to acknowledge a famous objection to virtue theoretic accounts: one can perform virtuous acts out of character and deserve credit for so acting. 30 This objection is only pressing if we are understanding virtues as robust or stable character traits that ensure the agent performs virtuous acts across many nearby possible worlds. If we understand virtues instead as dispositions or the grounds of dispositions, one can manifest some degree of virtue even when one acts out of character. It is a familiar point from the metaphysics of dispositions that one can have a disposition even while regularly failing to manifest the disposition. Dispositions can be masked, including by other, competing dispositions of the object. 31 And of course it will follow from this that one can have the properties that ground the disposition without regularly manifesting the disposition. As we will see, I understand virtues as certain grounds of a disposition to do the relevant act kind, so my view avoids this problem. With this worry out of the way, we can now see how my account gets the right results in my cases. It’s not hard to see how, on my account, both Kantian and RMF required motivations would be creditworthy in the case of courageous acts. The virtue associated with courageous acts is courage. Just as you can perform a courageous act, you can have courage. And the following is a plausible gloss on courage as a character trait: courage is a disposition to overcome your fears in pursuit of morally worthy ends. 32 Since the performance of courageous acts is a morally worthy 27 See Turri (2011) 28 [Footnote removed for anonymous review] 29 In a way, by assuming uniqueness, I make my task harder, since, if there are more virtues associated with each kind of act, there are more ways for the agents in the cases I developed earlier to be creditworthy. 30 This objection has been pressed by Herman (1981), Markovits (2010), Smith (1991), and Isserow (2021). 31 See Ashwell (2010) and Clarke (2010) 32 Note that this disposition itself grounds a disposition to do courageous things. 22 end, you can manifest this disposition by overcoming fear in pursuit of this end, and thereby get credit. That explains why you get credit for the Kantian-required desire. Since the normatively relevant features that help make acts courageous are also morally worthy ends, you can get credit for those as well. That explains why you get credit for the RMF-required desires when it comes to courage. The question, of course, is how we extend this to properties like supererogatoriness and rightness. I think the first step is generalizing Creditworthiness Virtue in the following way: Creditworthiness General: An agent A deserves credit for performing an act with normative property F iff the agent’s doing the F act manifests his possession of some degree of the virtue associated with acts that are F. The question to answer, then, is what the virtue associated with rightness and supererogatoriness are. It’s of course not as immediately obvious what they are, but I think, on reflection, it’s not difficult to identify them. With respect to rightness, it’s pretty straightforward: the virtue associated with rightness is just overall moral virtue. The question, then, is how either kind of motive can be creditworthy. Once again, this isn’t too difficult to see. Both the possession of RMF desires and possession of the Kantian desire are partly constitutive of moral virtue. After all, fully virtuous people care both about the things that make acts right and doing the right thing. 33 So all of these desires are parts of the virtue associated with right action. You can therefore manifest some degree of moral virtue by acting from any of these desires. The reason acting from one of these desires is necessary is that these are the only desires morally virtuous people have qua morally virtuous people, so these are the only desires constitutive of moral virtue. Desiring fame, for instance, is no part of being morally virtuous. So you can only manifest your moral virtue by acting from one of these desires. Supererogatoriness is less straightforward, but I think, once again, it isn’t too difficult to see what virtue is associated with the supererogatory. Since moral saints are the people most often associated with supererogatory acts, we should look to moral saints to figure out what kind of 33 This claim is contentious in the literature because of the moral fetishism objection. I find the arguments too unpersuasive to give up the compelling pre-theoretic intuition that one of the things virtuous people care about is doing the right thing. Johnson-King (forthcoming b) canvasses the reasons the objection is unpresuasive. Olson (2002) lends support to this intuition. 23 disposition is associated with supererogatoriness. The unifying trait of moral saints seems to be that their desire to realize morally worthy ends strongly outweighs their desire to avoid costs to themselves. 34 This suggests the following: your performance of a supererogatory act is creditworthy when it manifests your disposition to choose doing morally worthy things over avoiding costs to oneself. Both the Kantian motivation and the RMF motivations are acceptable for getting credit because going beyond what is required is a morally worthy end and the normatively relevant things that make some act an instance of going beyond what is required are morally worthy ends. One of these ends is required because, if something else is your end, you haven’t manifested your disposition to choose doing morally worthy things over costs to yourself; you’ve manifested your disposition to choose something else (fame or approval, perhaps) over costs to yourself. On my account, you deserve credit for doing the right thing if your right action manifests your possession of some degree of moral virtue. Since the manifestation relation is factive, this means that you must possess some degree of moral virtue to earn credit for doing the right thing. And since moral virtue is constituted by both the RMF-required desires and the Kantian-required desire, your act’s manifesting your moral virtue depends on your having an adequate collection of these desires. This, you might think, raises a worry for my view: if I must have a collection of desires in order to be creditworthy, my view is in fact stronger than both the Kantian and RMF view, so I face the same problems these views ran into because of my counterexamples. This worry presupposes that my account needs all of the desires to be active when I do the right thing for me to manifest my moral virtue. But nothing forces us to accept this presupposition: we can instead allow that these desires be idle and serve as background conditions on your manifesting your moral virtue. Just as, when you see the façade of a building, you count as seeing the building and not merely the façade so long as the background condition that there’s a whole building present is met 35 , you count as manifesting your moral virtue so long as you have enough other moral desires sitting idle in the background to count as possessing some degree of moral virtue. 34 This is continuous with the account in Wolf (1982). 35 See Husserl (1913). 24 You might worry that what holds for vision doesn’t hold for the manifestation relation, but in fact this seems to be right in line with how the manifestation relation intuitively works. When Allen Iverson crossed over Michael Jordan, he manifested not merely his ability to perform that particular crossover; he also manifested his general ballhandling skills. But, to have general ball handling skills, it’s not enough to be able to perform that very crossover: you must be able to perform other crossovers and ball movements. He counts as manifesting his ballhandling skills because he possesses these other abilities. Even though they’re idle in this situation, they are necessary background conditions for his manifesting the more general skill. Far from raising a worry, the fact that, on my account, the other desires must be present as background conditions is a virtue of my view. My view, as originally presented, avoided the objections to the dominant views because my view is more permissive when it comes to what conative components are necessary for being creditworthy. Whereas the original accounts allowed agents to get credit for only one kind of motive, my account allowed either of these kinds of motive to be creditworthy. One concern for views that are disjunctive in this way is that they avoid necessity problems by taking on sufficiency problems. In effect, by being less restrictive, they become too permissive. But we’ve now seen that the structure of the view I accepted to avoid the necessity problems for the RMF view and Kantian view commits me to something more than the simple disjunctive view: agents must have enough idle moral desires in the background to count as having some degree of virtue. This means that my view is more restrictive than the simple disjunctive view. And it turns out that this consequence of the structure of my view allows me to accommodate the most alarming forms of certain sufficiency objections to the RMF view and Kantian view. One such worry for the RMF view was raised by Sliwa. Sliwa worried that, if we deserve credit for merely acting on an RMF-required desire, then our performing the right action would be objectionably lucky. 36 If, for instance, I am rightly honest with someone because I care about the honest but am completely indifferent to doing the right thing, then I will do the honest thing even when it’s very bad to do the honest thing. And so, my having done the right thing looks completely 36 See Sliwa (2015). 25 lucky. Following Nathan Howard 37 , I will call this sufficiency luck. Sufficiency luck, as Sliwa argues, seems objectionable. My view avoids some of the most objectionable instances of sufficiency luck because it requires that other desires be idle in the background for you to be creditworthy. It’s not enough to just care about honesty to manifest goodness; you must have other desires as well, like other RMF- required desires and the Kantian desire. Given this, merely having the one desire to do the honest thing will not be enough to be creditworthy. Of course, my view still makes some luck possible. Agents who have enough of the RMF- required desires to count as being to some extent morally virtuous will still on occasion do the act that promotes one of their RMF-required desires when it is wrong to do the act. But my view prohibits agents from being creditworthy for the most disagreeable instances of sufficiency luck, and it eliminates systematic sufficiency luck. This same feature of my view allows me to avoid one of the most pressing objections to the Kantian view as well. As discussed previously, RMF supporters have argued that acting from a desire to do the right thing is too fetishistic to be creditworthy. They have justified this claim by appealing to cases where this seems true, and conclude that, in general, this motivation is not adequate to be creditworthy. There are many different variations on the fetishism objection. I believe my view can help with all of them, but, for lack of space, I will focus on one variation: David Shoemaker’s. Shoemaker worries that, if agents are solely concerned with doing the right thing, they will be “bloodlessly calculating morality machines.” 38 They will be totally preoccupied with doing the right thing at the expense of concern for the needs and rights of others. When someone’s rights and needs conflict with doing what’s right, they will be indifferent to those rights and needs. This does seem like a fetish, and it’s implausible that agents who are psychologically like this deserve credit for doing the right thing. So Shoemaker is right that the conative component that the Kantian requires is not sufficient for being creditworthy: if, conatively, acting from a desire to do the right 37 See Howard (2019). 38 See Shoemaker (2007). 26 thing were enough to be creditworthy, these bloodlessly calculating morality machines would be creditworthy. 39 But it falls out of my view, of course, that more is needed conatively for agents to be creditworthy for doing the right thing. It’s not enough that an agent act from a motive to do the right thing; agents must also have, idle in the background, lots of other moral concerns, concerns for the kinds of features that make acts right. Plausibly, this includes concern for things like the rights and needs of others. So my view prohibits bloodlessly calculating morality machines from being creditworthy. Conclusion The Kantian and RMF theorists started on the right track, correctly identifying motives that could have made Tim in Shopkeeper creditworthy. But they incorrectly held that those motives were necessary for creditworthiness. The debate has often assumed that we have to have it one way or the other: either we need RMF-required motives or we need the Kantian-required motive. Another option was always available: either motive is creditworthy, provided the right background conditions are met. This middle ground view was neglected because it was easy to convince oneself that the motive one had identified as necessary was enough to explain the full range of cases. This was easy because the common understanding of that range restricted the range to right actions. Given this and some reason for disfavoring the other kind of motive, like fetishism or accidentality, one could swallow denying the possibility of creditworthiness for the other kind of motive. But once we see that the range of cases is much larger and just how natural it is to give credit for either motive in many of these cases, this denial becomes unacceptable. Acting generously for the sake of being generous is creditworthy, and so is acting courageously in order to save lives. The only remaining impediment, after we appreciate this, to accepting the middle ground view is finding an account of creditworthiness that could explain its truth. I have offered such an account. We should seek the middle path. 39 Johnson-King (forthcoming b) points out that the RMF view runs into this problem as well. 27 Chapter 2: Credit for Dummies Readers of Calculus for Dummies, Quantum Physics for Dummies and Biochemistry for Dummies are not dummies in the colloquial sense of the word. What makes them dummies, in the technical sense of the For Dummies series, is that they do not yet know what they are doing. A popular group of views holds that dummies in this sense do not deserve credit for their successful performances. My aim in this paper is to undermine these views. As I will argue, dummies can deserve credit for their successful performances. It is easiest to see why people would resist giving dummies credit by considering a performance that is not credit-deserving. Suppose I am playing chess, and I move my knight so that it simultaneously threatens two of my opponent’s pieces, forcing him to decide which piece to save. This situation is called a fork. My opponent, recognizing the fork, might compliment the sequence of moves leading to it. Now suppose, in reply, I confess that I did not even realize the sequence would result in a fork. My opponent’s appraisal of the fork will change after this confession. Before my confession, he gave me credit for the move, which he thought I deserved. After, he will stop giving me credit. Somehow, my confession changes his mind about whether, in the terminology that has come to be accepted in epistemology and ethics, I was creditworthy. 40 What philosophers are concerned with in the first place when they talk of creditworthy performance is a kind of non-accidentality that we think is often required for performances to merit praise and be achievements. 41 My opponent changes his appraisal of my fork because my confession reveals that the act was too accidental to be the sort of thing that merits praise, which explains why he does not deserve credit. This use of “deserving credit” may be, to a certain extent, stipulative. After all, we sometimes credit people for things they do accidentally. 42 But it does not 40 Ernest Sosa, John Greco, and Wayne Riggs characterize the difference in these kinds of cases in terms of a difference in whether the agent deserves credit. See Ernest Sosa, “The Place of Truth in Epistemology,” Michael DePaul and Linda Zagzebski, eds., Intellectual Virtue: Perspectives from Ethics and Epistemology (Oxford: Oxford University Press, 2003), pp. 155-81, John Greco, “Knowledge as Credit for True Belief,” Michael DePaul and Linda Zagzebski, eds., Intellectual Virtue: Perspectives from Ethics and Epistemology (Oxford: Oxford University Press, 2003), pp. 111-34, and Wayne Riggs, “Two Problems of Easy Credit,” Synthese, CLXIX, 1 (2009): 201–16. 41 My use of ‘creditworthiness’ deviates from some philosophers’ use of the term as a synonym for praiseworthiness. I believe you can be creditworthy without being praiseworthy: routine, easy performances can be creditworthy because the success meets the right non-accidentality condition but fail to be praiseworthy for being routine and easy. I think this is true of many everyday right actions done for the right reasons. 42 This might explain in part why Sosa has preferred to describe the relevant relation as “aptness.” 28 stray very far from ordinary use, and it picks out a relation that shapes our appraisal of performances, as my case nicely illustrates. But what does it take for a performance to be non-accidental in the way required for creditworthiness? One might think, reflecting on this case, that, for a performance to be non- accidental, the agent must have been aware in some way of what he was doing. It is because I was unaware my move would result in the fork that I forked the piece accidentally and consequently do not deserve credit. This makes it plausible, in turn, that some kind of cognitive awareness of how you are producing the performance is required to be creditworthy for that performance. On this view, non-accidentality is not a primitive condition on creditworthiness; rather, creditworthiness entails non-accidentality because it imposes a more basic requirement of cognitive awareness, and cognitive awareness rules out the possibility of an accident. The claim that there is some requirement of cognitive awareness of what you are doing on creditworthy performance is the focus of this paper. I will call such requirements cognitive requirements or cognitive conditions on creditworthy performance. Since there are several different cognitive states that might be able to rule out giving me credit in the chess case, there are several candidates for what this condition could be. I will argue that no such cognitive condition on creditworthy performance is true. The argument is simple: all such conditions rule out crediting dummies, but dummies sometimes deserve credit, so all such conditions are false. I will begin by presenting a simple, straightforward cognitive condition and present cases that undermine this condition. I will then present more sophisticated conditions that philosophers have defended and offer variations on my cases that undermine these conditions as well. The falsity of cognitive conditions has serious implications in the philosophy of action and ethics. In the philosophy of action, several philosophers have argued for cognitive conditions on important action concepts by pointing out that those concepts have certain anti-luck conditions and that cognitive conditions rule out the relevant kinds of luck. But creditworthy performance has the same anti-luck conditions, and it does not have a cognitive condition, so these arguments are unsound. In ethics, arguments similar to those made in the philosophy of action have been used to motivate cognitive conditions on acting with moral worth. These arguments are also unsound because of the falsity of cognitive conditions. A more startling implication for moral worth is that the two most popular accounts of moral worth are at best extensionally correct: they do not tell us 29 what it is for an act to be morally worthy. After undermining the cognitive conditions in section 1, I explore these implications in section 2. Section 1: Against Cognitive Conditions First Pass: Knowledge A natural explanation for why I do not deserve credit for forking my opponent’s piece is that I did not know that what I was doing constituted forking my opponent’s piece. My lack of knowledge is what renders the act an accident. This suggests the following cognitive condition: Knowledge: An agent deserves credit for a performance only if he knows that what he is doing constitutes that performance. Knowledge counts as a cognitive condition because it holds that a cognitive state, knowledge, is a necessary condition on creditworthy performance. And Knowledge is an appealing cognitive condition because, intuitively, if you know what you are doing, then you do not do it by accident. I would not have forked the piece accidentally if I had known what I was doing. So knowledge is a state that plausibly rules out the possibility of agents deserving credit for accidental performances. Knowledge entails a non-accidentality requirement on creditworthy performance, which is exactly what we were looking for. Unfortunately, Knowledge is false, as we can see by considering the following cases: Pitching: Pat is a quarterback. His quarterback training has given him tremendous arm-strength for throwing footballs. One day, he sees guys pitching a baseball and decides to give it a try. He throws the ball as hard as he can. It turns out, to the shock of his peers, and unbeknownst to him, he throws the ball 100 miles per hour (MPH). Lifting: Eddie has grown up on his dad’s farm. Years of toiling away at farm chores has made him develop significant strength. One day, Eddie goes off to college, where he makes some friends. He sees his friends one day at the gym, where he has gone to run on the treadmill. His friends have some amount of weight that they are attempting to lift unsuccessfully. They jokingly ask Eddie to try lifting the weight. It turns out the amount of weight on the bar is 1103 pounds. His friends are 30 jokingly trying to set the deadlifting world record. Eddie goes up to the weight and deadlifts it successfully, completely unaware of how much weight he is lifting and that lifting this bar would constitute a world record. I have chosen two athletic examples, but it is important to note that nothing turns on this. We can construct analogous non-athletic examples. A singer who casually challenges herself to sing at higher and higher pitches might come to break the world record for highest musical note sung by a human without realizing it. And a starting counselor who has developed a lot of patience through raising his pesky siblings might deserve credit for spending the most time counseling a particularly recalcitrant criminal without quitting, even if he does not realize he has successfully spent the most time counseling this criminal. I focus on the athletic cases just because the relevant activities are more familiar to me. In both these cases, the agents are creditworthy for what they do. Their success does not seem accidental, since they exercised tremendous ability. On the contrary, Pat’s throwing a 100 MPH fastball and Eddie’s setting the deadlifting world record are enormous accomplishments, the kinds of things for which they are clearly creditworthy. But in neither case did they know that their actions constituted doing the thing for which they deserve credit. Weaker Conditions You might find this result unsurprising. Propositional knowledge looks like a very strong requirement, and you can probably think of much less extravagant cases that undermine this condition. What is interesting about my cases is that they undermine weaker conditions as well. We can see this by considering an account that weakens the cognitive state required to mere justified true belief. Gwen Bradford has defended a view in this spirit in the context of achievement. 43 She thinks that, for any achievement, the achievement is comprised of a process and a product, where the process competently causes the product. Competently causing the product is a matter of having justified true beliefs about how it produces the product. We can capture this using the language of creditworthiness rather than achievement as follows: 43 Gwen Bradford, Achievement (Oxford: Oxford University Press, 2015). 31 JTB: An agent is creditworthy for a performance A only if the agent has justified true beliefs about how the process by which they come to perform A results in their performing A. 44 It is not hard to see that my cases undermine this condition. To satisfy JTB, Pat would have to have some awareness of how what he is doing results in his throwing a 100 MPH fastball. And Eddie has to have some awareness of how what he is doing results in his breaking the world record. But neither of them has this kind of awareness because neither has any idea he is about to perform the relevant act. So JTB is not satisfied, even though each agent is creditworthy. Both Knowledge and JTB require that the agent be aware in some way of how what they are doing results in the performance that is creditworthy. JTB just makes the cognitive state weaker. But JTB still requires that agents be in a pretty demanding state, so one way to try to avoid these problems is to go even weaker on the cognitive state. Zoë Johnson King has defended a particularly weak condition. On her view, to be creditworthy for an action, you just need to have some idea that what you are doing constitutes that action. 45 Here is the view: Some Idea: An agent is creditworthy for a performance only if the agent has some idea that what they are doing constitutes that performance. Some Idea is undermined by my original cases. In Pitching, Pat has no idea what he is doing constitutes throwing a 100 MPH fast ball, and, in Lifting, Eddie has no idea he is about to set a world record, but they are still creditworthy for their actions. 46 44 JTB is not quite Bradford’s view, though it is how she initially characterizes her view. JTB clearly does not allow agents to get credit if they do not know the end product of their action. Bradford’s actual cognitive condition can be satisfied by agents who do not have any idea what the end product of their action is. But she notes that this is “almost impossible” on her view because it would require a very rich and detailed set of beliefs about the intermediate causal steps leading to the product. The agents in my examples do not have richly detailed beliefs about the intermediate steps, so my cases still pose problems for this view. I present the view as Bradford does in her first pass characterization of it for ease of exposition. 45 Zoë Johnson King, “Accidentally Doing the Right Thing.” Philosophy and Phenomenological Research, C, 1 (2020): 186-206. 46 David Faraci defends an account on which one deserves credit only if one is aware that one is following a set of norms. This account is also undermined by cases like mine. If Eddie is unaware that there is a set of norms prescribing lifting in the way he lifts, as there may be, he does not satisfy the condition but is still creditworthy. See David Faraci, “On Accomplishment,” Unpublished. 32 Knowledge How You might think that the reason all these accounts are unsuccessful is because the cognitive states they require make creditworthy performance too intellectually demanding. This is exactly what Errol Lord thinks. 47 Lord suggests, alternatively, that what is needed is knowledge how. This suggests the following cognitive condition: Knowledge-How: An agent deserves credit for a performance only if his action manifests his knowledge how to execute the performance. Knowledge-How is an attractive condition. Even if, as Williamson and Stanley think, knowledge how is just a particular form of knowledge that, it is a less cognitively demanding propositional state. 48 I can know how to do something without being able to articulate very precisely what I am doing. For this reason, it is plausible that Eddie manifests his knowledge how to deadlift even if he is unaware that he is deadlifting, and Pat manifests his knowledge how to throw a 100 MPH fastball. Knowledge-how views seem like precisely the right candidate to explain why these agents deserve credit for their performances. Unfortunately, there are slight modifications on my cases that undermine this condition. Here is one such modified case: Modified Pitching: While Pat has the arm strength to throw a 100 MPH fastball because of his past in football, he cannot yet throw it because he lacks the technique. The differences between throwing a football and a baseball are such that he cannot yet throw the baseball with full velocity. Pat is trying to figure out how to throw the ball as fast as he possibly can. So he is experimenting with various techniques one by one. Eventually, he stumbles on one that allows him to throw the ball 100 MPH, and so, for the first time, using this randomly selected technique, he throws the ball 100 MPH. 47 Errol Lord “On the Intellectual Conditions for Responsibility: Acting for the Right Reasons, Conceptualization, and Credit,” Philosophy and Phenomenological Research, XCV, 2 (2017): 436-64. 48 Jason Stanley & Timothy Williamson, “Knowing How,” The Journal of Philosophy, XCVIII, 8 (2001): 411-44. 33 The first time Pat throws the ball 100 MPH, he does not know how to throw the ball 100 MPH. His first time throwing the ball 100 MPH might be the moment at which he learns how to throw the ball 100 MPH, but he does not yet know how to throw the ball 100 MPH when he throws it. If this is the case, his act cannot manifest his knowledge how to throw the ball 100 MPH. For an act to manifest your knowhow, it has to be, in some sense, the product of your knowhow; your knowledge how to do the act has to explain in part why you successfully performed the act. But this cannot be the case if Pat does not already know how to throw the ball 100 MPH when he throws it. Thus, the pitch cannot manifest his knowledge how to throw the ball 100 MPH. And yet Pat still seems creditworthy for throwing the ball 100 MPH. This undermines Knowledge-How. Replies This completes my argument, but it is important to address responses that might make cognitive conditions look more plausible than I have suggested. There seem to be two strategies for replying to my argument. One is to deny that the agents I have described are creditworthy. The other is to try to finesse the details of the views I have rejected to accommodate the cases. The first strategy seems unpromising. The primary marker for creditworthiness, as I have said, is that the agent’s act satisfies the non-accidentality condition that is required for praiseworthiness. Neither Eddie’s success nor Pat’s seems like a fluke; their success is the result of tremendous ability. Their performances are significant accomplishments, and their ignorance is not the sort of thing that would disincline people to praise the acts. Eddie, for instance, would be celebrated by the powerlifting community as much as more typical world record breakers. 49 49 A way to defend this strategy is to claim that what Eddie deserves credit for is lifting this weight rather than setting a world record. There are two problems with this. To start, this conflicts sharply with ordinary practices of assigning praise: people would readily praise Eddie for setting the deadlifting world record, and no one would think this unfitting because he does not deserve credit for it. Additionally, we can construct counterexamples even to this proposal. We can imagine a case where Eddie has no idea what he is doing constitutes lifting this weight. Imagine Eddie actually is in on the joke. He knows the amount of weight on the bar adds up to a world record. What he is unaware of is that he has exceptional strength. He may think he is strong, but he might have no reason to suspect he is that strong, so he firmly believes he cannot lift it. So when he goes up to try to deadlift the weight, he goes up trying to lift it jokingly. To his complete shock, he lifts the weight. He might even say: I had no idea I could lift this weight. But, if he has no idea he could lift this weight, he also had no idea what he was doing constituted lifting this weight. And yet he still seems creditworthy for lifting this weight, not merely for attempting to lift it. Thanks to an anonymous referee for raising this worry. 34 So the strategy of finessing the views seems more promising. I will consider attempts to finesse Some Idea and Knowledge-How, and then draw a general lesson about why this strategy is unpromising. You might think that Some Idea is not really undermined by my cases because the relevant agents plausibly have some idea that what they are doing constitutes throwing a 100 MPH fastball or breaking a deadlifting world record. After all, they may have some very small credence that what they are doing could result in a world record or 100 MPH fastball. In that case, Some Idea is satisfied because having some idea of what you are doing requires very little. The problem with this line of response is that, once we make the condition this weak, it is not clear why we should accept this condition. Cognitive conditions are supposed to rule out agents meriting credit for acts they perform accidentally. Johnson King, for instance, introduces Some Idea to rule out the possibility of an agent deserving credit for finding buried treasure when their aim was merely to dig a hole. But, on this new way of understanding what it is to have some idea of what you are doing, the agent who digs up the buried treasure may have some idea of what he is doing; he may have some really, really small credence that what he is doing constitutes digging up buried treasure. So, if Some Idea is interpreted as merely requiring some really small credence, it does not explain the kinds of cases that made it look appealing in the first place. Once we weaken the condition in this way, there is not much pressure to accept it because it is not very explanatory any longer. One kind of reply on behalf of Knowledge-How runs into problems for similar reasons. Modified Pitching raised problems for Knowledge-How because, in Modified Pitching, Pat does not yet know how to throw a baseball 100 MPH when he throws it at this speed for the first time. He is experimenting with different techniques in the hopes of finding the one that will work. One way of responding is to suggest that his first 100 MPH pitch is the pitch through which he learns how to throw a ball 100 MPH. As such, this pitch can manifest knowhow. In ordinary cases, manifestation of knowledge-how might require that the knowledge-how explain your act, but maybe another way an act can manifest knowhow is by being the act through which you learn how to perform the act. The problem with this move is that, like with the reply on behalf of Some Idea, it undermines the motivation for accepting this cognitive condition. The main motivation behind 35 accepting a cognitive requirement of this sort is its ability to explain the contrast between cases where our doing something is a lucky accident and cases where it is genuinely creditworthy. But now this cognitive condition cannot explain what separates some cases of creditworthy action from lucky accidents, because both acts meet the condition in some cases. Consider the following case: Larissa: John spontaneously decides to travel down certain paths that lead out of his city, having no idea where they lead. It turns out, though, that the first path leads to Larissa. He, in taking that path, learns how to reach Larissa. On the new interpretation of the cognitive condition, the condition does not rule out John’s being creditworthy for making his way to Larissa. Since his making his way to Larissa for the first time results in his learning how to get there, his getting there, on this way of replying, manifests his knowledge how to get there, so he satisfies Knowledge-How. But this seems wrong: John’s reaching Larissa is not any sort of accomplishment. Though on future occasions when John travels to Larissa he will be creditworthy, this first time he gets there he is not. It is a total accident that he gets there. The revision to what it takes to manifest knowhow makes it so that Knowledge-How can no longer explain why, in cases like this, agents are not creditworthy. But that was the whole point of proposing such a condition in the first place. It was supposed to be an anti-luck condition that guaranteed agents like this do not deserve credit. Cases like these also undermine a natural way of combining Some Idea and Knowledge- How. You might think that the problem with Some Idea is that it is an ordinary propositional state, and the problem with Knowledge-How is that knowledge is too demanding. To avoid the problems facing both of these views, one could propose that creditworthiness requires that one manifest one’s possession of some idea how to perform an act. But modifying the aforementioned case undermines this condition. If John knows that the road to Larissa extends out of the north of his city, but he chooses randomly among many different possible roads, he does not deserve credit for making his way to Larissa even though he manifests his possession of some idea how to make his way to Larissa. I have considered ways of finessing particular views that I gave, but the problems these views faced illustrate why no cognitive condition will succeed in accommodating my cases. Eddie is cognitively no different from agents like the buried treasure finder. Since they are no different 36 cognitively, no cognitive condition will permit Eddie to be creditworthy and forbid the treasure finder from being creditworthy. But Eddie is creditworthy and the treasure finder is not, so no cognitive condition is true. Section 2: Implications I have argued that cognitive conditions are false. This result has direct implications for the philosophy of achievement, creditworthiness, and praiseworthiness. But it also has broader implications for the philosophy of action and debates about what it takes for acts to have moral worth. I will now explore those implications. The Cognitive Anti-Luck Strategy. In the philosophy of action, figuring out what it takes for a performance to count as non- accidental takes up central importance. This is because important concepts in this domain have non-accidentality conditions. Acting intentionally, acting for a reason, and achievement are all examples of central concepts in this domain that have non-accidentality conditions. You cannot act intentionally if you performed the act accidentally. If you accidentally went to the store, there is no reason for which you went to the store. Things you do accidentally are not achievements. To rule out accidentality for these and other concepts, a popular argumentative strategy has emerged: Notice that some concept to be analyzed has some anti-luck condition built into it. Affirm that some cognitive state rules out paradigmatic instances of the relevant kind of luck. Conclude that the notion to be analyzed requires that cognitive state. Call this the cognitive anti-luck strategy. If there is no cognitive condition on creditworthy performance, the use of this strategy to argue for cognitive conditions on several different concepts has been misguided. Creditworthy performance has the same kind of anti-luck condition as these concepts, and cognitive conditions rule out paradigmatic instances of the relevant kinds of luck in the case of creditworthy performance, but, as we have seen, there is no cognitive requirement on creditworthy performance. Something other than a cognitive state is what actually rules out the relevant kind of luck in the case of creditworthy performance. To be clear, this does not mean there is no cognitive condition on these other concepts. What it shows is that this popular strategy is not effective at establishing that there are such conditions for these concepts. More argument is needed to establish that these cognitive conditions 37 are required. This is important because this strategy has been taken to establish by itself these cognitive conditions and because it has been used to refute alternative views. If the strategy is illegitimate, we make room for these alternatives. What I have said so far may be a bit abstract. We can make it more concrete by examining cases of this strategy being deployed in the philosophy of action and seeing the consequences of undermining the strategy. The Cognitive Anti-Luck Strategy and Deviance. A famous problem in the philosophy of action is the problem of deviant causal chains. 50 It is easiest to see what the problem is by seeing an example of it. In trying to give an analysis of what it is to act intentionally, it is natural to propose that agents act intentionally just in case their behavior is caused by their intention, understood as a kind of mental state. But this analysis faces a problem. Suppose I am rock climbing with someone who I intend to kill. 51 As we climb, because I intend to kill him, I become nervous. My nervousness causes my hand to twitch, which results in my letting go of the rope holding him up. In this case, my intention causes my behavior, but I do not act intentionally. This undermines the simple analysis suggested. This kind of deviancy is a persistent thorn in the side of philosophers who have tried to give analyses of a variety of important concepts, not just acting intentionally. Notice that this deviancy is a kind of accidentality. I do not intentionally kill my climbing partner because it was accidental. So any concept that has a deviant causal chain problem is one that has a certain kind of non-accidentality condition. And so we see philosophers employing the cognitive anti-luck strategy to argue for cognitive conditions on various concepts in the philosophy of action that face deviant causal chain problems. This in turn seriously restricts the range of views of those concepts we can defend. Here are some examples of this argumentative strategy being deployed. 50 The problem was first developed by Donald Davidson. See Donald Davidson, “Freedom to Act,” Donald Davidson, ed., Essays on Actions and Events: The Philosophical Essays of Donald Davidson (Oxford: Oxford University Press, 1980), pp. 63-81. 51 The case is from Davidson (ibid). 38 John Gibbons argues for a knowledge condition on intentional action, and his central reason for imposing this condition is that it is needed for an agent’s action to exhibit the kind of control required for it to not count as an accident. 52 He uses cases of more ordinary luck and deviant causal chain cases to illustrate his point. In one such case, a man intends to kill his uncle, and he believes he will do so by pressing a button that detonates a bomb. His belief is true and justified, but he does not have much of an idea of how the process works. When he presses the button, the signal is intercepted by a government agency designed to prevent such attacks, thereby stopping the normal detonation process. The agency then warns the potential victim. Unfortunately, this time, the agency sends the warning on a frequency that detonates the bomb. The bomber justifiably and truly believes he will kill his uncle, but he does not do so intentionally. Gibbons thinks what is missing is knowledge. Errol Lord has argued that the way to avoid deviant causal chain problems for the concept acting for a reason is to analyze acting for a reason in terms of an agent’s knowhow, particularly in terms of their knowledge how to use some fact as a reason. 53 This is one of his central motives for accepting this view. Ofra Magidor and John Hawthorne also suggest that acting for a reason requires knowledge, and they motivate this view by appeal to its being necessary to rule out deviancy. 54 Suppose you are afraid of dogs. Now suppose you enter a room where a dog is being sculpted. You do not see the dog in the room; you only see the sculpture. But you confuse it for a dog, so you run out of the room. Intuitively, your motivating reason for running out of the room is not that there is a dog in the room. The proposition that there is a dog in the room plays an important role in your motivation and the fact that there is a dog in the room explains your action by transitivity 55 , but the causal chain is deviant. Hawthorne and Magidor think that, to ensure that we 52 John Gibbons, “Knowledge in Action,” Philosophy and Phenomenological Research, LXII, 3 (2001): 579-600. 53 Errol Lord, The Importance of Being Rational, (Oxford: Oxford University Press, 2018). 54 John Hawthorne and Ofra Magidor, “Reflections on the Ideology of Reasons,” Daniel Star, ed., The Oxford Handbook on Reasons and Normativity, (Oxford: Oxford University Press, 2018), pp. 113-41. Cf. Nick Hughes, “Is Knowledge the Ability to φ for the Reason that P?,” Episteme, XI, 4 (2014): 457-62 and Dustin Locke, “Knowledge, Explanation, and Motivating Reasons,” American Philosophical Quarterly, LII, 3 (2015): 215-33. 55 That there is a dog in the room explains why there is a dog sculpture in the room, which in turn explains why you run out. 39 do not count the agent as having acted for this reason, there must be a knowledge condition on acting for a reason: to act for the reason that P, you must know that P. 56 In the achievement literature, Gwen Bradford has argued that achievement requires competent causation, which turns out to be a cognitive condition. 57 She contrasts performances which, she points out, are too accidental to count as achievements with ones that are achievements. These too accidental performances are much like our chess case. She then argues that what makes these performances too accidental is that they are not competent and imposes a cognitive condition to rule out accidental performances. She arrives at her particular cognitive condition by pointing out that others face the deviant causal chain problem. This is of course a paradigmatic case of the cognitive anti-luck strategy. All these authors deploy the cognitive anti-luck strategy to motivate cognitive conditions, where the kind of luck they are trying to rule out is deviance. My argument against cognitive conditions for creditworthy performance undermines this strategy. There are deviant causal chain cases for creditworthy performance: a pitcher might try to throw a 100 mph fastball, slip during his windup in a way that would usually prevent him from reaching that speed, but ultimately reach 100 mph because of a fortunate wind that blows just as he releases the ball. 58 This is not creditworthy. So creditworthy performance, like these other concepts, is inconsistent with deviance. And cognitive conditions can rule out these paradigmatic instances of deviance in the case of creditworthy performance: the pitcher does not know he will throw the ball 100 mph, does not manifest his knowhow, since he did not throw the ball the way he knows how to throw it 100 mph, and does not competently cause the throw, since his belief about how he will throw it is false. 56 Other philosophers who have motivated similar conditions, sometimes in the same way, include Peter Unger, Clayton Littlejohn, and Christina Dietz. See Peter Unger, Ignorance: A Case for Scepticism, (Oxford: Clarendon Press, 1975), Clayton Littlejohn, “How and Why Knowledge is First,” Adam Carter, Emma Gordon, and Ben Jarvis, eds., Knowledge First: Approaches in Epistemology and Mind (Oxford: Oxford University Press, 2017), pp. 19-45, and Christina Dietz, “Reasons and Factive Emotions,” Philosophical Studies, CLXXV, 7 (2018):1681-1691. 57 Bradford, Achievement, op. cit. 58 An anonymous referee helpfully raises the worry that non-accidentality is built into the concept throwing a fastball: you cannot throw a fastball deviantly. I think this case puts pressure on this view, since it is natural to describe this agent as having thrown a 100 mph fastball, just not in a way that is creditworthy. But even if we grant that you cannot throw a fastball deviantly, this just moves the bump in the rug: there is a non-deviance condition on throwing fastballs, cognitive conditions can rule out deviant cases, but there is no cognitive condition on throwing fastballs, since Pat throws a 100 mph fastball. The strategy is still not valid. 40 But there is no cognitive condition on creditworthy performance. So the strategy by itself does not establish that there is a cognitive condition on these other concepts. It is worth emphasizing again that this does not mean that there are no cognitive conditions on properties like acting for a reason. It is very plausible that there are. The point is that the strategy is insufficient to establish cognitive conditions. If we know that creditworthiness is inconsistent with the same kinds of accidentality and there is no cognitive condition on creditworthy performance, then there are alternative ways of ruling out that kind of accidentality. So, to establish cognitive conditions, the defenders of this strategy must supplement this strategy with reasons for thinking the alternatives will not work for the concepts they are interested in or else they must independently motivate their cognitive conditions. Of course, because the activities I have used to resist cognitive conditions on creditworthy performance are less intellectual in nature, there is at least some reason to think the alternatives that rule out deviance for these cases will not be able to rule out deviance for more intellectual activities like acting for a reason without imposing cognitive conditions. But, once we know some such solution rules out deviance without imposing a cognitive condition in at least some cases, more still needs to be said to establish cognitive conditions on these other concepts. What is more, even if these alternatives end up needing to impose some cognitive condition for the intellectual activities, since these solutions can rule out some deviant cases without any cognitive condition, they may be able to do more with less: the cognitive conditions they impose on intellectual activities might be weaker than what seemed to be forced on us by the cognitive anti-luck strategy. 59 This is important because the cognitive conditions this strategy has seemed to commit us to can set very surprising and serious constraints on the views we can defend. By undermining these arguments for cognitive conditions, we open up the space of defensible views considerably. The case of acting for a reason makes this very vivid. The orthodox picture of rationality that we 59 For instance, an ability manifestation solution seems well-equipped to rule out deviant cases for physical activities. When one pitches a 100 mph fastball deviantly, the pitch fails to manifest one’s 100 mph fastball- throwing strength. And this solution has been proposed as a way to address deviant causal chain cases for acting for a reason by Susanne Mantel in Susanne Mantel, “Three Cheers for Dispositions: A Dispositional Approach to Acting for a Normative Reason,” Erkenntnis, LXXXII, 3, (2017): 561-582. But her solution does not entail a knowledge requirement unless virtue epistemology is true, so it opens up space to defend the primacy of reasons over knowledge. 41 get from decision theory, epistemology, and ethics holds that rationality does not require knowledge. This orthodox picture has emerged because, on the face of it, we can give rationalizing explanations of even ignorant people’s behavior and beliefs. Some have even thought that we can give non-factive rationalizing explanations of people’s acts. 60 If Hawthorne, Magidor, and Lord are right, though, the orthodox picture must be completely upended, and, since knowledge is factive, rationalizing explanations must be factive. This complete overthrow of the orthodox, internalist picture of rationality would be one of the great discoveries of philosophy. Our argument shows that it will take more to make this great discovery: the orthodox view cannot be dismissed so easily. There are other ways to rule out deviance, so Lord, Hawthorne, and Magidor have not established that we need a knowledge condition. Moral Worth and the Cognitive Anti-Luck Strategy. As mentioned, philosophers have used the cognitive anti-luck strategy in the moral worth literature to undermine one of the most popular views of moral worth. To see this, we first need to understand what it is for actions to be morally worthy. Kant pointed out that there is a difference in our appraisal of a right action done out of a concern for doing what is right and one done from self-interested motivations. 61 He contrasts a shopkeeper who gives correct change in order to keep his customers and increase profits with a shopkeeper who does so because it is right. Only the latter acts with moral worth. The former shopkeeper does the right thing too accidentally for his act to be morally worthy. He might have easily failed to do the right thing. 62 60 Mark Schroeder and Juan Comensana and Matthew McGrath offer prominent defenses of this view. See Mark Schroeder, “Having Reasons,” Philosophical Studies, CXXXIX, 1 (2008): 57-71 and Juan Comesaña and Matthew McGrath, “Having False Reasons,” Clayton Littlejohn and John Turri, eds., Epistemic Norms (Oxford: Oxford University Press, 2014), pp. 59-80. 61 Immanuel Kant, Groundwork of the Metaphysics of Morals, Allen Wood, ed., (New Haven: Yale University Press, 1785, repr. 2002). 62 The following passage from ibid is suggestive: “But I assert that in such a case the action, however it may conform to duty and however amiable it is, nevertheless has no true moral worth, but is on the same footing as other inclinations, e.g., the inclination to honor, which, when it fortunately encounters something that in fact serves the common good and is in conformity with duty, and is thus worthy of honor, deserves praise and encouragement, but not esteem.” 42 This case bears striking similarities to the chess case. Indeed, it is natural to think that what Kant is talking about when he talks about moral worth is what it takes for agents to deserve credit for their right actions. This is in fact how many philosophers working in the moral worth literature think of moral worth. 63 They think one’s right act has moral worth just when it is a right action for which one deserves credit. Because they think of moral worth as but an instance of this general category, a natural way for them to try to gain insight about the nature of moral worth is to look at the general category. Zoë Johnson King takes an approach like this to argue for the claim that you act with moral worth only if you have some idea that what you are doing is right. 64 She thinks that moral worth and praiseworthiness for doing the right thing have the same non-accidentality requirement. And she points out that, typically, lacking any idea of what you are doing makes your act an accident in a way that undermines its praiseworthiness. She uses the example of a person who finds a buried treasure while having no idea that her digging in the sand constituted finding the buried treasure. Such a person finds the buried treasure accidentally, and, for this reason, is not praiseworthy for finding the buried treasure. She is at most praiseworthy for digging a hole. Cases like this make it plausible that, in general, for your act to be non-accidental in the way that allows it to be praiseworthy, you must have some idea your more basic actions constitute that act. And if moral worth has the same non-accidentality condition as praiseworthiness, for your right action to have moral worth, you must have some idea that the more basic actions you perform constitute doing the right thing. Johnson King defends her view by arguing that, in general, for an act to be non-accidental in the way that is required for praiseworthiness, agents must have some idea of what they are doing. Consequently, there is a cognitive condition on praiseworthiness. And because moral worth has the same non-accidentality condition, moral worth has the same cognitive condition. Thus, we can evaluate her argument by thinking about whether this cognitive condition holds for praiseworthiness in general. Though I framed the conditions I posed in terms of creditworthiness, my cases are as effective at undermining analogous cognitive conditions on praiseworthy action. 63 Paulina Sliwa and Lord characterize moral worth this way. See Paulina Sliwa, “Moral Knowledge and Moral Worth,” Philosophy and Phenomenological Research, XCIII, 2 (2015): 393-418, Errol Lord, “On the Intellectual Conditions for Responsibility,” op. cit., and Errol Lord, The Importance of Being Rational, op. cit. 64 Zoë Johnson King, “Accidentally Doing the Right Thing,” op. cit. 43 There is no general cognitive condition on praiseworthy action, since Eddie is praiseworthy even though he does not satisfy any such condition, so Johnson King’s argument is unsound. Paulina Sliwa gives a similar kind of argument for a knowledge condition on acting with moral worth. She argues that acting with moral worth requires knowledge of what the right thing to do is rather than mere justified true belief because, if it only requires the latter, we run into deviant causal chain problems. 65 She imagines a case where a doctor prescribes a certain medicine based on a lab technician’s report that the patient is iron deficient. Unfortunately, the technician got it wrong: the patient is vitamin B12 deficient. Luckily, though, the medicine prescribed also contains B12, so the agent is healed. Sliwa suggests that the doctor does not deserve credit for prescribing the right medicine even though the agent had a justified true belief about what medicine to prescribe. She concludes that what is missing is knowledge, and so claims that you must know which act is right to act with moral worth. Unlike Johnson King, who appeals to a general cognitive condition on praiseworthy action, Sliwa motivates her cognitive condition by appeal to a deviance case for creditworthy right action in particular, so the argument from the first part of the paper does not show Sliwa’s argument to be unsound by directly falsifying a premise. But Sliwa’s argument is an instance of the cognitive anti-luck strategy, so my argument from the first part of the paper suggests that her argument does not establish its conclusion. Creditworthy performance is a concept that faces a deviant causal chain problem, but no cognitive condition rules out deviance for creditworthy performance. Therefore, there are ways of ruling out deviance that do not involve cognitive conditions. Sliwa must show that these alternative ways of ruling out deviance will not rule out deviance for creditworthy right action without requiring knowledge for her argument to establish its conclusion. Since she has not done so, there is space to reject her cognitive condition. That my account gives us space to reject Sliwa and Johnson King’s cognitive conditions on creditworthy right action is important, as they use these conditions to reject possibly the most popular accounts of moral worth: right-reasons accounts of moral worth, famously defended by Julia Markovits and Nomy Arpaly. 66 Right-reasons accounts hold that agents act with moral worth 65 Paulina Sliwa, “Moral Knowledge and Moral Worth,” op. cit. 66 Nomy Arpaly, Unprincipled Virtue: An Inquiry Into Moral Agency (New York: Oxford University Press, 2002) and Julia Markovits, “Acting for the Right Reasons,” Philosophical Review, CXIX, 2 (2010): 201-42. 44 just in case they are motivated to perform the act by the facts that make the act right, facts like its relieving suffering or upholding a promise. Arpaly and Markovits prefer these accounts over accounts that require that you be motivated by the bare fact that the act is right, which I will call Kantian accounts, in part because right-reasons accounts do not require you to have any idea your act is right. They and others are drawn to right-reasons accounts because they find it intuitively plausible that there is no such cognitive requirement on moral worth. Arpaly uses a case involving Huckleberry Finn to illustrate this. Huck Finn frees Jim from slavery even though he believes this is wrong. Huck uncritically accepts the moral views of his society, which hold that slaves are the property of their enslavers, so he believes what he is doing constitutes theft and is thus wrong. Yet he still frees Jim out of concern for his friend’s wellbeing. Because of the way Huck was socialized, he is far from realizing he is doing the right thing, but he still seems to deserve credit for doing the right thing. Cases like this make it appealing to allow agents to act with moral worth even while having no idea they are doing the right thing. If Johnson King or Sliwa is right, though, this renders one’s act too accidental for it to be an action that has moral worth. Because my argument gives us space to reject these cognitive conditions, it gives us space to defend the view that agents like Huck Finn can act with moral worth. Reasons-Based Accounts of Moral Worth Are Not Rock-Bottom. If I am right, right-reasons views are not refuted by Johnson King and Sliwa’s anti-luck arguments, because the falsity of cognitive conditions on creditworthy performance undermines those arguments. This does not, however, mean that I endorse a right-reasons account of moral worth. Right-reasons accounts of moral worth cannot be dismissed as quickly as Johnson King’s and Sliwa’s arguments would allow if sound, but, as I will argue now, neither a right-reasons account of moral worth nor a Kantian account can be the correct fundamental account of moral worth, if, as is widely accepted, moral worth is a species of creditworthy or praiseworthy performance. 67 67 Lord, Jonathan Way, Nathan Howard, Sliwa, Johnson King, Arpaly, Douglas Portmore, Keshav Singh, Grant Rozeboom, Jessica Isserow, and Joseph Cunningham all treat moral worth as either a form of praiseworthy action or creditworthy action. See Errol Lord, The Importance of Being Rational, op. cit., Jonathan Way, “Creditworthiness and Matching Principles,” Mark Timmons, ed., Oxford Studies in Normative Ethics, VII, (Oxford: Oxford University 45 This is not to say that neither account is true; they just are not rock-bottom. Once again, this follows from the falsity of cognitive conditions. Recall that, on right-reasons accounts, acting with moral worth requires that you act for reasons that make your act right. So, for instance, if what makes it right to help your friend mow his lawn when you promised to do so is that you would be keeping your promise, the reason for which you must act in order to act with moral worth is that it is a way of keeping your promise. On Kantian accounts, you deserve credit for doing the right thing when you do the right thing because it is the right thing to do. Both of these accounts hold that being creditworthy is about acting for a certain kind of reason. They just differ on what reason is required. But acting for a reason in general requires awareness of that reason when you act. My reason for going to the store cannot be to get milk if I do not have any belief whatsoever that there is milk at the store. So, for right-reasons accounts, in order to act with moral worth, you must be aware on some level that the fact that makes the act right obtains. And for Kantian accounts, you must be aware on some level that the act is right. These are both cognitive conditions on moral worth! So the two views we have discussed are committed to cognitive conditions on moral worth. The cognitive condition for the right-reasons view would look slightly different from the ones we have considered if it were a general condition on creditworthy performance. This cognitive condition would not require that, for you to get credit for some act F, you be aware, in some way, that it is an act of type F. It would only require that you be aware of the facts that make it an F. So, for instance, Eddie in Lifting would have to be aware in some way that he lifted 1103 Press, 2017), pp.207-29, Nathan Howard, “One Desire Too Many.” Philosophy and Phenomenological Research, CII, 2 (2021): 302-17, Paulina Sliwa, “On Knowing What’s Right and Being Responsible for It,” Phillip Robichaud and Jan Willem Wieland. eds., Responsibility: The Epistemic Condition, (Oxford: Oxford University Press, 2017), pp. 127–145, Zoë Johnson King, “Accidentally Doing the Right Thing,” op. cit., Nomy Arpaly, Unprincipled Virtue, op. cit., Douglas Portmore, “Moral Worth and Our Ultimate Moral Concerns,” Mark Timmons, ed., Oxford Studies in Normative Ethics, XII, (Oxford: Oxford University Press, 2022), pp. 276-98, Keshav Singh, “Moral Worth, Credit, and Non-Accidentality,” Oxford Studies in Normative Ethics, X (Oxford: Oxford University Press, 2020), pp.156-82, Grant Rozeboom, “The Motives of Moral Credit,” Journal of Ethics and Social Philosophy, XI, 3 (2017), Jessica Isserow, “Moral Worth and Doing the Right Thing by Accident,” Australasian Journal of Philosophy, XCVII, 2 (2019): 251-64, and Joseph Cunningham, “Moral Worth and Knowing How to Respond to Reasons,” Philosophy and Phenomenological Research (forthcoming). Faraci treats acting with moral worth as a member of the genus accomplishment in David Faraci, “On Accomplishment,” op. cit. My cases undermine cognitive conditions on accomplishment as well. 46 pounds, since this is what makes it a world record. But this condition is also false: even if Eddie does not look at the weight total before lifting, so that he has no idea how much it weighs, he still deserves credit for setting the world record. So both accounts of moral worth are committed to cognitive conditions that are false for the genus creditworthy performance and, we noted in our discussion of Johnson King’s argument, praiseworthy performance. And, as mentioned earlier, it is widely accepted that to act with moral worth is either to be creditworthy or praiseworthy for that right action. Given this, it would be tempting to conclude that these accounts are false. But this is too quick: accounts of creditworthy performance for particular performances could have cognitive conditions even if the fundamental, general account of creditworthy performance builds in no cognitive condition. It is plausible, for instance, that creditworthy chess performance involves satisfying a cognitive condition. Since there is no cognitive condition built into the fundamental account, this condition would have to follow from the general account together with unique features of chess. Something similar might also hold for moral worth. Reasons-based accounts could still be extensionally correct for moral worth, but the correct reasons-based account would have to fall out of some more fundamental explanation. My argument suggests, then, that, minimally, there is a more fundamental explanation of what makes acts have moral worth than the Kantian and right-reasons accounts. If there is a more fundamental explanation, we, as philosophers, should look for it. In searching for a more fundamental explanation, we may begin to formulate more sophisticated views, views that are ecumenical to both accounts of moral worth. For instance, instead of having to take hard stands on which kind of reason moral worth requires, we might find a way of allowing agents to get credit for either. We can begin to develop accounts that capture the appeal of both views. We will see this in closing. Conclusion You might dismiss my objection to cognitive conditions because you think that some cognitive condition has to be right. There must be some cognitive condition that explains the difference between lucky accidents and creditworthy performances. After all, how else could we explain the difference? 47 This is a natural reply, but it is not justified. That is because there is an account that seems able to explain the difference in a way that does not require a cognitive condition. As previously discussed, one famous account of creditworthiness, most famously defended by Ernest Sosa, holds that an agent is creditworthy just in case the agent’s act manifests the virtue(s) associated with the relevant activity. Virtues need not be intellectual, though they can be. They are abilities more broadly. This account can therefore differentiate lucky cases from creditworthy cases without appealing directly to something cognitive. We can start to see this by seeing how this account handles the cases we have discussed throughout this paper. This account seems to explain these cases well. Take Lifting. On this view, Eddie is creditworthy for setting the deadlifting world record if he manifests the virtue(s) associated with setting deadlifting world records. I submit that it is overwhelmingly plausible that the virtue associated with setting deadlifting world records is strength. Since Eddie’s setting the deadlifting world record patently manifests his strength, he is creditworthy for setting the deadlifting world record. The same holds for Pat in Pitching. It is overwhelmingly plausible that the virtue associated with throwing one hundred MPH fastballs is arm strength. Since Pat manifests his arm strength in throwing one hundred MPH, he of course is creditworthy for throwing the ball one hundred MPH. This also explains why intellectual conditions are not relevant for creditworthiness in these domains: the virtue(s) associated with these domains has nothing to do with anything intellectual, and so manifesting that virtue will not have to involve anything highly intellectual. But the account can also explain why certain intellectual conditions matter in certain cases, like the chess case. In those cases where the virtue(s) associated with a certain kind of act involves some intellectual capacity, creditworthiness will depend on manifesting the relevant intellectual capacity. In the case of chess, the virtue of chess is something like having the ability to come up with strategies and tactics in order to checkmate your opponent. Since this requires a high level of awareness about what you are doing, it is no wonder that creditworthiness for forking your opponent’s pieces requires having consciously decided to set that tactic up. The same goes for making your way to Larissa. Part of the virtue associated with going some place is knowing how to get there. So of course to manifest the virtue associated with getting to Larissa, you must know how to get there, and thus this is a condition on creditworthiness. A virtue-theoretic account can 48 thus explain our verdicts on all these different cases, and it can even explain why the features that the other views incorrectly held to be always relevant are merely sometimes relevant. This account also seems like a natural candidate for the more fundamental explanation of what makes right actions have moral worth. Obviously, the virtue associated with right action is moral virtue, so this account holds that an act is morally worthy iff it manifests moral virtue. The different views of moral worth we have seen can then be treated as giving necessary and sufficient conditions for moral worth by offering substantive accounts of the nature of moral virtue. But this account also leaves open the possibility that either of the ways these accounts tell us we need to be motivated in order to act with moral worth is a way of acting with moral worth. This enables us to defend a disjunctive view of the motives that are necessary and sufficient for moral worth without committing us to a fundamentally disjunctive theory. That is because, depending on the nature of moral virtue, acting from either motive may be a way of manifesting moral virtue. This view opens up space to defend views that capture the appeal of both of the original candidates. So there is, after all, an account on the table that seems able to handle the cases we have been considering while differentiating lucky accidents from creditworthy action. This means that we cannot just dismiss my objections because there must be some way to differentiate accidents from creditworthy performance and the only way to do so is through appeal to some cognitive condition. There is an alternative that does not appeal to cognitive conditions, so my objections, rather than simply raising puzzles, may be pointing us in the direction of the most plausible view of creditworthy performances. 49 Chapter 3: An Attempt at a General Solution to the Problem of Deviant Causal Chains The problem of deviant causal chains arises in many settings. Most famously, deviant causal chain problems wreak havoc in epistemology, the philosophy of action, and the philosophy of perception. 68 The most common approach to addressing these problems is to try to solve them individually or, at most, a few at a time. 69 My aim in this paper, as the title suggests, is to take a less travelled (perhaps because much more treacherous) path: I will endeavor to give a general solution to deviant causal chain problems. My approach will be straightforward. I’ll present the deviant causal chain (henceforth DCC) problem for apt performances and give my solution. I’ll then move on to different instances of the problem and show either how my solution applies to those instances or how it can be reformulated to apply to them. I’ll also discuss existing solutions to these problems in order to bring out the advantages of my solution and its relations to other solutions. The further instances I’ll consider, in the order in which I’ll consider them, are the classical Gettier problem, the DCC problem for acting intentionally, the DCC problem for seeing, and the DCC problem for manifesting a disposition. As will be explained, my solution generalizes to these cases because deviance arises in each of these cases when a disposition isn’t manifested and my solution gives us an analysis of disposition manifestation. So, if all deviance arises when a disposition isn’t manifested, my solution will be fully general. After discussing these cases, I will close by presenting and replying to some objections. DCC Problem for Apt Performance There’s a difference between hitting a bullseye aptly, or skillfully, and hitting it inaptly, or not skillfully. If I hit it by closing my eyes, guessing its location, and releasing the bow, I do not hit it skillfully. If I align my arrow with the bullseye, maintaining good technique as I release the 68 Gettier (1960) raised the problem in epistemology. Davidson (1973) raised the problem in the philosophy of action. Grice (1961) recognized that objects must cause experience in the right way for one to perceive. Sosa (2015) points out that these seem to be instances of the same problem. 69 Prominent solutions that focus on limited instances include Clark (1963), Goldman (1967), Harman (1976), Nozick (1981), Audi (1986), Peacocke (1979a and 1979b), Lewis (1980), Brand (1984), Mele and Moser (1994), and Sosa (2007). Sosa (2015) later proposes a general solution (see footnote 7). Grice (1961) and Goldman (1970) voice early skepticism that it’s philosophers’ task to solve DCC problems. See Lewis (1980) and Bishop (1989) for criticisms of this skepticism. 50 bow, propelling the arrow directly to the target, I hit the bullseye skillfully. 70 This suggests a natural way to distinguish these shots: only the apt shot hits the bullseye because of one’s skill. This analysis immediately faces a DCC problem. Suppose I aim at the target in such a way that, normally, my shot will hit the bullseye. But, because of a freak wind, the shot is directed away from the target. Fortunately, though, an angel nearby redirects the shot back towards the bullseye by manipulating the wind. And she does so out of pity that a shot that exercised so much skill was derailed off target. In that case, I hit the shot because of the angel’s help, and the angel helps me because of my skill. By transitivity, I hit the shot because of my skill, but I don’t hit the bullseye skillfully. 71 To solve this problem, we must add something to the analysis that ensures we don’t count these deviant cases as cases of hitting shots skillfully. A natural thing to try to add is a characterization of the non-deviant causal chain type. That is, we try to identify the repeatable causal sequence 72 through which your skill must cause your hitting the bullseye for you to hit it skillfully. That is what I will now attempt. Agents count as having some amount of skill at archery when they have a reliable disposition to hit bullseyes in certain circumstances. When such agents have a reliable disposition to hit bullseyes, this is because they have some traits that ground this reliable disposition in these circumstances. But how do these traits ground this reliable disposition? They ground this reliable disposition by enabling the agent to perform steps in causal chain types resulting in the activity. In other words, the grounds enable the agent to help complete causal chains that result in the activity. Of these causal chain types, the non-deviant causal chain type is the one that figures in the best explanation of how the grounds of the agent’s reliable disposition ground the agent’s reliable disposition. A causal chain token is non-deviant if it is a member of this type. 70 The apt/inapt performance distinction is from Sosa (2003). 71 Sosa (2007) raised the DCC problem for apt performance. Bradford (2015) raises a similar angel case to undermine certain analyses of achievement. 72 It might be worth it to say more here about what I have in mind by causal chain types. Causal chain types are sequences of event types each standing in a causal relation with some other event type in the sequence. These causal chain types shouldn’t be too mysterious. Two causal chain tokens starting with a plane exploding and ending with two different people dying may share some causal chain types in common while diverging with respect to others. If both explosions expelled the people from the plane, then then they share the following causal chain type in common: the plane explosion caused them to be expelled from the plane, which in turn caused them to die. But if one died of shock and the other from hitting the floor, then each causal chain token is a member of a causal chain type that the other is not a member of. 51 In the angel case, the archer deploys the grounds of his reliable disposition to hit the shot: he deploys his understanding of trajectories, aiming ability, poise, and technique. But the causal chain type, as it is specified, leading to the arrow connecting with the bullseye is not the causal chain type that figures in the best explanation of how the grounds of his reliable disposition ground his reliable disposition. The reliable disposition that constitutes skill is a reliable disposition to hit bullseyes across a wide variety of circumstances. Causal chain types that invoke the assistance of angels do not offer a good explanation of how the grounds ground a reliable disposition to hit bullseyes in this wide variety of circumstances, since, in many of these circumstances, angels are not present. Because the causal chain token is a member of only types like this, which do not provide the best explanation of how the grounds ground this reliable disposition, the causal chain token is deviant, and the shot is not skillfully hit. By contrast, when an agent hits the shot skillfully, the causal chain token is a member of the type that figures in the best explanation of how the grounds of the reliable disposition ground the reliable disposition. I can’t state every detail of this causal chain type, since that is a job for physicists and biologists, but here are some: Using his past experience with trajectories and shooting, the agent decides to shoot with a particular force and angle that tends to produce a trajectory leading to a bullseye. Having decided on this angle and force, his poise, technique, and aim allow him to shoot at this angle and with this force. This results in the arrow taking the expected trajectory, which in turn produces the bullseye. This partial specification of the non- deviant causal chain type reveals why the angel causal chain token is not a member: the angel- redirected arrow does not take the expected trajectory. That is my solution to the DCC problem for apt, or skillful, performances. Addressing some natural worries before moving on will help clarify the solution. Let’s start by getting a first pass analysis on the table: First Pass: A token causal chain is non-deviant just in case it is a member of the causal chain type that provides the best explanation of how the grounds of the agent’s reliable disposition ground the agent’s reliable disposition. There are two worries to address. The first is over what it takes for a causal chain type to provide the best explanation of how the grounds ground the agent’s reliable disposition. To meet the criterion, the causal chain type must first actually provide an explanation of how the grounds 52 ground the reliable disposition: the causal chain type must be one that the grounds enable the agent to perform steps in and must terminate in the right performance. It must then also provide the best explanation. The causal chain type that provides the best explanation of how the grounds of the reliable disposition ground the reliable disposition is the most informative causal chain type that successfully explains how the grounds ground the reliable disposition in the full set of circumstances in which the agent has the reliable disposition. In other words, the causal chain type that provides the best explanation is the causal chain type that strikes the best balance between depth and breadth of explanation. The angel-involving causal chain type, as I construed it, could not explain how the grounds grounded a reliable disposition to hit bullseyes in a set of circumstances that includes cases where no angels are present. But the skill-constituting reliable disposition to hit bullseyes includes such circumstances, so this causal chain type could not explain how the grounds of the skill-constituting reliable disposition ground the skill constituting reliable disposition. We could have of course construed the causal chain type differently. In my construal, one of the events in the sequence just is the angel’s intervention. But angel-involving causal chain tokens can also be members of types that make no reference to such interventions. My deviant causal chain token is, for instance, a member of the following causal chain type: the archer aims and lines his bow up with the target and releases the bow, which causes the arrow to hit the bullseye. This causal chain type is an explanation of how the grounds ground the skill-constituting disposition, so I need to be able to say why it isn’t the best explanation. This causal chain type isn’t best because it isn’t sufficiently informative. 73 We can give a more informative explanation of how the grounds ground the reliable disposition that still covers all of the target cases, and this more detailed explanation will rule out deviant causal chain tokens. For instance, the partial description of the non-deviant type that I gave was more informative, and it ruled out the angel case by requiring that the arrow take the expected trajectory that the agent 73 It should be fairly intuitive that different causal chain types can offer varying degrees of information about how something is brought about. They will vary when they vary in what events are included in their sequences and how they characterize those events. I might give two different causal explanations of how John died. One might specify that John died because of a heart attack while the other might specify that John died because of a heart attack that was itself caused by an overdose. The latter explanation is more informative because the latter causal chain type includes an additional event. An even more informative causal chain type might specify what kind of overdose he suffered from. 53 set up. In general, broad causal chain types that permit deviant causal chain cases get ruled out for being insufficiently informative. Of course, we must make sure to not include too much information in our construal of the non-deviant type. After all, if we include too much information, then we will rule out cases of skillful performances along with deviant cases. What guarantees that we do not include too much information is the requirement that the explanations be sufficiently broad: they must be able to explain how we have the reliable disposition in the full set of circumstances in which we have the reliable disposition. As we proceed, I will make sure to emphasize when causal chain types are ruled out for lacking breadth or depth so that we maintain our grip on how the solution works. The second worry has to do with having a reliable disposition in a set of circumstances. I said that the causal chain type involving the angel is deviant because it doesn’t offer the best explanation of how the grounds of the reliable disposition ground the reliable disposition. And that was because the relevant disposition to hit bullseyes was a disposition to hit bullseyes across a wide variety of circumstances: it is this wide-ranging disposition that constitutes skill. But, you might have worried, in what circumstances must agents be reliably disposed to hit bullseyes to count as skilled? Analogous questions have been asked of reliabilist epistemologies: in what circumstances must agents be reliable (or have a reliable disposition) for their beliefs to be justified? As we will see, similar questions can be pressed for my solutions to the DCC problem for intentional action and perception. I will not, in general, have an answer to this question. The problem being raised here is essentially a generalized form of the generality problem for reliabilist epistemologies. My task here is not to give a general solution to the generality problem but to give a general solution to the DCC problem. These are separate problems: agents whose beliefs are justified according to reliabilism because they are produced by a process that is reliable in the right set of circumstances may nevertheless lack knowledge because they are Gettiered. The generality problem, and other problems, will still need to be solved to vindicate theories of these concepts that rely on causation or dispositions, but solving the DCC problem would be substantial progress towards that goal. Because it is not my task to solve the generality problem, my description of the circumstances that help to individuate the relevant dispositions will be vague: I will say things like “One counts as having an intention to phi only if one is disposed to phi in some range of circumstances.” I won’t state the circumstances more concretely so as to stay neutral on the 54 solution to the generality problem. I will, however, rely on intuitions about how wide those ranges of circumstances are, which will allow us to rule out certain causal chain types as deviant, as I just did earlier. We can trust that the correct solution will include certain circumstances even if we do not know every circumstance it includes. With these clarifications in order, it will be helpful to contrast my view with others so as to see its promise. Ernest Sosa (2007) famously defended the view that the difference between the non-deviant and deviant case is that only in the non-deviant case is both the performance and the success due to skill. He thought that, in a deviant case in which an impersonal wind redirects the shot on target, the performance is caused by skill but not its success, and he hoped this might generalize. Unfortunately, cases like the angel case, in which the angel redirects one’s shot because of one’s skill, forced Sosa to abandon this solution in favor of a different solution. 74 The problem is that success can be deviantly caused as much as performance, but Sosa’s original solution can’t distinguish deviant successes from non-deviant successes. My solution, by contrast, does exactly that. Another popular solution, due to Jon Greco (2010), holds that the difference between deviant and non-deviant cases is that, in non-deviant cases, one’s skill is the salient cause of one’s successful performance. Salience, here, is context dependent; speakers’ interests determine salience. An attraction of this view is that it can successfully explain the deviance of the angel case: in the angel case, the angel, not your skill, is the salient cause of your success. In this respect, the view is an improvement over Sosa’s original solution. Unfortunately, though, this view has faced the problem that salient causation isn’t a necessary condition for non-deviance. If I give you a spectacular assist in basketball, my pass, and not your skill, might be the salient cause of your success, but you can still hit the shot skillfully. This problem is a general problem for Greco’s original solution to the DCC problem 75 : we can generate these kinds of cases for many different domains where DCC problems arise. Indeed, 74 Sosa abandons this solution in Sosa (2015) and proposes that a general solution to DCC problems can be provided by invoking disposition manifestation. As we will see, his newer account, which avoids this problem, is closely related to mine: I will argue that the account I am giving here works because it provides us with a successful analysis of disposition manifestation. 75 Greco (2012) revises the account of salience so that quality of causal contribution determines salience: an ability non-deviantly produces a success iff it produces the success in a way that can regularly be exploited for relevant practical purposes. I lack the space to do full justice to this proposal, but here are two important problems. One is that the solution is insufficiently general. Lightning has awesome killing power that can fail to be manifested because of deviance, but I doubt that whether lightning manifests its awesome killing power depends on whether it kills through 55 if you have studied virtue epistemology, this problem should already feel very familiar. The problem I presented isn’t novel. It’s Jennifer Lackey’s (2007) objection to Greco’s solution to the Gettier problem. I’ve just extended it to undermine his solution to the DCC problem for skillful performance. 76 Greco’s original solution is both inadequate for the case of skillful performance and unable to generalize to offer a solution to the Gettier problem. I believe my solution, by contrast, is a promising general solution. To begin to see this, we should check if my solution successfully generalizes where, as just observed, Greco’s original solution cannot. We should turn to the classical Gettier problem. The Classical Gettier Problem There’s a difference between knowledge and mere true belief. If I close my eyes and randomly form a belief as to the number of stars in the night sky, and my belief turns out to be true, my belief isn’t knowledge. If I look up and conduct a careful count, my belief is knowledge. This suggests a natural analysis of knowledge: to know that P is to form a true belief that P because of one’s good belief forming processes. Once again, though, our analysis immediately faces a problem. Suppose Sally runs into Nogot at the bank. Sally knows that, as late as a few hours ago, Nogot owned a Ford. So she believes, reasonably, that Nogot owns a Ford, and she concludes, reasonably, that someone in the bank owns a Ford. Unfortunately, unbeknownst to Sally, Nogot is at the bank to cash a check she received for the sale of her Ford. But, fortunately, Havit, who is also at the bank, owns a Ford. Consequently, her belief that someone at the bank owns a Ford is true. And it’s also the product of good belief-forming processes. But she doesn’t know. This is the classical Gettier problem, a route that can regularly be exploited for our relevant practical purposes. A second problem is that deviant causal chain tokens will always be members of some causal chain types that are practically exploitable for relevant practical purposes. Causal chain types that leave out information needed to rule out some deviance cases but that are tokened in lots of different practically relevant circumstances will be so exploitable. Though I am skeptical of Greco’s Greco (2012) solution to the DCC problem, I find the solution he offers there to the generality problem for knowledge more promising. It seems very plausible that which abilities are relevant for knowledge depends on our practical aims. So I believe Greco can pair my solution to the DCC problem with his solution to the generality problem for knowledge to get a promising account of knowledge. Indeed, Greco’s worry for solutions that appeal to disposition manifestation was that they treated this relation as a primitive and therefore were unexplanatory. Since my account provides an analysis of disposition manifestation, it avoids this worry and therefore should be much more palatable to Greco. 76 Lackey (2007) points out that the salient cause of a true belief may be the agent who provided testimony, not the belief-forming agent, so one can know even though one’s abilities don’t saliently cause the true belief. Greco (2007), Sosa (2007), and Riggs (2009) respond. Lackey (2009) addresses these responses. As discussed in the last footnote, Greco revises his account in Greco (2012). And he makes further revisions to his account of how testimonial knowledge is acquired in Greco (2020). 56 which I take to be distinct from the problem posed by Carl Ginet’s fake barn case. 77 The case, unlike fake barn cases, is structurally the same as DCC cases, so the problem can be treated as an instance of the DCC problem. My solution to the DCC problem for skillful performance can address the classical Gettier problem as well. We can start to see this by looking at the Ford case. Sally has a reliable disposition to form the true belief that something is F in a certain set of circumstances which includes her antecedently believing that A is F. The grounds of this reliable disposition is the cognitive architecture responsible for her being able to perform existential introductions. These grounds enable Sally to perform steps in causal chain types that begin with the input belief that A is F and terminate in the output belief that something is F. But the causal chain type that best explains how these grounds ground this particular reliable disposition doesn’t start with any old belief that A is F; it starts with the true belief that A is F. After all, performing existential introductions reliably produces true beliefs only because existential introduction is truth preserving. So only a causal chain type that specifies that the input belief is true will provide an adequately informative explanation. Therefore, for Sally’s true belief to be non-deviantly explained by this particular reliable disposition, it must be formed through a causal chain in which the input belief of the existential introduction is true. Since Sally’s initial belief is false, the belief that Nogot has a Ford, her belief is only deviantly the product of this reliable disposition. 78 Sally’s causal chain token is deviant because its first step is a false belief. This naturally draws to mind the failed attempts to rule out Gettier cases by requiring that agents base their belief on no false lemmas. 79 No False Lemmas accounts implausibly rule out knowledge that the number of people in a room is approximately 30 when an agent takes a rough head count, forms the off- by-one belief that the number of people in the room is 31 and concludes that the number is approximately 30. 80 My account does not have this implication. It’s not a part of the best explanation of how our approximating abilities ground our reliable disposition to form true beliefs that the number of 77 See Gettier (1960). Fake barn cases are from Goldman (1976), who credited the case to Ginet. 78 If Sally believed that other people in the room owned a Ford, and it was true of those people that they owned a Ford, then her belief that someone in the room owned a Ford would have been non-deviantly explained by this disposition. That’s because her belief would have been overdetermined by different causal chain tokens, some of which were non-deviant. 79 See Clark (1963). 80 Warfield (2005) originally states a case like this. Klein (2008) gives a different counterexample, but the point is the same: sometimes knowledge isn’t undermined by false lemmas. See Feldman (1974) for Gettier cases without false lemmas. 57 As is approximately N that we first take a count, then form a true belief about the number of As, then form a true belief about the approximate number. After all, we have this reliable disposition to form true approximations even in circumstances where we form false beliefs about the exact number, so this causal chain type provides an explanation that lacks sufficient breadth to be best. The causal chain type that best characterizes how the grounds ground this wide-ranging disposition goes in part as follows: we take a count, then form a belief about the number that is approximately true, then form a true belief about the approximate number. Since this is the causal chain through which you form this last belief in the presented case, the belief is non-deviantly the product of your reliable disposition to form true approximations. This highlights one of the attractions of my view. My view treats other attempted solutions to the Gettier problem, like No False Lemmas, as special cases that fall out of a more general, fundamental explanation of the source of deviancy in Gettier problems. We should want this from our solution. Things do go wrong for Sally because she inferred from a false belief, but this doesn’t always make things go wrong. My account delivers these verdicts. Since what makes it hard to solve the Gettier problem is the breadth of Gettier cases, we should consider other instances of the problem to see if my solution shows any promise of covering them all. In one famous case, an agent has just initiated the process of flicking a match against a matchbox. 81 Given her past experiences with matchboxes, she believes that the match will light. Unfortunately, though, subtle imperceptible features of the surface of the matchbox make it unable to light a match. Fortunately, though, an angel in waiting lights the match for the agent, making her belief true. In this case, the agent has a reliable disposition to form the true belief that an A will be F in circumstances that include her observing that past As have been Fs. The cognitive architecture that enables us to perform inductive inferences grounds this reliable disposition. Which causal chain provides the best explanation of how these grounds ground this reliable disposition across certain wide-ranging circumstances? It includes at least the following steps: the agent encounters a new A, which then leads to the recollection that past As have been F. This recollection together with the agent’s noticing similarities to the past then leads the agent to form the belief that this new A will be F. This belief is then made true by events unfolding as they did in the past when the agent noticed these features: induction is reliable because of regularities in nature. An adequately 81 From Skyrms (1967). 58 informative explanation of how the grounds of our inductive abilities ground these abilities will therefore make reference to these regularities. Since, in the matchbox case, the match did not light through the intervention of an angel in the past, the causal chain token through which the agent forms the true belief is deviant, and she doesn’t count as knowing that the match will light. My account can also address a case that has been notoriously difficult for other solutions that appeal to reliability. Suppose you are a farmer looking at what appears to be a sheep. In fact, however, it is a sheepdog in sheep’s clothing. But, just behind the sheepdog is a sheep. You, because you seem to see a sheep, form the belief that there is a sheep in front of you. You use a normally reliable process to form the belief: you believe things are a certain way because they visually appear to be that way. But your belief isn’t knowledge. 82 In this case, you have a reliable disposition to form a true belief that P in circumstances that include its visually appearing that P. Your visual system together with whatever enables you to base beliefs on its outputs ground this reliable disposition by enabling you to perform steps in causal chains leading to the true belief. The most crucial step is performed by your visual system: it represents some object to you. And you then form a certain belief, which is the final step. But which causal chain type offers the best explanation of how these grounds ground your reliable disposition? It is the causal chain that begins with the object reflecting light, which touches your pupils, which then triggers the processes of your visual system. That is, it is the causal chain that begins with your seeing the object: our visual processing systems ground the relevant disposition to form true beliefs in part because they enable us to see. Since, in the sheepdog case, you do not see the sheep, your belief that there is a sheep in front of you is deviantly the product of the relevant disposition, and you do not know. Because my solution yields correct verdicts for several Gettier cases, including notoriously intractable cases, I’m optimistic my solution is on the right track. But this isn’t the only source of my optimism: my optimism also stems from my solution’s ability to generalize to other DCC problems. My solution has been developed with a wide range of DCC problems in mind. In contrast, most solutions to the Gettier problem have been developed primarily with the Gettier problem in mind. Unsurprisingly, these views struggle to generalize No False Lemmas views and infallibility views are examples of such views. These views have no hope of generalizing. Whether one has false beliefs or whether one is infallible seems 82 The case is from Chisholm (1966). 59 irrelevant or inapplicable to other instances of the problem. The visual representations involved in seeing aren’t based on any beliefs. I can intentionally perform an act even though my intention doesn’t infallibly guarantee its performance. Other solutions, like those Greco and Sosa originally proposed, have some promise, but, as we saw, they too ultimately struggle to generalize. One of the most popular approaches to generalizing solutions to the Gettier problem to other DCC problems is to treat deviancy in the other cases of the problem as arising in some way because of a lack of knowledge. 83 On this view, there isn’t parity among the different DCC problems. Rather, the Gettier problem comes first, and our solution to the other problems is parasitic on our solution to this problem. Some have proposed, for instance, that, to act intentionally, you must know or be able to know what you are doing. 84 Others have proposed that, to act skillfully, you must manifest your knowledge how to perform that act. 85 In the case of perception, Timothy Williamson (2000) has argued that perceiving is just a way of knowing. This solution is very general, but unfortunately it isn’t fully general. For it to be fully general, all instances of the DCC problem must involve agents, since only agents know things. But, as we will see later, the DCC problem arises in contexts that don’t involve agency: my microwave can fail to manifest its disposition to boil water because it boils the water in a deviant way, but my microwave doesn’t know things. My solution, we will see, can explain what it takes to manifest a disposition. But, if we are causal theorists, seeing that will only be possible if we can solve the DCC problem for seeing. So we now turn to that problem. DCC Problem for Seeing We have visual experiences that represent the world as being certain ways. Sometimes, having those experiences constitutes seeing that the world is that way and sometimes it does not, as when the experiences are visual hallucinations. A natural attempt to distinguish hallucinations from seeing does so as follows: your visual experience constitutes an instance of seeing that the world is a certain way when that visual experience is caused by the world’s being that way. 83 Bach (1978), Lord (2018), Sliwa (2015), Unger (1975), Dietz (2018), Hawthorne and Magidor (2018) and Gibbons (2001) take this approach. Bach explicitly treats the DCC problem for intentional action as a special case of the Gettier problem. For some criticisms of examples of this strategy, see Hughes (2014) and Locke (2015). 84 Bach (1978) and Gibbons (2001) 85 See Stanley and Williamson (2017). Lord (2018) defends this for creditworthy performance, a closely related concept. 60 This proposal immediately faces a DCC problem. Suppose there is a toad in front of me that can shoot a hallucinatory gas as a defense mechanism, and it does so when creatures make eye contact with it. I inhale the gas and have a hallucinatory experience of that very toad being in front of me. The fact that there is such a toad in front of me causes me to have a visual experience that represents that toad as being in front of me. But I do not see that the toad is in front of me. The causal chain is deviant. A slightly altered version of my solution enables us to distinguish the deviant chain from the non-deviant chain. Our visual-perceptual system grounds a disposition to form accurate visual representations of objects when we are looking in the direction of those objects within a certain distance without obstruction. It does so because it enables us to perform certain steps in various causal chain types from the object’s having certain properties to our having representations of the object as having those properties. The non-deviant causal chain is the one that provides the best explanation of how our visual-perceptual systems ground this disposition. In the toad case, the causal chain begins with the object’s being in front of us. This causes the toad to then emit a poisonous gas which causes us to hallucinate that the toad is in front of us. This causal chain type does not provide the best explanation of how our visual systems ground our disposition, since they ground a disposition to represent objects veridically in a wide set of circumstances that characteristically do not involve toads. The causal chain type that provides the best explanation begins with the object being in front of us, reflecting light against our pupils, which then gets processed by our visual system to produce our visual experiences. Since, in the toad case, the visual experience of the toad is not produced by the processing of light by my visual system, the toad case is deviant. The main alteration I made in this case is a shift from talk of the agent performing steps in the causal chain to parts of the agent, in this case the agent’s visual system, also performing steps in the causal chain. This shift is natural, because we are now moving to processes that are sub- personal. In these cases, the grounds don’t ground the disposition merely by enabling the agent to perform steps in causal chain types; they also ground the disposition by enabling parts of the agent to perform steps in causal chain types. It remains true, though, that the non-deviant causal chain type is the one that, of these types, provides the best explanation of how the grounds ground the disposition. The other change was to drop talk of the disposition being reliable, which is only necessary to move us towards the final, fully general solution. We can still accommodate cases 61 where reliability is important by requiring, for those cases, that the relevant disposition be reliable. Here’s our second pass: Second (very slightly revised) Pass: A causal chain type is non-deviant just in case it is the causal chain type that provides the best explanation of how the grounds of the relevant disposition ground the disposition. Comparing my solution to other prominent solutions will help us see the advantages of my proposal. 86 In particular, it helps to consider the two most prominent approaches: the counterfactual approach and the functionalist approach. There are different counterfactual approaches to this DCC problem, but David Lewis’s (1980) is very representative and thoroughly developed, so I will focus on his. 87 On his approach, one counts as seeing just in case one’s visual experience matches and is caused by the relevant scene and there is a suitably robust counterfactual connection between scenes and matching visual experience. In my toad case, though one’s visual experience is caused by and matches the scene, there is no suitably robust counterfactual dependence of experience on scene: if the scene before my eyes had been different in certain ways, I would have still had the same hallucinatory experience. Christopher Peacocke (1979a) developed troubling counterexamples to counterfactual approaches, and Lewis’s view is subject to them. 88 If the toad before me only releases the poison if I am standing at a particular position relative to it and only when it is making a certain posture, then there could be robust counterfactual dependence even if I don’t see: if the scene had been different, the toad’s hallucinogen would have either not been released or would have missed, so my visual experiences would have been caused in the normal way. Lewis’s condition therefore is not sufficient. It is also not necessary: suppose the toad’s poison releases in a ring-like plume, so that, if I had been standing at a slightly different position, I would have hallucinated but, as I am 86 Though not a prominent solution, a suggestion Paul Coates makes in passing in Coates (2000) limited to the case of perception is the closest antecedent to my view. 87 More recent counterfactual approaches include Noë (2003) and McDonnell (2015). Tye (1982) and Jackson (1977) offer sensitivity strategies, which require systematic variation between scene and experience. These views are closely related to counterfactual views and face similar problems. See Campbell (2002) and Arstila and Pihlainen (2009) for some objections to these views. 88 So are the other counterfactual views mentioned in footnote 20. Though we didn’t discuss them, counterfactual solutions to the Gettier problem are well-known in epistemology. The most famous is from Nozick (1981). Safety conditions on knowledge, originally suggested in Sosa (1999), are another famous modal approach in epistemology. The worries mentioned here also apply to these solutions when they are made general. 62 standing, I don’t. I still see even though, if the scene had been altered slightly, I would have had non-matching visual experience. 89 Lewis’s view and other counterfactual views run into these problems because they don’t specify what causal chain must produce your visual experience for you to count as seeing. What matters is not whether there is robust counterfactual dependence in your scenario but how your experience is caused. Generally, if there is the relevant counterfactual dependence, your matching experiences will be caused in the right way, but the counterexamples show that this doesn’t always hold. The counterfactuals are only a marker of the correct causal chain. My proposal avoids these counterexamples because it gives us a way of picking out the causal chain that typically underlies the kind of counterfactual dependence Lewis is appealing to, which, if I am right, is the non-deviant causal chain. Usually, though not always, when this kind of counterfactual dependence obtains, it is because the agent has some wide-ranging disposition to represent objects accurately. My account extracts the right causal chain from this disposition itself instead of relying on counterfactual dependence directly. Carolyn Price’s (1996) functionalist solution to the problem does try to directly pick out the right causal chain. Functionalist approaches to the problem try to solve the problem by appealing to some device’s function. 90 For Price, the function of some device is whatever feature of its evolutionary ancestors accounted for their success, and the Normal way that they carry out this function is the way they went about fulfilling this function that explains their success. For our visual systems, their success is explained by their ability to enable us to represent middle-sized objects, and the way they did this is of course just the normal way we are all familiar with. So this is the causal chain through which objects must cause experiences for one to count as seeing them. One common problem for functionalist views is that a visual system can grant you sight even if it doesn’t have the relevant function. 91 For instance, a perceptual system built in a lab, which fails to have any function on Price’s account of a device’s function, can enable one to see. Another problem, particularly important for us, is that at least some functionalist solutions struggle to generalize to other DCC problems. Price’s is a good example of this. Price specifies a device’s function evolutionarily, but, for some DCC problems, the system that normally causes the relevant 89 Lewis responds to the necessity objection under a pseudonym in Le Catt (1982). He doesn’t respond to the sufficiency objection. 90 Other functionalist views include Davies (1983) and Enç (2003) 91 Both Davies and Price acknowledge the problem. Price bites the bullet. 63 event lacks an evolutionary function. This is especially vivid in the case of skillful performance. Many of our skills are not passed down to us from our ancestors through genetics, so they lack a function on Price’s account. Martin Davies’ (1983) functionalist account has features that enable it to avoid some of these problems. On Davies’ view, one sees just when one’s visual experience is produced by a mechanism whose function is to produce visual experiences that match one’s environment. His account avoids some of Price’s problems because he allows for different ways that something can become a mechanism’s function. He, for instance, allows that something can be a mechanism’s function because of its evolutionary role or because it is the way the mechanism has been used. Permitting multiple ways for a device to have a function enables the view to more successfully generalize, since, for instance, skills may now count as having functions. But this doesn’t address the most persistent problem for functionalists: devices without functions can enable seeing. The account also faces a new worry: it is disjunctive. In the absence of a unified account of having a function, it’s unclear what all these kinds of functions have in common that make them essential for ruling out deviancy. One is left wondering why all these disparate things referred to as functions play this shared role, and it is easy to start to suspect, given how different these functions are, that there must be some more fundamental explanation. My account gives us this more fundamental explanation without appealing to functions, thereby avoiding these difficulties. For all the different kinds of functions, when something has the relevant function, it is because it has a disposition to do that kind of thing. Devices acquire a use function when we use them to do something. But we of course use devices to do things because they are disposed to do that thing when we place them in the right stimulus conditions. Devices acquire an evolutionary function because they are passed down hereditarily because they did that thing. But of course, it’s not just because they did that thing, but because they were reliably disposed to do that thing that they get passed down. Functionalist accounts get close to the right solution by pointing us in the direction of the causal chain that best explains how the grounds of the disposition ground the disposition. But they err in trying to get us there through the notion of a function. Agents can see even if their visual systems have no function. And that is because, even if the systems have no function, they can still ground a disposition to represent objects accurately. 64 This completes my discussion of the DCC problem for seeing. There are two remaining cases that I wish to consider: the DCC problem for intentional action and for manifesting a disposition. I now turn to the former. DCC Problem for Intentional Action There’s a difference between doing something intentionally and doing it unintentionally. When I hug my mother because I’ve deliberated and decided I want to hug her, I hug her intentionally. When I hug her because of a convulsion, I hug her unintentionally. This contrast makes a certain analysis natural: one acts intentionally iff one’s act is caused by one’s intention. As you are probably expecting, this proposal immediately faces a DCC problem. If I form the intention to kill my mountain climbing partner and, as we are climbing, my realizing I have this intention causes me to nervously shake, so that I let go of the rope and thereby kill him, my intention causes my killing him, but I don’t intentionally kill him. 92 The causal chain is deviant. My proposal handles this case just as well as it handled the apt performance case. A mental state doesn’t count as an intention to perform an act unless the agent is disposed to perform the act in some set of circumstances upon acquiring the intention. Certain features of our brains and bodies, which it is the job of the special sciences to identify, ground this disposition. They do so by enabling our brains to perform certain steps in various causal chain types from some initial state (plausibly acquiring the intention itself) to the performance of the action. The non-deviant causal chain type is the causal chain type that, of these causal chain types, provides the best explanation of how the grounds of the disposition ground the disposition. Because I lack the scientific expertise which is plausibly necessary to adequately characterize this causal chain type, I cannot state many of its details here. But I know enough about it to explain why the rock climber case is deviant. Voluntary bodily movement is controlled by the somatic nervous system. So an adequately informative causal chain type will specify that our intention causes our behavior by causing certain motor signals to be sent through our somatic nervous system to the relevant muscle group, which then produces muscle contractions in the relevant body part. Since nervous shaking is produced by the sympathetic nervous system, and it is nervous shaking, not a motor signal being sent through the somatic nervous system, that causes the rock climber to let go of the rope, the rock climber case is deviant. 92 The case is from Davidson (1973). 65 Famously, though, not all cases of deviance are like this. The deviance, in this case, arises between the agent’s intention and a basic action, an action not performed by performing other actions. This is a case of basic deviance. In other cases, cases of non-basic deviance, the deviance occurs between a basic action and an intended result: an agent shoots at a person but misses and fires instead in the direction of a herd, which stampedes and kills the person the shooter intended to kill. 93 The agent intentionally shot the gun with the aim of thereby killing the person, but they did not intentionally kill the person. My account can explain why non-basic deviance cases are deviant. By definition, non- basic actions are actions that you perform by performing other actions. 94 Plausibly, then, the causal chain that best explains how certain features of our brains and bodies ground our disposition to perform non-basic acts we intend will include our performing actions that generate the non-basic action. But this is not all it will include, since this is not a particularly informative explanation. A more informative explanation will also point out that we perform these more basic actions as the realization of a plan we formulate in order to bring about the ultimately intended non-basic action. The best explanation of how certain features of our brains and bodies ground a disposition to perform non-basic acts we intend is that they enable us to effectively carry out a plan resulting in the ultimately intended act. So the non-deviant causal chain that these features enable us to perform steps in includes effectively carrying out one’s plan for performing the ultimately intended act. In Davidson’s stampede case, the agent does not effectively carry out his plan. His plan is to kill the agent by shooting him, but he ultimately kills him by causing a stampede. The causal chain is therefore deviant. The explanation I have given isn’t quite best: the best explanation cannot require that we carry out to a T the plan we formulate. After all, we are finite creatures in an environment that isn’t fully cooperative, so, if this were the best explanation of how these features grounded our disposition to act, our disposition to act would clearly be very narrow. So this explanation lacks adequate breadth: in lots of cases in which we act as a result of our intentions, we don’t carry out exactly the plan we intended. The best explanation of how the relevant features of our brains and bodies, given our finite nature and a not fully cooperative environment, ground our disposition to perform non-basic acts is that they enable us to carry out a sequence of acts that together 93 This case is from Davidson (1973). 94 Danto (1965) introduces the distinction. 66 approximately match our formulated plan. So, really, it is because Davidson’s shooter’s acts don’t even approximate his plan that he doesn’t act intentionally. This feature of my view is a virtue. Hector-Neri Castañeda (1979) drew our attention to the fact that acting intentionally does not require that we act exactly according to plan. To give a case like his, if I intend to kill someone by pressing a button that triggers a bomb with my finger, but my finger slips off and I press the button with the palm of my hand, I still intentionally kill the agent. Alfred Mele and Paul Moser have convincingly suggested that what is required is that the actual sequence of steps not diverge too far from the intended plan. 95 My account is in full agreement on this point. So my account provides us with a promising unified approach to ruling out cases of basic and non-basic deviance. This is significant. Many action theorists treat the two types of cases as raising distinct DCC problems, offering different solutions for the two types. This treatment has seemed attractive to them because the most promising way to rule out cases of non-basic deviance, requiring an approximate match between plan and actual performance, isn’t promising for cases of basic deviancy. 96 I don’t need to plan for how my intentions cause my basic movements to perform a basic action intentionally. To hold onto the progress for cases of non-basic deviance, these philosophers would rather invent two problems than look for a more general solution that entails the required match for the special case of non-basic deviance. Unsurprisingly, this approach struggles with cases of basic deviance. Myles Brand’s (1984) solution is a classic example of this approach. He rules out cases of non-basic deviance by imposing a requirement that one’s acts match one’s plans. To handle cases of basic deviance, he requires that intentions proximately cause the physiological chain leading to action. 97 Since, in Davidson’s climber case, the mental state of nervousness intervenes between the intention and the physiological chain leading to action, this case is ruled out. Unfortunately, as John Bishop (1989) has pointed out, this solution doesn’t work because we can find deviance in the physiological chain which the intention proximately causes. A case from a deeply insightful paper by David Pears (1975) makes this clear: Suppose an agent intends to kill someone and is about to carry out the act by pulling the trigger on his gun. Unfortunately, the nerve that connects 95 See Mele and Moser (1994). 96 See Schlosser (2007) for a discussion of this. 97 Mele (1992) endorses the strategy. Searle (1983) adopts a similar strategy. 67 his brain to his finger is severed. Fortunately, though, the buildup of charge at the point of separation attracts a lightning bolt that delivers the signal down his nervous system anyway. 98 In this case, the agent does not act intentionally even though the intention is the proximate cause of a physiological chain leading to the relevant behavior. Mele and Moser (1994) also adopt different solutions for the two types of cases of deviance. Their view nicely handles cases of non-basic deviance in part by requiring a degree of match between plan and actual performance. Their solution covers cases of basic deviance by requiring that intentions must sustain and guide the relevant behavior. 99 In Davidson’s climber case, because, at some point, the nervousness seems to cause the act independently of the working of the intention, it’s plausible that the intention either fails to sustain or fails to guide the behavior. To evaluate this proposal, we need to understand what guidance and sustenance require. Mele and Moser did not say much, but fortunately Mele (2003) later explained the sustenance and guidance conditions. 100 Unfortunately, though, it’s not plausible that Mele’s guidance condition is necessary for intentional action. On Mele’s account, one’s behavior is guided by one’s intention when certain feedback loops that promote the execution of the behavior are present. This condition ruled out Davidson’s nervousness case because such loops are not present in that case. But these feedback loops are also not present in some cases of intentional action. As discussed by Erasmus Mayr (2011), these feedback loops are facilitated by the central nervous system and can be eliminated by numbing drugs like anesthesia and disabilities that affect the central nervous system. 101 Yet intentional action is still possible for agents with these disabilities and under the influence of these drugs. These approaches are paradigmatic examples of ruling out cases of basic and non-basic deviance by treating them as instances of distinct problems in need of different solutions. This strategy was only motivated by the recognition that a promising solution for cases of non-basic deviance didn’t apply to cases of basic deviance. This makes the strategy look ad hoc. These 98 Peacocke (1979b) and Bishop (1989) claim that the shooter acts intentionally. Enç (2003) registers his agreement with Pears’ intuition. I agree with Enç and Pears, as do philosophers and non-philosophers I have probed. It’s noteworthy that Peacocke accepts that the shot is intentional “with serious unease” after giving an argument, which suggests he’s biting the bullet rather than reporting his intuition. 99 Alston (1986) also imposes these conditions. 100 The basics of this account were in Mele and Moser (1994), but they didn’t clearly explain how it could systematically address Davidson-style cases. 101 Mayr also pointed out that these feedback loops are not sufficient for non-deviance: they are present in some deviant cases. 68 philosophers’ solution to the DCC problem for intentional action was unsuccessful in cases involving basic deviance, so, without providing independent reason to think there are two problems, they posited that there really are two problems and tried to impose further conditions to rule out the cases for which their solution was unsuccessful. It’s natural to think that this ad hoc maneuver is what got them in trouble. So we should look at views that don’t make this maneuver. We should look at views that treat the two kinds of deviance as but two instances of a single problem and offer a solution to both instances. Because there’s already a solution that can rule out cases of non-basic deviance, a tempting way to develop such a view is to try to extend this solution to cases of basic deviance. This is the strategy Gilbert Harman (1976) has adopted. He has argued that intentions have self-referential content: part of an intention’s content is that that very intention lead to the relevant act in the normal way. Since, in cases of basic deviance, the intention does not produce the loosening of the rope in the normal way, there isn’t a match between the content of the intention and the events that unfold. Thus, the loosening diverges from the agent’s plan, and the behavior is deviant for the same reason it is deviant in cases of non-basic deviance. The problem with this solution, as Mele has pointed out, is that it makes intentional action way too intellectually demanding: acting intentionally, on this view, requires having complex beliefs about one’s own mental states. A more recent example of this strategy comes from Marcus Schlosser (2007). Schlosser suggests that, in Davidson’s famous case, the intention does not cause the act in virtue of content, since the nervousness is not a contentful state. He then considers the objection that, at the subpersonal level, our neurological states are not caused by intentions in virtue of their content and replies that this confuses different levels of explanation. 102 Causation in virtue of content need only occur at the psychological level of explanation. The problem with this is of course that, as Pears’ lightning case shows, we can generate deviance at the neurophysiological level, and so, since we don’t have causation in virtue of content at this level, Schlosser’s solution will not rule out these cases. These two approaches to the DCC problem for intentional action encounter these difficulties because they fail to face up to what cases of basic deviance tell us about the strategy used to rule out cases of non-basic deviance: there is a more fundamental source of deviancy than a mismatch between content and behavior. That these approaches have failed to identify the 102 He owes the point to Wedgwood (2006). See Tännsjö (2009) for a different objection. Schlosser (2010) replies. 69 fundamental source of deviancy becomes especially clear when we look at how hopeless they are at generalizing to the other problems we’ve discussed. Plans and guidance are completely irrelevant to, for instance, solving the DCC problem for seeing. Proximate causation also won’t help, since objects don’t proximately cause the physiological chains that produce our visual experiences of them. And objects don’t cause our experiences of them in virtue of their content, because ordinary objects don’t have propositional content. Unlike the solutions we’ve seen so far, my solution offers a promising way of ruling out basic deviance, entails the conditions others use to rule out non-basic deviance, and works well for other instances of the problem. It’s obvious why my solution is more promising than these solutions. What we need is not a solution that treats the two cases as distinct problems or tries to force the plan-act match requirement to do work it clearly cannot do, but a general solution that entails the requirement for the special case of non-basic deviance. We need a view that recognizes that there is a more fundamental source of deviancy. That is exactly what my solution does. 103 Manifesting a Disposition I have now explained how my solution differentiates deviant from non-deviant cases for four different DCC problems. If the solution works as well for these cases as I’ve suggested, we should be optimistic that this solution can serve as a general solution to the problem. But we shouldn’t just settle for optimism. We should try to evaluate this solution’s prospects for offering a fully general solution to the problem. And the first step to evaluating these prospects is understanding why this solution works in these cases. We can then think about what would have to be true in general of instances of the DCC problem for this solution to be fully general. To appreciate why my solution worked for these four instances, we must first notice that the deviance arises in each instance when some disposition fails to be manifested. In both the non- deviant archery case and the deviant archery case, the agent’s hitting the bullseye is due to his skill, but only in the non-deviant case does he manifest his skill. In the knowledge case and in 103 Berent Enç develops a more general solution in Enç (2003 and 2004). Unsurprisingly, his solution handles cases of basic deviance better. Unfortunately, his solution is functionalist in nature and runs into some of the problems that plague those solutions. And, for reasons I lack the space to explain, his solution struggles with cases of non-basic deviance. Sensitivity strategies are also more general. Morton (1975) and Peacocke (1979a and 1979b) are the earliest examples of the strategy. Peacocke acknowledges that his conditions aren’t sufficient for non-deviance, so he adds further conditions for individual DCC problems. Bishop (1989) and Shope (1991) offer variations of the sensitivity strategy which aim to avoid these difficulties. See Mayr (2011) for pressing objections to both the necessity and sufficiency of sensitivity conditions. A more recent solution comes from Shepherd (2021). Although I lack the space to evaluate this view here, it is worth noting that the solution Shepherd provides can only work for the special case of intentional action, as it makes essential reference to planning states. 70 Gettier cases, true belief is explained by cognitive ability, but only the non-deviant true belief manifests the agent’s ability. In seeing cases and the deviant hallucinatory cases, the agent might have a representational experience of an object because of a disposition to have those experiences when looking at the object (the toad’s poisonous gas may work only on beings with such a disposition), but only in the seeing case is this disposition manifested. For intention, we can understand both killings as being caused by the disposition (after all, it is because having an intention to kill disposes one to kill that the rock climber is nervous) but only the killing that is caused non-deviantly manifests this disposition. This is interesting because analyses of manifesting a disposition are also susceptible to DCC problems. In an illuminating discussion of the DCC problem for intentional action, Ralf Stoecker (2003) pointed out that a rotten wooden bridge that is collapsed by engineers tasked with removing rotten bridges collapses because of the disposition to collapse typical of rotten bridges, but it doesn’t thereby manifest that disposition. 104 In a more recent and well-known paper, John Turri (2011) contrasts a case in which a microwave boils water because its magnetron releases microwaves that travel through the water in its compartment, thereby boiling water, with a case in which the magnetron heats an improperly insulated wire in the control circuit to the point of catching fire, which then travels to the compartment and heats the water to boiling. 105 Only in the first case is the boiling power of the microwave manifested. So the deviance in the DCC problems we discussed arises when some disposition fails to be manifested. And disposition manifestation itself faces a DCC problem. We should not treat this as a coincidence. It’s more plausible, instead, that the second fact explains the first: the DCC problems we discussed are instances of the more general DCC problem for manifesting a disposition. So all of the instances of the DCC problem for which my solution, I hope, was successful were instances of a more general DCC problem. A natural explanation for why my solution succeeds for multiple instances of a more general problem is that it provides a solution to the general problem: it provides us with a way to analyze disposition manifestation. If my solution really does offer such an analysis, it’s clear how general my solution will be: my solution will 104 Stoecker (2003) and Stout (1996 and 2005) appeal to concepts similar to disposition manifestation to address the DCC problem for intentional action. 105 Turri thinks we can make use of disposition manifestation to rule out deviance and leave it to metaphysicians to understand disposition manifestation. Sosa (2015), Hyman (2014), Wedgwood (2006) and Mantel (2017) use the notion of manifesting a disposition to rule out deviance. 71 cover any DCC problem where disposition manifestation can be invoked to distinguish non- deviant cases from deviant cases. If it can be invoked to distinguish these cases for every instance of the DCC problem, then my solution is fully general. As I will now attempt to show, my solution offers us an analysis of disposition manifestation, so it will indeed solve any DCC problem that can be solved by invoking the notion of manifesting a disposition. The solution I offered previously, revised to refer to objects generally rather than just agents, offers a straightforward analysis of disposition manifestation: Disposition Manifestation: An act manifests an object’s disposition to so act in some set of circumstances iff the act is non-deviantly explained by the disposition. An act is non-deviantly explained by a disposition iff the act is the result of a causal chain token that is a member of the causal chain type that best explains how the grounds of the disposition ground the disposition. 106 The grounds of the disposition to boil in Turri’s microwave case are the geometry of the inside of the microwave and the magnetron. The best explanation of how these properties ground the disposition to boil is not that they enable the microwave to burn a wire in its control circuit, which then catches fire, traveling to the compartment and boiling the water. Enabling the microwave to perform steps in this causal chain wouldn’t give it the disposition to boil water in the wide-ranging set of circumstances in which it has this disposition because, in many of those circumstances, the wiring of the control circuit is properly insulated. The microwave does not lose the relevant disposition when it goes to an electronics repair shop. The best explanation is that the grounds enable the microwave to send microwaves into its compartment, which then penetrate the water and excite it to the point of boiling. This is the non-deviant causal chain. Therefore, the microwave only manifests its disposition to boil water when it sends microwaves through the water and thereby excites the water into boiling, as fails to occur in Turri’s deviant case. Objections and Replies Objection 1: You mentioned that Greco’s original solution faced a famous objection from Lackey, but Ian Church (2013) raised a parallel dilemma for solutions that invoke disposition manifestation. Ability manifestation comes in degrees. If you require that the agent manifest lots 106 To be clear, I am not here analyzing dispositions. I am instead analyzing what it takes to manifest a disposition. For the purposes of this paper, I am happy to treat dispositions as a primitive. I should note, though, that the manifestation relation I am here discussing is different from what is often referred to as a disposition’s manifestation in the disposition literature, which is just the event that the object is disposed to bring about. We must be careful to appreciate that, even if a certain event is a disposition’s manifestation in this sense, an object with that disposition may bring about that event while failing to manifest the relevant disposition. 72 of ability, then you avoid Gettier problems, but you make knowledge too hard to come by, since everyday knowledge often does not manifest tremendous ability. If you relax how much ability must be manifested, you avoid this at the cost of readmitting Gettier cases. Reply: This objection equivocates on two different ways ability manifestation can come in degrees: we can manifest to different degrees a certain amount of ability or we can manifest different degrees of ability. The solution doesn’t avoid Gettier problems by requiring that the agent manifest tremendous abilities. It requires that we fully manifest whatever abilities we exercise. Hawks may have better vision than humans, but both can see if they fully manifest their visual abilities. Objection 2: Your solution assumes uniqueness in two places. It assumes that there is a unique set of properties that count as the grounds of the disposition and a unique causal chain type that provides the best explanation of how the grounds ground the relevant disposition. But what if there’s overdetermination? Reply: Overdetermination by grounds does not require any revision to the view. If two distinct sets of properties determine the disposition through two causal chain types, then the one causal chain type provides the best explanation of how the one set of grounds grounds the disposition and the other provides the best explanation of how the other set grounds the disposition. So one can manifest the disposition through either causal chain type. And if two distinct sets of grounds ground the disposition through the same causal chain type, that causal chain type provides the best explanation for both, so the disposition is manifested through that causal chain type. Overdetermination by causal chain type requires a slight modification of the view. To handle cases where two causal chain types provide equally good explanations of how the grounds of the disposition ground the disposition, we need to restate the view to say that a non- deviant causal chain type is a causal chain type that provides the best explanation of how the grounds of the disposition ground the disposition. Since more than one causal chain type can be tied for best, one manifests the disposition when one produces the act through one of the causal chain types tied for best. Objection 3: Suppose that, at some other possible world, certain features of our brains and bodies ground the relevant disposition to act as we intend by enabling the agent to perform steps 73 in a causal chain that, at our world, seems deviant. Then, on your account, when they perform acts through this causal chain at these worlds, they act intentionally. But this is implausible. These agents don’t act intentionally just because, at their worlds, the best explanation of how the grounds ground the disposition is that they enable the agent to perform steps in this weird causal chain. Reply: It’s not clear that, if we fill in the details of the case, my view’s predictions will conflict with intuition. After all, we know that, at our world, strange causal chains involving prosthetic devices can enable agents to see and act intentionally, so long as the prosthetics help ground a wide-ranging disposition. 107 So, unless the details of these strange worlds can be filled in in a way that makes it clear that my view faces a problem, I’m inclined to just accept that such agents act intentionally. But even if such a problem case were developed, my view still has resources to address it: if the world is sufficiently strange, it may be that the agent isn’t disposed to act in the kinds of circumstances we are disposed to act in, because the possible circumstances at her world are radically different from ours. Though the agent may have a wide-ranging disposition to act, it might be a disposition to act in a different set of circumstances than those we are disposed to act in. If acting intentionally requires manifesting the disposition to act in circumstances like those we act in, agents at these worlds can’t act intentionally, so there’s no problem for my view. Conclusion In Paul Grice’s (1961) classic discussion of the deviant causal chain problem for perception, he suggested that philosophers should leave identifying the right causal chain to the scientific experts. Natural though this proposal was, it ran into the problem that different causal chains may underly different organisms’ vision. We therefore need something unifying to say about these right causal chains, and this looks like a task for philosophers. My solution accomplishes this task: what makes each causal chain right is that each best explains how the grounds of some object’s disposition ground that object’s disposition. But it also vindicates Grice: it is the job of the scientist to uncover, for each of these objects, the causal chain that best explains how the grounds of its disposition ground its disposition. I hope to have shown how this solution succeeds for four famous instances of the DCC problem. This is no guarantee that it will solve every instance of the problem. But it should, 107 See Enç (2003). 74 minimally, encourage us to check carefully whether individual instances are amenable to this treatment. We may find that giving this treatment is more challenging in some instances. If we do, we should face up to these challenges. The potential payoff is a fully general solution to the problem of deviant causal chains. 75 Chapter 4: Overcoming Difficulties with Virtue My aim in this paper is to provide a novel solution to one of the most difficult problems virtue epistemology has faced: Carl Ginet’s fake barn problem. This problem has seemed so intractable for virtue epistemologists because it seems to force them to deny that knowledge is creditworthy true belief, which is a core commitment of virtue epistemology. I will argue that virtue epistemologists have struggled with this problem because they have been working with an implausible account of virtue and, consequently, creditworthiness. Overcoming this difficulty will require accepting a new account of virtue: virtues are traits that overcome characteristic difficulties. A potential further benefit of accepting this account of virtue is that it may allow virtue epistemologists to avoid an important problem they’ve faced in extending their approach to creditworthy performance to moral worth. Background: The Origins of Virtue Epistemology Ernest Sosa famously drew attention to how our appraisals of agents’ performances differ depending on whether those performances are in certain ways accidental. We appraise the shot of the archer who hits the bullseye under normal conditions differently from the shot that is hit as a result of the intervention of wind. 108 One natural way of describing the difference is in terms of creditability or, as I prefer, creditworthiness. The agent who hits the shot normally deserves credit for the shot whereas the agent who hits it accidentally is undeserving. This pattern of appraisal arises in lots of different domains. Correspondingly, we can describe lots of different performances as creditworthy or not creditworthy. Musical performances, checkmates, cake-baking, and weightlifting are all types of performances for which agents can or can fail to be creditworthy. Even right action can be creditworthy. It is plausible that acting with moral worth just is being creditworthy for right action. 109 Most importantly for Sosa, and for those philosophers who followed him, agents can or can fail to be creditworthy for true belief. I deserve credit for forming the true belief that the number of stars in the night sky is currently 23 if I form this belief by a careful count. I don’t deserve credit if I form the belief by closing my eyes and guessing. These differences in 108 Sosa (2003) 109 This view is very popular in the moral worth literature. Lord (2018), Way(2017), Singh (2020), Rozeboom (2017), and Sliwa (2015) are examples of philosophers who hold this view. 76 creditworthiness seem to track differences in knowledge: I know that the number of stars is 23 in the first case but not in the second case. Consequently, Sosa and others who followed him have identified knowledge with creditworthy true belief. 110 This identification comes with a commitment: the correct account of knowledge must fall out of the correct account of creditworthy performance in general. The safest way to discharge this commitment is to develop one’s account of knowledge by developing one’s account of creditworthy performance. This is exactly what Sosa did. He noted that the common factor in creditworthy performance seems to be that the success is the result of virtues related to the activity that the agent possesses. The credit-deserving archer hits the bullseye because of his skill at archery. The credit-deserving true believer believes truly because of his cognitive abilities. So he identified knowledge with true belief due to virtue. Other philosophers followed suit, and this family of views has, accordingly, been called virtue epistemology. 111 All virtue epistemologists agree that knowledge is true belief that is due in some way to virtue. But different virtue epistemologists give different accounts of virtue. Which account they give depends on what attracts them to virtue epistemology. Linda Zagzebski, for instance, is drawn to virtue epistemology in part because she thinks virtue epistemology is needed to explain why knowledge is more valuable than true belief. 112 Because she thinks the problem is more persistent than even many virtue epistemologists have thought, she thinks we need a very rich account of virtue, one inspired by moral virtue, if we are to solve this problem. Her account of virtue, in turn, incorporates qualities that reflect on the personal worth of the agent, and she thinks agents must be motivated by a love of the truth to have knowledge. By contrast, Sosa is drawn to virtue epistemology by the striking parallels between knowledge and other creditworthy performances. 113 Because it’s plausible to him that what makes an act creditworthy is that it is related in the right kind of way to virtue, he consequently accepts a virtue epistemological view. But his account of virtue is, as a result of what attracts him to virtue epistemology, much more minimalistic than Zagzebski’s. For him, virtues are just the features of 110 Greco (2003) and Riggs (2009) are prominent virtue epistemologists who followed Sosa in making this identification. Sosa makes this identification in Sosa (2003). 111 Prominent defenses of virtue epistemology include Sosa (2015), Greco (2010 and 2012), and Riggs (2009). 112 See Zagzebski (1996) 113 See Sosa (2015) for an important discussion of the relationship between reliabilist virtue epistemology and responsibilist virtue epistemology. 77 agents that relate to their performance in such a way as to make them creditworthy. Skill at archery would be, on this account, a kind of virtue. Since skill at archery does not have much in common with moral virtue, does not reflect on the personal worth of the agent, and doesn’t involve a motivation to pursue bullseyes, Sosa’s virtues are nowhere near as demanding as Zagzebski’s. And neither are the virtues of those who have followed Sosa. Indeed, those who have been drawn to virtue epistemology for the same reason that Sosa has been drawn to virtue epistemology have all converged on what is possibly the most minimalistic account possible: virtues are just reliable dispositions to perform the relevant act. They are known as reliabilist virtue epistemologists. Because I am, in this paper, trying to identify what conception of virtue falls out of the account of knowledge as creditworthy true belief, my critical comments are directed towards this family of views. Just as different virtue epistemologists give different accounts of virtue, they also give different accounts of the relation the virtues have to stand in to the true belief in order for it to be knowledge. Some virtue epistemologists think that manifestation is the relevant relation, where this is understood as a primitive metaphysical relation. 114 Others think that the virtues have to be the most salient cause of the true belief. 115 It’s not important for the purpose of pressing the fake barn problem what exactly the relation is. This is because the objection to reliabilist virtue epistemology arises not because of the relation between virtue and true belief but because of the nature of virtue that the virtue epistemologists whose views we are interested in accept. It arises because these virtue epistemologists accept the view that virtue is just a reliable disposition. Background: The Fake Barn Problem With this in mind, it’s not difficult to see why fake barn cases create trouble for reliabilist virtue epistemology. Suppose I am on the side of the road, standing before what looks to be a row of barns. As it turns out, all but one are highly accurate holograms of barns. But I turn out to be looking at the one barn, and I consequently form the belief that there’s a barn in front of me, relying on all the normal mechanisms I use to form beliefs about what things are in front of me. In that case my belief is true and due to a reliable disposition, namely my ordinary visual and judgmental 114 Turri (2011) holds this view. 115 Greco (2010) defends this view. He later endorsed a revised view in Greco (2012). 78 capacities, but it doesn’t seem like I have knowledge. Hence, fake barn cases seem to be a counterexample to reliabilist virtue epistemology. This, however, doesn’t explain why these cases are especially troubling for reliabilist virtue epistemologists, since fake barn cases are a problem for almost all theories of knowledge. What makes them so troubling for virtue epistemologists is that they seem to show, as Duncan Pritchard has argued, that knowledge is not just an instance of creditworthy performance. Pritchard imagines a case in which an archer hits a bullseye in the normal way, but all around him are other targets that have forcefields in front of them. 116 If he had shot at any of those, he would not have hit the target. In this case, it still seems like he is creditworthy for hitting the shot: the shot manifests the virtues of archery. But, if this is right, since the case is analogous to the fake barn case, it seems to follow that Barney’s belief that there’s a barn in front of him manifests his intellectual virtue and is creditworthy. But of course Barney doesn’t know there’s a barn in front of him, so it seems knowledge is not creditworthy true belief. Replies to the Fake Barn Problem The two most popular kinds of solutions virtue epistemologists have offered to this problem are to impose some sort of safety condition on knowledge or to make reliability in one’s epistemic environment a condition on manifesting virtue. Sosa is an example of the former strategy: he denies Barney knowledge because Barney’s belief is not meta-apt. It lacks meta-aptness because Barney is not sensitive to risk in forming his belief, and so his belief is unsafe. 117 An example of the latter strategy is provided by Clayton Littlejohn. 118 He argues that, in forming beliefs about objects in our environment, the ability that constitutes virtue is an ability to discern those objects. Since, in fake barn country, you aren’t reliable at discerning fake barns, you don’t manifest this virtue in this environment. The problem with these solutions is that they seem to commit virtue epistemologists to rejecting the view that knowledge is creditworthy true belief, because they impose conditions on knowledge that don’t generalize to other creditworthy performances. Pritchard’s archer’s shot is 116 Pritchard (2009) 117 See Sosa (2015). Other examples of this strategy include Pritchard (2009), Turri (2016), and Pavese and Beddor (2020). Kelp et al. (2017) offer a variation of this strategy, but they ultimately conclude the account doesn’t explain why Barney fails to know. 118 See Littlejohn (2014). Greco (2010) and Berrocal (2017) are further examples of this strategy. 79 neither safe nor an exercise of an ability that is reliable in his environment, but, contrary to what some virtue epistemologists’ bullet biting would have you think, he is creditworthy. 119 So, the conditions that virtue epistemologists have imposed on knowledge don’t seem to hold for creditworthiness in general. Consequently, these theorists seem committed to holding either that knowledge is more than mere creditworthy true belief or that it simply is not creditworthy true belief. 120 But if we accept either of these, we are forced to give up what made virtue epistemology appealing in the first place. To start, one of the things that made virtue epistemology seem so compelling were the striking parallels between knowledge and other creditworthy performances. This made virtue epistemology attractive because it aimed to explain these parallels by treating knowledge as an instance of creditworthy success. But, on these revised views, knowledge isn’t really parallel to creditworthy success; it’s at best parallel to creditworthy success plus some further conditions. Consequently, it’s no longer clear the parallels Sosa observed lend all that much support to virtue epistemology. Theorists who adopt these responses, in effect, concede that there are limits to the parallels, thereby undermining their force. More importantly, though, giving up on the idea that knowledge is mere creditworthy true belief undermines virtue epistemology’s ability to elegantly solve the value problem in epistemology. The problem is to explain what makes knowledge more valuable than true belief and justified true belief that P. Knowledge that P is more valuable than true belief that P and merely justified true belief that P, even though a true belief that P and a merely justified true belief that P can have important practical benefits. This is a fact that needs to be explained. But, as highlighted by Zagzebski, there is a difficulty we run into in giving such an explanation that makes the problem even tougher. It’s not enough to appeal to the origins of the true belief in the case of knowledge to show that it is more valuable than true belief and justified true belief. This is because something’s origins don’t always make it more valuable. A good cup of coffee has the same value whether it comes from a reliable coffee machine or an unreliable one. So, if an account does try to explain the superior value of knowledge by appeal to its origins, it will have to be able to say something more. 119 Littlejohn (2014) and Berrocal (2017) bite the bullet. 120 Pritchard openly accepts that knowledge is more than creditworthy true belief. 80 Virtue epistemology has an elegant answer to the problem of explaining the value of knowledge, an answer that seems to successfully address Zagzebski’s concern for certain kinds of answers. Knowledge just is creditworthy true belief according to the virtue epistemologist. And creditworthy performance seems to be a status with distinctive normative significance. We value creditworthy performances over and above mere successful performances. This was pointed out by Aristotle. Having virtues can be valuable, as can be having successes. But we value success due to virtue above both these things. The life of activity in accordance with virtue is better than a life of mere virtue, which can be spent asleep. And a life of success not in accordance with virtue isn’t as valuable as a life of success in accordance with virtue because it is a life whose success is due to luck. We value actions that can rise to the level of full-blooded achievements. We can appeal to this fact to explain why knowledge is more valuable than true belief and justified true belief. All else equal, a merely justified true belief or true belief will never be quite as valuable as knowledge, because what distinguishes knowledge from justified true belief is that knowledge is fully creditworthy, whereas mere justified true belief involves some luck. An agent deserves more of the credit for forming a true belief when it is knowledge than when it is a merely justified true belief. Because we value fully creditworthy performances, the piece of knowledge has more value than the justified true belief. This solution also addresses Zagzebski’s worry. Something’s being an achievement or creditworthy depends on its extrinsic qualities, but in a way that clearly affects the final value of the product. So we’ve made sure that our explanation doesn’t put knowledge in the same category as good cups of coffee. On this account of the value of knowledge, what makes knowledge distinctively more valuable than true belief and justified true belief is that it is creditworthy. But, on the new accounts we have considered, some merely justified true beliefs, like Barney’s, are creditworthy. So, given this explanation of the value of knowledge and these accounts, Barney’s belief has the value that knowledge has. But this is implausible! Barney’s belief lacks the value that distinguishes knowledge from true belief and mere justified true beliefs. So, if we accept these responses, we must look elsewhere to explain what makes knowledge more valuable. Diagnosing the Problem with these Replies Pritchard’s argument shows that reliabilist virtue epistemologists must either concede that knowledge isn’t creditworthy true belief or that the virtues relevant for creditworthiness in general 81 aren’t mere reliable dispositions. Rather than directly face this choice, most virtue epistemologists have tried to revise their account of knowledge to avoid the fake barn problem. The consequence of ignoring the dilemma has been that they’ve adopted accounts of knowledge that seem incompatible with identifying knowledge with creditworthy true belief. Their revisions have inadvertently committed them to the first horn of the dilemma. Upon realizing this, some have retroactively tried to preserve the identification of knowledge with creditworthy true belief by suggesting that their account of knowledge generalizes to other creditworthy performances, but it has been pretty clear that their accounts cannot: these accounts do not even predict that the archer in Pritchard’s forcefield case is creditworthy. 121 I think the mistake virtue epistemologists have made is to be so narrowly focused on epistemology. This move makes sense if the lesson of Pritchard’s argument is to give up on the identification of knowledge with creditworthy true belief, as Pritchard believes. In that case, we need not worry whether our account generalizes, and so there is no risk in theorizing narrowly about knowledge. But, if the lesson of his argument is that we’ve been working with an inadequate general account of virtue and creditworthiness, we should be focusing on the nature of creditworthiness and virtue in general so that we can construct a more adequate general account. On reflection, it seems clear that the right lesson to take away is that we’ve been working with an inadequate general account of creditworthiness and virtue. Indeed, it’s a bit strange to draw the conclusion, as Pritchard does, that knowledge isn’t creditworthy true belief from Barney’s case, because, intuitively, Barney doesn’t seem creditworthy for his belief. His belief seems lucky, not skillful. By contrast, the archer’s shot seems skillful, not lucky, and so is creditworthy. The problem is with the account of virtue, not knowledge. That this is right becomes even more clear once we appreciate that we can generate fake barn-style cases for non-epistemic performances. Giving up on the identification of knowledge with creditworthy performance does nothing to help with these cases. With a little background on the table, it becomes clear that the sport of boxing provides such cases. In boxing, many skilled boxers are famously good at setting up traps for their opponents. These boxers will purposely let 121 Sosa is among the philosophers who, after recognizing that his account of knowledge seems to be in tension with the account of knowledge as creditworthy true belief, later suggested that the conditions he sets on knowledge hold in general for creditworthy performance. For sharp criticisms of his suggestion, see Sylvan (2020). 82 their guard down in certain ways to invite their opponent to try to attack the opening. When the opponent tries to do so, they are ready to slip or dodge the punch and offer a return punch, which is known as a counterpunch. In turn, an important skill in boxing is the ability to differentiate traps from non-traps. Given this background, now consider a case where boxer A is going up against boxer B, who is known for setting up lots of traps. Suppose that, at some point in the fight, frustrated by his inability to detect when boxer B is setting up a trap, boxer A decides to just throw his punch at the next apparent opening. As it happens, this is a genuine opening, and boxer A knocks out his opponent. Boxer A, in this case, definitely deserves some credit for landing the punch. After all, if boxer B is a good enough boxer to set up traps, landing a punch on him probably required speed, accuracy, a good grasp of distance, and power. But, I submit, boxer A isn’t fully creditworthy for landing the punch; it’s a bit lucky. Notice, though, that the boxer does manifest a reliable ability to land punches: speed, accuracy, and understanding of distance do reliably dispose one to land punches, provided one is facing an opponent who does not set up traps (which is most people, since this is an ability that requires extensive training). And so, facing yet another counterexample, we might try, as reliabilist virtue epistemologists did when confronted with fake barn cases, to explain the inadequacy of boxer A’s punch in terms of its being unsafe or not being an exercise of an ability that is reliable in that environment. This would run us once again into the problem raised by Pritchard’s forcefield case. But since the disanalogy here is uncontroversially between two cases of creditworthy performance rather than between knowledge and creditworthiness, the way out that Pritchard suggested, to deny that knowledge is creditworthy true belief, is no longer available. The problem Pritchard identified, then, was all along a problem with reliabilist virtue epistemologists’ account of virtue. We need an account of virtue and creditworthiness that explains why the abilities that one needs to exercise to be creditworthy in the boxing and fake barn case differ from those in the archery case. The most natural way to try to discover such an account is to return to the starting point from which Ernest Sosa developed virtue epistemology in the first place. Rather than narrowly focus on epistemology, we should focus on theorizing about creditworthiness and virtue in general. Once we do so, as I will now argue, it becomes clear that the reliabilist account of virtue was misguided all along: there are problems for the account that are totally 83 independent of the fake barn problem. But, fortunately, these problems point in the direction of a better account of virtue and creditworthiness, one that solves the fake barn problem. Problems for Virtue Epistemologists’ Account of Virtue To start to see the inadequacies of a reliabilist account of virtue, consider the following case. A soldier is afraid of fighting in war. He doesn’t want to die, and this desire not to die overwhelms him when he enters combat. He is scheduled to participate in a battle. His commanding officer, recognizing that he will freeze up in combat because of his fear, convinces him that some special potion the officer has concocted will prevent him from dying. As a result of drinking the potion, which is really water, the soldier willingly fights in the battle. In the battle, he successfully takes the enemy’s bunker. The soldier certainly deserves credit for taking the bunker. In doing so, he manifests a variety of the virtues of soldiery, like skill in combat and respect for the chain of command, but he doesn’t deserve full credit for taking the enemy bunker. This is because his taking the bunker is insufficiently due to him; it’s a bit lucky. By contrast, his peers who take the bunker in the full knowledge of the risk of death deserve full credit for what they do. But notice that each of these parties has as reliable a disposition as the other. The disposition is grounded in different features of the agent. The water drinker’s disposition is grounded in different characteristics of him, but they make for the same reliability. We can come up with another case like this. Suppose some agents are participating in some competition where they try to eat some disgusting food the fastest. Now suppose, for whatever reason, one of the competitors was born with an ability to turn off his taste buds. This competitor wins by turning off his taste buds. Intuitively, this competitor does not deserve as much credit for winning the competition as he would deserve if he had tasted the disgusting food and ate the food as quickly as a result of a tolerance for the taste. He deserves some credit, since he displays one of the important virtues, namely his ability to eat fast, but he doesn’t deserve full credit for what he did. And this is so despite the fact that his ability to turn off his taste buds gives him as reliable a disposition to eat the food as a tolerance for eating the food would give him. A different problem is posed by the case of moral worth. It is widely accepted in the moral worth literature that to act with moral worth is to deserve credit for one’s right act. 122 So, it would 122 See footnote 2 for some supporters of this view. 84 be nice if the virtue theoretic approach to creditworthiness could be extended to the case of moral worth. On the reliabilist virtue epistemological account of virtue, virtue is just a reliable disposition. So, extending this account to moral worth, it would be enough to have the virtue related to doing the right thing to be reliably disposed to do the right thing. This suggests the following schematic view of what it takes to act with moral worth: An agent’s doing the right thing is morally worthy just in case it is related in the right way to a reliable disposition to do the right thing. But this view is implausible. It has the result that agents like Kant’s shopkeeper, the paradigmatic instance of an agent who fails to act with moral worth, can act with moral worth if their instrumental desires line up more systematically with doing the right thing. After all, if the shopkeeper’s non-moral desires line up conveniently with morality systematically, so that he reliably does the right thing, he has a reliable disposition to do the right thing. But, certainly, he doesn’t act with moral worth. This last problem isn’t quite as serious as the problem raised by the first two cases. Though virtue epistemologists and philosophers working on moral worth have described their subject matter similarly, and though there are interesting parallels between the phenomena, it may turn out that there are important differences in their objects of study. Indeed, perhaps that is why there has been little contact between the two literatures. But I submit that, given the parallels between the domains, it would be an attractive bonus of an account of virtue if it held some promise of extending the virtue theoretic approach to credit to moral worth. The reliabilist approach clearly cannot secure this bonus. We don’t need the language of creditworthiness to raise the problems these cases present. We can state the problem just in terms of virtue, a concept more familiar to philosophers. The point is that one of the virtues of soldiery is a willingness to fight in spite of the knowledge that one is risking death. The mistaken belief that you don’t face a risk of death is not, by contrast, one of the virtues of soldiery. I suspect that any plausible account of virtue must make sense of this. Good soldiers overcome their knowledge of possible death; they don’t believe they are impervious to death when they aren’t. Similarly, tolerance for disgusting foods is one of the virtues of the kind of competition I described whereas an ability to turn off your taste buds is not. And having certain moral desires is a part of moral virtue whereas having instrumental desires that conveniently line 85 up with what it is right to do is not. Any plausible account of virtue must make sense of these facts about virtue. But the account of virtue as reliable dispositions cannot do so; it is a too coarse- grained account. Having a false belief that you won’t die makes you every bit as reliable at taking a bunker as being willing to fight in spite of the risk of death. Being able to turn off your taste buds makes you as reliable as tolerance for disgusting food. But these qualities are not every bit as much a virtue of the relevant activity. An Alternative Account of Virtue Fortunately, these cases point us in the direction of a better account of virtue. The first thing we must do to get to this better account is accept a different account of what kinds of properties constitute virtue. On the traditional account, virtue is constituted by a reliable disposition to succeed at the activity. But my cases show the properties need to be more fine- grained than this. The most obvious more fine-grained properties to identify with virtue are the grounds of the reliable disposition, so I propose that we identify virtue with some (but not all of) the grounds of reliable dispositions. You count as having virtue with respect to an activity when you have enough of certain properties that ground a reliable disposition. We can call these grounding properties individual virtues. So, for instance, courage is an individual virtue that partly constitutes moral virtue. Once we’ve identified having virtue with possessing enough individual virtues, we can explain why certain people with a disposition have virtue whereas some do not. Those that are virtuous have a disposition that is grounded by individual virtues whereas those that are deficient in virtue do not. In effect, once we’ve identified virtue with the grounds of the disposition, we can distinguish which grounds count as constituting virtue. And, fortunately, there’s a natural way to explain why grounds like the ability to turn off your taste buds and a belief that you won’t die don’t count as individual virtues. In both the food case and the bunker case, the agent deploys a trait that makes him or her reliable so long as he or she does not come face to face with one of the characteristic challenges associated with the relevant activity. That you know you are risking death is one of the things that famously makes it difficult to do things like take a bunker. And having to taste the disgusting food is of course one of the characteristic difficulties associated with the kind of activity I described. Neither of the agents faces these challenges in our cases. The soldier’s reliable disposition is 86 grounded in his false belief that he can’t die. This belief prevents him from coming face to face with the possibility of his death. And the food taster’s reliable disposition is grounded in his ability to turn his tastebuds off, which prevents him from having to face the horrible taste of the food. This suggests a natural way of distinguishing individual virtues from mere grounds of dispositions: the individual virtues are those properties that are called upon to overcome the characteristic difficulties associated with the relevant activity. A property is called upon to overcome a characteristic difficulty if having that property reliably disposes agents not to fail because of that difficulty when they face it. With respect to taking bunkers, one of the characteristic challenges is the risk of death. You come face to face with this difficulty when you know you are risking death. The trait that reliably disposes agents not to fail to take the bunker because of the risk of death when they knowing they are risking death is a willingness to risk death for the sake of taking the bunker. So this trait is a virtue of the relevant activity. By contrast, the false belief that you won’t die plays no role in overcoming any of the characteristic challenges associated with the activity, and so it isn’t one of the individual virtues associated with the relevant activity. In our food competition case, the characteristic difficulty that is important to us is the disgusting taste. You face this difficulty when you taste the taste. A tolerance for disgusting foods is one of the individual virtues associated with the relevant activity because it reliably disposes you not to fail to eat the food because of the taste even when you taste the disgusting food. A lack of taste plays no role in overcoming any such characteristic difficulty, and so it doesn’t count as a virtue. To develop this account, I have relied on the notion of a characteristic difficulty and the notion of facing such a difficulty. I should therefore say more about these notions. Characteristic difficulties are features of acts that make people who attempt to perform that kind of act fail. The characteristic difficulty may not always be present in every instance of the activity, but, to be a characteristic difficulty, it has to be one of the things that make people who attempt that kind of act fail. Often, when people fail to take bunkers, they fail because of a risk of death. This is a characteristic difficulty even if not every instance of taking a bunker poses a risk of death. I can’t say precisely what makes a difficulty characteristic, but I can highlight ways things can become characteristic difficulties. A difficulty can become characteristic because it is one of the challenges 87 the interest in overcoming of which motivated people to pursue the activity in the first place. Difficulties can also become characteristic when they are very often the cause of failure. As for facing difficulties, to face a difficulty, you just have to be in certain salient background conditions in which there is a substantial risk that people fail because of the difficulty. With respect to taking a bunker, you face the risk of death when you know that there is a risk that you die if you try to take the bunker. It is when people know that there is a risk of death that they tend to freeze up and fail to take bunkers. In our taste case, the disgusting taste of the food poses a risk of failure when you taste the food, so it is when you taste the food that you face the difficulty. The cases we presented don’t just suggest a distinct account of virtue; they also suggest modifications to what it takes to be creditworthy. Suppose the person in our disgusting food case has a tolerance for disgusting foods. He has simply chosen to turn off his taste buds. Even though he has virtue with respect to the relevant activity, and even though his act is explained in part by his possession of one of the individual virtues, namely his speed eating ability, he is not fully creditworthy because he does not face one of the characteristic difficulties and does not deploy the individual virtue that overcomes that difficulty. This suggests that, to be creditworthy, you have to face and overcome any of the characteristic difficulties that, in your environment, tend to cause people to fail. And you must do so by deploying the individual virtue that overcomes that difficulty. Manifesting virtue requires more than merely performing the act because of your virtue; you have to overcome the characteristic difficulties by use of the relevant individual virtue as well. This is not to say that only difficult performances can be creditworthy. 123 You can be creditworthy for routine performances. But when there is a substantial risk that attempters fail in your environment because of a difficulty, you have to face the difficulty, and you have to overcome it by utilizing the relevant individual virtue. Addressing Fake Barn Cases We’ve now given an account of virtue that falls out of consideration of general instances of creditworthy performance rather than a particular concern for epistemology. With this on board, we should return to epistemology, since my central claim has been that adopting an account of 123 Bradford (2015) argues that difficulty is required for achievement. But, as I conceive of creditworthiness, creditworthiness is distinct from achievement: it is the non-accidentality condition that makes performances eligible to be praiseworthy or achievements. 88 virtue that is more generally adequate will resolve the fake barn problem while retaining the account of knowledge as creditworthy true belief. To see how the solution works for the fake barn problem, we should first consider how it handles the analogous boxing case. If the account predicts that the boxer does not deserve credit while also predicting that Pritchard’s archer does deserve credit, the account should be able to resolve the problem posed by fake barn cases. One of the characteristic difficulties one faces in attempting to land punches in boxing is that an apparent opening may be a trap. Many failed punches fail precisely because the boxer fell for a trap. There is a substantial risk that you fail to land a punch because of the possibility of falling for a trap when your opponent is disposed to set up traps. The ability that reliably disposes you not to fail because of the possibility of falling for a trap when your opponent is disposed to set up traps is an ability to discern traps from genuine openings. And so, when one faces the difficulty, when one’s opponent tends to set up traps, full creditworthiness requires deploying this virtue in the process of landing one’s punches. By contrast, that some of one’s targets are unhittable is not one of the characteristic difficulties one faces in trying to hit bullseyes in archery. Very few failed bullseyes fail because the target the archer shot at was unhittable. And so, an ability to discern hittable targets from unhittable targets is not a virtue of hitting bullseyes. You can therefore be creditworthy without manifesting this ability. My account, therefore, allows us to distinguish the forcefield case from the boxing case. It distinguishes the fake barn case from the forcefield case in an analogous way. One of the characteristic difficulties one faces in forming true visual beliefs is that there are objects that seem to be something but are not in fact that thing. People who form false visual beliefs often do so because something seemed to be something it was not. This difficulty poses a substantial risk of failure to form true beliefs when there is a significant number of merely apparent instances of the thing intermingled with the actual objects in our environment. So we face this difficulty when we are in such an environment. What reliably disposes us not to fail because there are things that seem to be a thing but are not that thing when we are in such an environment? It is of course a capacity to differentiate real instances of the object from merely apparet instances of the object. And so, one of the virtues of forming true beliefs is an ability to differentiate real instances of an object from merely apparent instances. And, when agents are in a situation where merely apparent objects are intermingled with real instances of it, they must exercise this ability to be creditworthy: they 89 must deploy some method for differentiating the real objects from the fake objects. Barney, of course, is in such a situation: fake barns are intermingled with real barns. But he does nothing to differentiate the real barns from the fake barns. Consequently, he does not deserve credit for his true belief. He lacks knowledge. Accounting for Common Intellectual Virtues Because my view gets the right verdicts about certain virtues by restricting which grounds of a reliable disposition count as individual virtues, a natural worry for my view is that it will be too restrictive: it will not count certain knowledge-granting intellectual virtues as individual virtues. To alleviate this worry, I will now explain how my view classifies two different knowledge-granting capacities that we have as individual virtues: our perceptual capacities, and our inferential capacities. Let me start with our perceptual capacities, like our ability to see, hear and feel. These capacities ground a reliable disposition to form true beliefs in certain circumstances, and they are essential to acquiring knowledge. For my view to predict that these capacities are individual virtues, they must help us to overcome certain characteristic difficulties with forming true beliefs. I submit that one characteristic difficulty with forming true beliefs is that we sometimes lack good sources of information. The salient background condition in which this difficulty poses a substantial risk of failure is that our current information doesn’t support believing the proposition that is true. So, we face this difficulty when we are in this condition. What allows us to overcome this difficulty? That is, what reliably disposes us not to fail because we lack good information sources when our information currently doesn’t support the belief that is true? I submit that it is possessing good sources of information. Our perceptual capacities, so long as they are reliable, are good sources of information, and so they are individual virtues. Our inferential capacities help us overcome a different characteristic difficulty. Sometimes, when we attempt to form a certain true belief, we don’t fail to form it because we lack a good source of information; we fail to form it because we fail to infer it from our current information, which supports the true belief. We have gotten information from a reliable source of information, but we just don’t connect this information to the question we are trying to answer by forming a belief. Thus, one characteristic difficulty with forming true beliefs is that we fail to infer beliefs that are supported by our belief set. We face this difficulty when we can’t easily form the belief 90 through non-inferential methods; that is the condition in which failing to infer poses a serious risk of failure to form a true belief. What reliably disposes us not to fail to form a true belief because we fail to infer it when we can’t easily non-inferentially form the belief? It is of course reliable inferential capacities. Thus, our deductive, inductive, and abductive capacities are intellectual virtues. A Possible Bonus: Extending the Virtue-Theoretic Approach to Moral Worth I stated that an additional, though less pressing, problem for reliable virtue epistemologists’ account of virtue is that it yields implausible verdicts with respect to moral worth. In particular, the view seems committed to holding that you can deserve credit for doing the right thing even if you are not motivated by intrinsic moral desires, since instrumental moral desires can reliably produce right actions if your final ends are systematically conveniently aligned with doing the right thing. As far as I know, every existing account of moral worth denies this. It would therefore be nice if a virtue-theoretic approach to creditworthiness were not committed to this. Although more work would need to be done to determine if the account can really be extended to moral worth, I believe my account has a more promising start to explaining why only intrinsic moral desires are part of the virtue associated with right action. One of the characteristic difficulties associated with doing the right thing is lack of adequate intrinsic moral concern. When do we face this difficulty? That is, what is the salient background condition in which we tend to fail to do the right thing because we lack adequate intrinsic moral concern? It is when our non- moral desires are better promoted by failing to do the right thing. This is familiar from Socrates’s discussion with Glaucon and Adeimantus in books 2 and 3 of The Republic. Glaucon and Adeimantus recognize that agents will tend to do the right thing when they lack adequate moral concern if their non-moral desires are promoted by doing the right thing. It’s clear to them, though, that these agents avoid rather than face the difficulty posed by lack of adequate intrinsic moral concern. They want to know why we should overcome this difficulty when we face it. But what allows us to overcome this difficulty? It is of course possessing sufficiently strong intrinsic moral desires. If we are reliably disposed not to fail to do the right thing because we lack adequate intrinsic moral concern when our non-moral desires are better served through wrong actions, that is because we have intrinsic moral desires that motivate us to do the right thing even when our non-moral desires are better promoted by not doing the right thing. Thus, sufficiently 91 strong intrinsic moral desires are individual virtues. Because instrumental moral desires play no such role in overcoming any characteristic difficulties with doing the right thing, they are not individual virtues. Conclusion I have argued that virtue epistemologists have struggled to solve the fake barn problem because they are working with an account of virtue that is implausible across many different domains of creditworthy performance. Once we accept a more adequate general account, the problem goes away. A potential further, unexpected benefit of this account is that it may overcome an obstacle to extending virtue-theoretic approaches to creditworthiness to moral worth. These are the narrow lessons that I have hoped to draw out in this paper. But there is a much more important lesson that is the underlying theme in this paper: if you aim to develop an account of some phenomenon by treating it as an instance of a more general phenomenon, it is not enough to theorize about the more specific phenomenon. To develop an adequate account, one must also carefully and independently theorize about the more general phenomenon. I hope that, even if you do not agree with me in some of the details, you will at least agree that what I have done exhibits the promise of taking this approach. That way, you too will be more likely to take this approach in overcoming the difficulties you face in constructing your views. If I am right about the value of this approach and the nature of virtue, this will make you a more virtuous philosopher. 92 References Alston, W. (1986). An action-plan interpretation of purposive explanations of actions. Theory and Decision, 20, 275-299. doi:10.1007/BF00134042 Alvarez, Maria (2016). "Reasons for Action: Justification, Motivation, Explanation". The Stanford Encyclopedia of Philosophy (Winter 2017 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/win2017/entries/reasons-just-vs-expl/>. Arpaly, Nomy (2002). “Moral Worth”. Journal of Philosophy. 99: 223-45 —— Unprincipled Virtue: An Inquiry Into Moral Agency. New York, NY: Oxford University Press. Arstila, V. & Pihlainen, K (2009). The causal theory of perception revisited. Erkenntnis, 70, 397–417. doi: 10.1007/s10670-008-9153-7 Ashwell, L. (2010). Superficial dispositionalism. Australasian Journal of Philosophy, 88(4), 635– 653 Audi, R. (1986). Acting for reasons. Philosophical Review, 95, 511–546. doi: 10.2307/2185049 Bach, K. (1978). A representational theory of action. Philosophical Studies, 34, 361-379. doi: 10.1007/BF00364703 Baron, M. W. (1995). Kantian Ethics Almost Without Apology. Ithaca: Cornell University Press. Bedor, Bob and Pavese, Carlotta (2020). Modal Virtue Epistemology. Philosophy and Phenomenological Research. Benson, P. (1987). “Moral Worth”. Philosophical Studies, 51, 365–382. Bishop, J. (1989). Natural agency: An essay on the causal theory of action. Cambridge: Cambridge University Press. Bradford, Gwen (2015). Achievement. Oxford, UK: Oxford University Press. Brand, M. (1984). Intending and acting: Toward a naturalized action theory. Boston: MIT press. Broncano-Berrocal, Fernando (2017). A robust enough virtue epistemology. Synthese 194, 2147–2174. https://doi.org/10.1007/s11229-016-1043-y Campbell, S. (2002). Causal analyses of seeing. Erkenntnis, 56, 169-180. doi: 10.1023/A:1015633007571 Castañeda, H.N. (1979). Intensionality and identity in human action and philosophical method. Noûs, 13, 235-260. doi: 10.1007/BF00485491 Chisholm, R. (1966). Theory of Knowledge. Englewoods Cliffs, NJ: Prentice Hall. Church, I.M. (2013). Manifest failure failure: the Gettier Problem revived. Philosophia, 41, 171- 177. doi: 10.1007/s11406-013-9418-5 93 Clark, M. (1963). Knowledge and grounds: a comment on Mr. Gettier’s paper. Analysis, 24, 46- 48. doi: 10.1093/analys/24.2.46 Clarke, R. (2010). Opposing powers. Philosophical Studies, 149, 153–160 Coates, P. (2000). Deviant causal chains and veridical hallucination: a problem for the anti- causalist. The Philosophical Quarterly, 50, 320-331. doi: 10.1111/j.0031-8094.2000.00188.x Cunningham, Joseph (2021). “Moral Worth and Knowing How to Respond to Reasons” Philosophy and Phenomenological Research. Comesaña, Juan and Matthew McGrath (2014). “Having False Reasons.” in Littlejohn and Turri (eds.). Epistemic Norms. Oxford University Press, pp. 59-80. Danto, A. (1965). Basic actions. American Philosophical Quarterly, 2, 141–148. Davidson, D. (1973). Freedom to act. In T. Honderich (Ed.), Essays on Freedom of Action (pp. 137-157). London: Routledge and Kegan Paul. Davies, M. (1983). Function in perception. Australasian Journal of Philosophy, 61, 409-426. doi: 10.1080/00048408312341161 Dietz, C (2018). Reasons and factive emotions. Philosophical Studies, 175, 1681–1691. doi: 10.1007/s11098-017-0929-y Enç, B (2003). How We Act: Causes, Reasons, and Intentions. Oxford: Oxford University Press. —— (2004). Causal theories of intentional behavior and wayward causal chains, Behavior and Philosophy, 32, 149-166. Faraci, David (unpublished). “On Accomplishment.” Unpublished draft. Feldman, R. (1974). An alleged defect in Gettier counterexamples. Australasian Journal of Philosophy, 52, 68-69. doi: 10.1080/00048407412341051 —— (1985). Reliability and justification. The Monist, 68, 159–174. doi: 10.5840/monist198568226 Gert, Joshua (2012). “Moral Worth, Supererogation, and the Justifying/Requiring Distinction.” Philosophical Review 121 (4): 611-618. Gettier, E.L. (1963). Is justified true belief knowledge? Analysis, 23, 121–123. doi: 10.1093/analys/23.6.121 Gibbons, J. (2001). Knowledge in action. Philosophy and Phenomenological Research, 62, 579- 600. doi: 10.1111/j.1933-1592.2001.tb00075.x Goldman, A. (1967). A causal theory of knowing. The Journal of Philosophy, 64, 357–372. doi: 10.2307/2024268 —— (1970). A Theory of Human Action. Princeton, NJ: Princeton University Press. 94 —— (1976). Discrimination and perceptual knowledge. The Journal of Philosophy, 73, 771– 791. doi: 10.2307/2025679 Greco, John (2003). “Knowledge as Credit for True Belief.” in DePaul & Zagzebski 2003: 111- 134. –––(2009). “Knowledge and Success From Ability.” Philosophical Studies, 142(1): 17–26. ––– (2010). Achieving Knowledge: A Virtue-theoretic Account of Epistemic Normativity. ––– (2012). “A (Different) Virtue Epistemology.” Philosophy and Phenomenological Research 85: pp. 1-26. —— (2020). The transmission of knowledge and garbage. Synthese, 197, 2867-2878. doi: 10.1007/s11229-019-02090-3 Grice, H.P. (1961). The causal theory of perception. Proceedings of the Aristotelian Society, 35, 121-53. Harman, G. (1976). Practical reasoning. The Review of Metaphysics, 29, 431-463. Hawthorne, J. and Magidor, O (2018). Reflections on the ideology of reasons. In D. Star (ed.), The Oxford Handbook on Reasons and Normativity (pp. 113-141). Oxford: Oxford University Press. Herman, B (1981). ‘On the Value of Acting from the Motive of Duty.’ The Philosophical Review. Hills, Alison (2009). “Moral Testimony and Moral Epistemology”. Ethics 120: 94-127. Howard, N. R (2019). “One Desire Too Many”. Philos Phenomenol Res. 2019; 00: 1– 16. Hughes, N. (2014). Is knowledge the ability to φ for the reason that P? Episteme, 11, 457-462. doi: 10.1017/epi.2014.16 Hurka T (2010) Right act, virtuous motive. In: Battaly HD (ed) Virtue and vice, moral and epistemic. Wiley-Blackwell Husserl, Edmund. (1913, repr. 1983). Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy, First Book; General Introduction to a Pure Phenomenology. Trans. Martinus Nijhoff, The Hague. Hyman, J. (2014). Desires, dispositions and deviant causal chains. Philosophy, 89, 83–112. doi: 10.1017/S0031819113000685 Isserow, Jessica (2018). “Moral Worth and Doing the Right Thing by Accident.” Australasian Journal of Philosophy. —— (2020). Moral Worth: Having it Both Ways. The Journal of Philosophy, 117, 529–556. —— (2021). ‘Doubts about Duty as a Secondary Motive’, Philosophy and Phenomenological Research. 95 Jackson, F. (1977). Perception: A Representative Theory. Cambridge: Cambridge University Press. Johnson King, Zoë (2018). “Accidentally Doing the Right Thing”. Philosophy and Phenomenological Research 1-21. —— (2019). “Praiseworthy Motivations”. Nous —— (forthcoming a). “What Are We Praiseworthy For?”. To appear in a volume associated with the “Future of Jurisprudence” conference at Oxford. —— (forthcoming b). “Deliberation And Moral Motivation?”.Oxford Studies in Metaethics. Kant, Immanuel (1785, repr. 2002). Groundwork of the Metaphysics of Morals. Trans. Allen Wood, Yale University Press Keil, G. (2007). What do deviant causal chains deviate from? In C. Lumer & S. Nannini (ed.), Intentionality, Deliberation and Autonomy: The Action-Theoretic Basis of Practical Philosophy (pp. 69-90). London: Routledge. Kelp, C., Boult, C., Broncano-Berrocal, F., Dimmock, P., Ghijsen, H., & Simion, M. (2020). Hoops and Barns: a new dilemma for Sosa. Synthese, 197(12), 5187–5202. Klein, P.D. (2008). Useful false beliefs. In Q. Smith (ed.), Epistemology: New Essays (pp. 25- 63). Oxford: Oxford University Press. Lackey, J. (2007). Why we don’t deserve credit for everything we know. Synthese, 158, 345– 361. doi: 10.1007/s11229-006-9044-x –––(2009). Knowledge and credit. Philosophical Studies, 142, 27–42. doi: 10.1007/s11098-008- 9304-3 Le Catt, B. (1982). Censored vision. Australasian Journal of Philosophy, 60, 158-62. doi: 10.1080/00048408212340581 Lewis, D. (1980). Veridical hallucination and prosthetic vision. Australasian Journal of Philosophy, 58, 239-249. doi: 10.1080/00048408012341251 Lillehammer, Hallvard (1997). Smith on Moral Fetishism. Analysis 57 (3):187–195. Littlejohn, Clayton (2017). “How and Why Knowledge is First.” In Knowledge-First Epistemology ed. Adam Carter, Emma Gordon, and Ben Jarvis. Oxford: Oxford University Press Locke, D. (2015). Knowledge, explanation, and motivating reasons. American Philosophical Quarterly, 52, 215-33. Lord, Errol (2017). ‘On the Intellectual Conditions for Responsibility: Acting for the Right Reasons, Conceptualization, and Credit.’ Philosophy and Phenomenological Research. —— The Importance of Being Rational. Oxford: Oxford University Press. 96 Mantel, S. (2017). Three cheers for dispositions: a dispositional approach to acting for a normative reason. Erkenntnis, 82, 561–82. doi: 10.1007/s10670-016-9832-8 Markovits, Julia (2010). “Acting for the Right Reasons”. Philosophical Review 119(2): 201-242. Massoud, Amy (2016). “Moral Worth and Supererogation.” Ethics 126: 690-710. Mayr, E. (2011). Understanding human agency. Oxford: Oxford University Press. McDonnell, N. (2015). The deviance in deviant causal chains. Thought: A Journal of Philosophy, 4, 162-170. doi: 10.1002/tht3.169 Mele, A. (1987) Are intentions self-referential? Philosophical Studies, 52, 309–29. doi: 10.1007/BF00354051 ––– (1992). Springs of action: Understanding intentional behavior. Oxford: Oxford University Press. ––– (2003). Motivation and agency. Oxford: Oxford University Press. Mele, A. & Moser, P (1994). Intentional Action. Noûs, 28, 39-68. doi: 10.2307/2215919 Morton, A. (1975). Because he thought he had insulted him. The Journal of Philosophy, 72, 5- 15. doi: 10.2307/2025248 Noë, A. (2003). Causation and perception: the puzzle unravelled. Analysis, 63, 93–100. doi: 10.1093/analys/63.2.93 Nozick, R. (1981). Philosophical Explanations, Cambridge, MA: Harvard University Press. Olson, Jonas (2002). “Are Desires De Dicto Fetishistic?” Inquiry, 45, 89–96. Peacocke, C. (1979a). Deviant causal chains. Midwest Studies in Philosophy 4, 123-155. doi: 10.1111/j.1475-4975.1979.tb00375.x ––– (1979b). Holistic explanation: Action, space, interpretation. Oxford: Oxford University Press. Pears, D. (1975). The appropriate causation of intentional basic actions. Crítica: Revista Hispanoamericana de Filosofía, 7, 39-72. Portmore, D. (forthcoming). “Moral worth and our Ultimate Moral Concerns”. Oxford Studies in Normative Ethics. Price, C. (1996). Function, perception, and normal causal chains. Philosophical Studies, 89, 31- 51. doi: 10.1023/A:1004250931634 Pritchard, Duncan (2009). “Apt Performance and Epistemic Value.” Philosophical Studies. Riggs, Wayne D. (2009). “Two Problems of Easy Credit.” Synthese, 169(1): 201–216. Ross, W. D. (1930). The Right and the Good. Oxford: Oxford University Press. 97 Rozeboom, Grant (2017). “The Motives of Moral Credit.” Journal of Ethics and Social Philosophy. Schlosser, M. (2007). Basic deviance reconsidered. Analysis, 67, 186-194. doi: 10.1111/j.1467- 8284.2007.00673.x ––– (2010). Bending it like Beckham: movement, control and deviant causal chains. Analysis, 70, 299– 303. doi: 10.1093/analys/anp176 Schroeder, Mark (2007). Slaves of the Passions. New York: Oxford University Press. ––– (2008). “Having Reasons.” Philosophical Studies 139(1): 57-71. Searle, J. (1983). Intentionality. Cambridge: Cambridge University Press. Shepherd, J. (2021). The Shape of Agency: Control, Action, Skill, Knowledge. Oxford: Oxford University Press. Shope, R. (1991). Non-deviant causal chains. Journal of Philosophical Research, 16: 251-291. doi: 10.5840/jpr_1991_35 Silvan, Kurt (2020). “Can Performance Epistemology Explain Higher Epistemic Value?” Synthese 197: 5335–5356 Singh, Keshav (2020). “Moral Worth, Credit, and Non-Accidentality.” Oxford Studies in Normative Ethics (Vol. 10). Skyrms, B. (1967). The explication of X knows that P. The Journal of Philosophy, 64, 373-89. doi: 10.2307/2024269 Sliwa, Paulina (2014). “Praise Without Perfection: A Dilemma for Right-Making Reasons”. American Philosophical Quarterly 52(2): 171-185. ––– (2015). “Moral Knowledge and Moral Worth.” Philosophy and Phenomenological Research 93(2): 393-418. –––(2017). “On Knowing What’s Right and Being Responsible for It.” in Robichaud andWieland 2017: 127–145. Smith, H.M. (1991). ‘Varieties of Moral Worth and Moral Credit.’ Ethics. Smith, Michael (1994). The Moral Problem. Oxford: Blackwell. Stanley, Jason & Williamson, Timothy (2001). “Knowing How.” Journal of Philosophy 98: 411- 44. ––– (2017). “Skill.” Nous 51 (4): 713–726. Sosa, E. (1991). Knowledge in Perspective: Selected Essays in Epistemology. Cambridge, UK: Cambridge University Press. ––– (1999). How to defeat opposition to Moore. Philosophical Perspectives, 13, 137-149. doi: 10.1111/0029-4624.33.s13.7 98 –––(2003). “The Place of Truth in Epistemology.” In M. DePaul & L. Zagzebski (eds.), Intellectual Virtues: Perspectives from Ethics and Epistemology (pp. 155-180). Oxford: Oxford University Press. –––(2007). Apt Belief and Reflective Knowledge, Volume 1: A Virtue Epistemology. Oxford: Oxford University Press. –––(2015). Judgment and Agency. Oxford: Oxford University Press. Stanley, J. & Williamson, T (2017). Skill. Noûs, 51, 713–726. doi: 10.1111/nous.12144 Stoecker, R. (2003). Climbers, pigs and wiggled ears: the problem of waywardness in action theory. In S. Walter & H. D. Heckmann (eds.), Physicalism and Mental Causation: The Metaphysics of Mind and Action (pp.296-322). Exeter: Imprint Academic. Stout, R. (1996). Things that Happen Because They Should. Oxford: Oxford University Press. ––– (2005). Action. Chesham: McGill-Queen's University Press. Stratton-Lake, P. (2005). Kant, Duty and Moral Worth. London: Routledge. Tännsjö, T. (2009). On deviant causal chains — no need for a general criterion. Analysis 69, 469- 473. doi: 10.1093/analys/anp078 Thalberg, I. (1984). Do our intentions cause our intentional actions? American Philosophical Quarterly, 21, 249-60. Turri, John (2011). “Manifest Failure: The Gettier Problem Solved.” Philosophers’ Imprint, 11(8): 1-11. –––Forthcoming. Knowledge as achievement, more or less. In M. A. Fernandez (Ed.) Performance epistemology. Oxford: Oxford University Press. Tye, M. (1982). A causal analysis of seeing. Philosophy and Phenomenological Research, 42, 311-325. doi: 10.2307/2107488 Unger, P. K. (1975). Ignorance: A Case for Scepticism. Oxford: Clarendon Press. Warfield, T.A. (2005). Knowledge from falsehood. Philosophical Perspectives, 19, 405–416. doi: 10.1111/j.1520-8583.2005.00067.x Way, Jonathan (2017). “Creditworthiness and Matching Principles.” In Mark Timmons (ed.), Oxford Studies in Normative Ethics, Volume 7. Weatherson, Brian (2014). “Running Risks Morally”. Philosophical Studies 167(1): 141-163. Wedgwood, R. (2006). The normative force of reasoning. Noûs, 40, 660-86. doi: 10.1111/j.1468- 0068.2006.00628.x Williams, Bernard (1981). “Persons, Character and Morality”. In B. Williams, Moral Luck, Cambridge University Press, pp.1-18. Williamson, T. (2000). Knowledge and Its Limits. Oxford: Oxford University Press. Wolf, Susan (1982). “Moral Saints”. Journal of Philosophy, 79(8), 419–439. 99 Zagzebski, Linda (1996). Virtues of Mind. Cambridge: Cambridge University Press.

Abstract (if available)

Abstract I try to make progress on certain difficult questions in ethics and epistemology by thinking about the nature of creditworthy performances.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Positivist realism

PDF

Cognitive boundaries for rational coherence requirements

PDF

The case for moral skepticism

PDF

Beyond the sea that separates: francophone voices of Oceania

PDF

Toward a more perfect liberalism: perfectionism in Kantian political philosophy

PDF

A perceptual model of evaluative knowledge

PDF

Process-oriented rationality

PDF

PRMT1 controls subcellular localization of p53 via PARP1

PDF

Reasons, obligations, and the structure of good reasoning

PDF

Statistical methods for causal inference and densely dependent random sums

PDF

Uncertain knowledge

PDF

The virtue of reasonableness: on the normative epistemic implications of public reason liberalism

PDF

Claus Spreckels of California

PDF

Did you get all that? Encoding of amplitude modulations at the auditory periphery predicts hearing outcomes

PDF

Four stages of blame

PDF

Rethinking reductive realism in ethics

PDF

Tractable information decompositions

PDF

Units of agency in ethics

PDF

Development of a historical urban land use web application for the city of Hong Kong

PDF

Aggregation and the structure of value

Asset Metadata

Creator Ward, Shane William (author)

Core Title Extending credit

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Philosophy

Degree Conferral Date 2023-08

Publication Date 06/05/2023

Defense Date 06/03/2023

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag credit,deviant causal chains,fake barns,gettier problem,knowledge,moral worth,OAI-PMH Harvest,Virtue,virtue epistemology

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Schroeder, Mark (committee chair), Finlay, Stephen (committee member), McHugh, James (committee member), Ross, Jacob (committee member), Wedgwood, Ralph (committee member)

Creator Email shanewar@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113147791

Unique identifier UC113147791

Identifier etd-WardShaneW-11925.pdf (filename)

Legacy Identifier etd-WardShaneW-11925

Document Type Dissertation

Format theses (aat)

Rights Ward, Shane William

Internet Media Type application/pdf

Type texts

Source 20230606-usctheses-batch-1052 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu