University of Southern California Dissertations and Theses

The mental states first theory of promising

(USC Thesis Other)

The mental states first theory of promising

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content The Mental States First Theory of Promising Alida Liberman 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 2015 TABLE OF CONTENTS Acknowledgements Introduction 1 Chapter 1: The Validity Conditions on Promising 5 Chapter 2: The Problem of Marginal Cases 24 Chapter 3: The Speech Act Sincerity Pattern 64 Chapter 4: What the Sincerity Condition on Promising is Not 109 Chapter 5: What the Sincerity Condition on Promising Is 147 Chapter 6: What Are Resolutions? 181 Chapter 7: The Escape Conditions on Promises 220 Chapter 8: The Escape Conditions on Resolutions 251 Works Cited 278 ACKNOWLEDGEMENTS There are more people whose thoughtful comments, probing questions, and insightful criticism have influenced this dissertation than I will be able to adequately thank here. Many of the best ideas are theirs, and any mistakes are mine. Thanks are due first and foremost to my dissertation committee: to Steve Finlay, for his thorough, detailed, and extremely constructive criticism of nearly all of the ideas in this dissertation; to Gary Watson, for asking questions that helped me keep sight of the big picture; to Jake Ross, for coming up with a new counterexample for me to work through every time I thought I had something figured out; and to Wolf Gruner of the History department, for graciously agreeing to be my outside committee member. My deepest debt of gratitude by far is to my advisor, Mark Schroeder. Mark has challenged me, encouraged me, and held me accountable. He has stuck with me through a long and winding process of attempting to write a couple of different dissertations before settling on this topic. And he has shown an almost uncanny ability to take the mess of multiple ideas I give him and guide me towards whichever are the strongest and will be the most fruitful, and to ask the right questions to help me come to realizations on my own. Mark has taught me how to be a better writer, and been crucial in helping me learn how to be a philosopher in my own right, rather than a student of philosophy. His extraordinary intellect and breadth of philosophical interests, unceasing dedication and availability, and genuine care for his students and their ideas have made him as good an advisor as I could imagine, let alone ask for. While Mark and the other faculty at USC have taught me how to be a philosopher, my undergraduate professors at The College of New Jersey taught me how to love philosophy. I owe them immensely for that. In particular, I wish to thank Pierre Le Morvan, my undergraduate mentor. Without his guidance and encouragement, I would have had neither the courage nor the confidence to pursue graduate studies. Sincere thanks are also due to my current and former graduate student colleagues at USC. You have sat patiently through dissertation seminars while I presented new ideas, given me valuable feedback on my work at every stage, and been my friends. Because of the graduate student community at USC, I’ve both developed as a philosopher and had a lot of fun. In particular, I wish to thank the workgroup of Mark’s advisees who have helped me workshop various chapters of this dissertation: Nick Laskowski, Michael Milona, Caleb Perl, and Abelard Podgorski. I also wish to thank the cohort of graduate students who started with me, including Brian Blackwell, John Kwak, Ben Lennertz, Indrek Reiland, Justin Snedegar, and especially Sara Ash Georgi. Without your help—and marathon PHL 500 study sessions—I would not have made it through my first semester. Thanks are also due to the Pasadena crew of philosophers who showed me the ropes when I first arrived: to Brian Bowman, Geoff Georgi, Brian Talbot, Rob Shanklin, and especially Julia Staffel, who has been a constant source of both culinary and philosophical inspiration. I am very lucky to have the most loving and supportive family I could imagine. Thanks especially to my parents, Cass and Howard, who have always believed in me, been proud of me, and encouraged me to pursue my dreams—even when that involved moving across the country and staying in school for a very long time. Thanks to my sister Anne, who has been my ally since we were children, and whose creativity and passion inspire me. Thanks to my best friend Alyssa, who shares my soul and always manages to talk me down when I get worked up over abstract philosophical problems. And deepest thanks to my in-laws, who have welcomed me with open arms and been a true California family to me: to Billie and Greg, Sue and Bill, and Elizabeth and Zech. Finally, and most importantly, my deepest debt of gratitude is to my husband (and classmate) Josh Crabill. Josh is the love of my life, my dearest friend, my partner in everything, and my continual philosophical sounding board. Without his support, I would be neither the person nor the philosopher that I am today. ! ! 1! INTRODUCTION Most people—philosophers and ordinary folk alike—find it obvious that, all else being equal, we are morally obligated to keep our promises. What is less obvious is why this is so. There has been much contemporary interest in this topic, and philosophers since ancient times have offered a variety of theories of promissory obligation. These theories typically ground promissory obligation in some external features of the promise, such as the negative effects that breach of the promise would have on the promisee, or the rules of the particular social conventions governing promising, etc. In this dissertation, I defend a new kind of theory, which takes an internal approach and grounds promissory obligation in the mental state expressed by promising. I call this the Mental States First theory of promising, because the theory posits that the norms on promises are inherited or derived from the norms on the promises’ underlying mental states. These mental states are “first” in the sense that they are explanatorily prior. My main aim is to show that the Mental States First theory should be taken seriously, because it is independently plausible and resourceful, accommodates a wider range of cases than existing theories can accommodate, and situates promising as part of a broad pattern of speech acts that behave in similar ways. The Mental States First theory has two features that set it apart from other views. First are its explanatory ambitions. I am not trying to offer a complete or exhaustive account of all of the ways in which promises can morally obligate us. Rather, I grant that there are often multiple sources of moral obligation to keep any given promise. These sources are the morally important considerations that other theorists have pointed to, such as satisfying the promisee’s expectations, or sustaining a valuable ! 2! social practice of promising, or maintaining a relationship of trust between promisor and promisee, and so on. But this cannot be the whole story, as these considerations are not always present, or are only present to a greater or lesser degree, for any particular promise. Such “marginal cases” of promising—deathbed promises, desert island promises, promises where no harm results from a breach, and so on—lack these independent sources of moral obligation, but that nevertheless seem to obligate us. So I take a different approach, highlighting a source of promissory obligation that is both distinctive of promissory obligation as such, and always present for all promises—even though it is not the most robust or important source of obligation for most promises. This brings us to the second unusual feature of my view, which is the explanatory strategy I take: grounding the obligation to keep promises in the obligations incurred by the mental state expressed by promises. If we want to explain promissory obligation for all cases of promising, including the marginal ones, we need something that is the same for every promise. Although some philosophers claim that the search for such a unified theory is futile—either because they think that no plausible candidate is forthcoming 1 , or because they think that “promising is not an independent source of a distinct kind of moral duty” 2 — promising does on the face of it seem to be a unified and morally distinctive phenomenon, and it’s always worth seeking more unified theories if we can find them. One feature shared by all promises is that they all involve a communicative speech act of a certain sort. And one thing you do when you perform many kinds of speech acts is express that you are in a certain mental state, by which I mean that you at !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 1 See Liberto, ms. 2 Dworkin, 2011, p. 304 ! 3! least purport to be in that state, and really are in it if you are sincere. I argue that when you publicly promise to do something, you express that you have a serious intention to do so—that is, you express that you intend to do so, and are serious about sticking to this intention, even if you don’t feel like it at the time. I cash out this idea of a serious plan as a resolution. I argue that resolving to Φ commits you to Φing, unless you have a good excuse. It follows that promising to Φ likewise commits you to Φing, unless you have a good excuse. I propose that this is because we can derive the norms on promising from the norms on resolutions—that is, from the norms on the mental state expressed by promising. Moreover, I argue that promissory obligation is a special case of a more general phenomenon: the norms governing or commitments stemming from a broad range of speech acts can be derived from the mental states that are communicated by utterances of those acts. Since every promise expresses such a mental state, this theory is maximally broad. And because it grounds promissory obligation in something that is a component of the speech act of promising itself, it is distinctively promissory. In Chapter 1, I lay out the validity conditions on promising, or the conditions that are necessary for successfully promising in the first place. While I make reference to some of these conditions later in the dissertation, this chapter is preliminary background information, and is therefore inessential to the main thrust of the argument. In Chapter 2, I motivate the need for a new theory by highlighting the limits of existing theories, and explaining how each is subject to some sort of marginal case, or instance of promissory obligation that it cannot accommodate or explain. ! 4! I introduce the Mental States First theory in Chapter 3, where I situate it as part of a larger pattern with other kinds of speech acts. In Chapters 4 through 6, I argue for the particular claims that the theory relies on: that promises express resolutions, rather than expressing belief or the intention to obligate yourself, and that resolving rationally commits you to acting. Finally, in Chapters 7 and 8 I survey the sorts of conditions in which breaking promises and abandoning resolutions is permissible. This analysis illustrates how the Mental States First theory can explain not only why we are obligated to keep our promises, but also when and why we are permitted to break them. ! 5! CHAPTER 1: THE VALIDITY CONDITIONS ON PROMISES Before launching into my substantive project—which consists in framing a problem (Chapter 2), proposing a new theory as a solution to that problem (Chapter 3), defending the premises involved in the theory (Chapters 4 – 6), and highlighting upshots of it (Chapters 7 and 8)—I will lay out some preliminary background information about how I understand promises. In particular, I will be laying out the conditions that are necessary for a successful promise to be made in the first place. This is important for a proper and thorough understanding of what promises are and how they work. However, most of the main argument of my dissertation does not depend crucially on any of these preliminaries. Readers who are anxious to get to the main ideas may therefore skip ahead to Chapter 2. 1.! How Promises Obligate Making a promise creates a pro tanto obligation to perform the promised action. A classic treatment of promises as pro tanto obligations comes from W. D. Ross. Ross argues that we are subject to what he calls prima facie moral duties, but what are more helpfully refered to as pro tanto duties. 3 Pro tanto duties tell us what we are required to do in light of particular considerations. For example, in light of my promise to meet you for drinks, I ought to meet you; in light of the fact that I have a steady job and you have just been laid off, I ought to offer to pay for your drink. The pro tanto obligation generated by a promise is a directed moral obligation that the promisor owes specifically to the promisee. A broken promise may wrong others, as well. For example, Rawls has !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 3 As has often been noted, “prima facie” is potentially misleading, as the term seems to imply that the promise is not actually an obligation at all, but only apparently (or “at first appearance”) an obligation, rather than being an obligation “to that extent.” For example, see Kagan (1989: p. 17n.) ! 6! argued that breaking a promise impermissibly free rides on the beneficial social practice of promising, which denigrates the practice and thereby harms the everyone who participates in it. 4 But this is at most a secondary and less central wrong, for promissory obligations are owed not to society as a whole, but specifically to a particular promisee. All else being equal, a pro tanto obligation yields all-things-considered obligation, as well. All-things-considered obligations tell us what to do overall, or in light of all of the relevant circumstances, including all of our pro tanto obligations. Sometimes, these other pro tanto obligations override or undermine the pro tanto obligation in question, entailing that it not be kept. For example, in light of my promise to meet you for drinks, I ought to do so. But if I can save the life of an injured cyclist by driving him to the hospital instead of meeting you, then all-things-considered I am required to break my promise to you. Similarly, my having a job while you are unemployed generates a pro tanto obligation for me to offer to pay for your drink. But if I know that such an offer would deeply offend you, then all-things-considered I ought not make it. 5 A valid promise is one that successfully generates a pro tanto obligation. Invalid promises fail to generate pro tanto obligations, and thus fail to morally bind the promisor in even a provisional or overridable way. A promise is invalid when certain !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 4 See Rawls (1971). 5 My construal of pro tanto obligations here presupposes only the weak claim that there is some kind of normative demand—and it does not matter to me whether we call it a pro tanto obligation, duty, or reason—that is both (1) strict, or such that you are required to act on it when all else is equal, and you go wrong in some way if you do not, and (2) overridable, or such that sometimes all else is not equal, and the demand is outweighed or undermined by other demands. This type of demand is to be contrasted with (mere) reasons, which are not strict (for if you fail to act on a reason when all else is equal, you need not go wrong in any way), and categorical or all-things-considered requirements, which are non-overridable or inescapable. I do not need to take any stand on the contentious question of whether we can have conflicts between all-things-considered requirements. ! 7! validity conditions are not met. In the rest of this chapter, I will lay out these conditions. There are two ways in which we can conceive of invalid promises. First, it could be that in order for a declaration to count as a promise, the validity conditions must be met and a pro tanto obligation must be generated, meaning that invalid “promises” are not promises at all. Or, it could be that any agent who purports to promise succeeds in doing so in at least in an attenuated sense, but the validity conditions must be met in order for there to be a promise that yields a pro tanto obligation. The first option allows only those declarations that meet the validity conditions and yield pro tanto obligations to count as promises, while the second option is more permissive about what we count as a promise. 6 I prefer the former option, although this is more a matter of terminology than of substance. For the sake of clarity, I will refer to declarations that do not meet the validity conditions as “promises,” with scare quotes. 1.2.! Reparatory duties Barring one class of exceptions that will be discussed in Chapter 7, people who break valid promises owe reparatory duties of some sort to their promisees. As Judith Jarvis Thomson puts it, “I may not simply wash my hands of the matter as no concern of mine when I learn that I will be unable to keep my promise, or after I have already broken it.” 7 Promise breakers must explain themselves, or apologize, or do the next best thing, or otherwise make it up to the promisee in some way. I am construing reparatory duties extremely broadly. These duties need not be compensatory, as for some breaches !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 6 For helpful discussions of this distinction, see Shiffrin (2011 p. 160) and Owens (2012 p. 232 ff). 7 1990, p. 96. This quotation occurs in a discussion of the ways in which reparatory duties are owed. Thomson has a reliance model of promissory obligation that requires the payment of reparatory duties in a way that is similar to the way in which broken legal contracts require restitution as a way of making up for detrimental reliance. ! 8! of promise compensation in the most literal and straightforward sense is impossible (e.g., breaking a promise to attend your sibling’s only graduation ceremony) or inappropriate (e.g., breaking a promise to remain monogamous with your spouse.) The point of the reparatory duty is to in some way acknowledge that the promisee has not received something that she was owed, and to make amends for this failure in the whatever appropriate way you can. What sort of reparatory duty is owed will depend on—and always be proportional to—the particular situation. For example, if the promisee is relying on the promise, or has already sunk costs of some kind because of the promise, the reparatory duty will be more burdensome than if there were no reliance or sunk costs. And if the promise is part of a reciprocal exchange and the promisee has already performed her part of the bargain, the reparatory duty will be more stringent than if this is not the case, all else being equal. 8 If I promise to buy your drink tonight and then fail to do so, I ought to buy you a drink next time to make up for it. If I promise to buy your drink tomorrow in exchange for you buying my drink tonight, and you buy my drink tonight but I fail to buy yours tomorrow, I will owe you more, and should perhaps buy you two drinks next time to make up for it. If the pro tanto obligation incurred by a valid promise is not legitimately overridden or otherwise outweighed—say, if I bail on our happy hour plans because I am in pajamas and do not feel like getting dressed—the reparatory duties I owe you will be fairly demanding; I will be morally required to do something like sincerely !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 8 The “all else being equal” clause is necessary because there could be cases in which the breach of a relatively unimportant reciprocal promise entails weaker reparatory duties than the breach of a very important non-reciprocal promise. Reciprocity in promise-making is one of many features that makes promises more important, and thereby makes the reparatory duty more stringent in the case of a breach. ! 9! apologize, reschedule with you, and commit to making a genuine effort to be more considerate of you in the future. If the pro tanto obligation is outweighed—say, if I have bailed on our happy hour plans in order to save someone’s life—the reparatory duties I owe you will be much less demanding. But I am still required to pursue some sort of reparatory action, even if this is simply calling you and explaining why I could not be there. Since I did what I was morally required to do, all-things-considered, I am not blameworthy in the way I would be if I broke a promise without a good excuse. But this does not completely negate the existence of the directed obligation I had to meet you. I owed you something, and I failed to provide it. Even though this failure was justified all-things-considered, I must still acknowledge the fact that you did not receive something you were owed. But I do not owe any reparatory duties to the promisee if I fail to keep an invalid “promise.” For there is no directed obligation in place, which means that there is nothing that I owe the promisee that I have failed to provide. So there is no need to make amends, for I have not violated any obligations of even a pro tanto sort. 2. Validity Condition #1: Moral Responsibility The first validity condition on promising is a general constraint on being subject to moral obligations of any sort. In order to be bound by a moral obligation, an agent must be sufficiently morally responsible. While theories of moral responsibility vary greatly, there are some obvious common requirements. First, morally responsible agents must be at least minimally capable of rational thought and understanding, and must be aware of what actions they are performing and the likely consequences of those actions. Agents who lack such features are unable to morally bind themselves by ! 10! making promises, as they are unable to take on moral obligations of any sort. So a “promise” made by someone who lacks moral responsibility—for example, by a toddler or someone with severe dementia—will be invalid, and will not generate a pro tanto obligation. Similarly, “promises” made by agents who are temporarily incapacitated— say, inebriated beyond comprehension or delirious from fever—are likewise invalid. In order to be morally responsible for their actions, promisors must be aware of what they are doing. It follows that it is impossible to inadvertently make a valid promise. Mumbling “I promise to make pancakes for breakfast” to your partner while sleep-talking does not create a valid promise. If a non-English speaker who has a faulty translation manual says “I promise to buy this” while believing that she is asking how much something costs, she does not successfully promise. Finally, agents can be morally responsible only for those actions that they believe to be within their control, at least to some degree. Suppose a diver has just jumped off of a high-dive at the pool. As she plummets towards the water, she shouts to her friend “I promise to hit the water!” This might be a true prediction of what will happen, but it is not a valid promise. The diver will hit the water at this point whether she likes it or not, and “promising” to do so does not yield any additional, directed pro tanto obligation. 3. Validity Condition #2: Lack of Coercion A second validity condition is that the promise must not be coerced, for promises extracted under coercion fail to bind. For example, suppose Adam is walking down a lonely road when he is accosted by a highwayman who demands payment. Adam declares that he has no money on him. The highwayman brandishes a gun, yelling “Promise me that you will mail me $100 tomorrow, or I’ll shoot!” Because he wants to ! 11! live, Adam says “I promise to send you the money tomorrow.” Satisfied, the highwayman leaves. Has Adam incurred even a pro tanto obligation to mail the highwayman $100? It is intuitively obvious that Adam is not morally required all-things-considered to keep his promise to the highwayman. Moreover, allowing coerced promises to morally bind all-things-considered would provide a perverse incentive to potential coercers. But coercion is more than just an excusing condition. Rather, coercion invalidates promises, making it the case that they do not yield even pro tanto moral obligations. This is because promises are essentially voluntary obligations. Unlike categorical moral obligations, which apply to us regardless of our choices and desires, we must actively choose when and which promises to take on. And if a promise is coerced, it is not freely chosen. As Eric Chwang argues, “it is part of the moral principle about promissory moral force that the promise in question cannot be coerced” because “the point of promissory activity is to allow people to create and then assume obligations all by themselves. That coerced promises carry no moral force is built into this idea.” 9 Granted, what counts as coercion is a difficult question, to which there are a variety of answers in the literature. 10 But I need not settle on an answer here. A promise is invalid if it counts as coerced, or undermining of an agent’s voluntary choice, according to whichever theory of coercion turns out to be the most plausible. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 9 2011, p. 159. Chwang understands coerced promises as genuine promises that carry no moral force. In my terminology, Chwang understands them as real promises that yield no pro tanto moral obligations. I have been using different terminology, according to which all promises yield pro tanto moral obligations, and invalid “promises” are promises in name only. 10 For a good comprehensive overview of the topic of coercion, see Anderson (2011). ! 12! Margaret Gilbert disagrees, and argues that promises made under coercion count as genuine promises with moral force. She construes the Voluntariness Argument that I presented in the previous paragraph in the following way: 1.! Promises are by definition voluntary: they cannot come about against the will of either party. 2.! Coercion does not allow a promise to be voluntary. 3.! Therefore, coercion rules out the very possibility of promise. 11 Gilbert claims that the argument has a “spurious plausibility” due to an “ambiguity in the notion of voluntariness or of acting in conformity with one’s will.” She argues that the first premise is true only if by “voluntary” we mean something like “intentional,” granting that “it is indeed the case that, as a conceptual matter, one cannot promise without intending to do so.” 12 Gilbert refers to this as the “decision-for” sense of voluntariness. Coercion does not undermine decision-for voluntariness, and “one can be in coercive circumstances without being rendered witless by fear.” 13 So the second premise is true only if we assume a more robust sense of voluntariness, which would be ruled out by coercion. Gilbert’s criticism of the Voluntariness Argument succeeds only if it is true that “there is nothing about [promises] that entails that they must be voluntary in any sense !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 11 In (1993), Gilbert critiques the Voluntariness Argument as it applies to agreements, arguing that agreements made under coercion are genuine agreements with moral force. While she does not explicitly discuss the promise version of the Voluntariness Argument, she states that her arguments about agreements should apply mutatis mutandis to promises. I have replaced “agreement” with “promise” in presenting and discussing the Voluntariness Argument here; otherwise, the words are Gilbert’s (p. 684). In (2006), Gilbert explicitly commits to the assumptions about promises that are necessary to get the promise version of the argument off the ground (p. 228). 12 1993, p. 684 13 2006, p. 228 ! 13! other than the decision-for sense.” 14 But this is not true. Adam acts intentionally when he promises the highwayman that he will send money. But he does not act freely in the way required to generate a real promise. For part of what makes promissory obligations effective is the fact that they are freely chosen by the promisor—the promisee believes that the promisor will act because she knows that the promisor has specifically and deliberately committed to performing the promised action. The highwayman should have little confidence that Adam will send him any money, because Adam never voluntarily committed to doing so. And forced “commitments” are highly unreliable ways of predicting action. 15 Adam Smith disagrees about whether coercive promises are invalid for different reasons. In The Theory of Moral Sentiments, Smith raises the highwayman example discussed above. Smith grants that “as a question of jurisprudence” there is “no doubt” that coercive promises or contracts cannot be legally enforced. 16 He also grants that “no injury is done to the robber” and “no regard is due to [his] disappointment.” But he maintains that “if we consider the matter according to the common sentiments of mankind, we shall find that some regard would be thought due even to a promise of this kind.” This is because Smith claims that “whenever such promises are violated, though for the most necessary reasons, it is always with some degree of dishonour to the person who made them.” Although certain circumstances of the coercive promise might function as a mitigating factor—for example, if one promised to pay the highwayman a ruinous amount of money—Smith insists that the character of someone !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 14 1993, p. 686 15 Chwang (2011) also argues that coerced promises lack predictive value. 16 All quotations in this section are from Smith (1759), Section IV Part II. ! 14! who breaks a coerced promise “if not irretrievably stained and polluted, has at least a ridicule affixed to it, which it will be very difficult entirely to efface.” I do not share Smith’s intuitions, and I suspect many readers will agree with me. But these intuitions are easy to explain away for anyone who finds them compelling. Smith explains the dishonor of failing to keep one’s word in the following way: Treachery and falsehood are vices so dangerous, so dreadful, and, at the same time, such as may so easily, and, upon many occasions, so safely be indulged, that we are more jealous of them than of almost any other. Our imagination therefore attaches the idea of shame to all violations of faith, in every circumstance and in every situation. Smith’s point seems to be that keeping one’s word is so very important, and so easy to fail to do, that we must err on the side of caution, and be disposed to scrupulously act as we have promised, no matter what the situation. That way, we will cultivate good habits of promise keeping, and will be less likely to mistakenly break a promise when it would not be appropriate to do so. This insight grounds Smith’s intuition that even coerced promises are valid, and that we act dishonorably in breaking them. We can grant Smith’s insight that promise keeping is a beneficial habit, and that because we are disposed to err in our own favor we should use caution when determining whether a promise is invalid. For this intuition is fully compatible with the claim that coerced promises are in fact invalid. Because we have been conditioned to value promise keeping so highly, we mistakenly view even coerced promises as binding to some degree. Smith makes an analogy to chastity that, although outdated, is enlightening for understanding how this sort of intuition works. Smith argues that “breach of chastity dishonours irretrievably,” because “we are so nice [i.e., fastidious or scrupulous] in this respect that even a rape dishonours, and the innocence of the mind cannot, in our imagination, wash out the pollution of the body.” In other words, a ! 15! society that is conditioned to value chastity extraordinarily highly will view with suspicion even entirely innocent and non-culpable “violations” of chastity. Likewise, it seems plausible that a society that values promise-keeping very highly will view with suspicion even entirely innocent and non-culpable failures to keep invalid coerced “promises.” But this does not mean that such failures are indeed culpable. 3.1. Third party coercion We have seen that coercion is an invalidating condition when the promisee is the one who does the coercing. But what about cases in which the coercer is a third party, and the promisee is innocent of any wrongdoing? Suppose Bill threatens to reveal an embarrassing secret about Christina unless she promises their mutual friend Dan that she will water Dan’s plants while he is away. Because she does not want the secret to be revealed, Christina promises Dan—who knows nothing about Bill’s manipulation— that she will water the plants. This “promise” was not made fully voluntarily, for it was solicited under coercion and Christina would not have made it had she not been threatened. So it is invalid, and does not yield a pro tanto obligation. But this is not to say that Christina may permissibly leave Dan’s plants to die. Dan he has not wronged Christina in any way, and he is relying on her to water his plants. As a result of this reasonable expectation, Christina owes Dan some kind of moral consideration. She might decide to satisfy this moral debt by watering the plants after all. But she could equally well satisfy it by informing Dan that Bill coerced her into making the “promise,” and giving Dan advance warning that she will not be acting as she “promised” she would. Or she could arrange for someone else to water the plants instead. This is the same sort of moral consideration that Christina would owe Dan had she knowingly led him in some non-promissory way to expect that she would ! 16! water his plants. Had Christina’s obligation towards Dan been a promissory obligation, she would owe him moral consideration of a different sort. For Christina could not equally well satisfy her promise by watering the plants or by giving Dan advance warning that she would not be doing so. Only the former counts as keeping a promise. The latter action is better than doing nothing, but is still second best. 17 There might be cases in which independent features of the situation make it the case that an agent is morally required to act as she “promised” to under coercion. For example, suppose Adam discovered that the highwayman who accosted him is a super- villain who will set off nuclear weapons in major cities all around the world unless Adam follows through on his promise to mail him $100. All-things-considered, Adam ought to avert nuclear disaster by sending the villain the money. But this is required of him in order to avoid nuclear catastrophe, and not because he promised to do so. Moreover, Adam does not owe anything to the villain, but rather owes something to the millions of people whose lives are at risk and whom he is uniquely positioned to save. In summary, a promise is valid when it is offered intentionally by a morally responsible agent who is not acting under coercion; in Chapter 5, I will argue that valid promises must also be accepted by a promisee who does not subsequently release the promisor. All valid promises generate pro tanto moral obligations, although certain conditions make it the case that these pro tanto obligations are overridden or undermined. I will explore these considerations in more detail in the last two chapters. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 17 Chwang (2011) agrees that third-party coerced promises carry no moral weight, although the promisor might have independent reason to fulfill the expectations of the innocent, non- coercing promisee. He also argues by way of “distinguishing between the moral force generated by promises and that generated by expectations.” (162). ! 17! 4. Is Moral Permissibility a Validity Condition? Some philosophers have argued that moral permissibility is a validity condition on promising, and that immoral promises are therefore invalid. There is some intuitive pull to this idea; the thought that an immoral promise yields any sort of moral obligation, even of a pro tanto sort, can seem troubling. For example, Seana Shiffrin insists that a robber who “promises” her conspirator that she will rob a bank has no more reason to rob the bank than she did before she “promised” to do so, for “her declaring that she has promised to rob the bank does not alter her moral status with respect to robbing the bank at all.” 18 Gary Watson similarly argues that the claim that promising to perform an immoral action gives you a reason to perform that action implies, “absurdly, that I could make my actions less pernicious just by promising so to act,” for “morality would have at least something to say in their favor.” 19 The non-moralized version of this claim, in which a reason to perform an immoral action which you have promised to do arises from bare volitional commitment, is in Watson’s opinion “equally incredible.” For it would let us “transform any otherwise insane project into something with reason on its side.” 20 Watson claims that this is true not only of gravely immoral actions like a “promise” to murder, but even of slightly immoral actions, like “promising” to steal two dollars for your office mate’s wallet to buy your friend a coffee. In fact, minor cases make the invalidity of immoral “promises” even clearer. For there will presumably come a point !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 18 2011, p. 159. 19 2009, p. 168. 20 ibid p. 169. ! 18! at which the immorality is so minor that the promisor “should keep [the immoral] promise, given its inherent normative force.” 21 And this seems bizarre. However, there are good reasons to think that immoral promises do yield pro tanto moral obligations. It would certainly be absurd to claim that promising to perform some morally atrocious act could make the act become objectively less morally atrocious; the facts that make murdering innocents morally atrocious do not change or count for less simply because an agent has promised to murder innocents. However, there is an internal or subjective element to morality, as well, which has to do with the mental processes that an agent goes through, and with the reasons and motivations for which one acts. The outcomes of killing the snitch because you promised the mob boss to do so and killing the snitch because you don’t like the way he looked at you are the same, but the internal processes may be quite different. The mobster who kills the snitch because he promised to do so commits a morally atrocious action. But at least he displays something positive that the other mobster does not. As a procedural matter, there is something good about keeping your word, something that is not present in the mobster who kills the snitch for reasons that have nothing to do with being trustworthy. 22 Contra Watson, examples of trivial immorality can illustrate nicely the notion that promising to perform an immoral action can make a moral difference of some sort. Suppose Ed and Fred, a pair of teenaged best friends, are heading to the mall, where Ed plans to shoplift a small item from a major chain store. As an earnest (albeit very much !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 21 ibid p. 172. 22 Watson agrees that failing to keep an immoral promise is normatively problematic, insofar as your failure marks you as weak or irresolute. What he denies is that the promisor wrongs the promisee, or violates a pro tanto obligation against her, by failing to keep the immoral promise. I respond to this claim in more detail below. ! 19! misguided) test of Fred’s friendship, Ed asks him to promise that he will shoplift a small item, as well. Fred does so. Clearly, this promise does not make shoplifting morally permissible. But it seems that it would not be quite as bad for Fred to shoplift in this case as it would be in a case where there is no promise involved. For in this case, Fred at least follows the right sort of moral procedure, even if to the wrong sort of end. If we posit that promises yield pro tanto moral obligations, we have an easy explanation of why the procedure followed by the keepers of immoral promises is of the right moral sort: Fred is satisfying a directed pro tanto obligation that he owes to Ed, just as the mobster is satisfying a directed pro tanto obligation that he owes to the boss. This is not to say that Fred and the mobster act correctly, for these pro tanto obligations ought not to be satisfied, all-things-considered. But obligations that ought not be satisfied all-things-considered do not lose their status as pro tanto obligations, and it therefore can display the right sort of moral attitude to act in accordance with them. For example, if you are obligated to pay your student loan bill and your child’s hospital bill and only have enough funds to pay one, you ought all-things-considered to pay the hospital bill. But you still remain obligated in some sense to pay off the student loan. And if you were to pay off the student loan instead of the hospital bill, you would at least be paying what you owed, and therefore following the right sort of moral procedure. This case is analogous to the mobster who kills the snitch because he has promised to, and (mistakenly) acts in accordance with an overridden pro tanto obligation. Compare this to a case in which you don’t pay your student loan because you pay for a luxury vacation instead. This case is analogous to the mobster who kills the snitch because he looks at him funny, and is not acting in accordance with any pro tanto obligation, mistaken or not. ! 20! That a directed pro tanto obligation can stem from a promise to perform an immoral action can be seen most clearly when the promisee is innocent of any wrongdoing. Suppose Ed promises Fred that he will bring him a new pair of designer sneakers. Fred accepts the promise, and does not realize that the only way for Ed to get the sneakers is to steal them. Clearly, this promise does not make it the case that Ed is all-things-considered obligated to steal the sneakers. But it does generate a directed pro tanto obligation towards Fred, and Ed owes Fred some kind of apology or reparation when he does not fulfill the promise. Moreover, Fred could rightly feel upset, and believe that Ed failed to give him his due, if Ed does not bring the sneakers. Shiffrin grants that we expect a jilted promisee to feel angry and betrayed when an immoral promise is not kept. But she suggests that this can be explained by the fact that making an immoral “promise” is independently wrong, and the promisee is upset about being wronged in that way. Watson agrees that a jilted promisee may rightly take issue with a promisor who fails to keep an immoral “promise.” But what is cause for complaint is not the fact that the “promise” was not kept, but that the promisor “made a false or careless promissory proposal.” 23 Furthermore, the promisor might still be answerable in some sense to the promisee, who could rightfully complain that she was not warned in time about the non-performance. But, Watson argues, “the legitimacy of these complaints does not entail that [the promisee] ever had an obligation or reason to do as promised.” 24 Shiffrin and Watson are right that it is wrong to knowingly make an immoral promise, and that the promisee can rightly take issue with being party to such a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 23 2009, p. 171. 24 ibid, p. 171 ! 21! promise. It is also true that someone who makes an immoral promise is answerable to the promisee, and should warn the promisee in advance about non-performance. But these observations do not fully capture the ways in which Fred may rightly take issue with Ed for failing to bring him new sneakers. To illustrate this, compare Ed and Fred’s case to the case of a promise that is immoral to make, but not to keep. Suppose you promise your spouse that you will not make any promises to your ex-partner, who still carries a torch for you. You then promise your ex-partner that you will avoid future APA conferences where he is on the program, so as to avoid accidentally running into each other. Making this promise is immoral, as it entails breaking a more important promise to your spouse. But (let’s assume) keeping this promise is morally permissible, and probably a good idea. Suppose your ex discovers that your promise to him violated a prior promise to your spouse. He may rightly feel upset that you made him the recipient of a promise that was immoral to make. But he will not be upset in exactly the same way that Fred is upset when Ed fails to bring him sneakers. For Fred can be upset on two counts: that he was the recipient of a promise that was immoral to make, and that he did not receive something he was owed (albeit owed in merely a pro tanto way.) There are also a number of independent reasons for rejecting the idea that immoral promises are invalid. David Owens offers two criteria of pro tanto obligation: the fact that one is obligated to Φ can by itself make sense of an agent’s Φing, and guilt is apt for the agent who fails to Φ. He argues that immoral promises meet these criteria, and therefore count as valid. For example, we can make sense of Benedick’s plan to kill Claudio simply in virtue of his promise to Beatrice to do so. And Owens regards it “as obvious that a decent person will regret having to breach a solemn promise, however ! 22! wicked,” and “would feel this way even if they had no doubt that breach was the right thing to do” all-things-considered 25 . Moreover, morality can be quite complicated. It is not always clear whether a given promise is immoral to keep. Is a promise to write a positive letter of recommendation for a less than stellar student morally permissible? The answer will presumably depend on a number of factors particular to the situation, and may not be immediately apparent. It would be strange if we need to know whether writing the letter was morally permissible in order to determine whether a valid promise was successfully made in the first place. Finally, the claim that immoral promises yield pro tanto obligations gives us a good explanation of the following sort of case. Suppose your neighbor asks you to promise her daughter Gia that you will hire her for a job. You make the promise to Gia, even though you know that there are other, better-qualified candidates, and that it would therefore be wrong to hire Gia instead. Gia, who is unaware of the qualifications of the other candidates or of the request from her mother, happily accepts the promise and begins making plans to take on the new job. All-things-considered, you ought not hire Gia. But suppose the situation changes, such that hiring Gia is no longer immoral: the more qualified candidates all voluntarily drop out, leaving only candidates who are equally or less qualified than Gia. Hiring Gia is now morally permissible, although not independently morally required; all else being equal, it would be permissible to hire any of the equally well-qualified candidates. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 25 2012, p. 246. ! 23! It seems all else is not equal in this case, and that you are morally required to hire Gia in accordance with your promise. This is easily explained by positing that the immoral promise to hire Gia yields a pro tanto moral obligation, which is outweighed by the immorality of passing up better-qualified candidates in favor of someone less qualified. Once the moral situation changes, the pro tanto obligation is no longer outweighed, and can yield an all-things-considered obligation. If promises to perform initially immoral actions were invalid, we could not appeal to this kind of explanation. One might object that hiring Gia seems obligatory because of the normal courtesy due to people for raising their expectations. But you owe Gia more than this. Had you raised her expectations about getting the job in some non-promissory way, you could discharge your obligations towards her equally well by hiring her or by warning her well enough in advance that you would be hiring someone else. But if you’ve promised Gia that you will hire her, and doing so ceases to be immoral and becomes morally permissible, you cannot satisfy your obligation towards Gia simply by warning her in advance. Rather, you must hire her, as you have promised to. With these preliminaries in place about how promises obligate us and what the validity conditions on promising are, we are ready to move on to the main topic of my dissertation: motivating and defending a new theory of promissory obligation. ! 24! CHAPTER 2: THE PROBLEM OF MARGINAL CASES 1. Introduction In this chapter, I outline what I call the Problem of Marginal Cases, a problem faced by most contemporary theories of promissory obligation but avoided by the theory I will propose. In a nutshell, the problem is that each theory by itself is insufficient to explain the full range of promissory obligation. I argue that each theory faces cases that appear to be genuine instances of promissory obligation, but which the theory cannot accommodate or explain. I establish this claim by surveying a representative list of theories of promissory obligation. I give a sense of the broader philosophical dialectic by noting any important theoretical criticisms that are commonly levied against each view, and then lay out what kinds of cases the view cannot accommodate. This motivates my task of developing and defending a broader view, which does not face any marginal cases of its own, in the rest of this dissertation. 1.! Framing the Problem It is easiest to explain why we are obligated to keep our promises in what I call paradigmatic cases, or cases that are at the core of our practices of promising. These are the everyday person’s conception of what promises are, rather than the philosopher’s tricky, bizarre, or borderline cases. In paradigmatic cases, it is obvious that a promise has been made, that the promise is something promisee wants and expects will occur, that in making the promise the promisor intends to bind herself, that some harm would occur if the promise were broken, and that the promise occurs within some obvious social convention of promising. As cases fail to meet one or another of these criteria, they deviate further from the paradigm. Otherwise, paradigmatic cases can be quite ! 25! diverse: they might be between strangers or friends, might involve a reciprocal exchange or be one-sided, might be about something important or something slight, etc. It is harder to explain why we are obligated to keep promises in non- paradigmatic cases, which I call marginal cases as a term of art. A case C is marginal for a theory T if it is at least controversial among proponents of T and their reasonable and charitable critics whether or not T can explain or accommodate C as an instance of promissory obligation. For this to be so, two conditions must be met, each of which can be met in one of two ways. First, there must be good theoretical arguments—or an intuitive prima facie presumption with no good arguments to the contrary—that C is indeed an instance of promissory obligation. Second, there must be good theoretical arguments—or, again, an intuitive prima facie presumption with no good arguments to the contrary—that T cannot adequately account for or accommodate C. A promise does not count as marginal for T if T implies that there are circumstances under which breaking the promise is morally permissible, all-things- considered. For as we saw in Chapter 1, promises yield pro tanto moral obligations. And when all else is not equal, pro tanto obligations can be undermined or overridden by the agent’s other reasons or by conflicting pro tanto obligations. When I claim that C is marginal for T, I am claiming that T cannot explain why the promise has any moral force, or why there is even a pro tanto obligation to keep it. In laying out this problem, I am not claiming that existing theories of promissory obligation are without worth. For most theories capture some important feature of promises, a feature that indeed tells us something about why we are obligated to keep promises in many or even most cases. And as we shall see in Chapter 3, these features ! 26! are sometimes necessary to explain the strength and stringency of the moral obligation incurred by a particular promise. Rather, the Problem of Marginal Cases states that each of these theories is in some way incomplete. Each might be able to explain the source and depth of moral obligation present in paradigmatic cases of promising. But each is insufficient to explain the full breadth of cases in which we are obligated to keep our promises. We can visualize the Problem of Marginal Cases with the following flower- shaped chart, with paradigmatic promises in the center and marginal cases as “petals” around the outside. 26 The details of the particular marginal cases will be discussed in subsequent sections; the general shape of the problem is what I wish to get on the table: !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 26 The size of each petal should not be taken to indicate anything about the importance of each marginal case. I have drawn the petals as (roughly) equal size merely for ease of visual presentation, and do not mean to thereby imply that each sort of marginal case is equally important, or that failing to accommodate each case is equally problematic for any given theory. Some of the marginal petals are more obviously or uncontroversially instances of promissory obligation than others; accommodating them is therefore more important. Paradigmatic promises No promise expectation Breach involves a bare wronging No social practice exists No intention to obligate Immoral promise Deathbed! promise Breach!has! better!conse< queces Lack of mutual account- ability No social stigma over breach ! 27! Most independently plausible theories can easily capture the center of the flower and explain why we are pro tanto morally obligated to keep promises in paradigmatic cases. But different theories struggle to accommodate one or more petals, and to explain why we are pro tanto obligated to keep our promises in various marginal cases. Proponents of these theories tend to wield such cases as dialectical weapons against each other, by arguing that view A is inadequate because it cannot capture case X, or that view B should be preferred because it can. 27 This style of argument presumes that any complete theory should be able to explain every plausible case of promissory obligation. This makes sense as a desideratum for a good account; more theoretical breadth and explanatory power is usually better than less. Moreover, since promises appear to be a unified phenomenon, we want a unified account of them, if possible. I raise the Problem of Marginal Cases to motivate my search for a new theory that meets this desideratum of breadth and unity. I aim to offer a maximally broad account of the basic feature that unites promissory obligation in all cases. Such a theory will not be subject to marginal cases of its own, for it will capture the wrong-making feature that is shared by every illegitimately broken promise, and thereby accommodate the entire flower chart. 2.! Representative Theories of Promissory Obligation Although individual theories of promissory obligation are quite varied, we can helpfully categorize different kinds of theories along thematic lines. Here, I survey !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 27 Scanlon criticizes conventionalist views in this way, arguing that they cannot accommodate his proposed case of promising without a social practice (as discussed in section 2.2.2 below). Scanlon’s critics make similar sorts of arguments about his view, claiming that it fails because it is subject to particular counterexamples; see, for example, Cholbi (2002, discussed in section 2.3.2 below) and Southwood and Friedrich (2009). Philosophers who argue in such ways typically also offer theoretical objections against the view they are targeting. But they assume that being unable to accommodate a case is highly problematic for a view. ! 28! eight categories of views with contemporary influence: (1) act consequentialist, (2) self- interested conventionalist, (3) actual rule conventionalist, (4) ideal rule conventionalist, (5) epistemic expectationalist, (6) normative expectationalist, (7) hybrid, and (8) normative power accounts, including rights transfer and authority transfer views. (The names I have given to views (2) – (6) are my own terms of art, and are not used by the proponents of the views.) In walking through these views, I provide a sense of the dialectic by noting the various ways in which one view has arisen out of criticisms of another. This discussion is not meant to be exhaustive; of necessity, I neglect particular representatives of each view type, and I ignore views that do not have much current popular support or influence in the contemporary literature on promissory obligation, such as virtue-ethical approaches, intuitionist accounts, and Kantian deontology. 28 After laying out examples of each kind of view, I explain what marginal cases each is subject to. For each putative marginal case, I argue that there is a plausible !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 28 Ancient philosophers such as Aristotle (Nicomachean Ethics, iv. vii, 1127a-1127b) and Cicero (De Officiis I,8) insisted that promise keeping is important because it is entailed by proper exercise of virtues such as honesty, justice, and liberality (for Aristotle), or fidelity (for Cicero). However, linking promissory obligation to virtue in such a bare way does not provide an informative and deeply explanatory account of the sort that contemporary philosophers tend to strive for. More recently, Philippa Foot (2001) has offered a naturalist and neo-Aristotelian virtue-ethical account according to which we have reason to do whatever is necessary for our full flourishing as humans in a cooperative society; she argues that facts about our dependence on each other “make it necessary for human beings to be able to bind each other to action through institutions such as promising” (2001, p. 198). The plausibility of Foot’s account will depend on whether it captures the right sort of normativity; why would a (perhaps dubious) claim about what will help the entire species flourish ground an individual obligation to keep promises? Intuitionist accounts like that of W.D. Ross (1930) are even more uninformative that classical virtue-ethical accounts; Ross insisted that it was intuitively obvious that we are pro tanto (or in his terminology, “prima facie”) obligated to keep our promises, and that we need do no more than consult our intuition to confirm this. Kant (1785) grounds promissory obligation in the categorical imperative; although his account is explanatory and informative, its implication that we may never break promises no matter the circumstances is deeply counter-intuitive. ! 29! argument to be made for counting the case as a genuine instance of pro tanto promissory obligation, and then explain why the case is marginal for the theory. 2.1.! Act consequentialist views A perennial worry about act consequentialism is that it cannot account for a stringent obligation to keep promises. On a simple act consequentialist theory, there are no distinctive pro tanto moral obligations to keep promises. Rather, promise-keeping is morally required whenever doing so will lead to the best consequences (e.g., keeping your promise to proofread your colleague’s essay, because this costs you little and will benefit her greatly, and therefore lead to an agent-neutral net increase in happiness.) Likewise, promise-breaking is morally required whenever doing so will lead to the best consequences (e.g., reneging on your promise to pay your affluent friend the $20 you borrowed from her last night, because donating it to charity will produce more utility.) In response to this sort of worry, act consequentialists have argued that their theories can in fact accommodate the right sort of promissory obligation. For example, Alastair Norcross denies that we need a robust notion of promissory obligation that is divorced from the consequences associated with individual promise breakings or keepings. 29 He accepts that promises should not be kept when the results are not optimal and attempts to cast doubt on the claim that they should be, in part by attacking the methodology of appealing to under-described hypothetical cases to illicit intuitions about when promise-breaking is morally permissible. Michael Smith takes a different approach, assuming that promises should be kept in at least some non-optimal cases and arguing for a consequentialism that maximizes agent-relative values, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 29 2011. ! 30! including values other than welfare. Smith claims that this can provide a consequentialist gloss on the important aspects of promissory obligation that have been pointed out by other theorists; for example, perhaps Scanlon is right that expectations are important, because satisfying the expectations that you personally have created maximizes some kind of non-utilitarian, agent-relative value for you. 30 2.1.2. Problems for act consequentialist views It seems clear that there are cases in which a promise must be kept even when doing so yields worse consequences. Consider W.D. Ross’s example of a promise whose keeping creates 1000 units of good, and whose breaking creates 1001 units of good. It seems obvious that one ought to keep the promise in such a case. 31 Cases like these are marginal for act consequentialist theories, which imply that the required action is whatever leads to better consequences, even if better by only one unit. (Of course, the calculation will have to be nuanced enough to include not only, for example, your friend’s disappointment at losing $20, but also degree to which your breaking the promise to repay her undermines our societal promising convention, and the extent to which this undermining leads to a loss of net utility.) Norcross’s and Smith’s defenses of act consequentialism do not fully alleviate these worries. For Norcross’s defense of act utilitarianism seems to involve biting the bullet, and accepting a claim (that promises ought not be kept when the results are not optimal) that other theorists take to be a reductio against act utilitarian views. And Smith’s agent-relative consequentialist defense of promising is very revisionary, as a key component—indeed, what many take to be the most essential and attractive !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 30 2011. 31 1930, p. 34 – 35. ! 31! component—of most consequentialist theories is that they are agent-neutral. Moreover, as Brad Hooker (who is a rule consequentialist) notes, theories that attempt to force act consequentialism to have the same results as deontological theories must “postulate so much (e.g., agent-relative value, desert, etc.) in order to squeeze everything into act- consequentialism that the resulting theory is left explaining very little.” 32 2.2.! Self-interested conventionalist views Conventionalist theories of promising argue that we cannot successfully make promises without the existence of some kind of social practice or convention of promising. They then go on to explain why we are obligated to keep promises, typically by arguing that the convention itself is good or valuable for us. The most influential early proponent of such a view was David Hume, who argued that promises are necessary for attaining mutually beneficial reciprocal exchanges with others, such as you helping me with my harvest this week in exchange for me helping you with yours next week. My natural selfishness and limited inclinations towards gratitude and generosity entail that I will be inclined to refrain from helping you once I have received the benefit from you. Predicting this, you will not be inclined to help me in the first place, and we will both be worse off than we would be if we had cooperated. Inventing a social convention of promising gets us around this problem. When people participate in this convention (for example, by saying “I promise to Φ”), they convey that they are committed to Φing, and they subject themselves “to the penalty of never being trusted again in case of failure.” 33 Self-interest then motivates people to keep their promises, since never being trusted again prevents one from being able to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 32 2011, p. 241. 33 Treatise 3.2.5.10, SBN 522 ! 32! make promises in the future, thus severely limiting one’s chances for ongoing mutually beneficial exchange. Furthermore, people in a society with such a convention will eventually begin to empathize with those who are victims of broken promises, and to approve of the convention and disapprove of violations of it, even apart from whatever impact violations have on future trust. This will motivate at least some people to keep at least some promises, even in cases when their future trustworthiness is unlikely to be affected, and their self-interest is therefore not at stake. 34 Contemporary philosophers have developed more nuanced self-interested conventionalist accounts. For example, Hanoch Sheinman’s account is Humean in spirit, as it grounds promissory obligation in self-interested reasons stemming from social practices. 35 Sheinman argues that someone who sincerely promises to Φ communicates the intention to give herself a self-interested, motivating, practice-based reason to Φ, and in doing so thereby successfully gives herself such a reason. 36 37 A “practice reason” is present when Φing as promised is in the promisor’s self interest because failure to do so goes against the rules of a practice. Violating these rules is against one’s self-interest because it can subject you to disapproval and hostile reactions from others, and can result in a loss of future trust, as Hume claimed. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 34 See Cohon (2006) for an account of the (often underlooked) nuances of Hume’s view. 35 See Sheinman (2008.) 36 Sheinman’s account is more detailed than this; he also claims that promising requires the promisor to believe (or at least purport to believe) that Φing is in the promisee’s interest, and that both promisor and promisee must be aware of this belief and of the fact that by promising, the promisor has communicated the intention to give herself a practice reason. 37 One who insincerely promises to Φ does not actually give herself such a reason, but at least communicates the intention to do so. Sheinman cashes out the relevant distinction not as a matter of sincerity, as I have done, but as a matter of “perfectly” or “imperfectly” promising. ! 33! 2.2.2. Problems for self-interested conventionalist views Self-interested conventionalist approaches like that of Hume and Sheinman struggle to explain the particularly moral nature of promissory obligation. If I fail to keep my promise to help you with your harvest in exchange for you helping me with mine, I act immorally. But a conventionalist account grounds promissory obligation in the fact that it’s better for me if I help you, since doing so makes me seem trustworthy and enables me to be helped by you and others in the future. Failing to help you therefore seems imprudent or irrational, but not necessarily immoral. Moreover, approaches like these are built on empirical claims about society’s reactions to promise-breakers that seem on the face of it to stand on rather shaky ground. In the globalized and impersonal world in which we live, how often will breaking a promise really result in a loss of future liability to trust? And is the prospect of avoiding hostile reactions from others really likely to motivate someone to keep a promise who is not antecedently motivated by gratitude or goodwill? There seem to be many situations in which the gains to one’s interests achieved by promise breaking— think of a con artist who promises victims that he will repay them the money they give him with interest, then never does so—will outweigh the losses to one’s interests brought about by hostile reactions—why should the con artist care if people disapprove of him, or if victims he will never encounter again would not trust him in the future? Such cases are marginal for self-interested conventionalist views. It is prima facie obvious that even a traveling con-artist who benefits greatly from breaking promises to repay loans and who is extremely unlikely to ever be caught is nevertheless morally obligated to repay his victims as promised. It would be absurd to allow that ! 34! Con-Artist 1, whose scam relies on cultivating trust among repeat customers, is obligated to repay his loans as promised, while Con-Artist 2, who will soon leave the country and head to a new place where people have no reason to distrust her, is not. But if the obligation to keep promises is grounded in the need to be trusted again in the future, as self-interested expectationalist views claim, then Con-Artist 1 is obligated to repay while Con-Artist 2 is not, which is implausible. 2.3.! Actual rule conventionalist views Conventionalist approaches that are grounded in rules rather than narrow self- interest avoid the preceding worries about cases where no loss of reputation is at stake. John Rawls offers such an account, according to which the practice of promising is defined by a set of constitutive rules, the basic rule of which is that one who promises to Φ must do so unless certain excusing conditions are in place. This practice is valuable because it enables us to “set up and to stabilize small-scale schemes of social cooperation by publicly maintaining an effective schedule of penalties,” and such cooperation is of benefit to everyone. 38 Rawls argues that fairness requires us to act in accordance with the just institutions and practices whose benefits we voluntarily accept. To fail to do so is to unfairly (and, ipso facto, morally problematically) free-ride on others’ participation in the practice. Promising is a just and valuable practice, and making a promise involves voluntarily participating in that practice. It follows that promise-keeping is morally obligatory, lest one unfairly free-ride on a beneficial and voluntary social practice. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 38 1971, p. 346 ! 35! 2.3.2. Problems for actual rule conventionalist views T.M. Scanlon has objected to actual rule conventionalism, citing two main difficulties with such theories; these concerns are not marginal cases, but rather are independent theoretical difficulties. First, Scanlon notes that it is possible to lead someone to rely on your behavior in ways other than by making promises, and failing to follow through on such behavior seems wrong in much the same way that failing to keep a promise seems wrong. Actual rule conventionalist theories that appeal to a distinctive practice of promising cannot explain why a similar thing is wrong with other sorts of failures of reliance. Second, free-riding on a social practice wrongs the entire community that adheres to and benefits from that practice. For example, if I don’t pay my taxes while all of my neighbors do, I hurt all of my neighbors by unfairly taking advantage of their participation in our cooperative scheme. But breaking a promise is a directed and particular wrong to the promisee. When I break my promise to meet you for lunch, I wrong you in particular, rather than (or perhaps in addition to) generally wronging everyone who participates in the practice of promising. In addition to these independent concerns, actual-rule conventionalist accounts are, at least arguably, subject to a class of marginal cases. Such accounts presume that making promises is impossible without an established social practice of promising operating in the background. In fact, most societies do have clear social conventions of promising. 39 However, it seems at least plausible that we can successfully make promises, and in doing so successfully impose promissory obligations, even without such social conventions—and therefore, without free-riding on any established social !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 39 See Korn and Decktor Korn (1983) for an empirical argument that the indigenous society of the small island nation of Tonga lacks any social convention or practice of (or even closely resembling) promising. ! 36! practice that we are illicitly benefiting from. If this is so, Rawls’s account will not explain what is going wrong in such cases, just as it does not explain what goes wrong in non-promissory failures of reliance. Scanlon offers an example of two strangers separated by a river. A’s errant spear has fallen on B’s side of the river, and B’s errant boomerang has fallen on A’s side of the river. A and B are from entirely different communities, and share no social practices in common. By means of gestures, A successfully communicates to B that she is committed to returning B’s spear, if B will then return her boomerang in exchange. There is nothing incoherent about such an exchange, and it seems at least prima facie plausible that A has promised to return B’s spear, and that B has promised to return A’s boomerang—and that as a result of these promises, they are each obligated to return the other’s weapons. If such cases are possible, they will be marginal for both Rawlsian and self-interested conventionalist views, which will be unable to explain why they count as genuine promises at all, let alone why acting on them is obligatory. 2.4.! Ideal rule conventionalist views Rawls cashes out promissory obligation in terms of the conventions and rules that are actually followed by a society. Other theorists cash out promissory obligation in terms of ideal or hypothetical conventions and rules. On these accounts, we don’t have to worry about whether a social practice of promising actually exists. Nor is the source of the obligation to keep promises the demand that we refrain from free-riding on a valuable social practice. Brad Hooker offers a rule consequentialist view of this sort. His version of rule consequentialism states that An act is wrong if it is forbidden by the code of rules whose internalization by the overwhelming majority of everyone everywhere in each new generation has ! 37! maximum expected value in terms of well-being, with some priority for the worst off. 40 The code of rules is fixed not by whatever rules or practices people actually have internalized, but by the rules that would lead to the best consequences if people were to internalize them. Hooker proposes that such a code would include a rule that we are generally obligated to keep our promises. Moreover, he claims that the practice of promising demanded by this rule would meet six theoretical desiderata that are typically thought the province of deontological theories of promising. The first of these desiderata is most relevant for our present discussion; it states that promissory obligation is not contingent on benefits to the promisee or others. 41 Hooker claims that only a rule meeting these desiderata will be such that it would lead to the best consequences if it were internalized by everyone. 2.4.2. Problems for ideal rule conventionalist views As Hooker notes, rule consequentialism is often accused of being subject to worries about incoherence or collapse. If rule consequentialism is overridingly committed to maximizing expected value, but insists that we should stick to some rule even in a case in which doing so does not maximize expected value, the position seems incoherent. And if rule consequentialism maintains coherence by allowing that we should act to maximize expected value even when doing so is against one of the established rules, then the view is extensionally equivalent to act consequentialism. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 40 2011, p. 222 41 ibid, p. 249. Hooker’s other desiderata are that promissory obligations: “(2) are self-imposed (autonomy); (3) are backward-looking in the sense that they depend on events in the past; (4) are agent-relative; (5) confer rights on particular others; and (6) give only some others (the promisees) the status of being wronged if the promise isn’t kept.” ! 38! Hooker offers replies to these objections that seem at least prima facie plausible, although it is beyond my present scope to fully evaluate them here. First, he denies that rule consequentialism as a theory is overridingly committed to maximizing expected value. Rather, the theory is committed to selecting rules on the basis of expected value, and then allowing those rules to determine moral wrongness. Hooker argues that this is unproblematic, for “there is nothing necessarily incoherent in evaluating codes of rules differently from acts.” 42 However, Hooker does not consider a related objection about rule worship. Even if it is not incoherent to fail to maximize expected utility in order to satisfy a rule, it seems problematically fetishistic, for it seems to involve privileging a merely instrumental means (the rule) at the cost of a valuable end (maximizing expected value.) Second, Hooker argues that the collapse worry will be prevented by the constraint that the rules be internalized. The costs associated with internalization of the rules are factored in to the expected value. As rules become more complex or demanding, or as they multiply in number, the cost of internalizing them increases. There will come a point at which the additional cost of internalizing extremely complex/demanding/varied rules will outweigh any increases in expected value afforded by such rules. So it is extremely unlikely that the rules will be able to become so complex that they are extensionally equivalent to act consequentialism. But even if Hooker’s defenses against the incoherence and collapse objections succeed, and he is able to offer a good reply to the rule worship objection, I doubt that his view can capture the first desideratum above (i.e., the claim that the obligatoriness of promise keeping is not contingent on any benefits accrued to the promisee or others.) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 42 ibid, p. 247 ! 39! If I am right about this, and if the desideratum is indeed true, then cases in which promise-breaking does not lead to any benefits will be marginal for Hooker’s view. Hooker argues that the proper promising rule will accommodate cases in which keeping the promise is not of benefit to the promisee, but benefits a third party instead. However, this does not get at the crux of the issue, since promise-keeping in such cases presumably still leads to an expected overall gain in well-being. 43 What about cases in which no one benefits? Hooker argues that a rule according to which promises are binding only when keeping them is beneficial would be non-ideal, because it would induce insecurity in the promisee. Because “keeping a promise is very often disadvantageous to the promisor,” promisors are likely to weasel out of promises whenever they can; it follows that “promisees need assurance that promisors are not left with much room for talking themselves into believing that their promises aren’t binding.” 44 Hooker claims that if the promising rule permitted us to break promises whenever doing so failed to be beneficial, promisees would not have such assurance. His reasoning is not explicit here, but he seems to assume that if the promising rule contained a loophole allowing promises to be broken whenever they were not beneficial, self-interested promisors would too often exploit this loophole by convincing themselves that their promise lacked benefit and was therefore permissible to break, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 43 Hooker also discusses cases in which making the promise is valuable, but keeping the promise is of no additional benefit. Such cases indeed seem possible. (Hooker offers an example of someone who promises his in-laws that he will wear black as a symbol of respect should his spouse die before him; the in-laws find value in this promise being made, as it is a symbol of the man’s love, but everyone involved recognizes that his wearing black wouldn’t actually benefit his deceased spouse or his in-laws.) These are cases in which it makes sense for the rule to allow the making of promises. But this doesn’t tell us anything about why the rule should require the keeping of promises in such cases. 44 Ibid, p. 250 ! 40! even when this was not in fact the case. And the existence of such an easily exploitable loophole would decrease promisees’ confidence that their promises would be kept. Hooker’s argument presupposes that promisee assurance is so important for the success of the overall promising practice that it is worth protecting even at the cost of requiring promise-keeping in particular cases that do not maximize well-being. It also presupposes that promisors are so motivated to get out of their burdensome promissory obligations that they will be highly unreliable at assessing when a promise fails to benefit the promisee; we cannot give them any leeway in so assessing, lest we open the floodgates to promisors wrongly assuming their promises are of no benefit. I worry that these presuppositions prove too much. Intuitively, promises may be permissibly broken in a wide variety of cases, such as when keeping the promise is incompatible with satisfying a more important obligation, or when the promise was based on a false claim. 45 If promisors are so motivated to get out of their burdensome obligations, won’t they also be overly quick to wrongly assume that a promise was premised on false information, or that keeping it conflicts with satisfying a more important obligation? And won’t this too lead to promisee insecurity and lack of assurance? The same reasoning Hooker employs to justify constraining the promising rule so that promises must be kept even when the promise is of no benefit to the promisee might also justify constraining the rule so that promises must be kept even when they conflict with a more important obligation, or are grounded in a false belief. Hooker presumes that maintaining promisee security is so important that it is worth the cost to overall expected well-being of occasionally requiring actions that do !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 45 I survey the conditions in which promises may be broken in detail in Chapter 7. ! 41! not benefit anyone in particular cases. If promisee security really is so essential to the successful practice of promising, it is at least plausible that maintaining promisee security is also worth the cost to overall expected overall well-being of occasionally acting on a promise grounded in a false belief, or of failing to satisfy more important obligations in particular cases. But this is a bad result; we don’t want our promising rule to be so strict and constrained. This should lead us to think that worries about promisors exploiting loopholes and lack of promisee security will not entail that the ideal promising rule must require promise keeping even in cases that lack benefit. It therefore seems that Hooker’s rule consequentialism cannot explain promissory obligation in cases in which no one benefits from the promise being kept, after all. Recall that this will be a problematic marginal case for his view only if his first desideratum is true, and promise-keeping is not in fact contingent on benefit. To illustrate the plausibility of this desideratum, consider the starkest sort of case—that of putative promissory obligations involving what David Owens calls bare wrongings, or wrongs with “no wrong-base, no basis in facts about human interest which might help explain why it constitutes a wronging.” 46 These are cases in which not only would no benefit accrue from keeping the promise, but no objective harm would result from breaking it. An oft-cited example of a promise involving a bare wronging is the real-life case of Nicholay Miklukho-Maklay, a 19 th century anthropologist who promised his Malayan servant that he would not photograph him, because the servant feared that being !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 46 2012, p. 127. ! 42! photographed endangered his soul. 47 It seems to many that Maklay is obligated to refrain from photographing the servant even while the servant is sleeping and would never know it about it—in spite of the fact that, presumably, the servant would not be harmed in any way by the taking of the photograph, while the scientific community would benefit from the anthropological documentation. I doubt whether this is in fact a good example of a bare wronging, as I doubt that the presumption about lack of harm is justified. For I find it plausible that the servant has an objective autonomy interest in controlling what happens to his image, an interest that would be harmed by his being photographed even without his knowledge. Moreover, the uneven power relationship between Maklay and the servant might be illicitly swaying our intuitions; photographing the servant while he sleeps seems like an invasive overreach of Maklay’s power as an employer, and might seem impermissible on those grounds alone, regardless of whether it constitutes a breach of promise. To ensure that our intuitions are not being sullied by such considerations, imagine a variant on the case in which Maklay wants to photograph a traditional ceremonial object that is owned collectively by the local community. Most of the community members have no problem with this, but Maklay’s co-worker is a religious fundamentalist who believes that photographing the object will damage the spirits the object depicts. Maklay promises his co-worker that he will not photograph the item. Assume that there is no chance that the co-worker will ever find out if Maklay takes the photograph. Moreover, the co-worker has no special autonomy claim over the communally-owned ceremonial object. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 47 This example is cited in Peter Kropotkin’s Memoirs of a Revolutionist (1899: Houghton Mifflin & Co.); it was brought into the mainstream philosophical literature by Philippa Foot (2001, p. 47 – 50), and is also discussed extensively by David Owens. ! 43! Does Maklay have any promissory obligation to his neighbor not to take the photograph? Keep in mind that the existence of such an obligation is compatible with it being the case that Maklay ought to break the promise, all-things-considered (perhaps the anthropological value of the photograph would be massive, and the promise is not very important, because the co-worker’s religious beliefs are unreasonable.) It nevertheless seems quite likely that Maklay is indeed subject to a promissory obligation, and must take the neighbor’s desires into account in a way that he would not be obligated to had he not made such a promise. If this is so, promising comes apart from benefit, and such cases are marginal for ideal rule conventionalist views like Hooker’s. 2.5.! Epistemic expectationalist views Scanlon offers a view of a different sort, grounding promissory obligation not in an actual or ideal rule or practice but in the expectations of the promisee. Scanlon argues that a distinctive good of promising is the value of assurance; making a promise assures the promisee that the promisor will act as specified, and the promisee has independent reason to want such assurance. Breaking a promise is morally wrong because it frustrates the promisee’s expectations and violates this assurance. Scanlon explains the moral obligation to keep our promises by appeal to the following complex principle of fidelity, which he argues is independently justified on contractualist grounds: Principle F: If (1) A voluntarily and intentionally leads B to expect that A will do X (unless B consents to A's not doing so); (2) A knows that B wants to be assured of this; (3) A acts with the aim of providing this assurance, and has good reason to believe that he or she has done so; (4) B knows that A has the intentions and beliefs just described; (5) A intends for B to know this, and knows that B does ! 44! know it; (6) B knows that A has this knowledge and intent, then, in the absence of special justification, A must do X unless B consents to X's not being done. 48 If all of the conditions of Principle F are met, a valid promise has been made, and the promisor is obligated to act in accordance with it. I call this an epistemic expectationalist approach, because it is grounded in what the promisee reasonably believes will happen as a result of the promise having been made. Other epistemic expectationalist approaches ground the obligation to keep promises in different ways, which need not be explicitly contractualist. For example, Judith Jarvis Thomson argues that promising is morally obligatory because promising causes the promisee to rely on one’s word that the promise will be kept, rather than assuring her. (This view differs from Scanlon’s in that one can be assured that an action will occur without changing her behavior and relying on it in any way.) Thomson argues that the promisee relies on the fact that the promised action will occur in her planning; if the promise is not kept, such reliance might prove to be detrimental to her. 49 Daniel Friedrich and Nicholas Southwood cash out their epistemic expectationalist view not in terms of the promisee’s reliance on or assurance by the promisor, but in terms of the promisee’s trust of the promisor. They argue that In virtue of inviting her to trust him to [for example] be faithful to her, and having the invitation accepted (or at least not rejected), [the promisor] incurs an obligation not to betray the trust that he has invited in [the promisee.] The wrong involved in his breaking his promise, if he chooses to do so, is a matter of his violating this obligation, that is, of his betraying the trust that he has invited. 50 Friedrich and Southwood distinguish their view from Scanlon’s by noting that one can create assurance without trust, as when you impart relevant and reliable information about your plans to Φ without inviting your interlocutor to have faith that you will Φ. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 48 1998 p. 304 49 1990 50 2011, p. 278. ! 45! One can also trust someone to Φ without being assured that she will Φ, as when you agree to trust that a perpetually unfaithful spouse will not stray this time, not because you are assured of your spouse’s behavior but because it is important for your relationship that you at least attempt to trust her or him. What unites all of these epistemic expectationalist approaches is the focus on the effects that the making of the promise has on the promisee (e.g., inducing expectations, reliance, or trust) and on the negative effects that would befall the promisee if the promise is broken (e.g., unmet expectations, detrimental reliance, or betrayed trust.) 2.5.2. Problems for epistemic expectationalist views Usually, promisees expect promisors to act as they have promised; they also rely on such promises, and trust that the promisors will perform. Sometimes, however, the promisee does not expect the promisor to perform, and therefore is not assured by the promise. Scanlon himself discusses such a case, that of a “Profligate Pal” who is notoriously bad about repaying debts. The Pal asks you for a loan, and promises to pay you back. You fully expect that you’ll never see your money again, but you accept the promise to spare the Pal the embarrassment of receiving charity from you. Epistemic expectationalist views predict that the Pal does not have a promissory obligation to pay you back in such a case—for since you know that the Pal is unlikely to return your money, you are not assured by him, do not rely on him, and do not actually trust him. As Berislav Marušić notes, expectationalist views entail that if the promisor does “not succeed in forming expectations in the promisee . . . she will thus fail to incur promissory obligations; her promise, if it is one at all, won’t be binding.” 51 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 51 2013, p. 305. ! 46! This seems problematic, for promises that do not generate expectations still seem to yield pro tanto obligations. Surely the Pal is obligated to repay you; it would be worrisome indeed if you were justified in breaking a promise because the promisee assumed you were unreliable and did not expect you to keep your promise. It might even provide promisors with perverse incentives to cultivate unreliability, so as to get themselves off the moral hook for keeping their promises. Scanlon attempts to defuse this worry by arguing that we can account for the intuitive force of the claim that the Pal is obligated to pay back the loan by appealing to gratitude; from the point of view of the promisor, “what [the Pal] has received is a gift, tactfully described as a loan.” 52 If the Pal does not repay, this is problematic because he is failing to act upon the debt of gratitude he owes for the gift. However, as Michael Cholbi argues, the obligations of the Profligate Pal go beyond the way in which we typically assume gratitude obligates us. Suppose you give the Pal a lovely fountain pen you acquired while traveling. He owes some sort of duty of gratitude to you—he should thank you sincerely, and perhaps should mention the pen when appropriate in the future, and refrain from throwing it in the trash or using it as a door stop. But, Cholbi argues, he is “certainly not obligated, because [you] gave him the fountain pen … to give [you] a similar gift in kind.” 53 The Pal’s duty of gratitude to you “do[es] not involve any sort of reciprocation or analogous ‘repayment,’ akin to the duty Scanlon attributes” to him in the loan case. 54 If the Pal owed you a duty of gratitude, it would require him to do something like thank you, and use the funds for an appropriate purpose (e.g., paying his mortgage and not gambling.) But the duty of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 52 Scanlon, 1998, p. 314. 53 2002, p. 480. 54 ibid. ! 47! gratitude does not require him to repay you; because “you did not lend him the money with no strings attached…the loan did not as a matter of fact function as a gift.” 55 Furthermore, even if Scanlon’s appeal to duties of gratitude worked in the case of the Profligate Pal, there are plausible cases involving a lack of promisee expectation which do not involve gratitude, and therefore cannot be explained away in the same way. For example, suppose it is very important to you that the people you love drink good beer and try new things. Your cousin is in the habit of ordering a Bud Light every time you go to a bar together. After much discussion, he finally promises you that the next time the two of you are out together, he will order a local IPA instead. You don’t expect him to keep this promise, because you know what a creature of habit he is. So you are not assured by his claim, and you in no way rely on his ordering an IPA or trust that he will do so. Yet it seems clear that your cousin is morally obligated, at least all else being equal, to keep his to you promise and order an IPA. And he certainly does not owe it to you to order an IPA because he is grateful to you, or because he has an independent moral obligation to drink bitter and complex beers, or for any other reason. His obligation is clearly promissory, and the case is therefore marginal for epistemic expectationalist views. 2.6.! Normative expectationalist views Because epistemic expectationalist views are grounded in the promisee’s beliefs about what the promisor will do, they struggle to explain promissory obligation in cases in which the promisee doesn’t believe the promisor will follow through. Expectationalists can avoid this problem by grounding promissory obligation not in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 55 ibid. ! 48! what the promisee actually expects, but in what the promisee is entitled to expect. Such accounts focus on what the promisee can expect in a normative sense (“I expect you to study hard for your exam!”), rather than what they expect in an epistemic sense (“But I expect that you won’t actually start studying until the morning of the test.”) Michael Cholbi develops a contractualist version of such a normative expectationalist account in response to the Profligate Pal case, claiming that the fact that the promise recipient does not factually expect the promise to be kept does not affect whether the promise recipient may morally expect it to be kept. As Scanlon states, promising gives us a “right to rely” on others, but having a right to rely on them is distinct from believing in their reliability. 56 Cholbi cashes out his contractualist “Promising as Accountability” theory in terms of an agreement to hold accountable: the promisor agrees to be accountable to the promisee to perform the promised action, and the promisee agrees to hold the promisor accountable for acting as specified, by (for example) being disposed to blame, criticize, and express disapproval of the promisor if the promise is broken. This notion of accountability concerns taking action and “is a relational notion: it means that B is within her righs to subject A to moral condemnation, criticism, or scorn, and in institutional contexts, to appropriate punishment or sanctions.” 57 In the event of a breach, the promisee need not in fact hold the promisor accountable; for example, the promisee might refrain from holding accountable because she reasonably fears reprisal if she criticizes an angry promisor. But even if she has an independent reason not to criticize the promisor, the promisee must at least be “justified in not refraining from criticism, scorn, etc,” such that “the promise maker could not !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 56 2002, p. 483 – 484. 57 2002, p. 485. ! 49! object if subjected to the negative consequences of promise breaking.” 58 Ultimately, Cholbi claims, we are obligated to keep promises because failing to do so is “failing to perform acts that one has made oneself accountable for.” 59 2.6.2. Problems for normative expectationalist views I worry that Cholbi’s account demands too much of the promisee, and does not make room for forgiveness. Cholbi grants that promisees may permissibly refrain from holding accountable in particular cases in which harm would be caused by blaming or criticizing. But he insists that in general, in agreeing to be held accountable for a promise the promisor essentially says the following to the promisee I recognize you as a being who has reasons for wanting Φ to happen and I take on the burden of making the end served by Φing an end of my own. I hope you will grant me the same recognition, for in promising, I have manifested an attitude of respect for you, and it would be insulting, and even disrespectful, for you not to hold me accountable when I don’t Φ. To do that would be to deny my capacity to make your end mine in the first place. The promisee, in turn, says the following to the promisor: In acknowledging your promise to Φ, I show my respect for you as a being who is able to incorporate my ends into your own. If you fail in your duty of fidelity to me, you can expect that I will hold you up to moral scrutiny as a sign of my respect for you as a moral being capable of acting on reasons whose authority you recognize. 60 Under Cholbi’s contractualist account, the promisor must act as promised. And the promisee incurs an obligation, as well—to hold the promisor accountable by holding the promisee up to “moral scrutiny” by (for example) blaming or criticizing. However, it is not plausible to think that promisees are obligated to hold their promisors accountable in such a way. Consider a case involving gratuitous forgiveness. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 58 2002, p. 484. 59 ibid. 60 Both block quotations are from 2002, p. 489, and contain my emphases in italics. ! 50! Suppose your spouse promises you that he will remember to pick up your clothing from the dry cleaner on his way home from work, since you won’t be able to get to the cleaners yourself before they close. He arrives home without your dry cleaning. Assume this is not a case in which holding him up to moral scrutiny would lead to bad results; it wouldn’t damage a fragile relationship or saddle him with disproportionate guilt. Must you blame or criticize him for his failure to keep his promise? Or is it permissible—even morally supererogatory —for you to simply tell him that you forgive him? Maybe you shouldn’t forgive everyone all of the time for every promise they break to you. But surely it is at least sometimes permissible, and even desirable, to refrain from blaming or criticizing those who break their promises. Cholbi worries that to fail to hold the promisor accountable is to fail to respect the promisor as a moral agent capable of acting for reasons. But this doesn’t seem to be the case. For, as I will argue in Chapter 5, promises must be accepted; the promisor cannot make a promise unilaterally, but must have implicit or explicit uptake from the promisee. In accepting a promise in the first place, you treat the promisor as a moral agent capable of keeping her promise and taking your end as her own. You cannot accept promises from creatures who are not moral agents, such as dogs or toddlers. Voluntarily accepting a promise is therefore sufficient to demonstrate such respect, which entails that criticizing the promisor upon breach is not essential. Furthermore, Cholbi’s normative expectationalist account is subject to a marginal case. For there seem to be cases of genuine promissory obligation in which the promisee lacks the standing to hold the promisor accountable. If the promisor lacks the relevant standing, she will not even be justified in blaming or criticizing the promisee, and the reciprocal accountability on which Cholbi’s account is based will be ! 51! undermined. Suppose that Alyssa and Brendan are coworkers. It’s Sunday, and Alyssa promises to give Brendan a ride home from work on Friday. On Monday, Tuesday, Wednesday, and Thursday, Brendan promises to give Alyssa a ride home from work. But he breaks his promise every time, quitting early and driving off without telling her, four days in a row. When she confronts him about these broken promises, he does not repent or apologize, but simply promises to drive her home the next day. So Brendan repeatedly breaks his promises to Alyssa, with no signs of changing his ways. If Alyssa breaks her promise to Brendan to drive him home on Friday, Brendan will not have the standing to take action and hold her accountable by blaming, criticizing, or otherwise holding her up to moral scrutiny. Doing so would be massively hypocritical on his part, and Alyssa could rightly object to his pointing out the speck in her eye when he has a log in his own. But does this entail that Alyssa is entirely off the moral hook for driving Brendan home on Friday? That she owes him nothing, not even a pro tanto obligation, and is morally permitted to drive off alone without a second thought, in the way that she would be had she not promised to drive him home? Certainly, Alyssa’s breaking a promise to drive Brendan home when he has broken similar promises to her wouldn’t be as bad as breaking a promise to drive him home would be had he not made her any similar promises, or when had he made similar promises and kept them. But Alyssa does not seem to be entirely off the moral hook. It would be reasonable for her to demand that Brendan release her from her promise. But if she doesn’t seek and successfully attain such release beforehand, there is a sense in which she remains obligated to keep her promise. Allowing her to break the promise seems like allowing the fact that you did wrong to me to entail that I am ! 52! permitted to do similar wrong to you in return. If this is true, then cases like Alyssa’s and Brendan’s are marginal for Cholbi’s view. 2.7. Hybrid views Dissatisfied with purely expectationalist or purely conventionalist approaches, a number of philosophers have recently developed hybrid theories of promissory obligation, which combine some element of expectationalism with some element of conventionalism. Niko Kolodny and R. Jay Wallace have offered one such influential account as a reply to—or really, an elaboration of—Scanlon’s view. They argue that Scanlon’s view faces a circularity worry and cannot in fact provide assurance. Assurance is generated when all of the stages of Principle F are satisfied, but step (1) of Principle F—that the promisor lead the promisee to expect that she will Φ—cannot be satisfied unless the promisee is already assured that the promisee will act. Kolodny and Wallace appeal to the communicative act involved in the social practice of promising to solve this circularity worry and provide the sort of assurance Scanlon seeks: A’s uttering the formula, ‘I promise to do X,’ which is not yet A’s assuring B that A will do X, results in A’s incurring a practice-based [moral] obligation to do X, which only then assures B that A will do X. 61 They then argue as Rawls does that the practice-based moral obligation to act as you have promised stems from a prohibition on free-riding on others’ participation in the practice. This appeal to a social convention gives the promisee a reason to expect that the promisee will act, and therefore solves the circularity worry. And unlike pure conventionalist views, the appeal to expectations explains the similar wrongs involved in breaking a promise and in failing to satisfy non-promissory expectations, and grounds the directedness of the promissory obligation. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 61 2003, p. 149. ! 53! Other philosophers have offered different hybrid views, all of which draw on a combination of expectationalist and conventionalist theories. For example, Neil Tognazzini tweaks Kolodny and Wallace’s view, arguing that the appeal they make to Principle F is unnecessary, and appealing instead to Scanlon’s weaker principles M (forbidding manipulation) and D (requiring due care with regard to the way in which one creates expectations in others.) 62 2.7.2. Problems for hybrid views If there is a circularity worry for Scanlon, Kolodny and Wallace’s hybrid view gets around it. But the view may strike one as strangely piecemeal. Elinor Mason claims that the “account is not hybrid in any useful way.” 63 Rather, she argues that Kolodny and Wallace offer accounts of two separate wrongs: the wrong of violating expectations and misleading people (i.e., the wrong highlighted by Scanlon’s view), and the wrong of free-riding on a socially valuable practice (i.e., the wrong highlighted by the practice view.) Mason’s worry that the view is not meaningfully hybrid applies mutatis mutandis to Tognazzini’s modified view as well, and to other similarly structured hybrid views. By combining two styles of views, Kolodny and Wallace seem to be subject to a narrower range of marginal cases than either expectationalist or conventionalist views alone. For example, Kolodny and Wallace note that their account implies that the Profligate Pal “has a practice-based moral obligation to keep his promise to repay the loan” even though he does not have an assurance based obligation stemming from !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 62 2007. 63 2005, p. 39. ! 54! Principle F to do so. 64 But since their view essentially relies on the appeal to social practices to solve the circularity worry, it will be unable to accommodate cases in which there is no promising practice, like Scanlon’s boomerang/spear exchange case. There is an important insight in hybrid views, which I mentioned briefly in Section 1. Hybrid views like that of Kolodny and Wallace nicely illustrate the way in which different theories of promising capture distinct, important, but ultimately incomplete aspects of promissory obligation. Hybrid views recognize that expectations are morally important for explaining why we are obligated to keep our promises, as Scanlon and the expectationalists claim. They also recognize that such views are incomplete; they cannot by themselves capture everything there is to say about promissory obligation. Kolodny and Wallace attempt to fill this gap by supplementing Scanlon’s account with Rawls’s conventionalist view. In this dissertation, I suggest that the best strategy isn’t finding one theory and plugging the gaps in it with another. Rather, the best strategy is finding a more comprehensive and basic explanation. 2.8.! Normative power views The idea that we can create promissory obligations out of thin air simply by willing so has seemed puzzling to some. Hume famously found this one of the most mysterious and incomprehensible operations that can possibly be imagined, and may even be compared to transubstantiation or holy orders, where a certain form of words, along with a certain intention, changes entirely the nature of an external object, and even of a human creature. 65 This puzzlement led Hume and others since to posit a conventionalist account of promissory obligation. Recently, however, a number of contemporary philosophers !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 64 2003, p. 153 (footnote 27) 65 1740, p. 524. ! 55! including Joseph Raz, Gary Watson, David Owens, and Seana Shiffrin have embraced the idea that we have the ability to change the normative situation directly and by declaration by exercising a normative power. To exercise a normative power is to “change what someone is obliged to do by intentionally communicating the intention of hereby so doing.” 66 Normative power theories claim that promissory obligation stems from an exercise of such power, and that such powers exist because they are in some way valuable for us to have. What is distinctive about such views is not the fact that they involve the ability to bind yourself by declaration, but the transcendental style of argument by which their proponents argue for the power to so bind. 67 In general, these views claim that X is an important feature of our normative lives, and that the only way we can adequately possess X is by having the ability to bind ourselves to each other through promising. The transcendental step comes in with the conclusion that we therefore must possess the ability to bind ourselves to each other through promising, lest X be inaccessible. Normative power theorists cash out what the normative power consists in and what valuable thing it supports in a variety of ways. For example, Raz argues that there is an “intrinsic desirability of forms of life in which people create or acknowledge special bonds between them and certain other individuals,” and that sustaining such bonds requires the “promisor to be, in the matter of the promise, partial to the promisee.” 68 This justifies the existence of a general promise-keeping norm, which in turn entails that individual acts of promise-keeping are obligatory. The general norm !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 66 Owens (2012) p. 4. 67 Shiffrin explicitly construes her argument as a transcendental argument; Owens and Raz do not. However, they follow a similar argumentative style as Shiffrin, and for that reason are helpfully grouped together. 68 ibid, p. 227, 228. ! 56! requires someone who promises to Φ to exclude from consideration any reasons for or against Φing other than the reason to Φ that is provided by the promise; the promisor must “regard the claim of the promisee as not just one of the many claims that every person has for his respect and help but as having peremptory force.” 69 Seana Shiffrin cashes out the normative power to promise not in terms of exclusionary reasons but in terms of rights, arguing that in promising to Φ, “the promisor transfers his or her right to act otherwise to the promisee.” 70 Acting in accordance with this promise is morally obligatory because to fail to do so “is to act in a way the promisor has no right to do.” 71 Shiffrin argues that this power is valuable because it is necessary for minimally morally decent intimate relationships. The ability to make a promise represents the power promises give friends to navigate around the fact that they are not identically motivated and may not always share the same agenda. The ability to promise provides a crucial tool to permit mutual engagement among equals, who are nevertheless distinct and diverse, without either party feeling the pressure to homogenize. 72 Promising acknowledges the promisee’s “investment in the situation in a way that includes [her] rather than making [her] a bystander to what unfolds,” which “actively affirms [her] status as a free person, capable and worthy of exercising sound judgment about what is to be done.” 73 David Owens proposes a normative power based in what he calls our “authority interest,” or the interest we have in having a certain practical authority over others, the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 69 1977, p. 228. 70 2008, p. 517. 71 Ibid. 72 Ibid. p. 506 – 507. 73 Ibid p. 507, 508. ! 57! authority that being the recipient of a promise gives us. By making a promise, the promisor “give[s] the promisee the right to require performance of the promisor;” this serves our authority interest in being able to determine by declaration whether another person is obligated to perform an act that she has promised to do. 74 Owens argues that normative interests like this one not only exist, but are not derivative from our non- normative interests. That is, he claims that we independently value our abilities to change the normative situation, regardless of whether they affect the outcomes that occur in the world. Owens's account is essentially conventional, as well; he argues that the promissory obligations stemming from our authority interest must be publicly recognized against a background of social practice, claiming that breaches of promise “constitute wrongs only against a background of social convention,” since “for it to be a good thing for us that we are subject to a norm, that norm must have a certain social reality; its authority must be recognized by the people around us.” 75 All of these views have a similar structure. Some important good which involves a change in the normative situation is identified, whether this is the ability to create and acknowledge special bonds, the sustaining of minimally decent moral relationships through the transfer of rights, or our normative interest in possessing authority over ourselves and being able to cede or transfer it to others when we choose. It is then assumed that we can change the normative situation in this way by declaration—and that in doing so, we are exercising the normative power of promising. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 74 2012, p. 144. 75 Ibid, p. 9 – 10. ! 58! 2.8.2. Problems for normative power views Transcendental arguments as a class have been subject to much philosophical controversy. 76 But even if we are not concerned about the plausibility of transcendental arguments, and do not share Hume’s discomfort with the “mysterious” nature of brute exercises of normative power, there are still worries about marginal cases for normative power views. First, just as there is some sense in which we are pro tanto obligated to keep promises even in cases of bare wrongings, there is some sense in which we are obligated to keep deathbed promises, or promises that we make to someone before she dies and which are to be kept only after she dies. Consider the dutiful son who promises his mother that he will scatter her ashes at her childhood home, or the soldier on the front lines who promises her dying comrade that she will deliver a final letter to the comrade’s loved ones. The son who fails to scatter the ashes or the soldier who fails to deliver the letter seem to wrong the mother and the comrade, in an obvious and direct sort of way. And these wrongs seem to be different in kind than the wrongs committed by people who make no promises, and then fail to act as they know their deceased friends or family would have wanted them to act. As Elinor Mason notes, “to those who take promising seriously it seems that I have a reason to keep a deathbed promise, even when the act required is clearly ridiculous.” 77 Philosophers differ as to whether the dead can be harmed, and if so, how. 78 Promises to the dead are marginal for normative power views regardless of whether the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 76 See Barry Stroud’s “Transcendental Arguments” (1968). 77 2005, p. 41 78 Joel Feinberg (1987), Steven Lupor (2007), and George Pitcher (1984) each independently argue that the antemortem individual may be harmed by events which occur after her death. For skepticism about these claims (and about whether accounts of promissory obligation that ! 59! dead can be harmed. Suppose that promising to Φ involves transferring to the promisee the right to decide whether you Φ, as Shiffrin claims; the son transfers to his mother the right to decide what he does with her ashes. Rights require rights-bearers; once the mother ceases to exist, how can the right to decide reside with her? Similar arguments can be run mutatis mutandis with Owens's authority transfer view, on the reasonable assumption that there must be a person in whom the authority resides. Another marginal case stems from the fact that the normative power views under consideration all assume that in making a promise, the promisor must intend to obligate herself. In Chapter 4, I argue that one can successfully promise, and can thereby successfully impose a pro tanto promissory obligation, without deliberately exercising a normative power by forming such an intention. If I am right about this, then promises in which the promisor lacks the intention to obligate herself will be marginal for normative power views. 79 A third marginal case is that of immoral promises. Some normative power theorists like Shiffrin and Gary Watson argue that immoral promises are invalid, or such that they do not yield even pro tanto obligations to act as promised. Shiffrin’s rights- transfer view “seems to presuppose that the promisor had a right to perform or not to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! conceive of promises as directed obligations can accommodate the obligation to keep promises to the dead), see Dressel (2014). 79 Since normative power views understand promising as changing the normative situation by declaration, it is natural that they typically construe promising as involving an intention to obligate yourself. I don’t know of any normative power views in print that do not claim that promising involves intending to obligate yourself. However, it might be possible to construct such a view—in which case, it would not be subject to this particular marginal case, although it would be subject to the other marginal cases discussed in this section. ! 60! perform the promised action.” 80 But we do not have rights to perform immoral actions, and so cannot transfer a right that we do not have in the first place. Similarly, Watson argues that “my promissory proposal cannot have its intended normative effect unless I am entitled to make and keep the promise,” just as I cannot give you an item that I do not own. 81 Not all normative power views imply that immoral promises are invalid; for example, Owens argues that our authority interest is not undermined if a promise has immoral content, and so immoral promises can yield pro tanto obligations. I argued in Section 4 of Chapter 1 that immoral promises are valid, or such that they do yield pro tanto obligations. (If you skipped Chapter 1, and you don’t already agree with me that immoral promises are valid, you may want to review this section.) If I am right, then immoral promises will be marginal cases for views like Shiffrin’s and Watson’s. 2.9.! Summing up the Problem of Marginal Cases We can summarize the marginal cases we have discussed with the following chart, which notes which cases are marginal for which views. THEORY Breach = better con- sequences No social stigma No social practice of promising Bare wronging No promisee expectation Lack of accountability Deathbed promise 82 No intention to obligate Immoral promise Act util. (Norcross) MARGINAL Might be OK Might be OK Likely marginal Might be OK Might be OK Likely marginal Might be OK Likely marginal Self-int. convent. (Hume) OK MARGINAL MARGINAL OK OK OK OK OK OK !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 80 2011, p. 156 81 2009, p. 169 82 If the dead cannot be harmed, then deathbed promises will be cases of bare wrongings (although there might be instances in which the wrongings are not completely bare, because living people know about the promise and would be hurt or offended by its breach.) If deathbed promises are bare wrongings, they will be marginal for act (and probably rule) utilitarians, as are all promises involving bare wrongings. And deathbed promises will be marginal for expectationalist views regardless of whether the dead can be harmed—for a dead person no longer has any normative or epistemic expectations. ! 61! Actual rule conven. (Rawls) OK OK MARGINAL OK OK OK OK OK OK Ideal rule conven. (Hooker) OK OK OK MARGINAL Likely marginal OK Likely marginal OK Likely marginal Epistemic expect. (Scanlon) OK OK OK Maybe OK, maybe marginal 83 MARGINAL OK MARGINAL OK OK Normative expect. (Cholbi) OK OK OK OK OK MARGINAL MARGINAL OK OK Hybrid (Kolodny & Wallace) OK OK MARGINAL OK OK OK OK OK OK Rights tran. (Shiffrin) OK OK OK OK OK OK MARGINAL MARGINAL MARGINAL Auth. tran. (Owens) OK OK MARGINAL OK OK OK MARGINAL MARGINAL OK Some theories are able to accommodate more cases than others. But no theory is able to accommodate every kind of case; at least one case is marginal for each type of view. Moreover, each type of view faces theoretical difficulties of some sort, as well, as we saw in the discussions above. What matters for my present purposes is not whether any particular case is in fact marginal for any particular theory, or whether there are adequate responses to the theoretical objections I have canvassed for each view. Rather, what matters is that each theory is flawed or incomplete in some way—because it faces a serious theoretical difficulty, or because it cannot accommodate one or more case that it should be able to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 83 I do not include bare wrongings as marginal cases for expectational views, because I am not certain about whether someone’s assurance or trust can be violated even if he never finds out about it. When Maklay takes the photograph of the ceremonial object, does he violate his co-worker’s trust or disappoint his expectations somehow, even though the co-worker will never discover this? I don’t know. However, bare wrongings can be accommodated by a normative expectationalist view like Cholbi’s, because Maklay remains accountable to his co-worker even if he never discovers the breach, and the co-worker remains entitled to rely on Maklay and would be justified in criticizing him for the breach, were he ever to find out about it. ! 62! explain, or both. For the presence of any sort of flaw in each of the most influential contemporary theories is enough to motivate us to seek an alternate explanation of promissory obligation. This is the task I turn to next. 3.! Setting the Stage for a New Theory In framing the Problem of Marginal Cases, I aim to have pointed out a myriad of ways in which various popular theories of promissory obligation fail to accommodate one kind of marginal case or another, and therefore fail to be fully comprehensive. My goal is motivating the search for a theory with enough breadth to cover every plausible case of promissory obligation. One might be worried that the Problem of Marginal Cases is not a unified problem—that I am simply lumping together a variety of objections under one label. However, this is not worrisome for me, as I don’t claim that there is any deep unity to the problem. There seems to be at least some surface level unity, as each theory struggles with at least the same sort of difficulty—namely, a failure of maximal breadth. The ultimate source of these failures, and whether they are unified in any way, is unimportant for my purposes. My focus in this chapter has been on the limits of each theory. Focusing exclusively on the marginal cases that a theory cannot accommodate risks missing out what each theory does successfully explain. Each of the theories I have analyzed in this chapter points to something of moral importance relating to our practice of promising. The expectations we induce in others matter, both when it comes to keeping promises and anytime we knowingly induce expectations in others. Consequences matter, too; it is important to realize that promise-breaking often has bad consequences, while promise-keeping often has good consequences. Promising indeed presupposes ! 63! reciprocal accountability in most cases. And there is something unfair about free-riding on a beneficial social practice, a practice that is undoubtedly made stronger if people regularly respect it. Most of us do value the way in which promising allows us to maintain our intimate relationships, and give others control over a limited range of our actions. All of this gives us a fuller understanding of what our rich and varied practices of promising consist in, and why they matter to us. Where each of these theories errs, however, is assuming that any one of these features alone captures an exclusive and fundamental source of promissory obligation that is present for all promises, including those at the margins of our practices. In the rest of this dissertation, I propose that the most basic, broad, and essentially promissory source of the obligation to keep our promises stems not from assurance or expectation, or from a social practice, but from the mental state expressed by promising itself—a theory I call the Mental States First theory of promising. As we shall see in later chapters, the Mental States First theory will in some cases be insufficient to explain the full depth of moral obligation resulting from promises; we will sometimes need to appeal to other features, such as satisfying expectations or the maintaining of intimate relationships, to explain why a particular promissory obligation is so strong. But the Mental States First theory is necessary to explain the full breadth of promissory obligation, and to explain why we are morally obligated in at least a minimal way in all of the marginal cases from the flower chart above. ! 64! CHAPTER 3: THE SPEECH ACT SINCERITY PATTERN 1.! Introduction In Chapter 2, we saw how most theories of promissory obligation are subject to marginal cases. In this chapter, I introduce the Mental States First theory of promising (MSF) as an alternative that avoids this problem. According to MSF, we are obligated to keep our promises because we are obligated to act in accordance with the strong intentions, or resolutions, that we express when we make a promise. Because MSF grounds promissory obligation in the mental state that is expressed whenever someone promises, it can explain why there is at least a minimal sort of obligation in every case; this is fully compatible with there being further, more robust sources of obligation to keep promises in most cases. In this chapter, I will outline my view. Space constraints entail that I must make certain assumptions for the sake of argument in my initial presentation of it; in the next three chapters, I will defend those assumptions in detail. My main task in this chapter is motivating my proposal by arguing that promising patterns with a broad range of speech acts—including complimenting, criticizing, thanking, congratulating, forgiving, and apologizing—in being such that at least some of the norms governing the speech act are inherited from the mental state, for both sincere and insincere utterances of that act. This defuses an objection about whether MSF can accommodate insincere cases, and is also an independent proposal about how the normativity of speech acts depends on the normativity of the mental states communicated by those acts. ! 65! 2.! Expression and Communicative Sincerity In seeking a solution to the Problem of Marginal Cases, we should seek to ground promissory obligation in something that is present to the same degree for all promises. One thing all promises have in common is that they involve a communicative act of a certain sort. This is often a verbal speech act, but can be written or even non-verbal; a nod or a stern look can successfully communicate a promise in the right context. And one thing you do whenever you perform a communicative speech act of this sort is express that you are in a certain mental state. For a speech act S to express a mental state M is for S to convey or communicate that the speaker is in M, and for anyone who utters S sincerely to be in state M. Because being in M is necessary for uttering S sincerely, M is a sincerity condition on S. John Searle uses this notion of expression in his analysis of speech acts, claiming that “whenever there is a psychological state specified in the sincerity condition, the performance of the act counts as an expression of that psychological state.” In other words, “the sincerity condition tells us what the speaker expresses in the performance of the act.” 84 This is a common understanding expression and sincerity; as Michael Ridge notes, “Searle’s view of sincerity is very seductive and seems to have functioned as an implicit philosophical orthodoxy.” 85 This notion of expression does not imply that a speaker can express M only if she really is in state M. If you utter S, you succeed in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 84 Both quotations are from 1969, p. 65 85 2006, p. 488. See Green (2007), Davis (2008, p. 427–445), and Eriksson (2011, p. 215) for more representatives of this Searlean view. Ridge himself does not share this view. Rather, he argues that a speaker is sincere when she believes herself to be in the mental state that she believes her utterance expresses. This modification is meant to accommodate cases in which the speaker is mistaken about her own mental states. ! 66! performing the relevant speech act and expressing M, even if you are insincere and are not in fact in M. Call this Searlean notion of sincerity communicative sincerity, defined as follows: Communicative Sincerity: A speech act S that expresses state M and is performed by an agent A is communicatively sincere if and only if A really is in state M. To test whether S expresses M, we can appeal to intuitions about communicative sincerity. Can you sincerely utter S without being in state M? If not, then S expresses M, and M is a sincerity condition on S. If so, then S does not express M, and M is not a sincerity condition on S. Call this test the linguistic test of communicative sincerity. All promises involve the performance of a speech act, and all illocutionary speech acts express a particular mental state. If we can ground promissory obligation in the mental state that promises express, we will have a source of promissory obligation that is present for all promises, no matter how non-paradigmatic they are. In the next section, I outline such a theory. 3.! Introducing the Mental States First Theory To ground promissory obligation in the mental state expressed by promises, we first need to know what mental state promises express. I argue in Chapters 4 and 5 that promising to Φ expresses a resolution to Φ, conditional on the acceptance of the promise by the promisee. That is, promising to Φ expresses that you seriously plan to Φ, or that you have an especially robust intention to Φ, and that you will stick to this plan so long as the promisee accepts your promise and does not release you. I argue in Chapter 6 that resolving to Φ rationally commits you to Φing, all else being equal; someone who ! 67! resolves to Φ must do so on pain of irrationality, unless she has a good excuse. Here, I assume these claims for the sake of argument, in order to lay out the structure of the theory I wish to propose. So we are assuming for the sake of argument that promises express resolutions, and that agents are generally obligated to act as they have resolved. This leads us to a natural picture of promissory obligation: Mental States First Theory of Promising (MSF): The obligation to keep promises is inherited or derived from the norms on resolutions, which are the mental states expressed by promises. MSF tells us that one who promises to Φ expresses that she resolves to Φ, and in virtue of expressing that resolution is subsequently obligated to Φ, all else being equal. Crucially, MSF does not purport to exhaust the reasons for which we might be obligated to keep promises. Any given promise might be morally obligatory to keep for multiple reasons: because the promisor will be harmed if it is not kept, or because breaking it would involve unfairly free-riding on a valuable social practice, or because it would create a lack of trust between promisor and promisee, or because breaking it would mean failing to respect a deliberate transfer of authority, etc. But as we saw from the Problem of Marginal Cases, such sources of obligation are not present in every case. MSF proposes a minimal source of promissory obligation that is present in all cases, stemming from the resolution expressed by the promise. In the majority of cases, this will coexist with other, more important sources of obligation. Because every promise expresses a resolution—including promises involving a lack of promisee expectation, weak or non-existent conventions of promising and societal pressures, deceased promisees, bare wrongings, the lack of an intention to ! 68! obligate yourself, etc.—MSF has broader applicability than the theories discussed in Chapter 2, and can easily accommodate cases that these other theories cannot. This is because MSF derives promissory obligation from a core component of the speech act of promising itself, rather than from the ways in which a promise interacts with the world (e.g., by creating expectations in the promisee, or being part of a sanctioning social practice.) It is the nature of the speech act of promising itself that grounds promissory obligation, independently of the content of or external circumstances involving any particular promise. It follows that content-based and external factors like whether the promisor intends to obligate herself, whether the promisee’s expectations are raised, whether a social convention is in place, and whether anyone will find out about the breach will not affect whether the promise yields an obligation At first glance, it might seem obvious why promising to Φ obligates you to do so when your promise is sincere. For we are positing that sincerely promising involves actually forming a conditional resolution. And we are assuming that all resolutions, including promissory ones, are subject to the norm that one who resolves to Φ must do so, all else being equal. It follows that one who sincerely promises to Φ must likewise do so, all else being equal, in virtue of having formed a resolution. In other words, I am obligated to keep my sincere promise to you because I have resolved to do so, and I am in general obligated to act on my resolutions. On closer inspection, though, this explanation is incomplete. 86 Resolving to Φ indeed obligates you to Φ. But this obligation seems to be a personal commitment, rather than a promissory one—you’re obligated to Φ not because you have promised !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 86 Thanks to Gary Watson for helping me to fully appreciate this point. ! 69! someone else that you will do so, but because Φing is part of your plan. When you sincerely promise, you are indeed obligated on the basis of the personal commitment that results from having formed a resolution. But you are also morally obligated to act on the basis of having made a promise that publicly expresses such a commitment to the promisee. If I sincerely promise the department chair that I will take minutes at the meeting and I fail to do so, I have both let myself down (for I didn’t carry out my own plans) and let her down (for I broke my promise.) How does MSF explain the latter kind of obligation—what I will call distinctively promissory obligation, or the moral obligation stemming not from possessing the mental state itself, but from expressing the mental state to the promisee? The answer to this question is important, for it will provide an explanation of distinctly promissory obligation for insincere promises, as well. We need such an explanation because those who promise insincerely, and merely purport to have resolved to Φ without actually having so resolved, are still subject to a promissory obligation. The fact that your promise was insincere does not get you off the moral hook for acting as you have promised to; this would fly in the face of common judgments about promise-keeping, and might perversely incentivize people to promise insincerely. Rather, insincere promisors are required to behave as if they had promised sincerely—that is, to keep their promises under all and only those conditions in which they would be required to do so had they actually formed the conditional resolutions they purported to have. How can MSF explain this? Why should a promisor be obligated to act in accordance with the norms on a resolution that she has expressed but has not actually formed? ! 70! I tackle these questions in the rest of this chapter. I will argue that promises are part of a broad pattern of speech acts, all of which inherit norms from the mental states they express for both sincere and insincere utterances. The existence of this pattern defuses the worry about accommodating insincere promises. I take a “partners in innocence” approach to establish that the ability to account for both insincere and sincere utterances is a general feature of the way in which the norms on publicly expressed speech acts can be derived from the norms on their mental states, rather than a uniquely puzzling feature of promises. Moreover, the truth of the inheritance claim explains both how sincere promises generate distinctively promissory obligations (in addition to generating rational obligations stemming from resolutions), and how insincere promises generate distinctively promissory resolutions in the same sort of way (in spite of no resolutions being present.) Finally, the pattern suggests that we can explain the norms on a wide variety of speech acts in terms of their underlying mental states, which suggests a broader Mental States First theory of speech acts in general. 4.! The Speech Act/Mental State Insincerity Pattern The fact that the norms governing the mental state expressed by a speech act apply even when the act is uttered insincerely and the mental state is not present is not unique to promises. Rather, this is true of a broad range of speech acts of different kinds, which form a striking pattern. This pattern makes insincere promises look significantly less problematic than they might first appear. For if the pattern holds, insincere promises can be subsumed under a much wider phenomenon, and we should not be surprised that MSF applies to insincere cases, as well. ! 71! We can formulate a general version of the pattern I wish to establish as follows: Speech Act/Mental State Pattern (SAMS): When a literal speech act S expresses a mental state M, and M is (1) subject to some publicly relevant aptness constraint X or (2) committed to some publicly relevant norm X, then S is also subject/committed to X, even when S is uttered insincerely and M is not present. There are some components of SAMS that must be explained before we are able to properly analyze instances of the pattern. I will address these in the next two sections, and then work through the way in which SAMS applies to a variety of speech acts. 4.1. The literal constraint on SAMS SAMS applies only to literal utterances. I take literal utterances to be those that are subject to a presumption of communicative sincerity, meaning that an utterance of S is literal if we presume that someone who utters S should be in state M. There is a general presumption of communicative sincerity for promises; all else being equal, we expect agents to be sincere in their promise-making. 87 There will be a similar general presumption of communicative sincerity for all of the speech acts in the SAMS pattern. Speech acts that do not have this presumption of communicative sincerity—such as irony, sarcasm, mere conversational pleasantries, and recitations of lines in a play— will not fall under the pattern. For example, saying, “How are you?” often expresses that you take an interest in how your interlocutor is doing. Ordinarily, the expression of such an interest commits you to listening attentively while your interlocutor explains how she is doing. However, we have a custom of routinely saying, “How are you?” as a mere conversational pleasantry, which does not conventionally signal any genuine !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 87 All else is not equal if promising sincerely is problematic for independent reasons, such as violating a more important conflicting moral obligation. For example, if a murderer shows up at your door seeking the life of your neighbor, and the only way you can get him to leave is by promising to send the neighbor his way tomorrow, you are morally required to promise insincerely. Generally, though, you ought to be sincere in your promise-making. ! 72! interest in the welfare of the person you are greeting. To count as sincere in uttering such a conversational pleasantry, one need not be genuinely concerned with how the interlocutor is doing. Such pleasantries are therefore not subject to the SAMS pattern. Similarly, we do not suppose that someone who has terrible service at a restaurant and sarcastically utters, “I’ll be back again soon” is required to intend to return to the restaurant, as she might be if she sincerely claimed that she would soon return. 4.2. The public relevancy constraint on SAMS The public relevancy constraint is necessary to ensure that SAMS does not overgeneralize. To illustrate how this works, I will first say a few words about the general motivation behind the SAMS, as this will help us restrict the norms to which it applies in a non-arbitrary way. In a nutshell, SAMS entails that you must act as if you really are in state M when you express to others that you are in state M. The basic insight behind this is the following. People are not mind-readers, and we cannot expect them to be able to ascertain when our utterances are sincere and when they are not. It follows that we must be able to take what people say at face value if we are to have productive interactions with them, at least in typical circumstances. 88 Doing so enables us to function well in dialogue with them—to respond appropriately to their utterances in our words and actions, to predict what they might do and say next, to alter our behavior in light of theirs, etc. If we weren’t entitled to presume that people really were in the states they purport to be in, we wouldn’t be able to interact with them very effectively. For example, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 88 Circumstances are not typical when you have some reason not to take your interlocutor’s utterance at face value—for instance, if you believe your interlocutor to be a habitual liar, or to be manipulating you. ! 73! assume that saying “thank you” expresses that you are grateful (and see section 5.4.1 below for further discussion of this claim.) If I could not presume at least for the sake of conversation that someone who thanks me really is grateful, I wouldn’t know whether it would be appropriate to respond to her utterance by saying, “You’re welcome” or “You ungrateful so-and-so.” Our dialogue would be at a standstill until I figured out some way of ascertaining my interlocutor’s true inner mental state, which might be difficult or even impossible. To facilitate our interaction, then, I am entitled to proceed, or “play along,” as if she really is grateful to me. Moreover, it seems that my interlocutor ought to play along, as well—that she should speak and behave as if she really is grateful, which involves behavior like restricting her expressions of thanks only to situations in which gratitude is appropriate. Failing to do this will constrain my ability to take her utterance at face value. Similarly, if I promise to Φ, it is generally reasonable for my promisee to presume that I am committed to Φing, regardless of whether I really have so resolved. 89 She cannot know my inner mental state, and so it would be unreasonable to expect anything else of her. Accordingly, the promisee is entitled to interact with as me if I have in fact resolved to Φ—which, depending on the circumstances, might involve believing that I plan to Φ, structuring her future plans and behavior around my Φing, or simply responding appropriately to in the moment to my commitment to Φing. And it seems that as promisor I am committed to playing along, as well—that I should behave as if I really have resolved, and thereby enable my interlocutor to take my promissory !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 89 There are circumstances in which such presumptions are not reasonable, all-things- considered—for example, if you know someone to be lying. See below for further discussion of this point. ! 74! utterance at face value. The aim of this overall chapter is establishing that the SAMS pattern holds, and that MSF is therefore likely to be true; I am not able to offer a fully fleshed out account of why the pattern holds. But the preceding remarks give us a glimpse: it seems to have something to do with the fact that it would be unreasonable to hold sincere and insincere utterances to different standards. For conversational interlocutors cannot be expected to know when speakers are insincere, and we want to be able to engage in productive dialogue with others without subjecting conversational partners to the unreasonable burden of making such a determination. With this understanding of the purpose of SAMS in place, we are now in a position to see what the public relevancy constraint is and why it is necessary. When a norm directly concerns another person in some way—either because it involves a speech act that is directly addressed to someone in the context of a conversation (as with thanking), or because it concerns what you are going to do or say in the future in a manner that affects others in some way (as with promising)—then the norm counts as publicly relevant. Only publicly relevant norms are the subject to SAMS, because only publicly relevant norms concern the sort of interpersonal interaction that SAMS is meant to facilitate. This might seem at first glance to be similar to the epistemic expectationalist view of promising, or the idea that the normative ground of the obligation to keep promises is the assurance or reliance that the promise induces in the promisee. But that is not the case, for public relevancy can come apart from whether an interlocutor is relying on or assured by a particular utterance. Recall the kinds of cases that are counterexamples to ! 75! epistemic expectationalist views. These are cases like that of the Profligate Pal, the well- meaning but perennially unreliable friend who asks to borrow money and promises that he will pay you back this time. The Pal’s promise in no way assures you; you don’t expect him to repay, or rely on him to do so. Scanlon’s view entails that no distinct promissory obligation is generated in this case, because there is no reliance. But the Profligate Pal’s promise is still publicly relevant in the sense described above, and is therefore subject to SAMS. For in accepting the Pal’s promise, you are entitled to take his utterance at face value. You don’t in fact believe that he will, but his promise is the sort of thing that could license such a belief. Moreover, you can presume that he is sincere at least in intending to pay you back, even if misguided in his own ability to carry out this intention. And this presumption will affect how you respond to him—for example, you might appreciate his good intentions, which you would not be able to do if you weren’t able to take his expression of resolution at face value. This example shows us that whether a norm is publicly relevant in general—that is, whether it is the kind of thing that involves interaction of a sort that is hindered if we are not able to presume that our interlocutors really are in the states they purport to be in—is a separate question from whether a particular interlocutor is assuming that we are sincere, and is being assured by or relying on this. The latter is relevant for the epistemic expectationalist view, while only the former must be in place for SAMS to apply. The best way to illustrate what it is for a norm to be publicly relevant is to look at one that is not. For example, assume for the sake of argument that assertion expresses belief, and that believing that p commits you to believing the obvious and relevant consequences of p. You assert p, and you believe that p entails q. Does SAMSI predict that you are committed to believing q on the basis of this assertion, as well—even if you ! 76! are insincere, and don’t actually believe p? 90 For example, suppose you are in a public library, and you insincerely assert that the building is about to collapse, without really believing this. If the building is going to collapse, then you ought to exit it as quickly as possible. Are you required to believe that you should exit the building, on the basis of your insincere assertion that it is about to collapse? It is clear that you are not. This is because norms that govern only your private beliefs are not publicly relevant in the requisite way. It doesn’t matter to your interlocutor whether you believe p, because they will never know this, and will never have a need to adjust their behavior in any way in light of it. We can contrast this example with a norm on assertion that is publicly relevant. Assume for the sake of argument that believing that p and that p obviously entails q commits you to assenting to q when you are asked whether q. Suppose you insincerely assert that the building is about to collapse, and that the building’s being about to collapse entails that you should run away. Someone in the library who overhears your assertion asks you whether he should run away. Are you committed to saying yes? The best thing to do in such a scenario would be to retract your assertion—to say that you were only kidding, and that the building is not in fact about to collapse. But suppose you do not retract your assertion. Are you committed on the basis of having publicly claimed that the building is going to collapse to publicly claiming that you should run away? It seems that you are. It would be inappropriate to respond to the inquiry by saying, “Well, as I already said, the building is definitely going to collapse, and you should always run away from collapsing buildings. But you definitely should !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 90 Thanks to Abelard Podgorski for raising an example like this in discussion, and pushing me about whether my original formulation of the SAMS pattern (which lacked a public relevancy constraint) could accommodate it. ! 77! not run away.” And this is so even if you do not in fact believe that the building is about to collapse. This shows us that this norm on assertion is subject to SAMS— because unlike the private beliefs that follow from one’s assertion, the further assertions that follow from one’s assertion are publicly relevant in the requisite way. (I say more about whether and when norms on assertion fall under SAMS in Section 7.1.1. below.) In the rest of this section, I will first discuss how the SAMS pattern holds for speech acts subject to aptness constraints, and then how it holds for speech acts subject to norms of further commitment. My ultimate aim is to show that SAMS holds not only for promises, but also for speech acts of a wide variety of types. I will offer proposals about what mental state is expressed by each speech act, and about what norms those mental states and speech acts are subject to. I take these proposals to be at least prima facie plausible, if not intuitively compelling. But the SAMS pattern will not necessarily be undermined if I am wrong about a few of the details. For the pattern will be in trouble only if the correct account of these speech acts turns out to be incompatible with it. The burden is on the critic to argue both that the intuitively compelling proposals I offer are misguided, and that the correct accounts violate SAMS. 5.! Examples of the SAMS Pattern The SAMS pattern holds for speech acts that are subject to norms of two different kinds, which I will address in turn: aptness constraints, and norms of further commitment. 5.1.! The SAMS pattern for speech acts with aptness constraints Aptness constraints concern the appropriateness of forming a certain mental state or performing a certain speech act. I mean to capture an intuitive sense of aptness, according to which X is apt if it conforms to some X-relative standard, and inapt if it ! 78! deviates from that standard. This is a broad notion, which can apply to different sorts of actions or utterances across a variety of normative domains. For example, jokes are meant to be funny—funniness is part of the kind of thing that a joke is—and they fall short if they are not. An unfunny joke is inapt in the sense I have in mind here. The fact that X is apt with regard to one standard does not mean that it will be apt with regard to another; even a funny joke can fail in other ways (e.g., because it is sexist or racist.) In general, the aptness of a speech act can come apart from whether the agent ought to utter it, all-things-considered; it might be apt in one sense to tell a funny joke about death at a funeral, but not something you ought to do overall. 91 The details of how we conceptualize aptness constraints are not important, so long as there is some coherent and consistent way of capturing the intuitive concept. I am assuming that the aptness norms I will discuss in this section are publicly relevant. 92 When someone performs a speech act that is subject to an aptness constraint, we can reasonably presume that she really possesses the mental state expressed. And so we are entitled to expect that she will adhere to the conversational constraints that would be required were she really in the relevant state—which is to say, we are entitled !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 91 In a similar manner, Justin D’Arms and Daniel Jacobson (2000) distinguish between the correctness and propriety of emotional responses: correctness has to do with “whether the emotion fits its object,” while propriety has to do with other sorts of justification or endorsement, such as judging “that a feeling is expedient, that it is morally permissible, or that it is ‘what to feel’ all things considered.” When I speak of aptness, I am speaking of something like correctness in their sense. 92 It might be objected that aptness norms do not in fact meet the thin public relevancy constraint that I have outlined in this chapter. One might wonder in what sense thanking, for example, is truly publicly relevant. I don’t think this objection is on the right track, for the reasons outlined in Section 3 above when I first presented the constraint. But if need be, I could easily accommodate the objection by reformulating SAMS so that there is a public relevancy constraint only on norms of further commitment, and not on aptness norms. Aptness norms would then still be subject to the pattern—although the basic explanation of why would have to be different from the rough sketch I give about our needing to take our interlocutors at face value to facilitate productive interactions. ! 79! to expect that the utterance will be in line with the aptness norms on the relevant mental state. We shall see how this works in the rest of this section. 5.1.1.! Promising In addition to being subject to a norm of further commitment—that you act as you have promised—resolutions are subject to an aptness norm. Resolving to Φ is inapt if you do not believe that it is possible for you to Φ. In general, it is misguided to intend to Φ if you do not believe that Φing is possible, and the same constraint applies to resolutions, which involve forming intentions. 93 If you are in the hospital with two broken legs, it makes no sense to resolve to go for a jog tomorrow morning, as it is impossible for you to do so. Promises are subject to the same aptness norm; promising to Φ is incorrect or misguided if you do not believe that it is possible for you to Φ. There is something inapt about promising your running partner that you will join her for a jog tomorrow morning when you are laid up in the hospital with two broken legs. This is so even if the promise is insincere, and you don’t actually resolve to jog when making the promise, and intend to leave your jogging partner in the lurch. You go !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 93 Harman (1976) and Velleman (1989) offer strong cognitivist accounts of intention, according to which the intention to Φ is identified with the belief that you will Φ. A strong cognitivist will agree that that intending to Φ while believing that Φing to be impossible is irrational, because this is equivalent to believing that you will Φ while believing that you won’t Φ. Weaker cognitivists who claim that intending to Φ entails the belief that you will Φ will agree; for example, R. J. Wallace claims that “the commitment to realize an end is constrained by one’s beliefs about the possibility of realizing the end” (2006, p. 104). Non-cognitivists like Bratman agree, as well. For if intentions are part of plans, it makes no sense to intend what you take to be impossible; to do so is to make plans that you believe cannot be achieved. See Davidson (1978) p. 100 – 101 for the stronger claim that one cannot intentionally do what one takes to be impossible. Here, I assume only the weaker claim that it is inapt (or perhaps irrational) to intend to do what you take to be impossible. See Ludwig (1992) for argument that it is possible to intend to do what you believe to be impossible; Ludwig offers a series of counterexamples to Davidson’s claim, all of which seem to me to be indeed possible, but inapt. ! 80! wrong in promising insincerely, and you go wrong again and in a different way in promising to do something that you don’t believe to be possible. Just as SAMSI predicts, the aptness norm on resolutions applies to promises as well, in both sincere and insincere cases. 5.1.2.! Complimenting In general, complimenting expresses that you like, admire, or appreciate the thing being complimented. The Merriam-Webster dictionary defines “compliment” as “an expression of esteem, respect, affection, or admiration.” 94 As Janet Holmes notes, A compliment is a speech act which explicitly or implicitly attributes credit to someone other than the speaker, usually the person addressed, for some ‘good’ (possession, characteristic, skill etc.) which is positively valued by the speaker and the hearer. 95 I take it that when one “positively values” something, one finds it good in some way, broadly construed. 96 It is apt to appreciate X only if X is in some way good, admirable, or praiseworthy, broadly construed, with what counts as good in a particular case depending on the kind of thing being assessed; brightly colored floral patterns are a good feature of sundresses, but a bad feature of military uniforms. For example, suppose Alice is !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 94 http://www.merriam-webster.com/dictionary/compliment. 95 1995, p. 117; cited in Jucker (2009). 96 One might worry it isn’t always true that compliments express admiration, for there are situations in which compliments seem to function more as conversational pleasantries than they do genuine expressions of admiration. For example, it may be expected that a guest at a dinner party will tell the host that the food is delicious. But even routine compliments like these express some sort of appreciation. For there is a definite sense in which a guest who says “What a delicious meal!” when she thought the food was disgusting is lying, and in which the guest who says “What a delicious meal!” when she thought the food was fantastic is not. Such a lie might be forgivable, or even required as a matter of politeness. But this does not entail that conversationally expected compliments do not express appreciation. (Thanks to Andrei Marmor for pressing me to address examples like this.) ! 81! wearing a well-tailored and elegant suit, which has an accidental and ugly coffee stain on it. It would be misguided for her friend Bobby to ignore the beautiful suit and admire the coffee stain instead; good construction and pleasant design are proper objects of appreciation in garments, while inadvertent coffee stains are not. Likewise, a compliment is apt only if the thing complimented is in some way good, admirable, or praiseworthy, broadly construed. It would be misguided for Bobby say “Wow, what a fabulous coffee stain!” And this compliment would be inapt even if Bobby were insincere, and didn’t actually appreciate Alice’s coffee stain but was complimenting her in a misguided attempt to get on her good side. As SAMS predicts, the mental state and both sincere and insincere utterances of the speech act are all subject to the same aptness constraint. 5.1.3.! Criticizing The speech act of criticizing is very like that of complimenting, except it focuses on negative aspects rather than positive. (Even “constructive” criticism points out some faulty feature or area in need of improvement, even if it does so in a helpful and positive way.) Criticism expresses a dislike or disapproval of the thing you are criticizing; “criticize” is defined as “to express disapproval of (someone or something).” 97 Disapproval of something is apt only if that thing is bad or objectionable in some way. As with appreciating, what counts as bad or objectionable depends on the kind of thing being disapproved of, and the standards that determine good or bad instances of that thing. It is apt to disapprove of rambling personal reflections in an !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 97 http://www.merriam-webster.com/dictionary/criticize ! 82! academic essay (as these make for a bad essay), and inapt to disapprove of a concise and debatable thesis statement (as this is essential to a good essay.) Criticism of something is likewise apt only if that thing is bad or objectionable in some way. Professor Claire can properly criticize meandering autobiography in her student Dora’s essay on utilitarianism. But it would be inapt for her to criticize something like the essay’s clear and disputable thesis statement. And this criticism would remain inapt even if it were insincere—if Claire didn’t really disapprove of the thesis, but thought Dora was coddled and lazy, and wanted to criticize arbitrary aspects of the paper to get her attention. Again, the mental state and both sincere and insincere instances of the speech act are all subject to the same aptness constraint. 5.1.4.! Thanking It is commonplace among philosophers and others that the speech act of thanking expresses that you are grateful to X for Φing. For example, Searle states that that thanking “counts as an expression of gratitude.” 98 Coleen Macnamara notes that “if I do you a good turn, you may respond with gratitude and express that gratitude by saying ‘Thank you.’” 99 Gratitude presupposes that X is responsible for Φing, and is subject to the aptness condition that Φing benefits X in some way, broadly construed. 100 As the starting point for a comprehensive psychological treatment of gratitude, Michael !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 98 1969, p. 65 99 2013, p. 894 100 I do not wish to rule out appropriate gratitude towards someone who sincerely attempted to benefit you but failed. You may be properly grateful to a co-worker for attempting to get you an interview at a new company, even if the interview could not be scheduled; in such a case, the co-worker’s effort and good intentions can be understood as a sort of benefit to you. Thanks to Steve Finlay for pointing out the possibility of gratitude for an attempted benefit. ! 83! McCullough et. al. assume that “gratitude operates typically when people are the recipients of prosocial behavior.” 101 For example, it is apt for Evan to be grateful to his boss Frank for writing him a positive letter of recommendation. It would be inapt for Evan to be grateful to Frank for writing a letter filled with damaging and false rumors (say, a letter claiming that Frank is always late and has terrible handwriting.) Thanks are subject to the same aptness conditions as gratitude, whether they are sincere or insincere. 102 As Macnamara claims, an expression of thanks is “unwarranted if directed at targets . . . who did not do a kindness.” 103 Unless you are speaking ironically or sarcastically, it doesn’t make sense to thank those whom you take to have harmed you. Thanking Frank for spreading rumors is inapt, and would remain so even if Evan were insincere in his thanks, and was thanking Frank not because he was grateful but because he was afraid of looking bad in front of his coworkers if he didn’t say thanks Once more, the mental state and both sincere and insincere instances of the speech act are all subject to the same aptness constraint. 5.1.5.! Congratulating Congratulating A for achieving X expresses that you are happy (or proud, or pleased, etc.) that A achieved X. “Congratulate” is defined as “to express to (someone) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 101 2001, p. 252 102 There are cases in which uttering “thank you” seems to function more as a conversational pleasantry than as a sincere expression of gratitude. For example, saying “thank you” when the cashier hands you your change is something you do simply to be polite, and seems apt even if you are not in any way grateful to the cashier for fulfilling the requirements of her job. But this does not entail that the aptness condition on gratitude misguided. Rather, I assume that such uses of “thanks” are merely formulaic or ritualistic utterances that do not in fact express gratitude, and therefore are not subject to the aptness norms on gratitude. They are like “how are you?” when uttered as a mere conversational pleasantry. 103 ibid, p. 897 ! 84! admiration for his or her success or good fortune,” 104 and Searle notes that to congratulate is to express “pleasure (at H’s good fortune.)” 105 Taking pleasure in another’s achievement presupposes that the person has achieved X, and is apt only if X is something that is in some way positive or worth celebrating. 106 Perhaps this is why schadenfreude—the feeling of happiness or pleasure that comes from observing another’s pain or suffering—seems misguided to many people: it is inapt to be happy that something bad has happened to someone. Likewise, congratulating A for X, whether sincerely or insincerely, is apt only if X is something that is taken by the congratulator to be good or worth celebrating. It is inapt for Gina to congratulate Herbert on bombing an interview and failing to get a job, because this is bad for Herbert. These congratulations remain inapt if Gina is insincere, and is congratulating Herbert only because she wants to get a rise out of him. As before, the mental state and both sincere and insincere instances of the speech act are all subject to the same aptness constraint. 5.2.! The SAMS pattern for speech acts with further commitment norms Norms of further commitment are like the norms on promises and resolutions already discussed, according to which performing a speech act or forming a mental state (i.e., promising to Φ or resolving to Φ) normatively commits the agent to acting in a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 104 http://www.merriam-webster.com/thesaurus/congratulate 105 1969, p. 65 106 There are vicarious uses of “congratulations,” such as congratulating a sports fan on her favorite team winning a game. The fan had nothing to do with the win, so the presupposition of congratulations is not strictly met. But it can still be appropriate as a matter of convention, assuming our speech conventions are such that what you do when you congratulate her is acknowledge her team’s win, and perhaps express that you are happy on her behalf. This might be a merely formulaic or rituallistic use of “congratulations” that need not be subject to the aptness constraints, like thanking the cashier for giving you the change that you are owed. ! 85! certain way (i.e., Φing.) In general, an agent is normatively committed to Φing in virtue of being in state X whenever there exists some standard which the agent ought to buy into—not necessarily objectively, but at least from within the context the practice she is participating in—that demands that she Φ if she is in X. 107 This needs to be distinguished from being committed to Φing in a volitional sense, or being such that you really do care about Φing. The person who promises to quit smoking is normatively committed to doing so in virtue of having made a promise, even if she is not volitionally so committed because she doesn’t actually care about quitting smoking. 108 The SAMS pattern applies to normative commitments, rather than volitional commitment. Like the notion of aptness, the idea of further normative commitment is broad and flexible, and applies across a variety of normative domains. For example, baptizing your child in a Catholic church commits you to raising her Catholic, since this is what !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 107 We can understand norms of further commitment as having either wide or narrow logical scope. On a wide-scope understanding, the norm demands that the agent be such that she either lacks the state that incurs the commitment or carries out the commitment. On a narrow- scope understanding, the norm demands that if the agent has state that incurs the commitment, then she carry out the commitment. Although debates about the logical scope of conditional norms are independently interesting, it does not matter how we construe commitment for my present purposes. My claim that the norms of further commitment on the speech act and on the mental state it expresses are the same requires only that the scope of the commitment be the same for the mental state and for the speech act. (The wide-scope/narrow-scope debate was first formulated about conditional oughts; see Broome (1999) for a defense of a wide-scope view, and Kolodny (2005) for a defense of a narrow-scope view. For more on the scope of rational requirements, see Broome (2007), Kolodny (2007), Brunero (2010), and Lord (2011)). 108 What exactly normative commitments are, and whether they are best be construed as a phenomenon distinct from requirements, oughts, or reasons, is a controversial issue; for a defense of commitments as a distinct phenomenon, including a clear distinction between normative and volitional commitments, see Shpall (2013 and forthcoming); see also Liberman and Schroeder (forthcoming). For present purposes, all that matters is that we have an intuitive notion of normative commitment—a sense in which being in state X commits you to being in state Y, in virtue of some relevant standard. So long as this sense of commitment is the same for the commitment incurred by mental states and the commitment incurred by the speech acts expressing those states, the details of the view should not matter. ! 86! the standards of Catholicism require; playing a game of freeze tag commits you to freezing when you are tagged, since this is one of the rules of the game. The domain in question determines what the grounds of these commitments are: the further commitment norms governing Catholic baptism are religious norms, while those governing freeze tag are the invented rules of a game. As with aptness, claims about what an agent is normatively committed to doing can come apart from claims about what she ought to do, all-things-considered; you can be committed by the rules of the game to freezing when tagged even though all-things- considered you ought to run away from an oncoming truck. Similarly, you can be committed by religious norms to raising your child Catholic even if this is not what you have most reason to do overall. And what an agent is normatively committed to doing can also come apart from what she is volitionally committed to doing. A religiously apathetic person who baptizes her child in a Catholic church in order to mollify her devoutly religious family is normatively committed in some sense to raising the child Catholic, in accordance with the religious practice she has participated in. But she likely does not have a very strong volitional commitment to doing so. It is easier to see why norms of further commitment are subject to SAMS than it is to see why aptness norms fall under the pattern. For further commitment norms are publicly relevant in an obvious way, in that they all in some fashion concern with the speaker is going to do or say next (or in some cases—like the Profligate Pal—what the speaker is merely planning or committed to doing or saying next.) And the future actions and plans of our interlocutors might be relevant to us in many ways. ! 87! 5.2.1.! Forgiving Forgiveness expresses repudiation or forswearing of blame. 109 This is a common view of forgiveness; there is consensus in the psychological literature that a primary process of forgiving is “letting go of one's right to resentment and negative judgment.” 110 Repudiating blame of A for doing X normatively commits one to not only ceasing blaming A for X in the moment, but also to refraining from actively expressing blame towards A for X again in the future. Blame can be expressed in a variety of ways, most of which negatively affect the relationship between the blamer and the offender. One might express blame by reprimanding or chastising the offender, by holding the offender accountable (for example, by punishing her or demanding restitution), or by obviously dwelling on the failure (say, by bringing it up in future arguments with the offender, and interpreting the offender’s future actions in light of her failure), among other things. The expression of blame will take different forms depending on the circumstances and the relationship between the offended party and the offender. It is important to distinguish the deliberate repudiation of blame of the sort associated with sincere forgiveness from a mere cessation of the emotional state or attitude affiliated with blaming. People often feel angry or resentful when someone !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 109 The moral practice of interpersonal blame should not be conflated with the societal practice of legal accountability. A wrongdoer’s legal accountability can come apart from blame in this moral and interpersonal sense; a crime victim’s personally ceasing to blame a criminal does not commit the legal system to refraining from punishing the criminal. There are also cases in which a criminal is legally blameless but properly held morally accountable; perhaps the criminal did wrong and should be morally blamed, but cannot be found legally guilty because of a technicality, or because the wrong is private and not appropriately criminalized. 110 Lawler-Rowe 2008, p. 51 ! 88! wrongs them, and they are entitled to such emotions. However, these emotions do not always track neatly with blame in the sense of actively continuing to hold someone accountable. Anger and resentment might disappear of their own accord, even for a serious wrong for which blame in this more deliberate sense is still appropriate; perhaps the offender ought to be held culpable, but the wronged party is a devotee of mindfulness meditation, who habitually lets go of all negative emotions. The mindfulness meditator has no bitterness towards the offender, but may could still express blame by coolly noting her disapproval or demanding restitution. Conversely, anger and resentment might linger on beyond the point at which it is no longer appropriate to blame in the deliberate sense; perhaps the wrong was slight, and the offender apologized, but the wronged party suffers from an anger management problem and is prone to frequent and disproportionate resentment. Continued anger and resentment is compatible with repudiating blame; the person with anger management issues might deliberately decide to forswear actively blaming the offender, thereby committing to refraining from the expression of blame, while being psychologically unable to give up the anger she feels. To repudiate blame, she need not actually feel that she has ceased resenting and thereby mended her relationship with the offender. Rather, she must at least behave as if she has done so. Consider the following example to illustrate the way in which the repudiation of blame is subject to further commitment norms. Suppose it slips Ivy’s mind that today is her wedding anniversary, and she fails to meet her husband Jack for a celebratory dinner they have planned. Jack knows that he is likely to hold the mistake against her, and he doesn’t want to damage their relationship by doing so. So he decides to ! 89! foreswear blaming Ivy for her oversight. This foreswearing commits him to refraining from expressing blame towards Ivy. Interpersonal forgiveness likewise (and perhaps more obviously) commits one to refraining from blaming. As David Owens notes, forgiveness makes blame for a wronging inappropriate, involving not merely “the cessation of blame or guilt but an activity that changes the normative situation.” 111 Owens claims that forgiveness is irrevocable, by which he means that someone who forgives another is committed not only to ceasing to blame in the present, but also to refraining from taking up blame again in the future. Suppose Jack forgives Ivy for missing their anniversary dinner, and then later that evening angrily proclaims, “I can’t believe you stood me up on our anniversary! How could you do such a terrible thing?” Ivy can rightly object to this outburst. For once Jack has forgiven her—in the sense of expressing forgiveness, rather than in the sense of personally ceasing to be angry—he is committed to refraining from actively and openly blaming her. This is so even if Jack is insincere, and forgives Ivy not because he has ceased blaming her but because he wants to avoid conflict. His interpersonal utterance of forgiveness commits him to refraining from expressing this blame or acting in ways that would be licensed only if he continued blaming her (e.g., repeatedly bringing up Ivy’s error in subsequent arguments.) As with aptness !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 111 2012, p. 52. Owens argues that forgiveness involves the exercise of a distinctive normative power, through which we by declaration “change what someone is obliged to do by intentionally communicating the intention of hereby so doing” (p. 4). I do not wish to commit myself to the claim that forgiveness is a distinctive normative power; all I claim is that one who forgives is normatively committed to the cessation of blame. ! 90! constraints, the further commitment norm on the mental state applies to both sincere and insincere instances of the speech act. 5.2.2.! Apologizing Apologizing for Φing expresses that you regret or are sorry that you Φed. 112 This is regret in the sense of feeling shame or remorse for what you have done, rather than disappointment in an outcome that does not reflect on you as an agent (e.g., regretting that you did not bring an umbrella after it unexpectedly starts raining) or regret in the merely causal sense (e.g., regretting that you were involved in an unavoidable car accident.) What agents commit themselves to when they sincerely regret their actions varies from one kind of case to another. For example, a cheating spouse who genuinely regrets his infidelity commits to not straying again, while a party guest who sincerely regrets spilling red wine on a new area rug might incur a commitment clean up the spill, or to pay to replace the rug if it cannot be cleaned. What these examples have in common is a form of taking responsibility for one’s action, broadly construed. In general, we can say that someone who regrets Φing commits herself to taking responsibility for Φing. Philosophical and popular consensus is that one who apologizes for Φing commits herself to taking responsibility in the same way. For example, Mihaela Mihai notes that !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 112 As with congratulations, “I’m sorry” can be used in a different (and perhaps metaphoric) sense to express sympathy for something that wasn’t your fault, as when one says, “I’m sorry for your loss” to someone who has lost a loved one. And like thanks, expressions of apology can be used in merely ritualistic ways as a matter of convention; “I’m sorry,” can mean “I didn’t do that intentionally” rather than “I regret having done that,” as one someone accidentally bumps you on a crowded train and says, “I’m sorry.” I assume that these sorts of apologies do not express regret, and are therefore not subject to the same further commitment norms as standard apologies. ! 91! while philosophical accounts of apology vary, “there is a growing consensus that an authentic apology implies: an acknowledgement that the incident in question did in fact occur and that it was inappropriate; a recognition of responsibility for the act; the expression of an attitude of regret and a feeling of remorse; and the declaration of an intention to refrain from similar acts in the future.” 113 And an internet search for “sincere apology” brings up dozens of web pages that share this consensus. 114 At a minimum, apologizing is assumed to involve admitting one’s wrongdoing, without trying to explain away the mistake or offer excuses; this admission in turn often commits the apologizer to making some kind of remedy or restitution. 115 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 113 2013 114 For example: A meaningful apology communicates an acceptance of responsibility for your actions [and] a statement of your willingness to take some action to remedy the situation— either by promising to not repeat your action, a promise to work toward not making the same mistake again, a statement as to how you are going to remedy the situation (go to therapy), or by making restitution for the damages you caused . . . (Engel) Taking responsibility means acknowledging mistakes you made that hurt the other person . . . If there’s anything you can do to amend the situation, do it. It’s important to know how to apologize with sincerity, and part of the sincerity of an apology is a willingness to put some action into it. If you broke something of someone’s, see if you can replace it. If you said something hurtful, say some nice things that can help to generate more positive feelings. If you broke trust, see what you can do to rebuild it. Whatever you can do to make things better, do it. (Scott) Of course, these self-help websites do not have something like communicative sincerity in mind when they talk of what is necessary for sincerely apologizing. Rather, their focus seems to be on sincerity in the sense of being earnest or well-intentioned. But being sincere in this broader sense requires being communicatively sincere, as well; you cannot be earnest in your expression of S unless you are actually in the mental state you purport to be in. An earnest apologizer must at least regret her action, even if she might also need to have certain good intentions. 115 We can imagine cases in which an apology does not seem to commit the agent to anything. For example, suppose A is deciding between two suitors: B, who is wealthy but whom she doesn’t love, and C, who is poor but whom she loves. A turns down suitor C in favor of suitor B. She later regrets this decision, and apologizes to C. It’s not clear whether this apology commits her to anything. Since she is already married to B, it would be inappropriate for her to make it up to C by giving C any sort of special treatment now. Nor will she have the ! 92! Suppose Karen catches her student Larry plagiarizing a paper. Larry apologizes to Karen. His apology commits him to taking responsibility for his error—to doing something admitting that he was wrong, accepting his F grade without complaint, and avoiding plagiarism again in the future. This is so even if Larry is insincere, and apologizes for plagiarizing not because he regrets doing so but because he thinks this will mitigate his punishment. Karen could rightly object if, after apologizing, Larry complained about his grade or kept insisting that he had done nothing wrong. As with forgiving, the further commitment norm on the mental state applies to both sincere and insincere instances of the speech act. 5.3. Upshots of the SAMS pattern I aim to have established that there is a broad range of cases that are subject to the SAMS pattern—cases in which the speech act and the mental state it expresses are subject to the same publicly relevant norms of aptness or further commitment, regardless of whether the speaker is sincere and is actually in the mental state that is expressed. The mere existence of this pattern makes plausible the claim that the expression of a resolution generates a distinctly promissory obligation, in both sincere and insincere cases. For while this might at first glance appear difficult to explain, we should expect it to be true as part of a general pattern covering a wide range of speech act types. In the next section, I draw on the SAMS pattern to defuse a different worry for MSF. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! opportunity to avoid making the same decision in like circumstances in the future, as A is unlikely to find herself in such circumstances again. However, this is not a problem for my account of the norms on apologies. For A’s unexpressed regret does not seem to commit her to anything in this case, either. SAMS requires only that whenever there is a further commitment norm on regretting X, there is one on apologizing for X; similarly, whenever there is no further commitment norm on regretting Y, there is none on apologizing for Y. ! 93! 6.! Can MSF Accommodate Moral Obligation? It is not immediately apparent how MSF can explain the distinctively moral nature of promissory obligation. As I will argue in Chapter 6, one who resolves to Φ is rationally obligated to Φ, lest she be self-undermining and inconsistent in her long-term planning. But promises seem to be paradigmatic moral obligations, and one who promises to Φ is morally obligated to do so. How can we explain the moral nature of promissory obligation if we cash it out in terms of a rational obligation to act on our resolutions? It is indeed important to explain why promises yield moral obligations. But this worry is not as pressing for MSF as it first seems. For in most cases, there are multiple sources of obligation to keep a promise—e.g., keeping a particular promise might be morally obligatory both because breach harms the promisee and because you have publicly expressed the resolution to do so. These independent sources of obligation are grounded in obviously moral concerns like harm avoidance, fairness, trustworthiness, etc. The burden for MSF is explaining why promises yield moral obligations even in non-paradigmatic cases, in which none of these other, more typical sources of moral obligation are present. The worry about whether MSF can explain moral obligation in marginal cases is mitigated when we see that the moral force of obligations arising from such cases is significantly weaker than that arising from paradigmatic cases. For example, suppose Mary and Noah are airplane seatmates who are stranded on the tarmac for several hours because of a flight delay. Noah asks Mary to promise him that when she gets off the flight, she will write a 1-star Yelp review of the airline. Mary, who is as annoyed ! 94! with the airline as Noah is, makes the promise. Suppose that there are no independent sources of moral obligation to keep this promise—Noah will never be able find out whether Mary posted the review and so cannot be disappointed, no one else knows about Mary’s promise, and posting the review will not affect the airline’s overall rating or bottom line at all, as it will be lost in a sea of thousands of 5-star reviews from satisfied customers. Mary nevertheless seems to be pro tanto morally obligated to post the review, in virtue of having promised to do so. If she gets off the plane, has an opportunity to write a review, and fails to do so without a good excuse, she will not be acting entirely as she ought to, morally speaking. But neither will Mary be morally failing in a drastic way. For her obligation to keep the promise to her airplane seatmate does not seem to be an especially important obligation. It might easily be outweighed by other obligations, and Mary’s character will only be slightly impugned if she breaks her promise. Accordingly, MSF need not explain why promises yield particularly robust or important moral obligations. All it must do is explain why marginal promises yield moral obligations of a very minimal sort, like the obligation Mary has to write an airline review. The SAMS pattern gives us reason to be optimistic that there is such an explanation. For other speech acts in the pattern are also such that the norms on the mental state are in one normative domain, while the norms on the speech act are in another. For example, the aptness norms governing appreciation are at least sometimes aesthetic norms (e.g., Marc Chagall paintings are worthy of appreciation because they are aesthetically valuable; Precious Moments figurines are not because they are kitschy). 116 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 116 If the subject of appreciation is someone else’s behavior, the norms might be moral rather than aesthetic (e.g., morally appropriate to appreciate the firefighter’s brave rescue of the puppy ! 95! But the norms governing complimenting are also etiquette norms; you do not refrain from complimenting someone on her coffee stain simply because the stain lacks aesthetic value, but because doing so is rude. The same is presumably true of the norms on disapproval and criticizing. Similarly, the norms on regret are at least sometimes prudential norms governing how you should regulate your own behavior (e.g., not making the same mistake again in the future.) But the norms on apologizing seem to be at least interpersonal—if not explicitly moral—norms: it is at least rude and perhaps morally impermissible to apologize to someone and then behave as if you do not regret what you’ve done. Publicly relevant speech acts always involve another person. And so it makes sense that they are subject to social, interpersonal norms, even though the internal, intrapersonal mental states they express are not. Promises are likewise subject to social, interpersonal moral norms, even though the internal, intrapersonal mental states they express are not. Another pattern is beginning to emerge. Resolutions yield rational obligations. When resolutions are promissory and interpersonal, they yield moral obligations as well, because they essentially involve another person. Similarly, approval and disapproval are subject to aesthetic norms. When they are interpersonal, they are subject to etiquette norms, as well, because they involve another person. And regret is subject to a personal or prudential norm, while apologizing is also subject to an interpersonal or moral norm. Promises and resolutions are therefore not alone in being subject to the same norm across different normative domains. If we do not have a problem with cross-domain !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! from the burning building, morally inappropriate to appreciate the arsonist’s cowardly and callous starting of the fire.) ! 96! inheritance of norms in all of the SAMS cases, we should not have a problem with cross- domain inheritance of norms when it comes to promises. The SAMS pattern mitigates two potential objections to MSF, by helping us to see why it is unsurprising and unproblematic that MSF both extends to insincere cases and generates interpersonal obligations. The pattern itself is also worth further consideration in its own right. What, if anything, can SAMS tell us about the general relationship between speech acts and the mental states they express? 7.! The General Mental States First Theory 7.1. Inadequate responses to the SAMS pattern The SAMS pattern calls out for some sort of explanation. There are five possible ways to respond to it: (1)!Appeal to coincidence: deny that any explanation is needed, and chalk up the phenomena to mere coincidence. (2)!Rejecting the pattern: deny that there is in fact a pattern by rejecting the cases I have presented and/or offering cases that should fit into the pattern but do not. (3)!Common cause explanation: posit some independent third factor that causes both the mental state and the speech act to be subject to the same norms, for both sincere and insincere cases. (4)!External to internal directional explanation: explain the norms on mental states in terms of the norms on speech acts. (5)!Internal to external directional explanation: explain the norms on speech acts in terms of the norms on mental states. We do not need to take the first possibility seriously. While the similarities I have laid out in the previous section could all be due to coincidence, this is unlikely. Coincidence should be appealed to only as a last resort if no other explanation is forthcoming. ! 97! I will argue that the fifth option—the internal to external directional view—is the best response. The other views are not viable, for the pattern holds in spite of purported counterexamples having to do with assertion, the most plausible common cause view is insufficiently general, and the external to internal directional explanation is independently implausible. 7.1.1.! Rejecting the pattern: does SAMS apply to assertion? One response to the SAMS pattern is to deny that there is a pattern in the first place, which requires rejecting all (or at least most) of the cases I presented in the last section. I find the cases I have presented compelling, and suspect that it would not be easy for a critic to explain away all or most of them. An alternative way to put pressure on the robustness of the SAMS pattern is to claim that there are speech acts that we should expect SAMS to apply to, but to which SAMS does not apply; this would give us reason to doubt the generality and predictive power of the pattern. For example, I have not yet analyzed whether assertion is subject to a SAMS pattern. Do whatever norms govern the mental state expressed by assertion govern both sincere and insincere assertions, as well? I have not addressed this question for the simple reason that the norms governing assertion are too controversial to be given a quick analysis of the sort I have offered for the other norms in the pattern. There is a great deal of disagreement about what mental state assertion expresses, and about what makes an assertion apt and whether there are norms of further commitment on assertions. 117 Adjudicating among !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 117 For a good survey of the diversity of views about the aptness constraints on assertion, see MacFarlane (2011). For example, some philosophers think an assertion that p is apt only if p is true; see Weiner (2005). Others think an assertion that p is apt only if it is reasonable for the ! 98! these views in order to determine whether there is a SAMS pattern governing assertion is too large a project to be adequately handled here. However, we can see how SAMS can plausibly apply to assertion by assessing an illustrative view as an exercise. Suppose for the sake of argument that a belief that p is apt only if the agent knows that p. Someone who takes knowledge to be an aptness condition on assertion—such as Timothy Williamson—thinks that an assertion that p is apt only if the asserter knows that p, and that this aptness condition applies to both sincere and insincere assertions. If we make the natural and plausible assumption that assertion expresses belief, we will have a SAMS pattern: the mental state (belief) and both sincere and insincere instances of the speech act that expresses it (assertion) will all be subject to the same aptness constraint (knowledge.) 118 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! speaker to believe that p; see Lackey (2007) and Douven (2006). Another influential view is that an assertion that p is apt only if the speaker knows that p; see Williamson (2000), and also DeRose (2002) and Hawthorne (2004). It is not implausible that any one of these aptness constraints on assertion applies to belief, as well (i.e., that a belief is apt only if it is true, or reasonable, or you know it)—in which case, all of these views fall under the SAMS pattern, so long as assertion expresses belief (or a state that includes belief as a component.) Some philosophers argue that further commitment norms are essential to a proper understanding of assertion, claiming that someone who asserts that p takes on a special commitment to the truth of p, which can be cashed out as a commitment to justify or defend p to one’s interlocutors. See, among others, Peirce (1934), Searle (1969), Brandom (1983), Wright (1992), Watson (2004), and MacFarlane (2005). Beliefs seem to incur a similar sort of commitment: if you believe that p and are questioned about whether p, you seem committed to defending or justifying p, even if you haven’t asserted p. If this is so, then this is another instance of the SAMS pattern. 118 This is compatible with the claim that there are further, independent aptness conditions on assertion that have to do with conversational dynamics rather than with the norms on belief; e.g., an apt assertion must be germane to the conversation in question. Moreover, in offering this example, I do not mean to assume that anyone who accepts that belief and assertion are both subject to the same aptness constraint must also accept the claim that assertion expresses belief. For example, Williamson does not accept any claim about what assertion expresses, but rather thinks that assertion is constituted directly by the fact that it is correct if and only if you know it. All I mean to establish with this example is that there is a coherent and plausible account of assertion that allows us to generate a SAMS pattern for aptness norms. ! 99! Can SAMS accommodate further commitment norms on assertion, as well? Assume for the sake of argument that assertion expresses belief. 119 Are there norms of further commitment on belief? If so, are these inherited by assertions? Recall from Section 4.2 above that non-publicly relevant norms on belief—such as the requirement to believe whatever is obviously and relevantly entailed by your beliefs—are not subject to SAMS, and will not be inherited by assertions. We also saw in Section 4.2 that belief and assertion both seem to be subject to a norm about publicly assenting to the obvious and relevant consequences of your beliefs; one’s unretracted assertion that the building is about to collapse commits one to assenting to the claim that everyone ought to run away, even if the assertion is insincere. (Keep in mind that this is merely a pro tanto commitment rather than an all-things-considered obligation; it would be better to retract an insincere assertion than assent to the consequences of it.) One might worry that there is another norm of further commitment on belief that is not plausibly shared by assertion. 120 It might seem plausible that believing that p commits you to acting as if p were probable. But it might not seem that asserting that p commits you to acting as if p were probable when your assertion is insincere. For example, suppose I insincerely assert that the library is made of graham crackers. Am I required to behave as if it is probable that the library is made of graham crackers? It’s not immediately obvious what such behavior would consist in, but we can suppose it would involve something like treading carefully on the floors to avoid crushing the crackers, or attempting to break a piece off and take a bite. And it seems clear that my insincere assertion does not commit me to behaving in such ways. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 119 The same considerations will hold if assertion expresses a state that has belief as a component, such as justified belief or knowledge (at least, on reductive accounts of knowledge.) 120 Thanks to Jake Ross for pressing me on this worry and offering the graham cracker example. ! 100! However, this example is potentially misleading, and our intuitions about it are therefore not a reliable guide to whether an insincere assertion that p commits you to behaving as if p is probable. For it seems likely that the relevant norm on belief does not apply in this particular case. This is because the belief that library is made of graham crackers is unreasonable. And you do not seem to be obligated to act as if your unreasonable beliefs are plausible. Unreasonably believing that the library is made of graham crackers need not commit you to trying to bite the floor or spread peanut butter on the walls. In general, allowing unreasonable beliefs to commit you to what are presumably unreasonable further actions would be allowing one mistake to lead to another, which would be problematic. If this is so, then the graham cracker example poses no problem for SAMS. For if there is no norm of further commitment on the unreasonable belief that the library is made of crackers, we should not expect there to be a norm on the assertion, either. Our intuitions are likely to differ if we choose an example involving a reasonable belief. For example, suppose that a major earthquake strikes near the library. The building shakes, plaster rains down from the ceiling, and the roof starts to buckle in. It would be reasonable to believe that the building is going to collapse. But suppose that you are overly optimistic about the structural integrity of the library, and you do not form this belief. You then insincerely assert, “The building is about to collapse!” Are you committed to behaving as if it is probable that the building will collapse—say, by running outside and taking cover? Running outside is likely to be what you objectively ought to do in this scenario, all-things-considered. But it also seems plausible that your unretracted assertion normatively commits you in a more personal way—that there would be something problematically inconsistent about someone who kept staunchly ! 101! insisting that the building was going to collapse, but who didn’t behave as if this were the case. Someone could rightly approach you and criticize you for not running away. This is distinct from any criticism you are subject to for making an insincere assertion in the first place; a fellow optimist about the building’s integrity who insincerely asserted that the roof was going to collapse would not be criticizable in exactly the same way you are. You both are insincere, but at least she is also consistent. I realize that some readers may not share the intuition that we are committed to behaving as if our reasonable but insincere assertions are probable, in which case I will offer two debunking explanations of their contrary intuitions. First, the commitments incurred by assertions are always escapable in a way that the commitments incurred by other norms in the SAMS pattern are not. One cannot unilaterally revoke an act of apology or forgiveness or release oneself from a promise. Once you have apologized, forgiven, or promised—even if insincerely—there’s no going back, and you’re stuck with the commitment. But one can always retract an assertion, which entails that insincerely asserting p does not irrevocably stick you with the commitment to assenting to the consequences of p or behaving as if p is true. If you don’t retract, you’ll be saddled with the commitment. But you can—indeed, you usually should—go back and retract your assertion. It should therefore be unsurprising if we have the intuition that we are genuinely committed to following through on the further commitment norms stemming from apology, forgiveness, and promising, but are not genuinely committed to following through on the further commitment norms stemming from assertion. Finally, in many contexts failing to act on the belief expressed by your assertion is functionally equivalent to conveying that you have retracted the assertion. If I say, “The building’s going down!” and I stay put, my interlocutors will know that ! 102! something is wrong: either I’m suicidal, I’m massively akratic, or I was lying. In most circumstances, the best explanation of my inaction will be that I was lying, in which case my interlocutors can presume that my assertion was insincere. Conveying that your assertion was insincere is akin to retracting it; saying “I was lying about p” is a means of retracting your assertion that p. And retracted assertions are no longer subject to further commitment norms. Perhaps there are cases in which a speaker who fails to act in accordance with an insincere assertion thereby retracts the assertion, thus making it the case that she is not in fact obligated to act on it after all. If this is so, then it is not even possible to fail to follow through on the further commitment norm on assertion in such cases, because failing to follow through makes the norm no longer apply. We should not expect the SAMS pattern to apply in such cases; perhaps the reader’s contrary intuitions are all cases like this. 7.1.2.! A common cause solution: the Same Functional Role View It is not immediately clear what sort of independent feature could cause both the speech act and the mental state it expresses to be subject to the same norms, and thereby undermine the directional claim that we can derive the norms on one from the norms on the other. The most likely common cause explanation I can think of—in fact, the only common cause explanation I can think of—is the following: Same Functional Role View (SFR): A speech act and the mental state it expresses are subject to the same norms because they play the same functional role—the former in the inner, personal realm, and the latter in the public, interpersonal realm. SFR states that aptness and further commitment norms result directly from the functional roles of the phenomena in question, such that nothing about the norms on the speech act depends on anything about the norms on the mental state or vice-versa. ! 103! For example, because the function of resolutions is committing yourself to future action, you must act unless you have a good excuse. The function of promises is the same, only interpersonal: you commit to another to act in a certain way in the future, and so you must act unless you have a really good excuse. Generalized, SFR posits that we derive the norms on the mental state by figuring out what the function of that state is, and what norms are necessary to attain that function. We then do the same for the speech act. On this approach, the question of what norms govern a particular state or act is entirely empirical: we ask what in fact best serves (or is necessary to attain) the function of the state or act in question, for the kind of creatures we are. If the mental state and the speech act serve similar functions, it seems likely that the speech act and the mental state will be governed by similar norms in order to attain that function. Because the functions of promising and resolving indeed appear to be similar, SFR is a potential explanation of why the norms on promising and resolving are the same. But SFR will not provide a good explanation of the SAMS pattern as a whole, because it is insufficiently general: even if it can explain promising, it cannot explain all of the speech acts in the pattern. And if SFR is wrong about one or two of the cases in the pattern, let alone wrong about the majority of them, it will not be as good an explanation as a directional view that can explain all of the cases. For even if SFR could explain some cases, we would still need explanations of the others. And whoever attempts to give such an explanation (or such explanations, if different kinds of cases need to be explained in different ways) will also be burdened with explaining why the SAMS pattern presents us with an apparently unified phenomenon that is subject to multiple kinds of explanation. ! 104! SFR fails if the function of the mental state is different from the function of the speech act. This is obviously true for forgiveness. Repudiating blame improves one’s emotional well-being; in general, “letting go of old grudges reduces levels of depression, anxiety and anger. People who forgive tend to have better relationships, feel happier and more optimistic, and overall enjoy better psychological well-being.” 121 Repudiating blame and resentment can also improve one’s physical health in a variety of ways, 122 and studies have shown that “reduction in negative affect” is a primary source of these good health effects. 123 It therefore seems likely that the purpose or functional aim of repudiating blame is improving one’s mental and physical health. The primary aim of public forgiveness, on the other hand, is not improving the forgiver’s well-being but repairing the relationship between the forgiver and the offender. Telling someone that you forgive her can be an important step in fixing a broken relationship; for example, it has been shown that “in long-term successful marriages, spouses reported that the willingness to seek and grant forgiveness is one of the most important factors contributing to relationship satisfaction and marital longevity.” 124 The function of forgiveness is healing a relationship with another, while the function of repudiating blame is increasing one’s own well-being. Since forgiveness !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 121 Haupt (2012). 122 Empirical studies have demonstrated “higher levels of forgiveness correlated with better health habits, lower anxiety, lower anger, lower depression, and more task coping. In addition, people with higher levels of forgiveness had lower hematocrit levels, lower white blood cell counts, and higher TxPA levels. Lower forgiveness levels were correlated with higher T- helper/cytotoxic cell ratios” (Seybold et al. 2001). Although the studies I am citing speak about the benefits accrued by forgiveness, their focus is on what happens to physical health when people let go of anger and resentment. The health benefits discussed occur when people give up grudges, even if they do not express them publicly through forgiveness; the same benefits do not occur when people insincerely express forgiveness while still hanging on to anger, blame, and resentment. 123 Lawler et al. (2005); see also Lawler-Rowe et al. (2008). 124 Allemand et al. (2007); for further support of this claim, see Fennell (1993) and Fincham and Beach (2002). ! 105! and repudiation of blame have different functional roles, we cannot appeal to SFR to explain why the norms on them are the same. It seems likely to me that the functions of the mental states and speech acts are different for other speech acts I have discussed as well, and that SFR therefore fails to explain them (although for reasons of space, I will not engage in a substantive defense of this claim.) For example, it is plausible that the function of appreciation is something like helping people to orient themselves towards good things, while the function of complimenting is something like helping us create better relationships and smooth over social interactions with those whom we compliment. 125 It is also plausible that the function of regret is to draw our attention to the bad things we have done with the aim of helping us to avoid such behavior in the future, 126 while the function of apologizing is improving the relationship between the apologizer and the offended party. 127 Ultimately, though, it doesn’t matter whether SFR can explain the norms on any one particular instance of the SAMS pattern. What matters is whether SFR can explain all of them, and it is clear that it cannot. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 125 A Psychology Today article about complimenting assumes that appreciation orients us towards the good: “we need to cultivate awareness of the good developments that are all around us . . . focusing on and noticing the good qualities in the world around us . . . is a kind of cognitive training, a training of attention.” The same article states that compliments “ease the atmosphere around two people and kindly dispose people to each other” (Marano 2004). 126 A number of empirical studies have shown that anticipated regret—that is, predicting that you will regret a certain decision or behavior—influences decision making. Studies have also shown that experienced regret—that is, regret one actually feels after making a decision— influences future behavior, as well; “because experienced regret helps us to remember “wrong” decisions and motivate us to undo these decisions, it may help us to adapt to similar situations in the future” (Zeelenberg 1999, p. 336). 127 A meta-analysis of previously published empirical studies on apology found that apologizing significantly influenced the offended parties’ “(1) forgiveness, (2) attributions of positive qualities to the apologizer, (3) positive emotions toward the apologizer, (4) positive legal outcomes for the apologizer, (5) intentions to purchase goods from the apologizer, and (6) overall positive reactions and behaviors toward the apologizer” (Hill 2014). ! 106! 6.1.3. The external to internal directional view It does not make sense to attempt to explain the norms governing mental states in terms of the norms governing speech acts. For there is a clear sense in which mental states are independent of and prior to the speech acts that express them, and in which speech acts are not independent of and prior to the mental states they express. One can properly and without any insincerity be in a given mental state without expressing it via a speech act; it is no problem to conditionally resolve to act so long as another person agrees to your plan without promising that you will do so, to appreciate something without complimenting it, or to repudiate blame of someone without forgiving her. Since the mental states can properly function entirely independent of the speech acts, it would be strange to derive the norms on them from the speech acts. But the speech acts we have been discussing are not independent from mental states in the same way: they express that the agent is in a particular mental state, and it is problematically insincere to have the speech act without that mental state. This dependency makes it quite natural to derive the norms on the speech act from the norms on the mental state—which is the best explanation of the SAMS pattern. 7.2.! A good explanation of the SAMS pattern By process of elimination, the internal to external directional view is the best response to the SAMS pattern. This view is captured by a broader Mental States First approach to speech acts, which is a generalization of MSF about promising: General Mental States First Theory (MSF-G): For a literal speech act A that expresses a mental state M: Aptness: If M is apt only if condition C holds, then a publicly relevant performance of A (even if insincere) inherits this constraint, and is apt only if C holds. ! 107! Further commitment: If anyone who is in M is thereby normatively committed to Y, then anyone who performs A (even if insincerely) in a publicly relevant way inherits this commitment, and is normatively committed to Y. MSF-G predicts that the norms of aptness and further commitment that govern the mental state that a speech act expresses apply to the speech act as well, even when the utterance is insincere. That is, MSF-G predicts the SAMS pattern, and proposes that the pattern holds because the speech acts inherit their norms from the mental states. This is a claim about explanatory priority; it posits that the norms on the mental state explain the existence of the norms on the speech act, and fix what those norms are. It is important to note that I am not claiming that MSF-G can explain every kind of norm that applies to a given speech act. To the contrary, there are likely to be norms on speech acts that do not derive from their underlying mental states. For example, the public nature of speech acts might make them subject to norms that are not derived from the mental states expressed by those acts. Norms prohibiting rudeness presumably work in this way; it would be rude to compliment A in front of B if doing so would hurt B’s feelings, but a private appreciation of A cannot be rude. Finally, as with MSF for promises, MSF-G is compatible with there being additional explanations of the norms on any particular utterance of a speech act, which may be more important sources of obligation than the obligation stemming from MSF-G. For example, taking responsibility after apologizing might be important in a particular case not only because you have publicly expressed regret, but because you are independently morally required to make amends for your wrongdoing, regardless of whether you apologized. Because the mental state expressed by a promise is both distinctive of promises and present in all cases, MSF captures the minimal, essentially promissory obligation ! 108! that is always there, while being open to the pluralist idea that other theories explain why the obligation is stronger in many cases. MSF-G shows us that this is a promising strategy to adopt more generally, because promises are not the only speech act whose norms can be derived from mental states in this way. The success of MSF should therefore lead us to be optimistic about providing similarly structured and equally resourceful accounts cashing out the norms on a variety of speech acts in terms of their underlying mental states. ! 109! CHAPTER 4: WHAT THE SINCERITY CONDITION ON PROMISING IS NOT 1. Introduction In Chapter 3, I illustrated how we can derive the norms on a wide range of speech acts from the norms on the mental states expressed by those acts. I suggested that this can provide us with a sufficiently general account of promissory obligation, and assumed the following claims: (1) that promises express conditional resolutions, and (2) that resolving commits you to acting, unless you have a good excuse. In the next two chapters, I will argue in detail for the first claim. The overall goal of these chapters is determining what mental state promises express—which is to say, what the sincerity condition on promising is, or what mental state a promisor must be in if she is to be sincere. To make this determination, we should first think about what the concept of sincere promising involves in its most paradigmatic cases, or cases in which everything about the promise is as it ought to be. Such reflection shows us that paradigmatic sincere promisors meet the following three conditions: (1) They actually intend to act as promised, and are serious about this intention. (2) They know (or at least, believe) that they are going to act as promised. (3) They intend to obligate themselves. To test which of these conditions is necessary and sufficient for sincerely promising, we try stripping away each one to see if the promise still seems to be sincere without it; any condition that cannot be eliminated is necessary, and any condition(s) that alone suffice ! 110! for sincerity is sufficient. 128 In this chapter, I will argue that the second and third conditions on the list are neither necessary nor sufficient for sincerely promising. This will pave the way for my argument in the following chapter that the first condition is both necessary and sufficient, and that promises express conditional resolutions. 2. Does Sincerity Require Intending to Obligate Yourself? The view that promising expresses an intention to obligate yourself to act is common in the philosophical literature. 129 One prominent proponent of this account is David Owens, who argues that promising to Φ expresses the intention to change the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 128 One might worry that this style of argument is question begging. But that’s not the case; this is a process of legitimate conceptual analysis aimed at figuring out which components of the concept are essential, and which are typical but not necessary. The following analogy illustrates the plausibility of this method of conceptual analysis. Suppose we want to figure out what is essentially involved being a philosopher. We can make a list of the paradigmatic features of philosophers; this list might include loving wisdom, publishing in philosophy journals, teaching philosophy courses, having a Ph.D. from a philosophy department, and studying topics using a philosophical methodology. For each candidate condition, we see if someone who lacks this characteristic might still be a philosopher. If so, it’s not an essential component of the concept. For example, the person who has published in philosophy journals and loves wisdom might not be a philosopher (there are interdisciplinary issues of philosophy journals, and non-philosophers can love wisdom, too.) Someone who doesn’t teach and hasn’t published but does study topics with a particular methodology would count as a philosopher (such as a graduate student who has not yet published or taught courses.) We continue with such analysis until we arrive at the core concepts. This is the style of argument I will employ in determining the sincerity condition on promising. 129 Variations of this view can be found in a wide variety of authors; these citations are due to Sheinman (2008, footnote 6 p. 295): Rawls (1999: 305): “Promising is an act done with the public intention to deliberately incurring an obligation;” Raz (1982: 928): “All promises communicate an intention to undertake, by that very act of communication, an obligation;” Finnis (1980: 298f): “The giving of a promise is the making of ... a sign that signifies the creation of an obligation, and which is knowingly made with the intention of being taken as creative of such obligation;” Mackie (1977: 69): “someone who promises purports to put himself under an obligation;” Altham (1985: 9): “A promise ... is an act whereby one communicates an intention to undertake, by that act of communication, an obligation to do a certain action.” ! 111! normative situation by undertaking an obligation to Φ. 130 This is meant to be both a necessary and sufficient condition: you must intend to obligate yourself in order to promise sincerely, and all you must do to promise sincerely is convey that you have intended to obligate yourself. Owens attempts to illustrate the intuitive plausibility of this claim by offering examples of purportedly sincere promisors who intend to obligate themselves but lack the intention to actually keep their promises. He concludes that these cases establish that promisors can “succeed in putting themselves under an obligation to perform by deliberately communicating the intention of doing so without communicating the intention of actually performing.” 131 In the rest of this section, I offer cases meant to show that intending to obligate yourself is not in fact necessary for sincerely promising. I then lay out Owens's examples, and suggest that their seeming plausibility does not stem from the fact that they are communicatively sincere. Rather, Owens seems to be conflating communicative sincerity with sincerity in the sense of being forthright or non-deceptive. But communicative sincerity and lack of deception can come apart; I argue that Owens's cases are indeed non-deceptive, but are communicatively insincere. It follows that intending to obligate yourself does not suffice for communicative sincerity. 2.1. Why intending to obligate yourself is not necessary I will offer three putative counterexamples to the claim that intending to obligate yourself is necessary for sincerely promising. First, a sincere promisor need not even foresee that she will obligate herself, let alone intend to do so, if the chance of successfully creating an obligation is very slim. Second, a sincere promisor might !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 130 2008. 131 ibid, p. 738 ! 112! merely foresee that she will obligate herself without intending to do so, because she and the promisee aim at some goal that does not involve creating obligations. Finally, even a nihilist about obligation can still promise sincerely. 2.1.1. Sincerely promising without foreseeing obligation Suppose Andy likes to stand outside people’s offices not knowing if anyone is inside, and declare, “I promise to give $5 to anyone who is inside this office.” Since his co-workers are out of their offices 75% of the time, Andy never knows whether the conditions of his promise will be met, and whether an obligation to give $5 will thereby be created. (Suppose that Andy can assume his promissory offers will be accepted; his co-workers know that he likes to play this strange game, and are generally disposed to happily accept a random promises of $5, should it happen outside their office doors.) Andy intends to follow through and give away the money, if and when any obligations are created. But he does not seem to be intending to obligate himself every time he makes an office door promissory offer. Nor does he even seem to be foreseeing that he will obligate himself. For he will succeed in creating an obligation only if someone is inside the door, and you cannot foresee the occurrence of events that are up to chance. Yet Andy’s promises seem to be sincere, which means that intending to obligate yourself is not necessary for sincerely promising. 132 We can highlight the plausibility of cases like Andy’s by noting that they are analogous to what goes on when you form an intention to φ only if a certain unlikely condition is met. Such cases are not uncommon in real life; we sometimes see them as acts of social media bravado. For example, suppose Belle tweets that she intends to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 132 Thanks to Mark Schroeder for drawing my attention to examples like this one. ! 113! donate $10,000 to charity if she gets one million Twitter followers in the next 24 hours. Because her chance of getting so many followers so quickly is very low, Belle does not actually intend to obligate herself, or even foresee that she will do so. But so long as she is prepared to follow through should she get enough follows, it seems that her social media offer is sincere. The Andy example works in the same way. 2.1.2. Sincerely promising while merely foreseeing obligation A sincere promisor need not intend to obligate herself even when there is no uncertainty about whether an obligation will be created. This is because in many cases, it suffices that the promisor merely recognize or foresee that the normative situation will be changed by the making of the promise, without explicitly intending that this change occur. 133 Thomas Pink argues that in the vast majority of cases, the promisee does not care very much (or at all) about whether the promisor is morally obligated to perform the promised action. Rather, the promisee is typically concerned with whether the promisor actually will perform the promised action. For example: Suppose, for example, that I am a doctor faced by a worried patient who wants me to be present next week at a crucial operation. The worried patient begs: ‘Promise me doctor, that you will be there.’ And I reassure her: ‘I promise’. My goal in making this promise is, surely, to give my patient something she has requested and badly wants. But precisely for this reason my goal need not be to change what I am morally obligated to do. For my promisee need not be at all interested in what duties I am to be under. 134 Pink notes that we typically make promises in order to attain a variety of goals, none of which explicitly include altering our future obligations or changing the normative state !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 133 This assumes that a meaningful distinction can be drawn between intending and foreseeing. That much seems uncontroversial. What is controversial is the normative significance of such a distinction (e.g., whether the Doctrine of Double Effect is correct.) I can remain neutral about such questions, as my claims here do not rely on any particular understanding of the normative significance of the intention/foresight distinction. All I need is that there is such a distinction. 134 2009, p. 392 ! 114! of the world. Why should sincerity require us to take on intentions we do not otherwise have, and that our promisees don’t care about? As Pink claims about the doctor case, Even if as the doctor I am not at all concerned with the moral status of my future actions—I am concerned simply to provide honest reassurance to the patient that I will indeed be there—I am plainly not being in any way deceitful or insincere, nor am I being incompetent in my use of language. My utterance of the promise need not be in any way misleading to the patient, for she need not understand me as having the production of a moral obligation as my goal. My sincerity as a promisor she will understand as lying not in the genuineness of my intention to obligate myself to her, but in the honesty of my offer to her of my future presence at the operation. 135 However, a proponent of Owens's view might worry that the goal of reassuring the patient can only be attained because the patient believes that the doctor is deliberately undertaking an obligation. That is, perhaps the patient would not be assured by the promise unless she believed that the doctor was intentionally undertaking an obligation—in which case, intending to obligate yourself would be a necessary means to successfully assuring someone through a promise. And perhaps you must intend to take the necessary means to whatever end you intend to take. If so, then Pink’s doctor will need to intend to obligate himself in order to sincerely promise. I’m not sure whether assuring a patient is possible without the patient’s believing that the doctor is undertaking an obligation. But even if this is impossible, there are other sorts of cases where promising serves an aim that has nothing to do with undertaking an obligation. For example, suppose A and B have a long history of violating each other’s trust. They have come to a point where A is willing to trust B, but B still finds it hard to trust A. They want continue rebuilding trust in their relationship, and they realize that one good way to do so is by making, accepting, and keeping promises. And so A offers to promise something insignificant to B—say, that he will !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 135 ibid p. 393 ! 115! wear a red tie to work on Friday. The point of the promise isn’t assuring B about what will happen, for B doesn’t care what color tie A wears. Rather, the point is that B wants an opportunity to display a willingness to trust A by accepting a promise, and A wants an opportunity to follow through and solidify that trust. So A sincerely promises B that he will wear the red tie, and intends to follow through, lest the fragile trust between them be undermined. A does not explicitly aim at or intend to obligate himself. Yet his promise still seems to be sincere. Is this example susceptible to the same worry as the doctor case? That is, might B’s trusting that A will keep the promise only be possible because B recognizes that A is obligating himself? The answer is no, for we can assume that B doesn’t trust A because A has undertaken a moral obligation. Rather, B trusts A simply because A has invited him to do so by making a promise. If B were relying on the fact that A’s promise creates a moral obligation, and that A is generally morally conscientious and therefore likely to follow through, the promise would not serve to successfully rebuild trust between A and B. For A would be acting out of moral conscientiousness, rather than simply because he told B that he would. So A cannot aim directly at obligating himself, lest he undermine the trust-building goal of the promise. 2.1.3. Sincerely promising while not believing in obligation Consider someone who is a nihilist about all moral obligation, including promissory obligation; call him Niles. Niles recognizes that promises are a useful social convention that enables us to solve coordination problems and live more effectively together. But he is an error theorist about morality, and so doesn’t think that he’s morally obligated to keep promises, because he doesn’t think that any moral obligations ! 116! exist. However, Niles recognizes that it is in his own interest to fully participate in the useful social practice of promising, and that he can most efficiently do so by resolving to act whenever he makes a promise. Suppose a journal editor asks Niles to promise that he will complete his referee report by the end of the week. Niles resolves to do so, and promises to complete the report before the weekend. Niles’s promise seems sincere. But surely he is not intending to obligate himself, for he doesn’t believe that his promise actually will obligate him. One might worry that Niles is incapable of promising sincerely if he does not “buy in” to the promising practice. There are indeed some practices that you cannot participate in sincerely if you do not buy in to them. For example, an atheist who utters responsive prayers with a believer does not seem to be praying sincerely, even if the atheist has good social reasons for doing this. But there are other practices that you can participate in sincerely without buy-in. For example, suppose that you do not think there is anything genuinely normative about most wedding traditions. You are a bridesmaid in the wedding of a friend who values these traditions. Sincerely participating in the bouquet toss requires being an unmarried woman, since the presumption of the bouquet toss is that it determines which person will be a bride next. If you are an unmarried woman, you can participate in the bouquet toss sincerely, even if you don’t think there’s actually anything normative about it. Similarly, Niles seems to be able to sincerely participate in the promising practice by forming and carrying out resolutions, even though he doesn’t think there’s actually anything normative about it. If this is correct, then the sincere promises of nihilists are counterexamples to the claim that intending to obligate yourself is a necessary sincerity condition on promising. ! 117! 2.2. Owens's examples: why intending to obligate yourself is sufficient Owens offers multiple cases meant to show that intending to obligate yourself is a sufficient sincerity condition on promising. In the first example, I am low on cash and ask to borrow money from my wealthy and generous friend Janet. Because I am proud and do not want to be seen as taking charity, I promise Janet that I will pay her back. However, I know that she will never ask for the money back, and I am not sure whether I could repay her if she did. So I promise to repay her only on the assumption that I will never be required to do so. In promising, I recognize that I am creating an obligation to repay. But I do not expect to ever be held to this obligation, and so I do not intend to repay. (Assume that Janet knows that I never expect her to ask for repayment, lest our intuitions be thrown off by concerns that I am deceiving her.) Owens claims that my promise to Janet can be entirely sincere. Owens's second case has the same structure as the money borrowing case, except the promisor is more explicit about the fact that she does not actually intend to do the thing she promises to do. Suppose Junior is editing a book of essays. He asks Senior, who is a prominent and highly sought after academic, to contribute an essay within six months. Junior knows that Senior is unlikely to submit the essay in time, and plans to extend the deadline should Senior be late. But Junior “is requiring a promise from all the authors and he wants at least to begin by treating the well-known contributor as he treats the others.” Senior, who does not know whether she will have the paper ready in six months but who does know that the deadline will be extended for her if she does not, says to Junior: “I shall commit to six months like everyone else but I am sure you will not hold me to it: we will see how things are in six months time.” Owens claims that Senior can be sincere in making this promise, stating that “doubtless the ! 118! contributor is not behaving well in saying what she says but she could hardly be accused of insincerity.” 136 Note that, at least as I understand it, the social role that Owens takes the promise to play in this case puts pressure on the idea that promises express the intention to obligate. For Owens stipulates that Junior wants to treat Senior in the same way that he treats all of the other contributors. But all of the other contributors agree to turn in their papers on time. They do not agree merely to obligate themselves to turn in their papers on time; they might make such an agreement and fail to abide by their obligations, and then the volume would not be ready. If the deadline is to be effective, any contributor who commits to the deadline it must express that they intend to meet it. And so if Senior is to be (at least nominally) treated in the same was as everyone else, her promise must express that she, too, intends to turn in the paper on time. Third, Owens offers a case in which we are neighbors and I am tired of your filthy old car being an eyesore in your driveway. I have repeatedly asked you to wash the car, and you have repeatedly refused to do so. Perversely, I want to see you fail morally, in order to gain a sense of superiority over you. And so I ask you to promise to wash your car, knowing that you will never wash it. You wish to spite me, and so you say “Okay—I promise to wash the car, but I’m not going to do it!” Owens claims that the wrong you commit when you promise to wash your car and make it clear that you will not do so is a wrong of “blatant contempt” rather than insincerity. Finally, Owens offers a fourth case of another sort, which he calls a prophylactic promise. He is thinking of cases in which “a promisor cannot form the intention to do !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 136 All three quotations in this paragraph are from ibid, p. 751 ! 119! the thing he is promising to do but can still intend to do something that he believes will get him to fulfil his promise.” For example, suppose that I would like to quit smoking, but I love smoking so much that I cannot bring myself to intend to quit. However, being a person of my word I am usually motivated to keep my promises. And so I promise my spouse that I will quit smoking, hoping that “the making a promise to give up smoking is itself the act that will enable [me] to keep it; a valid promise is the very incentive the promisor needs to ensure performance.” 137 Owens once again claims that, intuitively, the promisor in this case is sincere. 2.3. Communicative sincerity vs. sincerity as non-deception In each of these cases, Owens asserts that the wrong in question is not a wrong of insincerity. What is the driving force behind these intuitions, and why does Owens find it so plausible that the cases described above are cases of sincere promising? Two textual clues point us to the answer. First, in an earlier discussion of whether a gift of a car that you expect will soon be stolen can be sincere, Owens argues that a speaker who is entirely explicit that she is giving a gift solely in order to attract the attention of thieves indeed commits a wrong. But he claims that this “is hardly a wrong of insincerity; the speaker is being far too blatant for that.” 138 The underlying assumption seems to be that insincerity must be subtle or deceptive, from which it seems to follow that a speaker who is blatant or obvious about her intentions must be sincere. Second, when discussing the prophylactic promise case in which I promise my spouse that I will quit smoking while not intending to do so, Owens claims that “there need be no wrong of insincerity here and I can remove any suspicion of such a wrong !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 137 Both quotations in this paragraph are from ibid, p. 747 138 ibid, p. 742, my emphasis ! 120! by being quite explicit about my situation.” 139 In other words, if I clearly tell my spouse that I do not (yet) intend to quit smoking, and that I do not know whether I will be able to do so in the future, then my promise to quit is sincere. So Owens seems to take explicitness to entail sincerity. This makes intuitive and pre-theoretical sense. We often think of sincerity as something closely akin to the virtue of honesty: the sincere person is one who does not deceive others or misrepresent her true self in any way. Indeed, the Oxford English Dictionary defines sincerity in this way, as “freedom from dissimulation or duplicity; honesty, straightforwardness.” We can call this sincerity as non-deception (hereafter referred to simply as “non-deception,” for the sake of clarity.) When we are non-deceptive, we do not pretend to be in a state that we are not really in. Rather, we are honest and forthright in our communication. Whether an utterance is deceptive must be distinguished from whether the utterance is communicatively sincere (i.e., whether the speaker is being tricky must be distinguished from whether she really is in the state she purports to be in.) Usually, communicatively sincere utterances are non-deceptive, and communicatively insincere utterances are deceptive. Sincerity without deception is your standard sincere promise; I promise my sister that I will take good care of the cashmere sweater she lent me, and I intend to do so and haven’t tricked her about this fact. Insincerity with deception is your standard insincere promise; I promise my sister that I will care for her sweater in order to trick her into letting me borrow it, when I really intend to sell the sweater on Ebay and keep the profits. However, it is possible to be communicatively sincere while being deceptive, and to be communicatively insincere while being non-deceptive. I address these possibilities in the next two sections. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 139 ibid p. 747, my emphasis ! 121! 2.4. Communicative sincerity with deception Cases of communicative sincerity with deception are rarer. But they are possible. For example, communicatively sincere assertions can be deceptive. Suppose you believe that I am a pathological liar, who regularly asserts the opposite of what she really believes. I am not in fact a pathological liar, but I know that you falsely believe me to be one. You ask me whether the bank at which I work is open on Saturdays. I assert that it is, and I genuinely believe that the bank is open on Saturdays. I am communicatively sincere, in that I communicate that I believe that the bank is open on Saturday, and I really do have this belief. But I know that you take me for a pathological liar, and will therefore assume from my statement that the bank is not really open on Saturdays. I make no effort to convince you that I am not in fact a liar. To the contrary, I intend for you to draw the wrong conclusion from my statement. There is a clear sense in which I am deceptive, and intend to mislead you. A communicatively sincere apology might be deceptive, as well. Suppose a group of children are teasing a younger child. Their teacher catches them in the act, and tells them that each of them must apologize. Four of the five children do not regret what they’ve done. One by one, each of the four walks up to the victim and says, “I’m sorry,” while visibly crossing their fingers—thereby indicating by ubiquitous playground convention that they are lying and their utterance is insincere. But the last teaser does regret his behavior, and truly feels bad about what he did. He walks up to the victim and utters “I’m sorry” while crossing his fingers, in order to (falsely) indicate that he isn’t really sorry, and thereby save face in front of his four compatriot teasers. This last child is communicatively sincere—he really is sorry as he purports to be—but also deceptive, for he implies by his finger cross that this is not the case. ! 122! Communicatively sincere promises can be deceptive, as well. For example, suppose I am a bibliomaniac who has a compulsion to acquire books. I ask to borrow a book from you, and you agree to lend it to me on the condition that I return it to you within six months. I promise to return the book to you within six months, and I really do plan to do this. But I know myself (and my bibliomanic compulsion) well, and past experience has taught me that 9 times out of 10, I can’t bring myself to return the books that I borrow. In spite of my good intentions to do so, I often change my mind when the time to return the book arrives, and decide to keep the book instead. My promise to return the book is communicatively sincere. But unless I inform you that I am not likely to actually return the book in spite of my resolution to do so, I am deceptive, for my promise implies that I am likely to return it. 2.5. Communicative insincerity without deception Assertions can be communicatively insincere even when they do not deceive. For example, suppose Bernie is a compulsive liar, who deeply regrets his lying habit and is undergoing therapy to rid himself of it. He hasn’t yet progressed to a point where he is able to refrain from lying. But he has progressed to a point where he is able to avoid deceiving people, by explicitly pointing out the fact that he is lying. Bernie’s landlord asks him whether he has paid the rent, which he hasn’t. Bernie replies, “Oh yeah, the check is in the mail . . . and you know that I’m a liar.” Bernie is communicatively insincere, for his utterance that the check is in the mail conveys that he believes that the check is in the mail when he does not in fact believe this. But he doesn’t deceive the landlord. ! 123! An agent might also make an insincere but non-deceptive assertion in order to be perverse, in the way that Owens's car washing neighbors are perverse. Suppose that the villain has captured and is interrogating Bond about where the key to the nuclear device is hidden. Bond wishes to defy and irritate the villain by openly lying to his face, and so snarls, “The key is in the desk drawer . . . and I’m not being honest with you!” As with Bernie, Bond’s assertion that the key is in the drawer is communicatively insincere but does not deceive. Apologies can be insincere and non-deceptive, as well. Suppose Carla is fighting with her sister Darla. Carla strikes Darla, and their mother intervenes and tells Carla that she must apologize. Carla grudgingly turns to Darla and says, “Fine, I’m sorry . . . but only cause Mom is making me!” In saying that she is sorry, Carla conveys that she regrets her action, and then immediately declares that she doesn’t really, making her utterance both insincere and non-deceptive. Forgiveness works in the same way. Suppose Mom tells Darla that she must forgive her sister. Darla says, “Fine, I forgive you . . . but I’m still really mad!” She is likewise insincere but non-deceptive. Cases of insincerity without deception should seem familiar, because we discussed many of them while employing the linguistic test of communicative sincerity in Chapter 3. If I say, “What a lovely coffee stain! I don’t appreciate it,” I am non- deceptive but communicatively insincere. The same basic phenomenon is happening in all of these cases: the speaker declares her own communicative insincerity in order to avoid being deceptive. ! 124! 2.6. Defusing Owens's counterexamples Like Bernie, Bond, Carla, and Darla, the promisors in Owens's cases seem to be sincere only in the sense of being non-deceptive, and not in the sense of being communicatively sincere. I do not deceive Janet into thinking that I will repay her; she knows that I am promising to do so only to save face. Senior does not trick Junior into thinking that she will turn in the essay by the deadline; rather, she is explicitly states that she may not. The perverse neighbor could not be clearer that he does not actually intend to wash his car as he promised to. And I openly inform my spouse that I am unsure whether I will be successful at intending to quit smoking, let alone successful at actually quitting. Owen invites us to see all of the promisors in his examples as intuitively sincere. If we agree with him, I suspect that this is because we recognize that the promisors in these cases are non-deceptive—they are not lying or tricking anyone, and are presenting themselves as they really are. They are clear and upfront about the nature and limits of their promissory intentions (or lack thereof.) But none of this entails that their utterances are communicatively sincere. To the contrary, by making promises, the agents in Owens's cases express that they plan to act in a certain way. They then explicitly declare that they do not plan on so acting—which is equivalent to avoiding deception by proclaiming their own communicative insincerity. 140 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 140 Keep in mind that someone who promises insincerely need not be morally blameworthy all- things-considered, depending on the situation. For thinking that insincerity is always blameworthy is likely to throw off our intuitions, because here are obvious instances in which sincerity is not required: if the axe-murderer at the door looking for you says he will not leave unless I promise to notify him when you return, I am morally required all-things-considered to make an insincere promise. And if insincerely promising my spouse that I will quit smoking is ! 125! I have argued that Owens's purported counterexamples are examples of non- deceptive but insincere promises. We cannot use intuitions about sincerity as non- deception to determine what is expressed by an utterance, for this can be done only by appealing to communicative sincerity. It follows that Owens's cases do not establish that intending to obligate yourself is sufficient for sincerely promising. And this shows us that the third of the three potential sincerity conditions proposed at the beginning of this chapter does not succeed. In the rest of this chapter, I will argue against the second condition: that sincerely promising requires knowing or believing that you will act. 3.! Does Sincerity Require Knowing that You Will? Imagine that your 5-year-old nephew has a bad dream, in which you jump to the moon and decide to live there, never to return. 141 He’s concerned that his dream will come true, and is very distressed by the prospect of your jumping to the moon. So you set his mind at ease by saying, “Don’t worry, I promise not to jump to the moon.” On the face of it, this seems to be a genuine promise. Moreover, it seems to sincere. You’re certainly non-deceptive: since you’re definitely not going to jump to the moon, you haven’t misled your nephew in any way. And you seem to be communicatively sincere, as well: you’re conveying that you plan to refrain from jumping to the moon, and you will surely refrain from doing so. Why is this promise sincere? It might not seem plausible that you have intended or resolved not to jump to the moon in this case. For intending to refrain from doing !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! the only thing that will enable me to quit, this very well may be what I ought to do all-things- considered, especially if I do not mislead my spouse. 141 Thanks to Jake Ross for suggesting this example to me in conversation, and pressing me to consider and respond to the Knowledge Account of sincerity. ! 126! something impossible seems superfluous at best, and possibly downright irrational; why waste your mental and volitional energy resolving to avoid something that you couldn’t do anyway, even if you wanted to? However, you do know that you’re not going to jump to the moon any time soon. This case therefore suggests the following potential sincerity condition on promising: Knowledge Account: Promising to Φ expresses that you know that you will Φ; to sincerely promise to Φ, it is necessary and sufficient to express that you know that you will Φ. The Knowledge Account explains why the promise to your nephew seems sincere: it doesn’t matter if you haven’t intended or resolved, because all you need to do to promise sincerely is know that you will act. This account is compatible with the claim that we form intentions or resolutions in most cases of sincere promising. For in standard cases in which failing to keep the promise is possible—say, promising your nephew that you will take him to the zoo—you know that you will act because you have intended or resolved to do so. In the rest of this section, I will argue that knowing that you will act is neither necessary nor sufficient for sincerely promising. I then argue that a proper understanding of the jumping to the moon case shows us that the promise involved is not plausibly construed as both a genuine and sincere promise; it therefore is neither a good motivation for the knowledge account, nor an objection to a resolution-based account of the sort I will develop in the next chapter. 3.1.! Why knowing is not sufficient Consider a case of promising is when you know that you will Φ through the agency of someone else, regardless of what you intend. Suppose that you are getting a ! 127! ride home with Eva tonight, and you know that she is going to stop at happy hour on the way home. A colleague asks whether you’ll be at happy hour, and you promise her that you’ll go. You know that you will be there, because you know that Eva will take you there, and Eva is your only option for getting home. But you don’t form any intention to go, let alone a resolution. And you don’t really care about whether you go. If Eva were to change her mind and skip happy hour, you would have no backup plan to get there yourself; you wouldn’t look for another ride, or try to talk Eva into going. In this case, you meet the sincerity condition posited by the Knowledge Account. But your promise does not seem to be sincere. Sincerely promising requires being in a state with some counterfactual stability. When you promise to Φ, you convey not just that you plan to Φ, but that you plan to Φ even in the face of barriers to action, or if the situation changes in a way that makes Φing uncertain or unlikely on independent grounds (e.g., if Eva decides not to go to happy hour after all.) A promise that is grounded in the agency of someone else does not display this sort of counterfactual robustness. It follows that promises must express something more than merely the predictive knowledge that you will act. Another type of case that satisfies the Knowledge Account but does not seem to be an instance of sincere promising concerns promising to do the inevitable. Suppose you are having an argument with your sister Farrah. You keep interrupting her, and she demands that you refrain from interrupting her again for the next three minutes. You then take a pill that, ten seconds hence, will make you unable to speak for three minutes. You say, “Okay, I promise not to interrupt you”—and then become incapable of speaking for three minutes. You know that you will not interrupt Farrah, because it ! 128! would be impossible to do so, even if you wanted to. But you don’t form any intention or resolution to avoid interrupting her. In fact, if the drug were to wear off early, you’re fairly confident that you would interrupt her. As before, this promise seems insincere, for it lacks the counterfactual robustness necessary for sincere promises. Since the Knowledge Account predicts that your promise to Farrah is sincere, the Knowledge Account must be insufficient. 3.2.! Why knowing is not necessary We have seen that knowing that you will Φ is not sufficient for sincerely promising to Φ. Is knowing that you will Φ necessary? It is generally up you whether you are sincere in making a promise, because a promise’s sincerity depends only on the inner mental state of the promisor. 142 Whatever state is required for sincerely promising should therefore be something we have control over. Knowing that p requires (at least) that you believe that p, that you are justified in this belief, that p is true, and that you are not in a Gettier situation (i.e., a situation in which your belief is justified but merely happens to be true for accidental reasons.) Some of these conditions on knowledge— particularly truth and the non-Gettiered constraint—are outside of our control. Whether p is true (or whether I have been Gettiered) depends at least in part on the way the world is, rather than exclusively on my inner state. Knowledge therefore does not look like the sort of thing that could be a sincerity condition on promising. 143 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 142 There might be cases in which someone is brainwashed or manipulated by a nefarious neuroscientist into having or lacking a certain mental state, in which case the sincerity of her promise would not be entirely up to her. But these cases need not trouble us, as they fail to meet the voluntariness validity condition on promising laid out in Chapter 1. 143 I wish to sidestep debates about volitionalism about belief, and about internalism vs. externalism about mental content. I am claiming only that beliefs about what you are going to do ! 129! For example, suppose that you promise to take your nephew to the zoo this weekend, and you resolve to do so, conditional on his continuing to accept your promise. You believe that you will take him to the zoo, and are justified in this belief because you typically succeed in following through on your resolutions. But unbeknownst to you at the time you make the promise, your belief is false: the zoo will burn down tomorrow, and you will be unable to take your nephew there. Your promise still clearly seems to be sincere, even though you don’t know that you will take your nephew to the zoo. 144 The same sort of counterexample could be run with a case in which you fail to know that P not because P is false, but because you are Gettiered. This shows us that promises do not express knowledge that you will act. However, there is a component of knowledge that does depend on your inner mental state. Whether you believe that you are going to Φ is something that is in your control, and isn’t typically affected by the way the world is. Perhaps, then, promising expresses not knowledge, but belief, such that sincerely promising to Φ requires believing that you will Φ. Such an account would be insufficient for the same reasons as the Knowledge Account is insufficient: your promise is insincere in cases in which you believe that you will Φ inevitably or through someone else’s agency, and promise to Φ without intending or resolving to do so. (The examples about Eva taking you to happy hour and not interrupting Farrah can be run mutatis mutandis appealing to belief !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! are within your control to at least some extent, and do not depend on external factors in the way that truth does. This is compatible with it being the case that, for example, we cannot believe at will in opposition to our evidence, or that my believing that I will bring you a glass of water depends in part on whether the clear wet stuff in front of me is H 20 or XYZ. 144 I’m assuming that we have some plausible account of what it is to have knowledge of the future, and that if your belief about the future is false, it fails to count as knowledge. ! 130! rather than knowledge.) Much more difficult is the question of whether believing that you will is necessary for promising sincerely. I turn to this question in Section 4 below. 3.3.! Why you can’t sincerely promise to jump to the moon We have seen examples of insincere promises with knowledge, and sincere promises without knowledge. What about the initial example, in which you seem sincere in promising your nephew that you will not jump to the moon because—and seemingly, only because—you know that you will not do so? This case does not give us a good reason to accept the Knowledge Account, because there are ways of explaining the case without relying on the claim that sincerely promising requires knowing that you will act. First, it seems quite likely that you are not actually making a promise to your nephew. Rather, you might be promising in an assertoric sense, as a means of expressing certainty about a claim rather than conveying a personal commitment to a course of action. Saying, “You’ll love this movie, I promise” conveys that you are confident that your interlocutor will love the movie, and not that you are going to cause her to love it; perhaps saying, “I promise you that I won’t jump to the moon,” conveys that you are absolutely confident that you will not jump to the moon, and not that you plan to refrain from doing so. Or you could be merely behaving as if your jumping to the moon were possible, and pretending to promise on this basis. Perhaps your promise makes sense only in the context of playing along with your nephew that jumping to the moon is a live option, in order to assure him and ease his mind. To motivate this explanation, consider a related case. Suppose your 12-year-old nephew dreams that you will jump to the moon, and tells you that he is worried that you will do so. You explain to him why jumping to the ! 131! moon is impossible, and he is competent enough with basic scientific concepts to fully understand you. But he wants more assurance than this, and says, “Promise me you won’t jump to the moon.” You might humor him by pretending to promise. But it would be reasonable for you to refuse, and say something like “Don’t you understand? I don’t have to promise you I won’t jump to the moon; I can’t do it anyway!” I suspect that any sense that the promise is sincere in the original case might stem from the fact that the 5-year-old promisee believes the promise is necessary and non-superfluous, because he mistakenly believes that it is possible to fail to act as you promise. Perhaps your nephew is too young to fully grasp that you cannot jump to the moon, in which case the only way to assure him and ease his mind is by playing along with him, and “promising” as he requests. But this does not make it a genuine promise, and so it is not a case in which knowledge is sufficient for sincerely promising. Moreover, we can accommodate the case without appealing to the Knowledge Account even if we do construe it as a genuine promise. There are several ways of doing so. First, it could be that you genuinely promise, but do so insincerely, and do not form a resolution to refrain from jumping to the moon. Insincerely promising is typically wrong, but it might be the best thing to do in this case: your nephew is distraught without a promise from you, and so you promise insincerely in order to attain the greater good of alleviating his distress. Second, perhaps you genuinely and sincerely promise, and you form an irrational resolution to refrain from jumping to the moon. This will be the case if your resolution is unnecessary or superfluous, and if forming superfluous resolutions is irrational. And it’s not implausible to think that someone who makes such a wacky promise—who ! 132! bothers to make a promise to refrain from doing something that she takes to be impossible anyway—might not be behaving entirely rationally. Finally, perhaps you genuinely and sincerely promise, and the resolution you form is not in fact irrational. Maybe it is rational to form superfluous resolutions when they are exceptionally easy to maintain, and there is an independent reason in favor of forming them. Refraining from jumping to the moon requires no effort on my part, which makes intending to do so come what may very easy. And there is an independent reason in favor of my so resolving: my expressing such an intention helps my nephew sleep at night, and my actually forming such an intention prevents me from being guilty of insincerity. If any of the above possibilities are plausible explanations of our intuitions in the moon jumping case, then the case will not provide any serious motivation for the Knowledge Account. 4.! Does Sincerity Require Believing that You Will? Berislav Marušić defends a belief sincerity condition on promising, cashing it out in two different places as a negative and positive formulation. Call these formulations together the Belief Account: Negative formulation: If we believe that there is a significant chance that we won’t Φ, then we are insincere in promising to Φ. 145 Positive formulation: Our promise is sincere only if we believe that we will do what we are promising to do. 146 According to the Belief Account, sincerely promising requires believing that you will do what you promise to do. You can fail to be sincere by failing to have this belief, either !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 145 2013, p. 299 146 2012, p. 4 ! 133! because you believe that you outright won’t do what you’ve promised, or because you believe that there is a good chance that you won’t do so. Marušić claims that this is a necessary condition on sincerely promising. 147 He does not presume that it is a sufficient condition, as well; he notes in a footnote that someone could endorse both his belief sincerity condition and the claim that sincerely promising requires intending to act. 148 My assessment of the Belief Account will proceed in three stages. First, I will offer counterexamples to the necessity of believing that you will, or cases in which it seems that an agent can promise sincerely even though she doesn’t believe that she will act. In discussing these examples, we will see that what might at first glance appear to be a failure of sincerity is sometimes in fact a failure of another sort. These examples also provide evidence that the resolution-based account I will defend in Chapter 5 is on the right track. Second, I will argue that the account does not pass the linguistic test for communicative sincerity, and offer a debunking explanation of any sense we may have that it does. Finally, I will respond directly to Marušić’s positive argument for the Belief Account, which appeals to the relationship between promising and asserting. 4.1. Counterexamples to the belief account If it is possible to intend or resolve to act without believing that you will, it will be possible for someone to promise to Φ and resolve to do so without believing that she will. If such a person can be sincere, we will have a counterexample to the necessity of the Belief Account. Consider the promise Peter makes to Christ in Matthew’s Gospel: !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 147 Marušić’s project in these papers is not articulating a sincerity condition on promises for its own sake, but rather doing so in the context of arguing that it is possible to sincerely promise to Φ or rationally believe that you will Φ even you have good evidence that you will fail to Φ (e.g., sincerely promising to remain married and believing that you will remain married even though you know that 50% of all marriages end in divorce. ) 148 ibid p. 14 footnote 52. ! 134! Peter said to him, “Though all become deserters because of you, I will never desert you.” Jesus said to him, “Truly I tell you, this very night, before the cock crows, you will deny me three times.” Peter said to him, “Even though I must die with you, I will not deny you.” 149 Assume that Peter is making a promise to refrain from denying Christ, and that he indeed resolves to do so: he plans to refrain from denying, and is extremely serious about this plan (serious to the point of death, even.) We can assume that Peter takes Jesus to be a reliable epistemic authority, and accordingly takes Jesus’ testimony to provide good evidence that Peter will deny. The most natural reading of the story is that Peter fails to believe that he will deny Christ, in spite of being presented with what he presumably takes to be excellent evidence for this belief; perhaps he engages in irrational wishful thinking. It is also possible that Peter rationally fails to believe that he will deny Christ; Marušić argues that it can be practically rational for you to believe that you will Φ even when this belief is epistemically irrational, because “as agents, we can, and should, take a different view of our future actions than observers would . . . we should look to our practical reasons, not to our evidence alone, to settle what we will do; we should decide what to do, not predict what we will do.” 150 However, let’s assume for the sake of argument that Peter does believe on the basis of Jesus’ testimony that he will deny Christ. The promise he makes to refrain from doing so is misguided in one of two possible ways. First, Peter might be being irrational. It is plausible that intending to act while believing that you will not do so is irrational for any intention, and thereby for any resolution (since intentions are an !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 149 New Revised Standard Version, Matthew 26: 33-35 150 2013, p. 294 ! 135! essential component of resolutions.) 151 152 If this is so, then Peter’s intending to refrain from denying while nevertheless believing that he will deny is irrational. 153 Second, Peter’s promise might be deceptive, as well. Peter’s insisting that he will refrain from denying Christ even to the point of death implies that he will not deny Christ, and licenses his audience to infer this. But by hypothesis, Peter believes that he will deny Christ. So unless Peter explicitly negates the implication that he will refrain from denying—for example, by saying something like “Even though I must die with you, I will not deny you . . . but I don’t believe I’ll succeed, so don’t take that to mean that I won’t in fact deny you”—he is being deceptive. (Of course, if Christ has divine foreknowledge, he won’t be deceived by Peter’s promise—or by anything, for that matter. But this is irrelevant, as one can be deceptive without successfully deceiving.) Peter’s irrationality or deceptiveness make his promise problematic. But they do not necessarily make his promise communicatively insincere. For we saw in the previous section that sincerely promising seems to require conveying that you are committed to acting in a counterfactually robust sort of way. And Peter is committed to refraining from denying come what may, even though this commitment is irrational or !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 151 A number of philosophers take a strong view that intending to Φ requires believing that you will Φ (see, for example, Anscombe (1963), Audi (1973), Harman (1976), Davis (1984), and Ross (2009).) My argument here turns on the weaker claim that rationally intending to Φ requires believing that you will Φ. 152 Strictly speaking, I take resolutions to be intentions combined with second-order desires; for more on this view, see Chapter 6. 153 Marušić assumes it is possible but irrational to intend without believing you will; this shows us “that the Sincerity Condition must be refined as follows: if we believe that there is a significant chance that we won’t Φ, and we avoid irrationality, then we are insincere in promising to Φ.” (2013, p. 300). ! 136! the expression of it deceptive. It therefore seems possible for Peter to promise sincerely without believing that he will act. Peter’s case is one of promising to Φ while believing that you definitely won’t succeed. Much more common are cases of promising to Φ while believing that you might not succeed. For example, suppose an ancient sailor, Gaius, is about to set off across the ocean using new navigational techniques, which work in theory but have never been tested in practice. His daughter Helen, a sailor in training who regularly accompanies her father, asks to come along. Gaius refuses, telling her that this trip is more risky than his usual voyages. Helen requests that Gaius stay home if the journey will be dangerous. Gaius tells Helen not to worry, and promises her that he will return. We can assume that Gaius resolves to return, and is strongly volitionally committed to doing so. And he is at least somewhat confident that he will succeed; the navigational techniques work in theory, and he is going to do his very best to ensure that they work in practice, as well. But we can also assume that Gaius is not completely confident he will succeed, and believes there to be a reasonable chance of failure. For he would take Helen along if he were completely confident in a safe return, and he refuses to take her. So Gaius is confident enough to promise Helen in good faith that he will return, but not confident enough to bet Helen’s life on it. Is Gaius’s promise sincere? To answer this question, we must address the broader question of whether we can sincerely promise to do that which we recognize has some chance of failure. Of course, every promise has some chance of failure, even if miniscule: I can sincerely promise my sister that I will meet her for lunch tomorrow while recognizing that I might fail because I get kidnapped en route to the restaurant. ! 137! Surely sincerely promising to Φ does not require absolute certainty that there is no situation under which Φing will not occur. If it did, we couldn’t sincerely promise anything. The relevant question is one of degree: is there a point at which I take failure to be so likely that I do not outright believe that I will succeed, but so unlikely that I can nevertheless sincerely promise to succeed? I’m not sure where this balance point is, but it seems likely that such a point exists, at least in cases in which my doubt about whether I will succeed is due to worries about the external circumstances, while my confidence that I may succeed is due to the strength and robustness of my volitional commitment. Gaius recognizes that there is a decent chance that he will not return, not from a failure of willpower but due to the vagaries of fortune. Yet he believes there to be a good chance that he will return, because doing so is possible, and he is dead set on it. Gaius seems to be overly optimistic to a fault: he is clear-sightedly, passionately, and perhaps foolishly—or perhaps admirably?—committed to a completing a certain course of action, in spite of his belief that he may not succeed even with his best efforts. For lack of a better term, we can call Gaius quixotic. As with Peter, he might be deceptive: maybe his utterance makes Helen think that he believes that he is certain to return, when in fact he does not believe this. But Gaius nevertheless seems to be sincere—he conveys that he is volitionally committed to returning, and he really is so committed in a counterfactually robust way. His quixotic idealism and potential deception do not alter this commitment, and therefore do not undermine his sincerity. Compare Gaius’s case to one in which the source of doubt is not external circumstances, but one’s own level of commitment. Suppose that another sailor, Indira, is tired of the nasty and brutish life she leads in her home country, and is about to set ! 138! off on a safe and routine trading mission to an exciting, prosperous, and highly cultured civilization across the sea. She feels very tempted to settle permanently in this new land, and hasn’t yet made up her mind about whether she will do so. Helen asks Indira to promise her that she will return from the voyage. Indira promises, in spite of her uncertainty over whether she will decide to stay. This promise does not seem to be sincere. But the insincerity does not stem from the fact that Indira fails to believe that she will return. Rather, Indira is insincere because she lacks the right sort of volitional commitment—she hasn’t in good faith resolved to return. And Indira is likely deceptive, as well: Helen is likely to think on the basis of the promise that Indira has made up her mind to return, whereas Indira is in fact still undecided. So cases in which you are volitionally committed but do not believe you will seem sincere, albeit irrational (if you believe you won’t) or quixotic (if you believe that you may not because of external circumstances.) And cases in which you do not believe that you will because you are not fully committed seem insincere. We have also seen several distinct ways in which promises can be flawed: they can be insincere, as Indira’s promise is; irrational, as Peter’s promise is; or quixotic, as Gaius’s promise is. Any particular insincere, irrational, or quixotic promise can also be either deceptive or non- deceptive; we saw in the discussion of Owens's view how deception can come apart from sincerity, and in this section how irrationality or quixoticness can come apart from deception, as well. I have suggested that someone who promises to Φ but fails to believe that she will do so is irrational or unduly optimistic, but is not insincere. If this is right, we have a counterexample to the Belief Account. However, it can be difficult to assess the source or reliability of our intuitions about sincerity, and especially about the fairly technical ! 139! notion of communicative sincerity. For that reason, I will turn in the rest of this chapter to a direct assessment of arguments for the Belief Account, with the aim of showing that they do not establish that believing that you will is an independent and distinctive sincerity condition on promising. 4.2.! The linguistic sincerity test In the last section, I focused on whether it is problematic to promise to Φ while lacking the belief that you will. In this section, I will focus on whether it is problematic to promise to Φ while proclaiming yourself to lack the belief that you will. In Chapter 3, we used a linguistic test to determine what mental state a given speech act expresses. Recall that the test works as follows: for a speech act S and mental state M, is performing S while declaring yourself to lack M straightforwardly communicatively insincere? If so, then S expresses M—or, equivalently, M is a sincerity condition on S. By “straightforward,” I mean that the explanation of communicative sincerity is direct, in that it need not appeal to any speech acts apart from S. Suppose S pragmatically implicates some other speech act S*, and one cannot utter “S*, but not M” without insincerity. This will show us that M is a sincerity condition on S*, but will not establish that M is also a sincerity condition on S. For example, saying, “John has excellent handwriting and is always prompt” in a letter of recommendation implicates that John is bad at his job. One could not assert, “John is bad at his job, but I don’t believe that John is bad at his job” without at least part of that assertion being insincere (either you don’t really believe that John is bad at his job and the first part of your assertion is insincere, or you do believe that John is bad at his job, and the second part is insincere.) But this doesn’t show us that “John has excellent handwriting and is always ! 140! prompt” expresses that John is bad at his job, or that believing him to be bad at his job is necessary for sincerely uttering “John has excellent handwriting and is always prompt.” We can use the linguistic sincerity test to construct an argument for the Belief Account. Suppose I say to my nephew, “I promise to take you to the zoo, but I believe that I might not take you there.” This utterance is problematic in some way; there’s clearly something wrong with promising to Φ while proclaiming that you don’t believe you will do so. If this utterance is problematic because it is straightforwardly communicatively insincere, this will show us that believing that you will is a sincerity condition on promising. I will argue that although such utterances are always flawed in some way, and can imply that the speaker is communicatively insincere, they need not be straightforwardly communicatively insincere. Consider the following Unsure Promise: I promise to Φ, but I believe that I might not Φ. There’s something problematic about Unsure Promise. First, it might imply that the speaker is communicatively insincere. For in many cases, a speaker believes that she might not act as promised because she lacks any intention to do so. Saying, “I promise to quit smoking, but I don’t believe I’ll succeed” could imply that you haven’t seriously resolved to quit, in which case the utterance would be straightforwardly communicatively insincere. However, this implication can be overridden, for an utterance like Unsure Promise need not involve a lack of a resolution. Suppose Kim is a smoker who has tried and failed to quit in the past. Kim’s roommate is tired of their apartment smelling like smoke, and asks Kim to quit. Kim promises with a caveat, telling her roommate, “I promise to quit, and I fully intend to do so. But quitting smoking is really hard, and ! 141! I’ve failed at it in the past. So I believe that I might fail again this time.” Kim makes a promise, and proclaims that she believes that she might not succeed. Her promise is overly optimistic or quixotic, perhaps problematically so. But her promise does not seem to be communicatively insincere, because she explicitly overrides any implication that her lack of believing she will stems from the lack of an intention to act as promised. Unsure Promise is also independently problematic, apart from the way in which it might imply communicative insincerity. The speaker performs a speech act (promising) that normatively commits her to acting in a certain way (doing as promised.) In general, performing a speech act S that makes you subject to a normative commitment of type C implies that you are volitionally committed to C. For example, if your chess opponent has put your king in check, the rules of the game state that you must attempt to move it out of check. And so someone who says, “My king is in check,” during a chess game implies that she is going to attempt to get out of check. Similarly, being a vegetarian normatively commits you to refraining from eating meat. Telling someone at dinner that you are a vegetarian implies that you are not going to order steak. Saying, “I promise to quit smoking” therefore implies that you are volitionally committed to quitting smoking. To immediately follow this with “but I believe that I might not quit” is to distance yourself from this volitional commitment, and display a lack of confidence that you will actually do what you are normatively committed to doing. In general, to imply that you are normatively committed to C and then publicly doubt that you will fulfill this commitment is to be in a problematic state that resembles hypocrisy. If you say to your chess opponent, “I realize that you’ve put my king in check, but I don’t believe that I’ll move it out of check,” you show that you are not taking the game seriously. If your dining companion says, “I’m a strict vegetarian for ! 142! ethical reasons. . . and I believe that I might have a steak tonight,” you could rightly accuse him of being hypocritical. And if you say, “I promise to quit smoking, but I believe that I might not actually do so,” you display a similar sort of hypocrisy or lack of seriousness. (Note that I am using “hypocrisy” not in the distinctively moralized way in which the word is often used, but simply to refer to someone who publicly proclaims a commitment to some standard, but does not seem to actually be committed to adhering to that standard.) Being hypocritical is not the same as being communicatively insincere. Someone who says, “I’m a vegetarian, but I believe that I might have a steak,” expresses that she is a vegetarian, and that she thinks that she may violate her vegetarian principles. So long as she really is a vegetarian, and really does think it likely that she’ll break her own dietary rules, she is communicatively sincere. Similarly, someone who says, “My king is in check, but I don’t believe that I’ll move it” expresses that her king is situated such that the rules of the game require her to move it, and that she might not follow these rules. So long as this is the case, she is communicatively sincere. And when Kim says, “I promise to quit smoking, but I believe that I might not do so,” she expresses that she has resolved to quit, and that she might not carry out this resolution. If she has indeed resolved to quit, and is doubtful about whether she will succeed, Kim is communicatively sincere. Other speech acts in the SAMS pattern work in the same way. For example, suppose Kim breaks her promise to quit smoking. She apologizes to her roommate with a caveat, saying, “I’m sorry I smoked, and I really do regret it. But I believe that I’ll probably do it again in the future.” Kim is being hypocritical: her apology implies that she is volitionally committed to refraining from making the same mistake again, ! 143! yet she proclaims herself to not be so committed. But she is not insincere: her apology expresses that she regrets her mistake, and she really does regret it. These examples show us that the core problem with Unsure Promise is not insincerity. Rather, the real problem with Unsure Promise is hypocrisy. And the fact that “S, but not M” is hypocritical does not show us that M is a sincerity condition on S. The linguistic sincerity test therefore doesn’t establish that sincerely promising requires believing that you will. 4.3.! Marušić’s assertion argument Marušić argues directly for the necessity of the Belief Account by drawing a connection between promising and asserting: I hold it to be an uncontroversial thesis that belief is a necessary condition for sincere assertion . . . . ’I promise to φ’ and ‘I will φ, I promise’ are two equivalent ways of making the same promise. Yet in saying, ’I will φ, I promise’, one explicitly asserts, ’I will φ.’ Hence, whatever is a necessary condition for sincerely saying, ’I will φ’ is also a necessary condition for sincerely saying, ’I will φ, I promise.‘ 154 This argument seems to presume that any promise can be equivalently expressed by making an assertion, and that the sincerity condition on assertion will therefore also be a sincerity condition on promising. I’m not sure whether all promises can in fact be equivalently expressed by assertions. But even if they can be, this does not establish that believing that you will act is a sincerity condition on promising. For we must ask: why does the fact that you could equivalently promise by means of an assertion entail that promising inherits the sincerity condition on asserting? This claim seems to presume the following general principle: if S can be equivalently expressed by S*, and M is a sincerity condition on S, then M must be a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 154 2012, p. 5. The same argument is also presented in 2013, p. 299. ! 144! sincerity condition on S*, as well. But this is not true as a general rule. Consider the following case. Suppose you are teaching your young cousin how to drive. You could say “You need to stop!” as you approach a red light. Or, you could convey this claim in an equivalent way by asserting, “There’s a red light up ahead.” (Assume that she knows that stopping at red lights is required.) Believing there to be a red light up ahead is necessary for sincerely asserting that there is a red light up ahead. But believing there to be a red light ahead is not necessary for sincerely asserting that your cousin needs to stop the car. Similarly, you can tell your cousin to put on the turn signal by asserting, “You’re about to turn” in the right tone of voice. But believing that your cousin is about to turn does not seem to be necessary for sincerely telling her to use her turn signal. Marušić’s assertion argument nevertheless illustrates an important point. All promises that are in the form of assertions—which will be many of the promises we make—are subject to a belief sincerity condition qua assertion. To sincerely promise by saying “I’ll do it, I swear,” requires believing that you will do it, because this involves asserting that you will do it, and sincere assertion requires belief. But Marušić would need to tell us more to establish that believing that you will is a direct and straightforward sincerity condition on promising, simply because it is a sincerity condition on asserting and we can convey the content of a promise via an assertion. We can conclude from this discussion that believing that you will act is not a necessary condition on sincerely promising. There are cases like that of Peter and Gaius in which you can sincerely promise without believing that you will. The linguistic test does not establish that believing that you will is necessary for sincerity. And appealing to assertion establishes at most that believing that you will is an indirect sincerity condition on promises qua assertions. It follows that the second proposed sincerity condition from the beginning of this chapter fails, as well. ! 145! CHAPTER 5: WHAT THE SINCERITY CONDITION ON PROMISING IS 1. Sincerely Promising and Intending to Act We saw in Chapter 4 that promises express neither the intention to obligate yourself, nor that you know or believe that you will act. So what do promises express? A very natural idea is that promising to Φ communicates that you plan to Φ, and are serious about carrying out this plan. That is, someone who promises to Φ conveys that she really will Φ, even if she doesn’t feel like doing so at the time, or a better option arises, or Φing is more difficult than expected, etc. For example, when I promise you that I will attend your band’s big show tomorrow, I’m telling you that I really am going to be there, even if faced with barriers to action that might otherwise prevent me from going (e.g., I get an invite to a show I prefer, or realize that your band plays death metal, or that the venue is west of the 405 when I live east of the 101, etc.) If my promise is sincere, I really do have such a plan. If I am insincere, I express to you that I am serious about going to your show without actually being so committed. How can we cash out this notion of a serious plan? We cannot appeal merely to desires. For desires aren’t normatively committing in the way that promises are: promising to Φ pro tanto morally obligates you to Φ, but desiring to Φ does no such thing. Moreover, we often sincerely promise to do things that we don’t desire to do. A chastised student caught plagiarizing sincerely promises to turn in original work next time, although he has no desire to write his own essays; you sincerely promise your department chair that you’ll attend the next faculty meeting, even though this is not how you desire to spend your Friday afternoon. ! 146! A more plausible suggestion is that promising expresses an intention to act. Searle claims that “S intends to do A” is a sincerity condition on promising, noting that “the distinction between sincere and insincere promises is that, in the case of sincere promises, the speaker intends to do the act promised; in the case of insincere promises, he does not intend to do the act.” 155 (Recall that the notion of communicative sincerity we’ve been working with is a Searlean notion.) As we saw in the previous chapter, intention seems to pass the linguistic sincerity test: it seems insincere to say, “I promise to go to your show, but I have no intention of going.” Searle takes this a step further, and claims that a promisor “could not say without absurdity, e.g., “I promise to do A but I do not intend to do A.” 156 Furthermore, as Michael Bratman has argued, intentions are unlike desires in that they are a mental state with some inertia or stability. 157 If I intend to go to your band’s show tonight, I have settled the matter of what I am going to do. I should not continue to engage in deliberation about it, or revise my plans for no reason. Similarly, once I promise you that I will go to your band’s show tonight, I have settled the matter of what I am going to do. As Owens emphasizes, the normative control of whether I go is now in your hands, since you hold the power of release. It would be inappropriate to continue deliberating about what to do, or to revise my plans unilaterally. Intentions therefore seem to be a decent candidate for the mental state expressed by promising. However, communicating that you have intended to act isn’t sufficient for sincerely promising. For intentions alone are not stable enough to capture the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 155 1969, p. 60. Searle lays out a long list of conditions that are necessary for sincerely and “non- defectively” promising. His list includes both the intention to act and the intention to obligate yourself. 156 Speech Acts, p. 62. 157 See Bratman, 1987, pp. 18–20 ! 147! seriousness of the plan and the strong sense of commitment involved in promising. For example, suppose that I intend to go to your band’s show tonight because I am a fan of the opening act. You attempt to solicit a promise from me to attend the show. I say, “I promise to be there. And as of now, I intend to go. But you should know that I’m not committed to refraining from revising that intention. My plans might change between now and then, especially if the opening act drops out. ” Such a statement does not seem to be a sincere promise. As with knowing that you will act, mere intentions don’t have the degree of counterfactual robustness that we assume promises have. My promise to come to your show binds me even if the opening act changes, whereas I would likely abandon my intention to go if I discovered that I won’t get to see the scheduled opener. Furthermore, even though intentions have some inertia or stability, they are not stable enough to ground promises. For it is not at all problematic for me to intend to go to your band’s show tonight and then abandon my intention because I feel like staying home and watching Netflix instead, or because a slightly more appealing offer comes along. But such circumstances would not license my breaking a promise to you to see your band play. Grounding the obligation to keep promises in mere intentions would overgenerate cases of permissible promise-breaking. 2. Sincerely Promising and Resolving to Act We have seen that intentions are on the right track to be the mental state expressed by promises, but are not robust enough. Is there some more robust intention- like state that we could appeal to instead? Consider the special sorts of intentions that we often form at the start of a New Year—what we typically call resolutions. These are serious and stable intentions that we’re strongly committed to, intentions to accomplish ! 148! important goals that we expect might be difficult to attain. I take resolutions to be a special kind of intention that is especially robust or resistant to revision. (I say much more about why we need resolutions and what I take them to be in Chapter 6; by way of preview, I argue that a resolution is an intention to act paired with a desire not to abandon that intention.) Resolutions are necessary when you plan to act and care about whether you do so, but suspect that some temptation or other barrier to action (such as laziness, aversion to an unpleasant task, apathy, etc.) might cause you to abandon this plan were you not to bolster it somehow. Resolutions are one means by which you can bolster your plan, and more effectively resist temptation. 158 It seems plausible that when I promise to Φ, I express that I have resolved to Φ— that I plan to Φ, and care enough about whether I do so that I will not reconsider or abandon that plan, even in the fact of temptation, laziness, better offers, and the like. This brings us to the following account: Resolution Account: Promises express resolutions; sincerely promising to Φ requires resolving to Φ. In promising to attend your show, I express that I have resolved to do so. If I am sincere, I really have such a resolution. In general, it is problematic to abandon a resolution on a whim, or because a slightly better option presents itself. So the Resolution Account won’t overgenerate the cases of permissible promise-breaking in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 158 Of course, there are other means by which you can resist temptation, which can be more effective than resolution-making. The best way to refrain from the temptation to drink tonight might be not form a resolution, but to lock the liquor cabinet and give the key to a reliable friend. I don’t mean to claim that resolutions are the only effective means of resisting temptation. I will discuss this further in Chapter 6. ! 149! the way that intention-based account would. 159 Furthermore, resolutions have the counterfactual stability that the knowledge and belief accounts lack. Suppose I am taking my friend out to dinner tonight, at a restaurant of her choosing. I resolve to have a salad for dinner, because I am vitamin deprived and would benefit from consuming some leafy greens. My resolution commits me to having a salad tonight if doing so turns out to be inevitable (because my friend chooses a raw vegan place that only serves salads), or if it turns out to be independently likely (because she chooses a restaurant that is renowned for its salads and panned for everything else), or if it turns out to be independently unlikely (because she chooses a steakhouse with rather uninspired salad entrees.) Resolutions therefore seem to be a more viable candidate sincerity condition than the other views we have considered so far. However, the Resolution Account as it stands is incomplete in an important way, which I turn to next. 3. Promisee Acceptance The Resolution Account describes an activity that can be performed unilaterally. But promises require a second party, and the acceptance or uptake of the promisee is essential to making a promise. 160 The sincerity condition on promises must accommodate this. In the rest of this chapter, I first outline why promisee acceptance is important, and what it takes to accept a promise; the claims I make here will be !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 159 In Chapters 7 and 8, I analyze what sorts of cases constitute permissible promise-breaking, and explain how my Conditional Resolution Account generates all and only the right cases. 160 In claiming that promises must be accepted by a second party, I am not ruling out by fiat the possibility of promises to oneself; you could qua promisor make a promise that is accepted by yourself qua promisee. I argue elsewhere (Liberman ms) that so-called “promises to the self” are not genuine promises. ! 150! important for my discussion of the “escape conditions” on promises and resolutions in Chapters 7 and 8. I then articulate a Conditional Resolution Account, according to which sincerely promising requires resolving to Φ, conditional on the continued acceptance of the promise by the promisee. Finally, I respond to two objections. 3.1. Why promisee acceptance is important Hanoch Sheinman notes that: When A promises B to do X, A does not simply commit to do X; A commits to B to do X. Of all the people in the world, B is the one agent to whom A’s commitment is made. . . . While a promise is first and foremost the act of the promisor, it is not the kind of act the promisor can perform all by herself. The act has to be performed as part of an activity in which the promisor and promisee engage together, as a group. A promise in other words is an instance of shared activity or agency. As such, it must involve the agency of the promisee. 161 I take it that the activity the promisee must contribute is accepting the promise. Prima facie, it is highly intuitive that promises require acceptance. For example, Darcy telling Elizabeth that he promises marry her does not generate any special moral obligation for him to do so unless Elizabeth accepts his offer. And Darcy incurs no reparatory duties if he fails to act on a promissory offer that Elizabeth has not accepted. If acceptance were not required, agents could generate directed, voluntary moral obligations without the consent of the person to whom the obligation is directed. Elizabeth has an interest in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 161 2008, p. 292. Although Sheinman argues that “perfect” promises require an agential contribution from the promisee, he denies that they require acceptance. Sheinman distinguishes acceptance—which is “an intentional communicative act (which in special cases includes omission, but always more than a belief)” (293n)—from uptake, which involves having the “belief that the promisor has the communicated intention (the intention she purports to have)” (302). I use “acceptance” and “uptake” interchangeably to refer to the more active notion that Sheinman calls acceptance. The substantive disagreement between us is about whether promises require at least non-rejection of the offer (I argue that they do while Sheinman claims that they do not; see footnote 166, below.) Moreover, Sheinman seems to imply that uptake requires believing the promisor to be sincere. I argue in Section 3.3.1 that you can accept a promise even if you believe the promisor is insincere (e.g., for strategic reasons, to be kind to an emotionally fragile promisor, or because you wish the promisor to be bound to you to perform.) ! 151! being able to prevent Darcy from being even pro tanto morally obligated to marry her should she so wish, and she can do so by refusing to accept his promissory offer. Were no acceptance required, Elizabeth might be owed all sorts of things from all sorts of people that she wishes not to be owed or to receive. Promisee acceptance is a communicative rather than psychological notion. There is a sense of acceptance that is an entirely internal, psychological matter, which can come apart from the agent’s external expression of that acceptance (or lack thereof.) If you make a promissory offer to come to my dinner party, I might communicate my acceptance by saying, “Great, see you there!” while secretly thinking “Oh no! If you come we’ll be an unlucky thirteen guests at dinner . . . I sure hope you don’t show up!” This counts as accepting the promise. Similarly, if I say, “Please don’t come; your estranged twin will be there and it would cause a scene for the two of you to be at the same table,” while secretly hoping you will show up anyway and make an amusing mess of things, I count as rejecting your offer. This is because promisee acceptance, like promising itself, depends on what is communicated, and not on what the inner mental state of the communicator happens to be. For it would not be very fair to the promisor, or a very useful social convention, if we determined whether a promise had successfully been made on the basis of the promisee’s private thoughts. Rather, we must make this determination on the basis of the promisor’s reasonable beliefs about whether the promisee has accepted. 162 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 162 For a similar line of thought, see Erin Taylor (2014, p. 671), who writes: “the actual internal mental states of participants (if these are different than their outward indicators) are inaccessible to typical participants in a convention. Conventionally-mediated requirements, including promises, will therefore not depend on them. The claim that internal mental states are inaccessible in as far as they are different from their outward indicators (such as behaviour) expresses the commonsense view that in many cases our access to the details of others’ motives, intentions, and other internal mental states typically proceeds straightforwardly through such ! 152! W.R. Carter offers an example that nicely illustrates why we need a communicative rather than a psychological account on acceptance. Suppose there are two rival families, and the patriarch of one has been wounded by a member of the other in the most recent flare-up of violence. Having preached the importance of revenge his whole life, his sons reasonably believe that their father would want them to avenge his injury by inflicting an injury of the same sort on a member of the rival family. However, being mortally wounded has given the patriarch a sudden change of heart, and he wants his sons to forgive the rival. But his wound prevents him from speaking or otherwise communicating with his sons. The sons say to him, “We promise to avenge you!” The sons take themselves to have made a promise. This promise is valid, because the sons reasonably believe that their father wants the promise to be kept, even though they are in fact mistaken. 163 3.2. Lack of promisee release as continued acceptance After someone has accepted a promise, she retains the moral power to change her mind and release the promisor. When a promisee releases a promisor, she makes it the case that the promisor is no longer bound to her in even a pro tanto way. For example, even if Elizabeth accepts Darcy’s promise of marriage, she can change her mind at a later date and release him. In doing so, she stops accepting the promise, and !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! outward indicators. Where those indicators are absent or the access they provide is unreliable, then conventions will not generate requirements that depend either on participants’ actually having certain mental states or on participants’ knowing the details of the actual mental states of others.” 163 See Carter, 1973. A potential complication with this example is that in promising to take revenge, the sons have promised to perform an immoral action, which is an independent escape condition—meaning that if they form a valid promise, it is one they ought not keep, all-things- considered. We can ignore this complication for present purposes, as the focus is on whether a promise was made at all and not on whether it ought to be kept, all-things-considered. ! 153! Darcy is no longer obligated to marry her or to perform reparatory duties is he fails to do so. It might be rude or cruel of Elizabeth to back out of their engagement if she does not do so in a kind and timely manner (e.g., if she releases him from his promise at the altar in front of their assembled families.) But barring such independent considerations, Elizabeth remains free to invalidate Darcy’s promise at any time and for any reason. Darcy does not have the same power: the promisor may request release, but cannot unilaterally release himself from the promise. 164 This shows us that promisee acceptance is not a one-and-done deal. Rather, acceptance must be continued, in the sense that the promisee does not subsequently release the promisor. This doesn’t mean that the promisee must perform some kind of active acceptance of the promise for its entire duration. Rather, what I mean when I refer to continued acceptance is initial acceptance combined with failure to release; I take it that releasing the promisor is equivalent to ceasing to accept the promise. 3.3. Validity conditions on accepting a promise In Chapter 1, I outlined some basic validity conditions on promising, or conditions that are necessary for a promise to “fire” and to generate a pro tanto obligation. Promisee acceptance is subject to validity conditions, as well. The following conditions must be met if there is to be legitimate acceptance of a promise. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 164 Usually, engagements to be married involve exchanges of promises: Darcy promises to marry Elizabeth, and Elizabeth promises to marry Darcy in return. For the purposes of this discussion, however, I am focusing only on Darcy’s promise to marry Elizabeth by itself, and not on his promise as part of a reciprocal exchange of promises. ! 154! 3.3.1. Understanding the promise Suppose someone comes up to you on the street and says a string of words in a language you do not understand. Caught off guard, you say, “Um . . . okay.” It turns out that this person was saying, “I promise to follow you around all day.” Have you accepted this person’s promise? Certainly not. A promise cannot be accepted if the promisee does not understand what is going on—for example, if the promisee does not speak the promisor’s language, or is a stranger from a society that lacks any semblance of a promising convention. 165 To be able to appropriately respond to any communicative speech act the responder must understand both what the speaker is saying, and any normative implications that follow from so saying. Moreover, accepting a promise that requires action from both parties (such as a promise to meet for lunch) obligates the promisee to do her part to facilitate the keeping of the promise. It would not be fair to hold people to obligations that they didn’t realize they were getting themselves into. What exactly must the promisee understand in order to accept the promise? In typical cases, the promisee will believe that the promisor is going to Φ in accordance with her promise. This is because in most cases, the promisee can reasonably assume that the promisor is sincere, and really does plan to Φ. However, there are deviant cases in which the promisee accepts the promise even though she believes that the promisor is insincere and not in fact planning on Φing. Perhaps the promisee and promisor are engaged in a calculated business negotiation, and the promisee does not wish to reveal that she has discovered the promisor’s insincerity. Or perhaps the promisor is a substance abuser who is trying to rebuild her trustworthiness as part of a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 165 See Korn and Korn (1983) for an empirical argument that the indigenous people on the small island nation of Tonga lack any social convention of (or closely resembling) promising. ! 155! 12-step process. You have been burned by this promisor many times before, and don’t believe that she will succeed in keeping her promise. But you graciously accept the promise anyway, to give her an (undeserved, perhaps) opportunity to rebuild trust with you. At a minimum, then, the promisee must believe that the promisor has expressed that she plans to Φ. I have been arguing that promises express resolutions. But the promisee need not specifically believe that the promisor has resolved to Φ in order to validly accept a promise. It is implausible to claim that the promisee must have in mind the particular concept of a resolution, much less the more sophisticated concept of a conditional resolution. So it suffices for acceptance that the promisee believe that the promisor has expressed a serious plan to Φ, however roughly that notion of a plan is cashed out. Even small children recognize that someone who makes a promise to do something isn’t supposed to haphazardly change her mind at a later point. This sort of intuitive grasp of what is involved in promising suffices to understand the promise. 3.3.2. Basic moral responsibility We saw in Chapter 1 that promisors must be morally responsible in order for a promise to bind. The promisee must be a morally responsible agent in order to validly accept a promise, as well. Accepting a promise is akin to consenting or agreeing to the promisor’s being morally bound to you with regard to a particular future action. Such acceptance requires basic moral agency. For example, I might “promise” my stir-crazy dog that I will take him for a long walk next time I leave the house. But since the dog is not a moral agent, I do not generate a directed pro tanto obligation towards him. ! 156! Parents sometimes make “promises” to their infant children, and these might seem like valid promises. However, such declarations bind not in the way of promises but rather in the way of vows or firm resolutions. For example, suppose you regret the fact that you never visited Disneyland as a child. So you “promise” your newborn son that you will take him to Disneyland before he turns 10 years old. This “promise” expresses your resolution or commitment to taking your son to Disneyland. All else being equal, you should act on this commitment, and it speaks poorly of you as a coherent agent if you do not. Perhaps changing your mind about such commitments without good reason also speaks poorly of your character. But you do not wrong your son if you fail to take him to Disneyland; since your son never accepted your “promise,” no directed obligation was generated. Compare this to a case in which you promise your 9-year-old son that you will take him to Disneyland before his tenth birthday, and he gleefully accepts. All else being equal, you clearly wrong him if you do not take him to Disneyland within the year, for you violate a promissory obligation. We’ve seen that in order to validly accept a promise, promisees must be competent moral agents who understand what they are doing. In the next few sections, I will argue that what accepting a promise consists in varies depending on the circumstances: sometimes, lack of explicit rejection suffices for acceptance, while other times, explicit acceptance is needed. 3.4.! Lack of explicit rejection Suppose two strangers are sitting near each other on an airplane. Alfred is sandwiched in the middle seat between two cranky children, while Bonnie has an aisle seat and a row to herself. Alfred is jealous of this seat placement, and begins scowling at Bonnie and muttering complaints under his breath. Bonnie notices his distress, and ! 157! says, “I see that you’re upset. So I promise to help you out by letting you borrow my noise-cancelling headphones after we reach cruising altitude.” Alfred angrily replies that he is indeed upset, and wants nothing to do with her promise. If Alfred later expresses a desire to take Bonnie up on her promissory offer, she does not seem to be obligated to give Alfred the headphones in the way that she would be had Alfred not rejected the promise. This example shows us that explicit rejection invalidates a promissory offer. The example is deliberately set up with two strangers who owe no special obligations to each other. There might be cases where someone is morally obligated for independent reasons to act as they have promised, even if the promisee rejects the promise. But considering cases like those will not lead to reliable intuitions. Suppose Alfred and Bonnie are not strangers on a plane but siblings who have recently had a falling out. Alfred is about to get married, and Bonnie says, “I promise to attend your wedding.” Alfred angrily replies that he is upset with Bonnie, and wants nothing to do with her promise. Is Bonnie obligated to attend the wedding? If it seems so, this is likely because we are assuming that siblings are independently obligated to attend each other’s major life events in virtue of their special family relationship, and not because Bonnie has successfully made a valid promise. Promises are voluntary obligations, which entails that a promise must be freely chosen, by both promisor and promisee. We saw in Chapter 1 that coercion undermines the validity of a promise, because it makes the promise non-voluntary on the part of the promisor. If the promisee explicitly rejected a promise yet it remained in force anyway, this would be akin to coercing the promisee’s role in the promise, and would problematically undermine the promissory transaction’s voluntariness. ! 158! 3.5.! Implicit and explicit acceptance Is understanding plus lack of rejection all that is needed to count as accepting a promise? It depends on the circumstances. If the promisee explicitly agrees to the promise, it is obvious that she accepts it. If I e-mail you saying, “I promise to donate $100 in your name to political cause X” and you reply, “Thanks, that’s super,” you have clearly accepted my promise. But what about cases without such explicit acceptance? Suppose I know that you are a very strong supporter of X; you’re always sending around petitions in support of X and you frequently solicit donations for X from your friends. Your birthday is coming up and I’d like to do something nice for you. I send you an email saying, “I promise to donate $100 to X in your name.” I know that you have a habit of reading but not replying to emails, and you don’t reply to this one. But it is reasonable for me to believe that you have read my message, that you want me to give money to X in your name, and that you would have replied and told me otherwise had you not been on board with my offer. Have I successfully made a promise? It seems that I have, and that if I failed to make the donation you could rightfully object. 166 In this case, your silence together with a lack of explicit rejection counts as an implicit acceptance of my promise. Actively soliciting a promise signals implicit acceptance in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 166 Sheinman discusses a case with the same structure as part of his argument that promises do not require active acceptance: “A and B are friends. They believe that A’s doing X is in B’s interest. A knows that B does not know whether A will do X, and would appreciate if A promised him to do X. Wanting to help B, A emails B: “Promise to do X.” A and B know that B normally reads but does not reply to email messages, and that A does not read much into B’s failure to reply. As expected, B fails to reply. A fails to do X (303).” Sheinman claims that “in my intuition, A makes a perfectly good promise to B to do X, then breaks it. And yet . . . B performs no intentional communicative act (or omission) before or after reading A’s message. As both parties understand, A’s promise is complete as soon as B reads A’s message” (303). In this case, A’s email counts as a promise because A reasonably believes that B wants her to promise to do X, and will read the message. As we shall see in the next section, implicit agreement is permissible only when we can make such assumptions. ! 159! the same way. If I say something like “So can I count on you to be at my party?” or “Promise me you’ll come!,” I pragmatically implicate my acceptance of your promise to come to the party. For were I not on board with the promise, I would not have solicited it in the first place. If you are worried that it strains the relevant concepts to consider non-rejection a form of acceptance in some cases, keep in mind that there are analogues for lack of explicit rejection counting as implicit acceptance in other forms of cooperative enterprise. Suppose we are in a group of people trying to solve the problem of what to have for dinner. You say, “Pa Ord has amazing duck noodle soup.” I comprehend your statement, and don’t reject it by saying something like, “No, their noodles are terrible!” For the purposes of our conversation, I have accepted the proposition that Pa Ord has great duck noodle soup. Since something is on the line—what we eat for dinner—I would have rejected your claim if I didn’t like the noodles. My silence therefore implies that I agree with you. Requests work in a similar way, at least in some contexts. Suppose I am preparing for a party, and delegating tasks to everyone. I say, “Abed, would you put up the decorations? Britta, please go buy the wine. Chang, you can bake the cookies.” If my fellow party preparers understand what I am saying, and do not explicitly reject my requests—by asking for a different task, or telling me that they cannot perform the task I’ve assigned them—they have implicitly accepted the tasks I have given them. And I can rightly be upset with them if they do not perform their tasks as I requested. There are other kinds of cases in which lack of explicit rejection does not entail implicit acceptance. Suppose Chaim is considering getting a tattoo of his friend ! 160! Donovan’s face on his arm. He has no idea whether Donovan would appreciate this gesture. Since having your face tattooed on your friend’s arm is not something most people desire, it would not be reasonable for Chaim to assume that this is something Donovan wants or would agree to. Chaim emails Donovan: “I promise to get a tattoo of your face on my forearm tomorrow.” Donovan does not write back, although Chaim is confident that Donovan has read the email and had an opportunity to respond. Is Chaim promissorily obligated to Donovan to get a tattoo of his face on his arm? It seems clear that he is not, and that if Chaim didn’t get the tattoo, he would not be violating an obligation he owed Donovan. Furthermore, if Chaim did get the tattoo and Donovan objected to it, Chaim could not defend himself by saying he was only keeping his promise—for Donovan could reasonably reply, “But I didn’t want you to make me that promise in the first place!” Nor does lack of explicit rejection always entail implicit acceptance when the promise requires action by the promisee. Suppose I text you “I promise to buy you lunch at Azla Ethiopian at noon tomorrow before the colloquium.” I know that you love Ethiopian food, and have been wanting to try Azla for a long time. But you do not write back to me. Promises like this require action from not only the promisor (who must act as promised), but also from the promisee (who must either act so as to enable the promise to be kept, or release the promisor from the obligation.) I cannot keep my promise to meet you unless you show up at the restaurant at noon, as well. Since my promise is going to obligate you to go to Azla at noon, I need to be absolutely certain that you are on board with the plan, lest I unfairly burden you. Explicit acceptance would make me certain of this; lack of explicit rejection usually will not. In typical cases involving joint action, then, the promisor cannot safely assume from the promisee’s lack ! 161! of explicit rejection that she has accepted the promise. We can imagine cases in which the promisor can be absolutely certain that lack of explicit rejection constitutes implicit acceptance, even when the promise requires action from the promisee. For example, suppose we have a long-standing habit of going to lunch at noon at the Mercado La Paloma food court before every colloquium, so I can reasonably expect you to be there at noon no matter what. And I know that you’ve been wanting to try Azla Ethiopian for a long time, so I can reasonably expect you to agree to dine there. I also know that you aren’t averse to other people buying lunch for you, so I can predict that you will not decline my offer to treat you. In such a case, lack of explicit rejection would suffice for implicit acceptance, so long as there was an opportunity to reject that you did not take advantage of. What is the principled difference between cases in which explicit acceptance is necessary and those in which it is not? If you have solicited a promise from me, or if I am promising to do something that I am completely confident that you want me to do for you, your silence counts as acceptance of my promise, so long as you had an opportunity to reject the promise if you wanted to. But if I am promising to do something that I’m uncertain whether you want me to do, or that requires you to undertake an action you would not otherwise take, your silence cannot count as acceptance. What determines whether explicit acceptance is needed is whether the promisor can reasonably assume that the promisee wants him to be bound to her to act as promised. If it is reasonable to make this assumption, lack of explicit rejection suffices. If not, explicit acceptance is needed. Why is this so? ! 162! 3.6. The implication of interest Promises are generally assumed to be in the promisee’s interest in some way, broadly construed. This is why lack of explicit rejection suffices for acceptance when we can presume that the promisee is interested in at least the promise being made, and typically in the promise being kept. When we cannot assume that the promisee so interested, we don’t know whether this “promisee preference” condition is met. Promises that do not meet this condition seem to be problematic. And so we can make valid promises only when we can be assured that they meet this condition—either because we can reasonably assume they do so, or because the promisee’s explicit acceptance assures us of this when we cannot otherwise so assume. The idea that the promisee must want the promise to be made and the promised action to occur is very natural and has been theoretically influential. John Searle articulates this as an essential felicity condition on the speech act of promising, arguing that it must be the case that “H [the promisee] would prefer S’s [the promisor’s] doing A to his not doing A, and S believes H would prefer his doing A to his not doing A.” 167 As an empirical matter, Searle’s constraint is surely true in most cases. 168 Promises are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 167 1969, p. 58 168 An empirical study by of linguistic intuitions about what counts as a promise confirms that ordinary speakers tend to assume that Searle’s felicity condition is true (Gibbs and Delaney, 1987). The explicit aim of the study was to test Searle’s felicity conditions on promising. Subjects were asked to rank whether various utterances seemed to them to be a promise. Subjects gave significantly lower scores to scenarios in which the hearer did not want the future action to occur compared to scenarios in which this was not the case; the authors interpret this to mean that ordinary speakers implicitly assume that promises must be about something the promisee wants to occur. In another experiment, a different group of subjects read a promise- like scenario, and were asked whether the speaker believed the hearer wanted the action performed, among other questions. The mean ratings for this “Hearer Preference” question were very high (6.15 out of 7, where 7 is an affirmative answer). Gibbs and Delaney state that ! 163! typically cooperative enterprises. Why would someone make a promissory offer to Φ, if not because she thought the promisee wanted her to Φ? And why would a promisee implicitly or explicitly accept the promise unless she wanted the promisee to Φ? But as a conceptual matter, Searle’s condition is too strong. David Armstrong articulates a more modest and plausible version of Searle’s constraint, according to which a promise “involves the speaker in believing that his audience would prefer the fulfilment of the promise to its non-fulfilment,” regardless of whether the speaker actually would so prefer. 169 Armstrong’s version mirrors the distinction between the psychological and communicative notions of acceptance; what is at issue is not what the promisee really prefers, but what it is reasonable to believe that the promisee prefers. One more adjustment is needed to the Searle/Armstrong condition. For there are cases in which the promisee is interested in the promise in a broader sense than preferring the fulfillment of the promise to its non-fulfillment. Vera Peetz notes that a promisee might refrain from rejecting a promise for reasons other than wanting the promised action to occur. For example, suppose my neighbor is making homemade jam, and I express mild interest in the process. My neighbor eagerly makes a promissory offer to give me a jar of the jam. I don’t like jam, but I accept the promise because I don’t want to offend my neighbor. Peetz concludes that Now although it might seem that, in the end, what I do not reject and what I want are the same thing, this is not so. What I am promised is damson jam, which I do not want; what I do want is not to hurt my neighbour's feelings, but this is not what I am promised, although it is the reason why I do not reject her promise (579). !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! “these data suggest that subjects viewed the Hearer Preference condition as being extremely important in making a promise” (118). 169 1971, p. 446 ! 164! In this case, even though I am not interested in the promise being kept, I am interested in the promise being made—for accepting the promissory offer serves my interests by preventing me from hurting my neighbor’s feelings, which is something I want to do. 170 This shows us that we should broadly construe the way in which acceptance of a promise implicates that the promisee is interested in the promise. Usually, the promisee wants the promised action to occur. But in special cases, she might just want a valid promise to be created. 4. The Conditional Resolution Account The fact that promises must be accepted shows us that the Resolution Account should be modified as follows: Conditional Resolution Account (CRA): Promises express resolutions conditional on the continued acceptance of the promisee; sincerely promising to Φ requires resolving to Φ, conditional on the promisee’s continued acceptance. Making the Resolution Account conditional brings the promisee and her interests directly into the content of the resolution. When I make a sincere promissory offer to attend your band’s show, I resolve to go so long as you accept this offer and do not release me. A valid promise is successfully created, and a pro tanto obligation generated, only if you accept my offer. How should we understand the conditionality involved in the Conditional Resolution Account? It might seem that conditional resolutions are resolutions about !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 170 A promisee might also want a promise to be made but not kept for more nefarious reasons. Consider Owens's (2008) case of A feuding with his neighbor B about the washing of B’s filthy car. A has repeatedly asked B to wash the car, to no avail. A does not expect that B will wash the car, even if he promises to do so. But A is feeling mean-spirited, and wants to see B fail morally. So A asks B to promise him that he will wash the car; B does so, and A accepts this promise. A wants the promise to remain in force, because he wants to see B fail morally. But A does not actually want the promised action to occur, because if B did wash the car, A would not have the perverse satisfaction of seeing B commit a moral wrong. ! 165! conditionals, such as: resolving that (if you continually accept, I will go to the show.) The content of the resolution on such an account is a conditional claim: Res(P!Q). Call this the Wide-Scope Account. Because Res(P!Q) is logically equivalent to Res(~P v Q), the conditional will be satisfied if either the promisor acts as promised (i.e., Q), or the promisee fails to accept the promise (i.e., ~P). 171 A sincere promisor must resolve that this condition is satisfied—meaning she must resolve to either act as promised, or for it to be the case that the promisor does not accept. The Wide-Scope Account is problematic, for it doesn’t make the resolution genuinely conditional on the promisee’s acceptance in a robust enough way. For example, suppose Enid hates driving in the city, and prefers commuting by bus. Her environmentalist friend Fiona asks her to promise that she will take the bus to work every day instead of her car. Enid promises, and resolves to take the bus. She satisfies the conditional posited by the Wide-Scope Account because she satisfies one of the disjuncts: she does intend to take the bus. Suppose Enid resolves to take the bus because and only because she dislikes city driving. She doesn’t care about whether Fiona accepts the promise, or consider Fiona’s acceptance of the promise in any way when forming her resolution. And if Enid were to discover that she in fact loved city driving, she would abandon her resolution to take the bus. Enid’s promise does not seem to be properly conditional on Fiona’s acceptance; she made it for entirely independent reasons having nothing to do with acceptance, and it does not have counterfactual robustness. But the Wide-Scope Account entails that Enid has formed the conditional resolution necessary for sincerely !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 171 These formulations are equivalent on the assumption that we are working with the material conditional. ! 166! promising. The Wide-Scope Account is therefore not a good candidate for capturing the kind of conditional resolution posited by the CRA. It might instead seem that conditional resolutions are resolutions that occur once a certain condition has been met, such as: if you continually accept, I resolve that (I will go to the show.) The content of such a resolution is unconditional, and the resolution gets triggered by acceptance: P!Res(Q). Call this the Narrow-Scope Account. This account is also implausible, for it over-generates cases of sincere promising. For example, suppose Garvey says to Harvey, “I promise to send you comments on your paper by tomorrow, so long as that would be helpful for you.” When Garvey makes this offer, he is insincere, and has no intention of sending Harvey comments. Harvey then accepts the offer with massive gratitude and relief in his eyes. As he thinks about Harvey’s reaction later, Garvey feels overcome with guilt, and resolves to send Harvey comments. 172 Intuitively, it seems that Garvey started out by making an insincere promise, and then later changed his mind and formed the relevant resolution at some point after the promise was accepted, thereby retroactively making his promise sincere. But Garvey has the conditional resolution posited by the Narrow-Scope Account: Harvey accepted the promise, which was the triggering condition, and Garvey then formed the resolution. The Narrow-Scope Account problematically entails that Garvey’s promise is sincere all along; the account has no room for claiming that it was insincere at the beginning, and then later becomes sincere. 173 The Narrow-Scope Account is therefore !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 172 Thanks to Mark Schroeder for suggesting an example like this, and pressing me to clarify the way in which conditional resolutions are conditional. 173 If the gap in time between Harvey’s acceptance and Garvey’s formation of the resolution bothers you, think instead about a case in which Garvey is immediately stricken by guilt over Harvey’s emotional reaction as he accepts the promise, and forms the resolution to send ! 167! not a plausible interpretation of the CRA as a sincerity condition. It is a general feature of conditional attitudes—such as conditional beliefs or credences, conditional preferences or desires, and conditional intentions or resolutions—that they run into problems if they are understood as either attitudes about conditionals (as the Wide-Scope Account posits), or as unconditional attitudes that are triggered by other claims (as the Narrow-Scope Account posits.) 174 As Richard Bradley observes about conditional desires and beliefs, An agent’s conditional beliefs and desires should not be identified with what she is currently disposed to believe or desire in the event that the condition were realized. . . . A conditional belief or desire that X given Y is not a counterfactual attitude to X (an attitude that one would take), but an attitude that one presently takes to the truth of X on the assumption or hypothesis that Y. 175 The same sort of idea applies to conditional resolutions: a conditional resolution to Φ given promisee acceptance is not a counterfactual attitude to Φing, but an attitude that a sincere promisor presently takes to Φing, on the condition that the promisee accept. We can formulate this as: Conditionally Resolve (If you continually accept, then I will go to the show), or Cond. Res (P!Q). The resolution does not take wide or narrow scope over a conditional. Rather, the conditionality is built right into the resolution. This is meant to be a formalization of the intuitive way in which we understand conditional intentions and resolutions (such as conditionally intending to jog tomorrow if it is not raining, or conditionally resolving to focus exclusively on publishing if you are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! comments immediately after the promise is accepted. Such a case still seems to be initially insincere, but the Narrow-Scope Account cannot accommodate this. 174 For an explanation of why these options aren’t a plausible account of conditional credences, see Lennertz (2015), sections 2.2. and 2.4. Lennertz offers counterexamples to these accounts of conditional credences as part of an overall defense of quantificational credences as a distinct and non-reducible phenomenon. 175 1999, p. 24 . . . 25. ! 168! granted a sabbatical.) It is beyond my present scope to offer a more detailed account of exactly how we should cash out this notion of a conditional resolution. However, this should not matter for my present purposes. So long as there is some plausible story about how resolutions can themselves be conditional—likely modeled on similar views about how beliefs, credences, desires, preferences, and intentions can themselves be conditional—we will have what we need for the CRA. I leave it to others to figure out what the best story is. 4.1. CRA and the linguistic sincerity test Does the Conditional Resolution Account pass the linguistic sincerity test? Recall that the test asks us to imagine an utterance of “S, but not M,” and ask whether the speaker is straightforwardly communicatively insincere. If so, then M passes the test and counts as a sincerity condition on S. Consider the following utterances: 1: I promise to Φ, but I don’t have a stable plan to do so. 2: I promise to Φ, and right now I intend to do so, but I may change my mind. 3: I promise to Φ, and I am going to do so even if the promisee does not agree to this plan. Each of these utterances seems to be communicatively insincere, which means that they all pass the linguistic sincerity test. Someone who promises to act without planning to do so, or who forms only a tenuous or fragile plan, does not seem to be promising sincerely. Nor does someone who promises to act while recognizing that she may decide not to follow through. And someone who forms a stable plan to act that in no way depends on what the promisee wants or agrees to may be forming a resolution, but does not seem to be making a sincere promise. To be sincere, a promisor must have formed an intention that the promisee is on board with, and that is not going to be ! 169! abandoned unless there is a good excuse. The basic idea of a stable and agreed upon intention is intuitive, and is something that competent promisors could employ without possessing a theoretical articulation of the concept of resolutions. One might worry that the theoretical articulation of this idea does not pass the linguistic sincerity test. Consider: 4: I promise to Φ, but I haven’t resolved to do so conditional on your acceptance. If we have a good grasp of what is meant by this utterance, it seems insincere in the same way as utterances 1 – 3 above—for resolving conditional on promisee acceptance is a more precise way of describing the process of forming an intention that the promisee is on board with and that is not going to be abandoned without a good excuse. Granted, someone who lacked a theoretical understanding of what conditional resolutions are wouldn’t be able to assess the sincerity of the promise. But this is not a problem, and does not entail that the utterance fails the linguistic sincerity test. For what matters for the test is whether someone would be inclined to think the utterance was insincere if they were to fully understand what it meant. We can highlight the plausibility of this claim with an analogy. Consider an utterance of “I am certain that p, but I don’t know that p.” This seems to be insincere, which shows us knowledge is a sincerity condition on expressions of certainty. What about an utterance of “I am certain that p, but I do not have a justified true belief that p that is not subject to a Gettier condition” (or, “I do not have a sufficiently high credence in the truth of p,” or whatever your preferred theoretical articulation of the concept of knowledge is?) If we have a good grasp of what is meant by this utterance, it likewise seems insincere. Of course, someone who lacked a sophisticated theoretical articulation ! 170! of the concept of knowledge would likely be unable to assess the sincerity of the utterance. But this doesn’t entail that it fails the linguistic sincerity test, for such a person would count the utterance as insincere were she to understand what it meant. The same is true of the linguistic sincerity test applied to the CRA. 5. Objection: Differentiating Between Promises and Resolutions The Conditional Resolution Account states that promising involves the public expression of a resolution. One might worry that a resolution-based account of promissory obligation cannot adequately distinguish between those resolutions that count as promises and those that do not. For you can form resolutions that depend on other people’s interests without promising; resolving to mow my elderly neighbor’s lawn does not entail promising her that I will mow the lawn. And you can publicly state that you have a resolution without thereby making a promise; to announce to you that I have resolved to quit drinking for a month is not to promise you that I will do so. The CRA has an easy response to such cases. It posits that promissory resolutions must be both publicly expressed and conditional on the acceptance of the promisee. The resolution to mow my neighbor’s law is not publicly expressed, while the resolution to abstain from alcohol is not conditional on someone else’s acceptance. Therefore, these resolutions are obviously not promises. However, this doesn’t completely resolve the worry. For we can imagine publicly proclaimed resolutions that are explicitly conditional on someone else’s acceptance or agreement. Suppose I tell my personal trainer that I resolve to lift weights with her five times a week, but only on the condition that she agree to work with me and continue so agreeing. She agrees. I have ! 171! announced that I have resolved to train five days a week, conditional on her acceptance of this resolution, and she has accepted. But I don’t seem to have made her a promise. There are a couple of principled ways to distinguish genuine promises from announcements that one has a conditional resolution. If at least one of them succeeds, the personal trainer case above will be unproblematic. First, it seems plausible that a necessary precondition for validly promising is recognizing that you are making a promise. In general, successfully engaging in an intentional action that alters the normative situation seems to require realizing that you are doing so: you must recognize that you are granting permission in order to grant consent; you must be aware of the fact that you are transferring property in order to make a gift, etc. When I promise my trainer that I will work out five times a week, I recognize that in doing so, I am making a promise. When I merely report my conditional resolution, I don’t have any such recognition. Perhaps I need to take myself to be promising in order to successfully promise. If don’t take myself to be promising when I make non-promissory conditional resolutions, I’m not promising. Similarly, perhaps the promisee must be aware of the fact that she is accepting a promise, as well. Suppose you tell your trainer, “I’ve resolved to lift weights five times a week, so long as you think this is a good idea,” and she says, “Okay, sure.” If the trainer does not take herself to be accepting a promise and thereby generating a valid promissory obligation, then what she is doing is agreeing with a resolution, and not accepting a promise. Second, recall that the Searlean notion of expression we are working with—the notion of expression that grounds the SAMS pattern from Chapter 3—states that a speech act expresses a mental state iff it is impossible to sincerely perform the speech act without being in that mental state. To promise your trainer that you will lift weights ! 172! five times a week is to express that you have conditionally resolved to do so; you are insincere if you do not have this resolution. But announcing that you have a resolution is not the same as expressing that you have one, in this sense. For the sincerity condition on announcing that you have a resolution is not having the resolution, but believing that you have that resolution (at least, on the plausible assumption that announcing is the same as asserting, and that a sincerity condition on asserting that p is believing that p.) We can motivate this claim by imagining a case in which you falsely believe yourself to have a resolution. Suppose that I have no desire or intention to lift weights. But I want to be the kind of person who enjoys lifting weights, and I am prone to self- delusion. As a result, I falsely believe that I have resolved to lift weights five times a week, without in fact having formed any intention or desire to do so. I announce to my trainer that I have formed a conditional resolution to lift. This announcement seems to be sincere. Compare this to a case in which I falsely believe that I have resolved to lift weights, have no intention or desire to do so, and then promise my trainer that I will lift. The promise does not seem to be sincere. If this is so, telling my trainer that I have conditionally resolved to lift weights five times a week is not the same as making a promise. For promising involves expressing a resolution, and announcing that I have resolved is not to express a resolution but merely to assert it. This response seems more plausible when we notice that other mental states in the SAMS pattern work in a similar way. For example, we have been assuming that forgiveness expresses the repudiation of blame (which is to say, that the sincerity condition on forgiving someone is that you have repudiated blame.) The sincerity condition on asserting, “I have repudiated blame” is believing that you have repudiated blame. Suppose you falsely believe yourself to have repudiated blame: you’ve been ! 173! taking meditation classes to let go of your anger, and you think that they’re working, but they’re not actually making you any less angry or wrathful. Under such circumstances, an announcement that you have repudiated blame would count as sincere. But a performance of the speech act of forgiving would not be sincere in such a case. We have a principled way to distinguish forgivings from proclamations of the repudiation of blame: only the former expresses repudiation of blame. 6. Objection: Can CRA Explain the Directedness of Promissory Obligation? It is commonly thought that promises are directed obligations, or that a promise is not just a generic moral obligation to act in a certain way, but an obligation owed to the promisee, who is uniquely wronged should the promise be broken. One might worry that the Conditional Resolution Account cannot accommodate the directedness of promissory obligation. Theories that ground promissory obligation directly in the well- being or moral rights of the promisee can explain directedness in a straightforward way. For example, promissory obligation is directed according to epistemic expectationalist views because failing to keep your promise harms the promisee by violating her expectations, and we owe it to people not to harm them. Under normative power views, promissory obligation is directed because you have transferred a right to the promisee to determine how you will act, and we owe it to people not to infringe their rights. How can a theory that grounds promissory obligation in a resolution of the promisor explain directedness? We don’t usually suppose that someone who resolves to Φ owes it to someone else to Φ. So why should promissory resolutions be directed? ! 174! 6.1. Determining directedness Before attempting to answer this question, we should get clearer about what it is for an obligation to be directed, or owed to a particular person. One way to track the directedness of an obligation is to ask whether anyone has special standing to resent you for a violation, and if anyone is uniquely positioned to forgive you. For example, suppose a passenger on an airplane has a medical emergency. The flight crew asks if any medical professionals are on board. There is a doctor on board (of medicine, not of philosophy) who could safely and easily assist the person in distress, but who fails to do so because she prefers not to miss any of her in-flight movie. When this fact comes to light, the other passengers on the plane might criticize or blame the doctor for her callous behavior. But the distressed passenger has standing to blame the doctor in a distinctive way. He is entitled to hold reactive attitudes towards the doctor that the other passengers are not entitled to, such as resenting the doctor, where resentment is understood as an negative emotional and evaluative response to a personal violation. And only the distressed passenger has standing to forgive the doctor. The flight attendant or a fellow passenger might cease to blame the doctor, but they are not in a position to forgive her and thereby make future blaming inapt. This shows us that the doctor owes it to the distressed passenger to help him; this is not a generic obligation to be beneficent, but a directed obligation owed to a particular person. We can apply this test to promising to see exactly how promissory obligation is directed. Suppose Johnna is a big fan of Taylor Swift, and Isaac promises Johnna that he will buy her tickets to a Taylor Swift concert for her birthday. He fails to do so. Johnna has special standing to resent Isaac and hold him accountable, and unique standing to forgive him. Suppose instead that Johnna’s brother Kirby is the Taylor ! 175! Swift fan. Isaac promises Johnna that he will buy Kirby tickets to the Taylor Swift concert, and fails to do so. In this case, it is still Johnna who has special standing to blame Isaac and unique standing to forgive him. Outside observers might blame or criticize Isaac for breaking his promise, but they will not be entitled to resent or forgive him. If Kirby knows about the promise, and has formed expectations in response to it, he might also be entitled to resent Isaac for the breach, because it caused him harm and Isaac owes it to Kirby not to harm him. But this is an independent obligation grounded in expectations, and is not the same as the distinctly promissory obligation that is incurred and is owed to Johnna rather than to Kirby, even when Kirby is the beneficiary of the gift. 6.2.! Promisee acceptance as the ground of directedness I argued in Section 3 that all promises require the acceptance of the promisee, and that such acceptance implicates that the promisee is interested in at least the making of the promise, and typically in the keeping of the promise, as well. Promisee acceptance is what grounds of the directedness of promissory obligation under the CRA. In a nutshell, the idea is that while most ordinary resolutions are non-directed, some resolutions yield directed obligations. The fact that promissory resolutions must be conditional on someone else’s acceptance ensures that they are always directed. If you resolve unconditionally and in secret to mow your elderly neighbor’s lawn as a favor, the rational obligation your resolution incurs is not directed to any other person. If it is directed at all, it is self-directed, and you owe it only to yourself to do so. And if you fail to keep the resolution, there is no person who is uniquely positioned to resent you or forgive you, although there may be a sense in which you could rightly ! 176! resent or forgive yourself. 176 If you tell your neighbor that you have resolved (unconditionally) to mow her lawn, you will incur expectation-based obligations to either follow through or alter her expectations. But the obligation stemming from your resolution will not be directed towards your neighbor, and will rather be undirected or self-directed. If you fail to carry out the resolution without violating any of your expectation-based obligations (say, by warning the neighbor in advance that have abandoned your resolution), your neighbor will not have any special standing to resent you or forgive you. But if your resolution is explicitly conditional on your neighbor’s interests—if you tell her that you have resolved to mow the lawn, so long as she wants you to—it might seem plausible that the obligation it generates is directed, and that the neighbor would be within her rights to resent or criticize you for failing to mow her lawn when she still desired that you do so. For your resolution conveyed to your neighbor that you plan to mow her lawn, and that this plan depends not on your whims or desires but on hers. It might seem that you owe it to her to carry out this plan, and that abandoning it for your own reasons wrongs your neighbor by failing to give her interests their proper due. And if your resolution is explicitly conditional not only your neighbor’s belief or desire that you mow her lawn but on her acceptance of this resolution—that is, on her active and continued agreement with your plan—it seems even more plausible that you owe her a directed obligation. For in that case, you have conveyed to her that you plan to mow her lawn, and that this plan depends not on your whims or desires but on her volition. The conditionality of the resolution essentially places the plan in her hands, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 176 I do not need to take a stand here about whether unconditional resolutions are self-directed, or about whether it is possible to resent yourself, and if so what self-resentment consists in (e.g., is blame a form of self-resentment?) These are interesting and complicated questions, but they are not relevant for my present argument and therefore are beyond my scope. ! 177! and to alter the plan while she doesn’t want you seems to wrong her. At least, it seems that she is entitled to resent you or forgive you if you fail to follow through, which is evidence that you owe her a directed obligation. It therefore seems that directedness is not a unique feature of promissory resolutions. Rather, it is a general feature of resolutions that if you resolve to Φ conditional on the agreement, acceptance, or uptake of X, you owe a directed obligation to X to Φ. Promissory resolutions are always directional, because they are always conditional on the acceptance of the promisee. 6.3.! Can promises be owed to someone other than the promisee? For most further commitment norms, the object of concern—that is, the person who is the beneficiary of the action, or whom the action is about—cannot come apart from the subject of the directed obligation, or the person to whom the obligation is owed. For example, if Laura wrongs Marco, Marco can forgive Laura and thereby make it the case that he is obligated to refrain from blaming her. Laura is both the object of concern—the one who will no longer be blamed—and the person to whom Marco owes the obligation. But if Marco forgives Laura for wronging him, he cannot owe it to Oliver to cease blaming Laura. Nor can Marco forgive Laura for Oliver’s wronging of him, and owe it to Laura to cease blaming Oliver. The same is true of apology. If Laura wrongs Marco, she can apologize to him and thereby make it the case that she is obligated to take responsibility for the wrong. This responsibility-taking both concerns Marco and is owed to him. But Laura cannot apologize to Marco and owe it to him to make reparation to Oliver instead. Nor can Laura apologize to Oliver for having wronged Marco, and thereby owe it to Oliver to make reparation to Marco. ! 178! For promises, however, the object of concern can come apart from the subject of the directed obligation, and this is likely to lead to confusion if we are not careful. The subject of the directed obligation—the one to whom the promise is owed—is the promisee. Often, the promisee is also the person the promise concerns, or the beneficiary of the promise. But this needn’t always be the case. For example, Laura might promise Marco that she will take care for their grandfather Oliver when his health declines. This promise concerns Oliver, but is owed to Marco. And it is Marco who must accept the promise, not Oliver. In accepting it, Marco conveys that he is interested in the promise being kept, because he is interested in Laura’s benefitting Oliver. And it is Marco, not Oliver, who is uniquely positioned to resent or forgive Laura if she fails to perform. One might worry that there could be cases in which A promises B—that is, cases in which A expresses a resolution to B, and owes B a promissory obligation—but in which the resolution is conditional on C’s acceptance in some way. If there were such cases, they would be problematic for the CRA. For A would owe B a directed obligation without B having accepted the promise, which would mean that acceptance could not be the key to directedness after all. 177 Are cases with such a structure possible? Suppose Phil tells his daughter Quinn that he promises to teach her son Randy how to knit, on the condition that Randy accepts this offer. One natural construal of such a case is that Phil is telling Quinn about a promise that he has made (or plans to make) to Randy, conditional on Randy’s acceptance. This is clearly not a promise to Quinn. Rather, the promise is made to Randy, the obligation is owed to him, the resolution is conditional on his acceptance, and Randy has unique standing to resent or !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 177 Thanks to Jake Ross for raising this worry. ! 179! forgive Phil should he break the promise. Understood this way, the utterance is no problem for the CRA. Another plausible construal of the case is that Phil is promising Quinn that he will teach Randy how to knit, conditional on Randy’s agreeing to this plan. Phil’s promissory resolution is indeed conditional on Randy’s wanting to knit; if Randy doesn’t agree, the condition will not be met and no valid promise will be formed. But the resolution is also conditional on Quinn’s acceptance. For she could reject the promissory offer if she so chose, in which case no valid promise would be formed. Moreover, Phil has an expectation-based moral obligation not to disappoint Randy, and to follow through with the promise or warn him in advance if he will not. But his distinctly promissory obligation is owed to Quinn. If Phil breaks the promise without violating his expectation-based obligations, Quinn will be uniquely positioned to resent or forgive Phil, whereas Randy will not be. So the promise is made to Quinn, the obligation is owed to Quinn, the resolution is conditional on Quinn’s acceptance, and Quinn has unique standing to resent or forgive. The case so understood presents no problem for the CRA. The only way for a case like this is a problem for CRA is if we understand it as Phil making a promise to Quinn—and thereby incurring a directed obligation to her— that is conditional on Randy’s agreement, but not conditional on Quinn’s acceptance. However, this seems to be conceptually confused; in such a case, either the promise is in fact made to Randy, and Phil is just reporting it to Quinn, or Quinn must at least implicitly accept the promise. Suppose that Randy tells Quinn, “I promise you that I will teach Randy to knit, but only on the condition that he accepts my offer.” Quinn replies, “Don’t bother; I prefer to teach Randy myself.” Phil could not properly respond ! 180! by saying, “Well, I’m promising you anyway, because what matters is whether Randy wants to knit, and not whether you want me to teach him.” Promissory offers that are explicitly rejected cannot generate valid promises. And so a promise that is made to X must be conditional on X’s acceptance—even if it is also conditional on the interests or acceptance of Y. Ultimately, the Conditional Resolution Account seems to be the best candidate for the sincerity condition on promises. It captures the right degree of commitment and inertia, is sufficiently counterfactually robust, accommodates promisee acceptance, and can explain both the difference between promissory resolutions and other sorts of resolutions and the directedness of promissory obligation. ! 181! CHAPTER 6: WHAT ARE RESOLUTIONS? 1.! Introduction When introducing the Mental States First theory in Chapter 3, I suggested that resolving to Φ commits you to Φing, unless you have a good excuse. In this chapter, I will say more about the way in which resolving rationally commits, and about what I take resolutions to be. In Willing, Wanting, Waiting, Richard Holton lays out a detailed account of resolutions, arguing that they enable agents to resist temptation. 178 Although others philosophers have offered accounts of resolutions, Holton’s is the best developed and most influential, and will therefore be my point of departure for discussion. 179 In Section 2, I offer a pair of examples to motivate the claim that resolutions rationally commit us to acting, and I say more about what rational commitment is. I then devote the rest of the chapter to developing a detailed criticism of Holton’s substantive view, and constructing an alternative that does not fall prey to the same !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 178 2009. 179 For example, see McGuire (1961), who argues that you can decide to do something only if you know that you are able to do it, but can resolve to do something even when you recognize that there is a chance of failure. McClennan (1990) outlines a notion of resolute choice according to which the chooser commits herself to a plan that seems best at the time of initial choice, and then sticks with this plan regardless of whether it continues to seem best to her; this notion is elaborated upon and modified by Gauthier (1997). Cohen and Handfield (2010) argue that resolve is a distinct rational capacity to cease or suppress deliberation, and sketch a non- consequentialist, virtue-based account of resolutions. It is a common idea among non-philosophers that a resolution is a promise that you make to yourself. Several philosophers who defend the existence of promises to the self note in passing that resolutions are a type of self-promise. (See, for example, Dannenberg (forthcoming, p. 1), Habib (2009, p. 545 n. 10), and Rosati (2011, p. 124 – 5).) However, the existence of genuine promises to the self is controversial. Some argue that self-promises do not exist because we have no moral obligations to ourselves at all; see Singer (1959). Others argue that we can have moral obligations to the self, but that the possibility of promisee release makes promissory obligations to the self implausible; see Hills (2003) and Liberman (ms). ! 182! critique. (This section is inessential to the main argument of the dissertation. As long as we have some plausible account of resolutions that entails that they rationally commit agents to acting, the Mental States First should succeed; the particular details of the account we adopt should not affect the plausibility of MSF or the CRA. Readers who are not independently interested in the details of what resolutions are may therefore skip sections 3 – 8 of this chapter.) In Sections 3 – 8, I first elaborate upon Holton’s account of temptation, and explain why temptation is likely to lead to judgment shift in the first place; Holton needs such an explanation to motivate his account. However, I argue that the explanation I offer can be extended, using principles Holton accepts, to establish that Holtonian resolutions are unable to block temptation after all. Thus, the defense of Holton’s account of temptation entails that resolutions cannot play the role Holton attributes to them. In conclusion, I sketch and briefly defend an alternative account of resolutions that is not undermined in the same way. 2. Resolutions and Rational Commitments As I explain in more detail in the next section, Holton argues that the primary function of resolutions in our psychology is enabling us to resist temptation, and thereby prevent problematic shifts in judgment that lead to over-hasty abandonment of resolutions. This is a good starting point for an account of resolutions, because abandoning a resolution over-hastily indeed seems problematic. Consider the following case. Suppose you are generally hesitant to try new foods, and are deeply entrenched in the habit of eating pizza for lunch every day. You very much want to expand your culinary horizons, but are such a creature of habit that you are unlikely to ! 183! do so unless you force yourself into it somehow. So you resolve to go to a Thai restaurant for lunch, knowing that if you don’t form this resolution you are likely to fall back into your pizza habit. When you leave your office to eat lunch, you abandon your plan to go to the Thai place and head to the pizzeria instead, deciding that you might as well just eat pizza today, since it’s easier for you to order from a familiar menu. Something seems wrong with this picture. Resolving to eat Thai food in order to expand your culinary horizons and then changing your mind without good reason seems problematically self-undermining. That is, there seems to be something irrational or incoherent about resolving to eat Thai food because you care about broadening your horizons, and then deciding to stick with your pizza habit instead because it is easier. It’s not that eating pizza every day is in any way independently irrational; having pizza all the time is rationally permissible, even if boring. Rather, it’s that it is irrational to resolve to eat Thai food, and then abandon this resolution for no good reason. Compare abandoning a resolution to eat Thai food to a case in which you merely desire to have Thai food and eat pizza instead. This seems perfectly acceptable; failing to act on a particular desire is not irrational. Or compare it to the case in which you intend to have Thai food but don’t really care about whether this plan changes. You do not display any irrationality if you change your mind because you suddenly have a craving for grilled cheese; you’ve simply made a permissible change of intention on the basis of a whim. Changing your resolution on a similar basis is not so innocuous. We can illustrate the way in which abandoning a resolution without good reason is irrational by comparing a pair of similar cases. Suppose that you and I each have a reason of strength X to eat Thai food and expand our respective culinary horizons. And ! 184! suppose we both care about expanding our culinary horizons to the same extent. But only I take action about it: I resolve to eat Thai food, making it a part of my plan and adopting it as one of the concerns that I decide to actively focus on. I seem to have more of a reason to eat Thai food than you do, simply in virtue of my resolution. One might worry that the commitment involved in such cases is psychological or volitional commitment, rather than normative commitment—that having formed a resolution might motivate me to eat Thai food, but doesn’t in any way rationally require me to do so. 180 But this is not the case, for resolutions can normatively commit you even when you are not psychologically committed. For example, suppose I want to be as healthy as I can be, and so I care about flossing regularly, drinking eight glasses of water a day, and wearing sunscreen whenever I go outside. I care about each in roughly equal amounts; each matters to me somewhat, but none is especially important. Assume for the sake of argument that the health risks I face if I do not accomplish these goals—tooth decay, dehydration, and sun damage—are all roughly equivalent. After hearing about a friend who developed melanoma, I resolve to begin wearing sunscreen religiously. I haven’t learned that my risk for skin cancer is increased, or anything of the sort—skin protection has simply become salient to me, so it’s the goal I choose to focus my attention on. I seem to be normatively committed to wearing sunscreen on the basis of this resolution in a way that I am not committed to flossing or drinking enough water. This is so even if I’m not psychologically committed to wearing sunscreen all the time—if I don’t actually feel compelled to slather on sunscreen when I’m heading out for a short walk. If I change my mind about wearing sunscreen today simply because I find it unpleasant, I go wrong in a way that I do not if !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 180 Thanks to Steve Finlay for helping me appreciate this worry. ! 185! I change my mind about flossing tonight simply because I find it unpleasant. My independent reasons to floss and to wear sunscreen are by hypothesis equivalently strong, and I’m equally psychologically committed to each goal. It’s the resolution itself that makes a normative difference. Over-readily abandoning a resolution seems problematic because resolving to Φ rationally commits you to Φing. 181 The easiest way to get a sense of what I mean by rational commitment is to consider the way in which holding one belief can commit you to holding another. Philosophers frequently talk about the ways in which our beliefs commit us: because Kant believes that the categorical imperative is the fundamental moral rule, he is rationally committed to believing that you are never permitted to lie; because the act utilitarian believes that maximizing utility is the fundamental moral rule, she is rationally committed to believing that it is permissible to sacrifice an innocent person to prevent a massive riot. Such commitments can come apart from what you objectively ought to believe, all-things-considered: it is possible (indeed, likely) that Kant should not all-things-considered believe that you are forbidden from lying in every circumstance, and that the utilitarian should not all-things-considered believe that it is okay to prevent chaos by sacrificing an innocent. I am claiming that a resolution commits you to acting in the same way that believing that P commits you to believing the obvious and relevant consequences of P. Someone who believes that P and that P!Q but does not believe that Q when the question of whether Q is salient in some way is irrational in a sense: she is failing to act on her commitments, and as a result her overall set of beliefs is not as complete and !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 181 For more on the concept of commitments (as distinct from obligations or reasons), see Shpall (2013 and forthcoming), as well as Liberman and Schroeder (forthcoming.) ! 186! coherent as it should be. Similarly, someone who resolves to Φ at a particular time and then fails to intend to do so at that time because she over readily abandons this resolution is irrational in a sense; she is failing to act on her commitment, and as a result she is not as effective a planning agent as she could be. 182 When I abandon my resolution to go to the Thai restaurant for lunch, I am undermining my own endorsed goals and plans: I value culinary diversity, and adopt this as one of my goals, but do not succeed in attaining it because I over-hastily abandon my resolution. This is not to say that it is never permissible to abandon a resolution. Sometimes, all else is not equal—you resolve to have Thai food and then an old friend unexpectedly offers to meet you for lunch at the pizza place, or you realize that Thai food often contains ingredients to which you are allergic, etc. 183 However, it seems that your resolution rationally commits you to acting unless there is a good excuse in place. 2.1. Holton’s view: resolutions as a means of resisting temptation Holton takes resolutions to be a special type of intention. He builds on Michael Bratman’s prominent account of intention, according to which intentions are non- reducible mental states that aid in inter- and intrapersonal planning and coordination. Typically, intentions are effective at bringing agents to act. Sometimes, though, temptations of various sorts threaten to sway an agent from her intended course of action. Holton argues that temptation “frequently works not simply by overcoming one’s better judgement, but by corrupting one’s judgement,” noting that “as a matter of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 182 Time indexing is necessary to avoid over-generating cases: if I resolve to go to eat Thai food at noon, it’s no problem that I haven’t yet formed the intention to do so at 9 AM. But it would be problematic if lunchtime rolls around and I haven’t formed such an intention. 183 See Chapter 8 for a detailed description of the kinds of scenarios in which all else is not equal, and it is permissible to abandon a resolution. ! 187! empirical fact, temptation normally induces judgment shift.” 184 In other words, an agent who succumbs to the temptation to smoke another cigarette in spite of an intention not to do so does not typically maintain her judgment that she ought not smoke while akratically lighting up anyway. Rather, in many cases her judgment about smoking shifts; she changes her mind about whether to smoke, making smoking in line with her current judgment about what is best. Holton argues that this sort of judgment shift leads to an agent’s judgment becoming “either corrupted or powerless” in such a way that it “cannot be the motor of resistance” to temptation. 185 He proposes that what can function as the source of resistance in such cases is a resolution, or a pair of intentions, one first-order and the other second-order. Resolutions consist in “both an intention to engage in a certain action, and a further intention not to let that intention be deflected.” 186 So an agent who resolves to quit smoking both intends not to have any more cigarettes, and intends not to reconsider the intention not to have any more cigarettes. This second-order intention prevents judgment shift. For if the agent does not allow herself to reconsider whether she should hang on to her intention to avoid cigarettes, her judgment will not have the opportunity to shift, and she will be likely to act as she originally intended. An important and insufficiently appreciated insight in Holton’s account is that temptation leads to judgment shift not merely through some brute causal process. Rather, there is a mental process involved that, while not fully rational, allows the agent who succumbs to temptation to make sense of her action; as a term of art, I call this a sense-making process. Holton suggests that one major cause of judgment shift is that !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 184 2009, p. 97. 185 ibid, p. 98 186 ibid, p. 11. ! 188! agents typically foresee that they are likely to succumb to temptation. They want to see themselves in a good light, and avoid the cognitive dissonance of believing that it would be best to refrain from Φing while believing that they will Φ anyway. An easy way to resolve this dissonance is changing the belief that it would be best to refrain from Φing. This method of resolving cognitive dissonance by shifting judgment is not rational, but it at least allows the agent to make sense of her own action. However, more needs to be said. Holton’s appeal to the avoidance of cognitive dissonance leaves unanswered the more fundamental question of why agents predict that they will succumb to temptation in the first place. Why are we likely to act as we are tempted to, even when we have explicitly intended not to so act? This question is independently interesting; temptation is a frequent and frustrating part of our everyday experience, and we want a complete philosophical understanding of how temptation works. But the question is of particular importance for Holton, for the answer will help explain why temptation leads agents to revise their first-order intentions, and Holton’s account of resolutions as second-order intentions not to reconsider is motivated by the need to prevent such revision. I return to this question in Section 5. 3. The Efficacy Objection One might wonder why adding a second-order intention makes any difference for preventing judgment shift. We can call this the efficacy objection: why is a resolution more effective at resisting temptation than a first-order intention alone would be? Holton’s answer is an appeal to willpower as the means by which agents resist temptation and achieve strength of will. For Holton, willpower is “substantial; it is at least a skill and perhaps a self-standing faculty, the exercise of which causally explains ! 189! our ability to stick to a resolution.” 187 Exerting willpower requires effort. Like other faculties, willpower can be built up by exercise or depleted by overuse. But even if we grant that agents possess some faculty of willpower and supposed that we perfectly understood how this faculty works, a worry about efficacy would remain. Sarah Paul characterizes Holtonian temptation as “essentially involv[ing] the tantalizing action's petitioning of one's awareness with a practical proposal.” 188 She argues that this seems to be a problem for Holton, for “surely this already suffices for the practical question to be raised.” And once the practical question is raised, the answer is obvious; for “once temptation has set in, of course the pleasure afforded by [for example] the wine outweighs all other considerations. Revision will be well-nigh automatic.” Paul goes on to suggest that because temptation presents the alternative action so vividly, it will push for revision of not only the first-order intention but also the entire resolution. In other words, it seems that the second-order intention should succumb to the same temptation to which the first-order intention is susceptible. John Maier articulates this worry as follows: For any considerations in favor of revising one’s first-order intention would presumably also be grounds for revising one’s second-order intention to not let that intention be deflected. But then it does not seem that a higher-order intention of this sort will be guaranteed to be any more rationally robust than a first-order intention – it is, as it were, just another intention. 189 It might seem that not even willpower can prevent the transmission of temptation from the first-order intention to the second-order intention. Why don’t the very same considerations that tempt you towards watching yet another episode of your favorite !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 187 ibid, p. 112. 188 2011; all citations in this paragraph are from p. 891. 189 2010, p. 361; my emphasis ! 190! TV show—say, your burning desire to find out what happens next, and your aversion to working—also tempt you to reconsider your resolution to turn off the TV and get to work on your paper, regardless of whether you possess willpower? There’s something intuitively compelling about this objection. Philosophers before me have noticed it, and it seems prima facie plausible. But as of yet, the objection is underdeveloped. For we haven’t seen why the temptation that pushes on the first- order intention pushes on the second-order intention, as well. What causes the temptation to transmit in this way? I will answer that question in the next two sections of this chapter, thereby making the intuitive objection more rigorous. In doing so, I will fill in the explanatory gap mentioned in Section 2.1 above, and explain why agents are likely to succumb to temptation at the first-order level. When thinking about how to precisify the efficacy objection, it’s important to realize that the first- and second-order intentions involved in a resolution are about different things: one is an intention to Φ, and the other is an intention not to reconsider the intention to Φ. So we cannot simply assume that a temptation to refrain from Φing will also be a temptation to reconsider an intention to Φ. For at least in principle, these two intentions could be susceptible to entirely different temptations. And if this were so, a temptation that leads to judgment shift at the first-order level will not automatically lead to judgment shift at the second-order level, as well. Rather, I will construct a Temptation Transmission argument to show that a temptation to abandon a first-order intention will automatically also yield an equally strong temptation to abandon any second-order intention about maintaining that first order intention. The argument relies on certain background principles, which I lay out ! 191! in the Section 4 before constructing the argument in Section 5. The first stage of Temptation Transmission explains why the tempted agent feels pressure to act contrary to her first-order intention in the first place, thereby answering the question posed at the end of Section 2.1. Holton must accept an explanation like this one in order to motivate his account. But the second stage of the argument entails that any temptation that puts pressure on an agent’s first-order intention necessarily also puts pressure on her second- order intention—which means that the second-order intention will not reliably block the very temptation it was designed to circumvent, and the resolution will be no more effective at preventing judgment shift than a first-order intention alone would be. Since both stages of the argument employ similar reasoning and draw on the same plausible principles, Holton cannot easily resist the second stage without also rejecting the first. 4. . Background Principles for the Temptation Transmission Argument 4.1. Temptation The starting point for Temptation Transmission is the following empirical claim: Temptation Claim: temptation works by altering the appearances in favor of there being a reason to do the tempting thing from the agent’s perspective. The Temptation Claim is a general formulation of the empirical understanding of temptation to which Holton is committed. Holton claims that what causes people to yield to temptation is desire in a certain sense. The sense of desire he has in mind is very like Scanlon’s concept of desire in the directed-attention sense, according to which a “person’s attention is directed insistently toward considerations that present themselves as counting in favor of” the desired object or action 190 . Assuming (as Scanlon does) that a reason for X is a consideration that counts in favor of X, what desire in this sense does is !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 190 1998, p. 39. ! 192! direct the agent’s attention towards apparent reasons for X. Since temptation leads to desire in this sense, temptation draws the agent’s attention to apparent reasons for the tempting thing. 191 The Temptation Claim captures the phenomenology of temptation well. Sometimes, temptation clues the agent in to genuine reasons that she has to do the tempting thing; an agent who is tempted to abandon her resolution to go to the gym at 7:00 AM because she feels extremely sleepy when her alarm goes off really does have a reason to sleep in, albeit likely a weaker reason than the reason to go as planned. In such a case, the temptation might make the reasons to stay in bed seem stronger to the sleepy agent than she takes them to be upon reflection, when not being tempted. But temptation can also make it seem to the agent like there is a reason for her to do the tempting thing when by her own lights she has none. For example, consider a recovering alcoholic who is committed to maintaining his sobriety. He is feeling tempted by a bottle of whiskey. He does not take himself to actually have a reason to drink. Rather, he takes himself to have strong reasons not to. But in the moment of temptation, it nevertheless seems to him as if he has a reason to drink. The appearances of what he has reason to do have altered, even though his actual reasons have not. In general, we can assume that if an agent saw the need to form a resolution in anticipation of a certain temptation, then she foresaw that the temptation would alter !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 191 Holton notes that the features of desire that Scanlon points out “may be necessary for desire to arise; but they do not constitute it” (102). He suggests that “what is missing in Scanlon’s characterization is the idea that desire pulls me to a certain course of action,” concluding that the sense of desire he is after “is a state that preoccupies an agent’s attention with an urge to perform a certain action” (102). The Temptation Claim makes sense of this: the agent’s attention is directed towards an apparent reason, and the apparent reason is what creates the urge to act. Since agents (typically) act for reasons, this is a natural picture. And if we did not cash out the urge in terms of reasons, it is unclear what else could be motivating or driving the agent. ! 193! the appearances in favor of there being a reason for her to do the tempting thing. If the temptation did not seem to provide the agent with reasons, then the agent would not have needed to form a resolution in the first place; merely intending to act would have sufficed. Any temptation that is serious enough so as to move the agent to form a resolution will seem to present that agent with reasons for acting. 4.2. Rational Means Reasons Transmission Temptation Transmission relies in large part on the philosophically commonplace and highly intuitive observation that reasons for ends transmit to the necessary means to those ends. 192 Since a necessary means/end reasons transmission principle is widely accepted in one form or another, I will assume that some version of it is true. In particular, I will rely on the following version of the principle, which is broader than most in a way that will play an important role in the arguments to come: Rational Means Reasons Transmission (RMRT): Where E is an intentional action, if (1) there is a reason of strength X for an agent A to attain end E, and (2) M is the only rationally permissible way to attain E, then there will be a reason of strength X for A to do M. RMRT states that if an agent has a reason to take some end—say, a reason to travel from LA to New York for a long weekend—then the agent has an equally strong reason to take whatever means are necessary for attaining that end in a rationally permissible way—say, buying a plane ticket. In essence, RMRT allows that a means !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 192 Versions of a necessary means reasons transmission principle have been formulated and defended in slightly different ways by Nagel (1970 p. 34), Darwall (1970 p. 16), Kolodny (2007 p. 251), Schroeder (2009 p. 245), Bratman (2009 p. 424), and Way (2012 p. 494), among others, none of whom formulate the principle in the broad way I do here. For arguments that reasons transmit to sufficient means as well as necessary means, see Bedke (2009, 687 n. 10) and Way (2010, 224). For a detailed discussion of whether means/end reasons transmission principles should be necessary or sufficient, see Kolodny (forthcoming), who concludes that the means must be non-superfluous, which rules out merely sufficient means. ! 194! counts as necessary for the purposes of reasons transmission if the means is the only rationally permissible option. This means that RMRT allows reasons to transmit in a broader range of cases than do typical necessary means reasons transmission principles. Buying a plane ticket is not strictly necessary to get from LA to New York for a long weekend—you could drive, or board the plane with a stolen ticket. But neither of these other options makes any sense; driving takes far too long to make a weekend trip worthwhile, and flying with a stolen ticket is far too difficult and risky. RMRT also includes a commonly accepted constraint that the strength of the reason transmitted from end to means remains constant, such that a reason of strength X for the end yields a reason of strength X for the sole rationally permissible means. 193 We can highlight the independent plausibility of RMRT by considering a couple of examples. Suppose a just country requires all of its citizens to either join the military for two years or perform alternate service. Smith has a reason to satisfy this legal requirement. Because Smith is a principled and committed pacifist, joining the military is rationally impermissible for her. Accordingly, the only rationally permissible way for Smith to attain the end of satisfying her legal requirement is by performing alternative service. It follows that Smith has a reason to perform alternate service. 194 Or suppose !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 193 Kolodny (forthcoming) argues that reasons will not always transmit full strength to necessary means. But his cases turn on features that do not matter for the ways in which I will be applying RMRT. While this complication is important for a general account of reasons transmission, I may gloss over it here without problem. 194 The case of pacifist Smith choosing between joining the military or performing alternate service has been discussed by Horty (1993, p. 73), Goble (2004, p. 80), and Nair (forthcoming, p. 2). Following Horty, these authors all formulate the example as an instance of a pattern of good or valid reasoning involving disjunction: if you have reason to do A v B, and you have reason not to do A, then you can conclude that you have reason to do B. RMRT as formulated restricts reasons transmission only to means. But the principle can be seen as an instance of the following broader disjunctive principle—call it Disjunctive Reasons Transmission, or DRT: if (1) you have a reason of strength X to do (A v B), and (2) A is rationally impermissible, then (C) you have a reason of strength X to do B. ! 195! that you have a reason to bring your new dining room table inside the house. It won’t fit through the door, and the only ways you can get it inside are by taking the door of its hinges or cutting the legs off of the table with a chainsaw. Cutting off the table’s legs makes no sense, as you have a reason to take the table inside only insofar as you are able to use it, and you cannot use a legless table. So the only rationally permissible way for you to get the table inside is to take the door off its hinges. It follows that you have a reason to take the door off its hinges. RMRT allows reasons to transmit to a means that are is not strictly necessary for attaining an end only when that means is the sole rationally permissible way to attain the end. This prevents reasons from transmitting too widely. For example, suppose you have a reason of strength X to pay for an item at the store. You can pay with cash, credit card, or check. Your credit card is maxed out, leaving paying with cash or check as reasonable options for you. RMRT does not predict that you have a reason of strength X to pay by check and a reason of strength X to pay by credit card. RMRT strikes a balance between being more permissive than typical necessary means reasons transmission principles, but less permissive than sufficient means reasons transmission principles. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! If we allow reasons transmission to range over any disjunction, we will face worries about reasons explosion. If it is possible for there to be a reason to A and for A to be rationally impermissible, DRT will run into an explosion problem: premise (1) that there is a reason of strength X to do (A v B) will be true because of the reason to do A, and premise (2) will be true because A is rationally impermissible. From this, we can conclude that (C) there is a reason of strength X to do B, for any B. Explosion is avoided if we either restrict the principle to means (thereby preventing any B from being included in the argument), or if we assume that it is not possible for there to be a reason to A if A is rationally impermissible. In this paper, I take the first approach, restricting RMRT to means. ! 196! 4.3. Transparency We have already noted that temptation provides agents not necessarily with reasons to act, but with the appearance of reasons to act. If we are to employ RMRT in a Temptation Transmission argument, then RMRT must apply not only to the transmission of reasons, but to the transmission of the appearance of reasons, as well. This means that we need the following constraint on RMRT: Transparency: RMRT is apparent: if it appears to an agent A that: (1) there is a reason of strength X to attain end E, and (2) M is the only rationally permissible way to attain E, then there will appear to A to be a reason of strength X to do M. Transparency is intuitively compelling. If it seems to Smith that she has a good reason to obey the laws of her country, and it seems to her the only rationally permissible way to do so is performing alternate service, then it will seem to Smith that she has good reason to serve. This will be so even if Smith’s country is unjust and she does not in fact have good reason to obey her country’s laws, or if serving is not in fact the only rationally permissible way for her to obey the law (perhaps Smith mistakenly believes that her religious beliefs commit her to pacifism when in fact her religion commands her to fight in just wars.) The transparent version of RMRT is not about what reasons there are, but about what reasons agents take there to be. RMRT states that the reasons an agent takes there to be for the end transmit to what she takes to be the sole rationally permissible means to that end. Transparency is also theoretically useful, for it offers us a good explanation of what is going on in cases like the following. Suppose a recovering alcoholic is clear- sighted enough to recognize that he does not really have any good reason to drink whiskey. But the tempting scotch in front of him makes it seem to him like he does. If he believes that opening the bottle is the only rationally permissible way for him to ! 197! have a glass of scotch, it will also seem to him like he has a good reason to open the bottle. It’s not clear how we would explain this if we deny Transparency. Temptation Transmission starts with the Temptation Claim, and then applies the transparent RMRT principle to a series of independently compelling claims about which mental states are the only rationally permissible means to rationally holding other mental states. The agent who follows the steps laid out by Temptation Transmission is not ideally rational, but proceeds in a way that allows her to make sense of what she does. As we shall see in the next section, the argument concludes that temptation necessarily transmits—not only to first-order intentions, as Holton claims it does, but to resolute second-order intentions, as well. Moreover, seeing the exact way in which temptation transmits will show us what sort of view we need to stop the transmission. 5. The Temptation Transmission Argument The following three claims form the basis of Temptation Transmission: 1)! When Φing is an action about which an agent has formed a resolution, an apparent reason to Φ (stemming from temptation) necessarily leads to an apparent reason to intend to Φ. 2)! An apparent reason to intend to Φ necessarily leads to an apparent reason to reconsider the intention to avoid Φing (call this intention “Intention 1.”) 3)! An apparent reason to reconsider Intention 1 necessarily leads to an apparent reason to reconsider the intention not to reconsider Intention 1 (call this intention “Intention 2.”) If these three claims are true, Holton’s account of resolutions is in trouble, for Intention 2 will not successfully block temptation and prevent reconsideration, after all. The first two claims by themselves are not problematic for Holton. To the contrary, Holton needs claims of this sort: they answer the question from Section 2.1, and explain why ! 198! temptation is likely to lead to judgment shift and the abandonment of the first-order intention in the first place, such that we need something additional to prevent such shift and abandonment. The third claim is what puts pressure on Holton’s account—but as we shall see, this claim is difficult to resist while accepting the first two, as it invokes similar reasoning and draws on principles Holton accepts. The three claims above can be defended by appealing to RMRT, Transparency, and the relationships between means and rationally permissible ends among various sorts of mental states. I will work through these claims one at a time, before putting the entire argument together. (While I lay out my discussion in terms of intending to Φ in spite of a resolution to avoid Φing—e.g., resolving to avoid eating sweets and being tempted to eat a donut—we could just as easily cash things out in terms of a positive resolution and a temptation to do something else—e.g., resolving to eat only healthy food and being tempted to eat a donut.) 5.1. Defending Claim 1 The empirical claim about temptation states that a temptation to eat a donut makes it seem to the agent like she has a reason to eat the donut. RMRT states that reasons transfer from ends to sole rationally permissible means. If intending to eat a donut is the only rationally permissible way to eat one, and this is obvious to the agent, we will have established Claim 1: that an apparent reason to eat the donut will always lead to an apparent reason to intend to eat the donut, as well. Surely intending to Φ is not always necessary for Φing; many of our actions are unintentional. Nor is it always the case that intending to Φ is necessary for rationally ! 199! Φing; habitual actions such as using a turn signal while driving can be rationally performed unintentionally. But when Φing is an action about which an agent is making a resolution, it is plausible that intending to Φ is the sole rationally permissible means to Φing. This is because it makes sense to make resolutions only about actions that are in your direct control; if you know that you will automatically reach out to deflect a ball that is thrown at you, it is silly for you to resolve not to deflect it, and superfluous for you to resolve to deflect it. Furthermore, it makes sense to make resolutions only when there exist no more effective or less costly means of attaining your ends. You need not resolve to refrain from eating donuts if you have a robust and deeply ingrained habit of avoiding donuts, or if you have been hypnotized or classically conditioned to avoid donuts. This means that coherent resolutions can only be made about actions that must be performed intentionally; it follows from this that one must intend to act in order to rationally carry out such resolutions. Sometimes, bringing yourself to act in unintentional ways makes the most sense for an agent, all-things-considered; a diabetic who undergoes hypnosis to avoid donuts because she has been unable reduce sugar intake using more standard methods of self- control acts well. But refraining from eating donuts because you were hypnotized does not count as a rationally permissible way of satisfying a resolution to avoid donuts. Rather, it is an alternate means of attaining the same goal, in a manner that does not involve forming and acting on a resolution. The only rationally permissible way to carry out a resolution to refrain from eating donuts is to intend to do so. Finally, since the agent had to intend to Φ in order to resolve to Φ in the first place, the fact that rationally Φing requires intending to Φ will be transparent to her. ! 200! 5.2. Defending Claim 2 Claim 2 states that an apparent reason to intend to Φ necessarily leads to an apparent reason to reconsider the intention to avoid Φing. The argument for this claim proceeds in two steps. First, I argue that giving up any intention that you have not to Φ is the only rationally permissible way to intend to Φ. By RMRT and Transparency, an apparent reason to intend to Φ will therefore lead to an apparent reason to abandon the intention not to Φ. Second, I argue that reconsidering an intention is the only rationally permissible way to abandon it. And so again by RMRT and Transparency, an apparent reason to abandon the intention not to Φ will therefore lead to an apparent reason to reconsider that intention. The first means-end claim is true because of the following commonly accepted and intuitively obvious constraint on intentions: Intention Consistency Constraint: It is not rational for an agent to intend to Φ at time t and to simultaneously intend not to Φ at time t. 195 When intentions work properly, they bring agents to action in accordance with them. Since it is not possible to simultaneously eat a donut and not eat a donut, it does not make sense to simultaneously intend to do both of these things. An agent who simultaneously intends to eat a donut and intends not to eat a donut must therefore abandon at least one of these intentions if she is to be rational. It follows from this that giving up any intention to avoid eating a donut is the only rationally permissible means !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 195 Philosophers ground this constraint differently; for example, cognitivists about intention (such as Gilbert Harman and David Velleman) will offer different explanations than will those with planning theories of intention, like Bratman. But there is a general consensus that there is some such consistency constraint. See Bratman (2009a) for a more detailed discussion of this constraint and how it may be grounded. ! 201! to intending to eat the donut. Giving up the intention to avoid the donut might not be something that the agent must do in order to intend to eat the donut at all; it is (presumably) possible for someone to hold conflicting intentions. But since it is not possible to rationally do so, we can apply RMRT to conclude that a reason to intend to Φ yields an equally good reason to abandon the intention not to Φ. The second means-end claim states that an agent cannot rationally revise or abandon an intention unless she first reconsiders that intention. This is true to our everyday experience. Suppose that when I wake up, I intend to go to a yoga class at 8:00 PM. But I procrastinate for much of the day and do not finish the work that I hoped to complete. It would not be rational for me to go immediately from being in this state to being in the state of abandoning my intention to go to yoga. To the contrary, I have to think about the new state of affairs, temporarily shelving the intention to go to yoga and deliberating anew about what to do. Do I go to yoga as planned and not get any work done today? Or do I skip yoga to do work? If I didn’t go through some sort of deliberative process like this, my revision would be entirely out of the blue. If I am lucky, this revision will be in accordance with the rational course of action. But since I won’t always be lucky, this is not a rational disposition to have in general. Some sort of reconsideration must always occur, lest the agent revise inappropriately. Reconsideration of an intention need not be a long or drawn out process, and it need not involve difficult deliberation. If I intend to bake a cake and my oven catches fire while it is preheating, I need not deliberate about whether to attempt to bake a cake in the flaming oven; I can decide immediately that baking a cake no longer makes sense. But I will go through some process of revision, even if very quick. One might worry that it is possible to rationally abandon an intention without first reconsidering it. ! 202! Sometimes our intentions simply disappear or fall away; for example, you might intend to go to a concert at the end of the week and forget about it before the day arrives, without any irrationality. But when we are dealing with an intention about which you have made a resolution, this is not the case. If you resolve to go to the concert at the end of the week, something does go wrong if you simply forget about it. So we can apply RMRT to conclude that a reason to abandon the intention to Φ leads to a reason to reconsider the intention to Φ. So far, Temptation Transmission has explained why a temptation to Φ necessarily leads, via a sense-making process, to the appearance of a reason to reconsider the intention not to Φ. This answers the question from Section 2.1 about why agents predict that they will be likely succumb to temptation. When an agent intends not to Φ and is tempted to Φ, there appears to her to be a reason to abandon her intention. She predicts that she will succumb to temptation because she predicts that she will respond to this apparent reason. Temptation Transmission starts with an account of temptation that Holton accepts, and then relies on plausible and widely accepted general principles to establish an explanation of the sort Holton needs to motivate his account of resolutions. Up until this point, then, Holton should be sympathetic to Temptation Transmission. 5.3. Defending Claim 3 The core of the defense of Claim 3 is the claim that rationally reconsidering an intention not to Φ (or what I call Intention 1) requires abandoning the second-order intention not to reconsider that intention (or what I call Intention 2). From this claim, we apply RMRT to establish that an apparent reason to reconsider Intention 1 will ! 203! always lead to an apparent reason to abandon Intention 2. As we saw in Section 5.2, abandoning an intention in a rationally permissible way requires first reconsidering it. This will establish Claim 3: that an apparent reason to reconsider Intention 1 necessarily leads to an apparent reason to reconsider Intention 2. Why should we believe that the only way to rationally abandon Intention 1 is to reconsider the second-order intention not to reconsider Intention 1? We can best answer this question by thinking about what it would be like for an agent to reconsider an intention without first giving up her intention not to do so. Suppose that I intend not to eat any dairy products, and I intend not to reconsider this intention. As I am grocery shopping, I see that my favorite kind of cheese is on sale. While actively maintaining my second-order intention not to reconsider whether to intend to avoid dairy, I go ahead and reconsider anyway. That is, even though I intend not to reconsider the intention to avoid dairy, I begin questioning whether I really want to maintain this intention after all. Such a state seems rationally impermissible, and with good reason—for it is an example of a paradigmatic sort of akratic irrationality. Akrasia can happen at two stages. 196 First, and most frequently discussed in the philosophical literature, there can be an akratic break between what an agent judges best and what she intends to do. Second, there can be an akratic break between what an agent intends to do and what !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 196 See Tappolet (forthcoming) for a helpful taxonomy of philosophical views on weakness of will and akrasia. Rorty (1980) notes that an akratic break can occur at multiple places, including between the formation of an intention and action. Wiggins (1978) sketches a view under which weakness of will has to do with failing to persist with one’s intentions. Mele (1987) treats acting against one’s intentions as one form of weakness of will, cashing it out as a special case of acting against one’s best judgment. ! 204! she actually does. 197 If I intend not to have a second glass of wine with dinner but I do so anyway, I am irrational by my own lights, for I am not successfully carrying out my own plans. When I intend not to reconsider my intention to avoid dairy but then reconsider this intention anyway, I am akratic in precisely the way that I am when I intend to not to have a second glass of wine but then have one anyway. 198 If I am to reconsider my intention to avoid dairy without being guilty of akratic irrationality, I must first abandon any intention I have to avoid engaging in such reconsideration, just as I would have to abandon my intention to refrain from a second glass of wine in order to rationally indulge in another glass. And as we saw in the previous section, rationally abandoning an intention requires reconsidering that intention. This shows us that rationally reconsidering the intention not to Φ requires reconsidering the second-order intention not to engage in such reconsideration. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 197 Holton uses “akrasia” in a narrow sense to refer only to judging that Φing is best, and then failing to intend to Φ. What I am calling second-stage akrasia will not always count as akratic in Holton’s narrow sense, for it might not involve going against your best judgment: you might judge that Φing is best, and then intend to ψ and akratically fail to do so. However, it does not matter whether we refer to this gap between intention and action as akratic. What matters is whether a gap like this is problematic, and (as I shall argue below) Holton seems to presume that it is. 198 Some important qualifications are necessary for a proper understanding of the general norm forbidding akrasia between what an agent intends to do and what she actually does. We do not want to predict that an agent who intends to Φ at time t+1 is akratic because she fails to Φ at some earlier time t. So the proper formulation of the general norm will have to include some reference to tense, or perhaps be formulated as a synchronic rather than diachronic norm. Nor do we want to predict that an agent who intends to Φ but is prevented from Φing by entirely external circumstances acts is guilty of akratic irrationality, which means that the general norm will have to be formulated so as to rule out such cases. Hammering out this general formulation is beyond my present scope. However, we need not worry about ruling out external circumstances for my present purposes. I employ the intention/action anti-akrasia norm with regard only to higher-order mental states: the agent intends not to reconsider her intention, and then reconsiders that intention anyway. Such higher order mental states are entirely internal, and—at least barring the interference of nefarious neuroscientists—are therefore fully in the agent’s control. ! 205! It would be very difficult for Holton to reject this appeal to anti-akrasia norms, because such norms form an important part of his own argument. Recall that Holton claims that a major cause of judgment shift is that agents foresee that they are likely to succumb to temptation, and want to avoid the cognitive dissonance of believing that it would be best to refrain from Φing while believing that they will Φ anyway. The fact that agents feel dissonance when they predict that they will Φ while intending not to is evidence that they recognize that being in this state is irrational. Were it rationally permissible to intend to Φ and not act on that intention, agents would not feel any cognitive dissonance. If they did not feel dissonance, they would not feel pressure to shift their judgments, and we would not need resolutions to block such shift. For ease of reference, I summarize the Temptation Transmission argument in the following table; Intention 1 refers to the agent’s intention to avoid Φing, while Intention 2 refers to the agent’s second-order intention not to abandon Intention 1. First stage Transmission Second stage Why? 1 Temptation to Φ Necessarily leads to . . . Appearance of a reason to Φ Empirical claim 2 Appearance of a reason to Φ Necessarily leads to . . . Appearance of a reason to intend to Φ Intentional action requires an intention 3 Appearance of a reason to intend to Φ Necessarily leads to . . . Appearance of a reason to give up Intention 1 Intention consistency constraint 4 Appearance of a reason to give up Intention 1 Necessarily leads to . . . Appearance of a reason to reconsider Intention 1 Abandoning an intention requires reconsidering it 5 Appearance of a reason to reconsider Intention 1 Necessarily leads to . . . Appearance of a reason to give up Intention 2 Anti-akrasia norm 6 Appearance of a reason to give up Intention 2 Necessarily leads to . . . Appearance of a reason to reconsider Intention 2 Abandoning an intention requires reconsidering it ! 206! The first four stages of the argument answer an independently important question about how temptation leads via a sense-making process to the appearance of a reason to abandon one’s first-order intention. This is a claim that Holton can accept. The last two stages, boxed in black, draw on an anti-akrasia norm that Holton in fact accepts to form an argument that Holton cannot accept: that temptation leads via a sense-making process to the appearance of a reason to abandon one’s second-order intention, as well. The efficacy objection is now back in force: since the temptation to Φ leads to the appearance of an equally strong reason to abandon the first-order intention to Φ and to abandon the resolute second-order intention, the second-order intention cannot do any meaningful work in blocking temptation and preventing judgment shift. 6. Can We Block the Efficacy Objection Another Way? It is important to note the limits of Temptation Transmission. I am not claiming that agents who have formed Holtonian resolutions will never be able to resist temptation and prevent judgment shift. I am simply claiming that means of resistance that Holton proposes—the second-order intention, specifically aimed at preventing reconsideration of the first order intention—is susceptible to the very same pressure that it was designed to resist, and is therefore unreliable. This is compatible with the claim that in certain cases, agents will have other sources of resistance that enable them to avoid reconsidering the second-order intention, and thereby prevent inappropriate judgment shift in some other way. Holton’s argument will still be in trouble if such other means of resistance are sometimes possible on a case-by-case basis, so long as there is no means of resistance that bolsters second-order intentions across the board. ! 207! What might such a broad boost to our second-order intentions consist in? I have argued that a temptation that pushes with strength X on Intention 1 will also push with strength X on Intention 2, such that all else being equal, Intention 2 will be no more effective at resisting judgment shift than Intention 1 alone would be. But if all else is not equal, there might be general, independent reasons for the agent to maintain Intention 2 that do not also apply to Intention 1. This defense requires that these reasons to maintain Intention 2 are not ipso facto also reasons to maintain Intention 1, lest Intentions 1 and 2 be equally well-supported, and therefore equally susceptible to judgment shift. Are there any such reasons? These reasons would have to be apparent to the agent, for reasons to maintain Intention 2 that the agent does not know about will obviously not help the agent maintain Intention 2. What apparent reasons might an agent have for maintaining an intention to avoid reconsidering another intention? One answer is whatever reasons she has for holding that other intention in the first place. That smoking is unhealthy is a strong reason to intend to quit smoking. The unhealthiness of smoking seems to be an equally strong reason not to reconsider the intention to quit. But this strategy of deriving the reasons for Intention 2 from the reasons for Intention 1 will not help, for the reasons supporting Intention 2 would be of the same strength and type as the reasons supporting Intention 1, making the two intentions equally well supported. The reasons for Intention 2 cannot be so general that they necessarily support Intention 1, as well, lest the two intentions remain equally resistant to judgment shift. Nor can the reasons for Intention 2 be too specific. We are seeking reasons that are general enough that they always, or at least very often, apply to Intention 2. For if agents occasionally had strong enough independent reasons to maintain their second- ! 208! order intentions, Holtonian resolutions would succeed at resisting temptation only occasionally, and Holton’s account would remain in trouble. To attain the requisite generality, the reasons for Intention 2 should be independent of the particular content of Intention 1. One candidate for such a reason is the thought that an agent’s having an intention not to reconsider Intention 1 makes her maintaining Intention 1 more likely than if she did not have such a second-order intention. So long as the agent had a general reason to maintain her first-order intentions, she would have reason to form second-order intentions that bolster those first-order intentions. But this proposal is question-begging, for the candidate reason—that the presence of Intention 2 makes maintaining Intention 1 more likely—is precisely what is up for debate. Second, even if this proposal were not question begging, it is not obvious that agents really do have broad reasons to maintain their first-order intentions that are independent of the content of those intentions. In general, we have reason to be resolute in our particular intentions only to the extent that those intentions are good to have. Sometimes particular resolutions ought to be abandoned, as when new information comes to light that entails that the resolution no longer makes sense. If agents had very a strong general reason to remain resolute independently of the content of their particular resolution, they might not be sufficiently responsive to situational factors that override or undermine particular resolutions. We do not seem to have a strong candidate for a reason for Intention 2 that is narrow enough so as not to always support Intention 1 as well, but general enough to support Intention 2 across the board. If we cannot find general reasons that support Intention 2 but not Intention 1, perhaps we can bolster our second-order intentions by appealing to willpower. It is not entirely clear exactly how Holton takes the willpower mechanism to work. But suppose ! 209! willpower functions as a source of friction or stickiness once a decision has been made and an intention formed: the resolute agent intends to Φ, intends not to reconsider this intention, and employs willpower to buckle down and follow through on each of these intentions, even if buffeted by temptation or other barriers to action. Might willpower enable an agent to resist reconsidering Intention 2, even when temptation presents a reason to reconsider it that is just as strong as the reason to reconsider Intention 1? Suppose that an agent who intends to Φ employs willpower to degree X to sustain her intention. If she is subject to a temptation-based reason of strength X+10 to reconsider this intention, her willpower may fail her. To prevent this, she can form a second-order intention, and employ willpower to degree X to sustain that intention. If she is not faced with a similar reason of strength X+10 to reconsider, this second-order intention is likely to remain effective where the first-order intention failed. But what if temptation transmits, and the agent is faced with a reason of strength X+10 to reconsider the second-order intention, as well? Then the second-order intention is likely to fail, as well. We have no good reason to think that an agent’s degree of willpower will as a rule be any greater or more effective at the second-order level than it is at the first-order level. For both intentions are aimed at the same goal, and the degree of willpower that the agent is willing and able to exert for any given intention is presumably a function of how important the goal is to her. If the amounts of temptation-based pressure and the degrees of willpower employed are the same at each level, then the Intention 2/Intention 1 pair will not be any more effective than Intention 1 alone would be. ! 210! The Temptation Transmission argument will retain some force even if we grant that agents are on a case-by-case basis able to increase their degree of willpower at the second-order level, and are therefore on a case-by-case basis able to remain firm in their Holtonian resolutions in spite of the transmission of temptation. For this represents an occasional further source of resistance to judgment shift that is neither systematic nor fully general. In the next section, I turn instead to an account of resolutions that defuses Temptation Transmission entirely, rather than circumventing it on a case-by-case basis. 7.! An Alternative Account of Resolutions If Temptation Transmission cannot always and easily be accommodated by Holton’s view, perhaps it can be blocked in another way. One strategy for blocking the transmission of temptation can be dismissed immediately. Adding a third-order intention to the resolution—an intention not to reconsider your second-order intention not to reconsider your first-order intention not to Φ—will not suffice. This is because we can draw on RMRT and the anti-akrasia norm to apply Temptation Transmission mutatis mutandis to a third-order intention, as well as to intentions of progressively higher orders. Nor do we want to give up on the plausible fundamental background principles that underlie Temptation Transmission. Rather, the Temptation Transmission argument shows us that we must appeal to a mental state of a different sort to block temptation—specifically, to a mental state that is not subject to an anti-akrasia norm. Desires can successfully play this role: Second-Order Desire Account (SODA): resolving to Φ involves intending to Φ, and desiring not to reconsider the intention to Φ. 199 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 199 A precursor to SODA can be found in Kovach and Fitzpatrick (1999). They assess whether resolutions might be understood in terms of broad second-order desires, which they take to be ! 211! SODA is meant to lay out necessary conditions on resolving to Φ, and does not purport to lay out a full set of sufficient conditions. According to SODA, an agent who resolves to Φ must at least intend to Φ, and must desire that she carry out that intention, even in spite of contrary inclinations. Essentially, an agent who resolves to Φ must care about whether she maintains her intention to Φ. In the rest of this section, I explain why SODA is appealing as an alternative to Holon’s view. First, SODA does not fall prey to Temptation Transmission. Second, it can explain data about resolutions that intention-based views may have a hard time accommodating. I then address worries that Holton raises about whether a view like SODA can adequately capture the phenomenology of making and keeping resolutions. Finally, I address a worry about whether SODA problematically entails that it is not !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! desires “the content of which makes some essential reference to some first-order desire of the agent, without hinging merely on the presence of such desire” (166). Kovach and Fitzpatrick claim that a view from Bigelow, Dodds, and Pargetter (1990) that “analyze[s] temptation as the co-existence in an agent of a first-order desire to X and a second-order desire that the first-order desire to X not be operative” (167) is on the right track. But they argue that Bigelow et al. are wrong to emphasize the structure, and that the content of the resolution is what really matters: they offer a process-specific second-order desire account, according to which the process by which a resolution is implemented is essential: “resolutions motivate not because they are an intrinsically motivating type of mental state,” but “in virtue of being projects which involve certain constitutive processes” (168). These are processes of self-control, or aligning one’s actions with one’s judgment, and self-transformation, or changing one’s fundamental desires. By contrast, taking a control pill that managed ones actions would not involve the right sorts of processes, and therefore would not be an instance of a resolution. Kovach and Fitzpatrick are right that a resolution can have two kinds of content, because it might aim at self-control or self-transformation. However, I maintain that the structure of the resolution is essential, for second-order desire component plays an ineliminable role in resisting Temptation Transmission. Moreover, contra Bigelow et al., I claim that the first-order element in SODA is an intention to Φ; we can grit our teeth and resolve to do things for which we have no first-order desires. ! 212! possible to resolve at will, and whether this understanding of resolutions is problematic for the Conditional Resolution Account of promising. 7.1.! Why SODA Avoids the Temptation Transmission Argument SODA will not fall prey to the same Temptation Transmission argument that troubles Holton’s view. For the argument to be effective against SODA, we would need to employ an analogue of Claim 3 that was about desire rather than intention: Claim 3*: Reconsidering any desire to avoid reconsidering the intention not to Φ is necessary for rationally reconsidering the intention not to Φ. If Claim 3* were true, we could use RMRT to argue that the appearance of a reason to reconsider the intention not to Φ necessarily leads to the appearance of a reason to reconsider your desire not to engage in such reconsideration. But Claim 3* is not true, because desires are not subject to the same rational norms forbidding akrasia that intentions are subject to. It is not necessarily irrational to desire that you Φ while nevertheless failing to Φ, or to desire that you refrain from Φing while Φing anyway. We all have very many desires that we do not act on, and we display no irrationality in doing so. I can desire to spend the day at the beach, and rationally fail to do so because I have to spend the day at work instead. Similarly, I might desire to avoid an unpleasant task such as filing my taxes, but I display no irrationality in filing them anyway. It can be rational to do what you desire not to do in a way that it is not rational to do what you intend not to do. Structurally speaking, making the move from second-order intention to second-order desire is exactly what we need to avoid Temptation Transmission. For the argument’s success depends on the existence of a second-order intention that is subject to an anti- ! 213! akrasia norm requiring the agent to avoid the state of intending to do Φ while not Φing. If we replace this intention with a mental state like desire that is not subject to the same anti-akrasia norm, we cannot generate the same argument. 7.2.! How SODA explains variations in the strength of resolutions Resolutions come in degrees; some resolutions are stronger than others. A resolution is stronger to the extent that an agent takes it more seriously, and a greater excuse is needed to license breaking it. My resolution to read Don Quixote is not very strong, because this is a nice but inessential goal, whereas my resolution to pursue my research project is very strong, because this is an important part of my career plan. Appealing to desire as a necessary component of a resolution gives us an easy and efficient explanation of how resolutions can vary in strength. For desires clearly come in degrees; some desires are stronger than others. We can explain the strength of a resolution as a direct result of the strength of the agent’s desire to avoid reconsideration: the stronger the desire, the stronger the resolution. Since intentions do not come in fine-grained degrees, Holton’s intention-based account cannot offer the same sort of explanation. 200 Explaining the strength of resolutions in terms of the strength of the agent’s desire to avoid reconsideration makes intuitive sense, as well. In general, the degree to which an agent is resolute in Φing seems to depend not on how strong the temptation to Φ is, but on how much the agent cares about whether she Φs. Suppose I resolve not to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 200 In Chapter 2 of Willing, Wanting, Waiting, Holton outlines a notion of ‘partial intention.’ This is not a fine-grained strength or degree of intention, in the way that there are fine-grained strengths of desire. Rather, “an intention to F is partial iff it is designed to achieve a given end E and it is accompanied by one or more alternative intentions also designed to achieve E” (36). ! 214! eat any cookies at a party tonight, but I really don’t care very much about whether I keep this resolution. Even cookies that are not very tempting (say, store-bought sugar cookies) will be likely to sway me to break my resolution. But suppose I resolve not to eat any cookies at a party tonight, and I care very much about whether I keep this resolution. In such a case, it seems that even extremely tempting cookies (say, homemade French macarons) will not be very likely to sway me to break my resolution. What explains the difference between these two cases is that the resolution is stronger in the second case, when my desire to avoid reconsideration is stronger. This variation in strength of desires gives SODA another explanatory advantage over Holton’s view. One might be worried that temptation puts pressure on an agent’s resolution not just by making it appear that there is a reason to reconsider her resolute intention, but also by creating a desire in her to reconsider this intention. After all, desire aims at the good, and the tempting thing presents itself as good. SODA can explain when resolutions are able to resist desire-based temptations of this sort. If temptation ever works in this way, agents are faced with conflicting desires: a temptation-based desire to reconsider, and a resolution-based desire not to reconsider. Resolutions sometimes fail; neither Holton nor I want to claim that our respective accounts of resolutions enable agents to resist in every instance. According to SODA, whether a given resolution is effective when faced with a conflicting temptation-based desire will depend in part on how strong the resolution-based desire is, and whether it is strong enough to outweigh the temptation-based desire. It would not be as easy and straightforward to weigh the strength of a second-order intention against that of the temptation-based desire. ! 215! 7.3.! Why SODA Can Capture the Phenomenology of Resolutions SODA is a natural option to consider once the possibility of a second-order account of resolutions has been raised. The view is no surprise to Holton; in Willing, Wanting, Waiting, Holton considers and rejects a higher-order desire account. Does Holton’s rejection of higher-order desire views bring to light problems of comparable gravity for SODA as the efficacy objection creates for Holton’s view? Holton considers the following higher-order desire view: suppose you resolve to Φ and “you have a strong desire to be resolute: a strong desire to stick to your resolutions.” 201 This could enable the agent to resist temptation so long as “when the date for implementing the resolutions comes, provided that your desire to be resolute is stronger than your desire to [give up], you have a desire-driven way to give up.” Holton calls this the further desire approach, presenting a problem for it in a footnote: The further desire approach seems to involve attributing to the strong-willed agent a desire for resoluteness that approaches a fetish. It is surely crazy to want to be resolute for its own sake, especially if, as a result of judgement shift, the agent comes to believe that the resolute course of action is the less desirable. 202 It might indeed be problematically fetishistic to want to be resolute simply for the sake of resoluteness. But according to SODA, agents do not desire to be resolute in general. Rather, they desire to be resolute with regard to specific intentions, which we can assume they take themselves to have independent reasons to act on. There is no reason to think that a desire to be resolute is fetishistic if it is restricted to particular cases in this way. A more serious problem Holton finds with the further desire approach is that it “completely misrepresents the phenomenology of strength of will:” !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 201 Both quotations in this paragraph from from 2009, p. 116. 202 ibid, p. 118. ! 216! If these accounts were right, then sticking to a resolution would consist in the triumph of one desire (the stronger) over another. But that isn’t what it feels like. It typically feels as though there is a struggle. One maintains one’s resolutions by dint of effort in the face of the contrary desire . . . by and large, maintaining strength of will requires effort. 203 There is an important insight here. Sticking to a resolution in the face of temptation indeed feels like a struggle, and maintaining strength of will requires effort. But rather than lead us to dismiss SODA completely, this worry should help us realize that desire- based accounts of resolutions need to be more nuanced than they might first appear. An account like SODA need not be committed to a simple Humean picture in which desires do all of the motivating work and the agent is simply along for the ride, at the mercy of whichever desire is strongest. For SODA does not claim that a second- order desire is by itself sufficient to bring an agent to act resolutely. Rather, it claims only that the second-order desire is necessary. This is entirely compatible with it being the case that serious mental effort on the part of the agent is also necessary for resolute action. Even a distinct faculty of willpower of the kind Holton describes is compatible with SODA, so long as desire plays an essential role in triggering the agent to employ that willpower. Holton grants that this is often the case. In a discussion about willpower, he notes that whether an agent revises a resolution “will depend on, amongst other things, the strength of their desire to maintain those resolutions in particular, and the strength of their desire to maintain their resoluteness in general.” 204 He also notes that “to be effective, these desires need not be the strongest. If the agent’s willpower is sufficiently !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 203 ibid, p. 118. 204 Both quotations in this paragraph are from ibid, p. 133. ! 217! strong, a weak desire to be resolute might be all that is needed to keep it in place when it wavers in the face of a strong contrary desire” stemming from temptation. SODA does not require that the desire to be resolute is the strongest desire that the agent has, capable of overwhelming all of her other desires. Only one claim about desire is essential for SODA: that the desire present in resolutions is not subject to the anti-akrasia norm that drives Temptation Transmission. Any account of desire that meets this constraint can block the transmission of temptation. Since intentions of any order are subject to anti-akrasia norms, they cannot block the transmission of temptation all by themselves. Something more is needed; I have suggested that something is desire. The primary point of disagreement between Holton’s view and SODA is not about whether desire can play a role in helping agents to remain resolute; Holton grants that it can. Rather, SODA and Holton’s view differ over the role played by second- order intentions. Holton claims that they are both necessary and sufficient for maintaining resolutions. I have argued that they are insufficient, because they leave resolutions vulnerable to the efficacy objection. By more thoroughly developing this objection, we have not only seen why temptation transmits in such a way that Holton’s view fails. We have also fleshed out the previously under-explained mechanism by which agents succumb to temptation, and discovered that what is needed to avoid the objection is a mental state like desire that is not subject to anti-akrasia norms. 7.4.! Must SODA accommodate resolving at will? One might worry that SODA is implausible if we can always form resolutions at will, since we cannot form desires at will and resolutions require desires. However, I do not think we can in fact form resolutions at will when we lack the necessary desires. ! 218! For example, imagine that you do not care at all about whether your intention to avoid smoking remains intact. That is, you have intended to quit smoking, but could not care less about whether you maintain or abandon this intention. How could you form a resolution in such a case? It is plausible to think that an agent who lacks any desire to remain consistent in her intention to Φ will be unable to genuinely resolve to Φ. In the majority of cases, an agent who attempts to form a resolution to Φ will desire to maintain the intention to Φ; she would not be attempting to form a resolution about Φing in the first place if she did not have such a desire. Granted, cases in which an agent attempts to form a resolution but cannot do so because she lacks the requisite desire are possible in theory; perhaps an agent suffering from depression intends to Φ and aims to turn this intention into a resolution, but cannot bring herself to care about whether the intention is effective. But we do not expect agents in such situations to be able to successfully form resolutions. Nor are such cases at all common, at least among psychologically well-functioning agents. I’ve argued that because resolutions require desires, we cannot form resolutions at will in every scenario. If sincerely promising requires resolving, it follows that we cannot sincerely promise in every scenario. Is this a problem? It doesn’t seem to me to be. For keep in mind that forming a resolution is not necessary for promising as such. Rather, it is necessary only for sincerely promising. I can of course insincerely promise my dentist that I will floss my teeth, even if I cannot successfully resolve to do so because I lack any desire about whether to floss or maintain an intention to floss. SODA and claim that promises express conditional resolutions jointly entail that I cannot sincerely promise my dentist that I will floss my teeth. But this is not a bullet- ! 219! biting response. To the contrary, this is what we should expect. For how could my promise to floss be sincere? Promising to floss expresses that I have a serious and stable plan to floss. But my total lack of concern about flossing (as well as about the efficacy of any intentions I might form with regard to flossing) means that any plans I have to floss are not at all serious or stable. I am not likely to succeed, and am very likely to abandon my plan as soon as I experience minor discomfort. And so I cannot truthfully convey that I have a serious and stable plan to floss. In this and the previous two chapters, I aim to have argued for the two main claims that I was assuming while laying out MSF in Chapter 3: that promising expresses a conditional resolution, and that resolving commits you to acting, all else being equal. In the final two chapters of this dissertation, I will explore an attractive positive upshot of MSF: the claim that it can explain both when we are permitted to break our promises, and when we are permitted to abandon conditional resolutions. ! 220! CHAPTER 7: THE ESCAPE CONDITIONS ON PROMISES 1. Introduction Up until this point in the dissertation, I have motivated the need for a new theory of promissory obligation by framing the Problem of Marginal Cases, introduced the Mental States First theory as a view with the right sort of structure to give a sufficiently general and distinctively promissory account that accommodates both sincere and insincere cases, and defended the particular claims that MSF relies on about the relationship between promises and resolutions and about the nature of resolutions. In these last two chapters, I will outline a positive upshot of MSF: that the view can explain both when and why we are permitted to break promises. As Henry Sidgwick writes in The Methods of Ethics, “reflection seems at least to disclose a considerable number of qualifications of the principle [that we are morally required to keep our promises]; some clear and precise, while others are more or less indefinite.” 205 Because promises are pro tanto moral obligations, we are not always morally required to keep them, all-things-considered. Sometimes, it’s obvious when breaking a promise is permissible; if saving a man’s life requires that I break a promise to you to meet you for happy hour drinks, all-things-considered I ought to skip our meeting. And sometimes, it’s obvious that breaking a promise is not permissible; if I fail to meet you for happy hour because I’ve been in my pajamas all day and I don’t feel like getting dressed, I act immorally. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 205 1907, p. 304 ! 221! But there is a wide range of cases that are not so clear-cut. What if my car unexpectedly breaks down, and it would be expensive for me to meet you by cab and inconvenient for me to meet you by bus? What if I agreed to meet you for happy hour because I believed the bar to have great prices on local craft beer, but it turns out that they only serve generic beer at a steep mark up? My aim in this chapter is to clearly articulate the types of situation in which it is morally permissible for an agent to break a promise. 206 As a term of art, I call such situations “escape conditions.” Each section of the chapter outlines a different escape condition. In the next chapter, I will lay out the kinds of situations in which it is rationally permissible for an agent to abandon a resolution. We shall see that the escape conditions on promises and on resolutions are strikingly similar. MSF tells us why this is so: the escape conditions on promises are roughly the same as those on resolutions because promising involves the expression of a conditional resolution. Just as promissory obligation can be derived from the underlying mental state expressed by the promise, so can an understanding of kinds of cases in which promise-breaking is permissible. 2. Unexpected More Important Conflicting Reason or Obligation (MICRO) In a discussion of the way changes in duty can arise due to changes in circumstance, Cicero offers the following example: If you have made an appointment with anyone to appear as his advocate in court, and if in the meantime your son should fall dangerously ill, it would be no breach of your moral duty to fail in what you agreed to do; nay, rather, he to whom your promise was given would have a false conception of duty, if he !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 206 I speak of “breaking” promises for the sake of convenience, although there is a danger that some readers might take this terminology to presuppose wrongdoing. If you dislike talk of permissibly breaking promises, substitute with permissibly “failing to fulfill” promises, instead. ! 222! should complain that he had been deserted in his time of need. 207 Cicero is right; it is quite clear that all-things-considered, you do not make a moral error if you fail to appear in court. To the contrary, you would be making a moral error if you left your dangerously ill son at home and went to court instead. This is because unexpected and more important conflicting moral obligations such as this are obviously escape conditions on promising. What might not be so obvious is how broad this kind of escape condition can be. Sometimes, an unexpected and highly important non-moral obligation can override a pro tanto promissory obligation and count as an escape condition. For example, suppose you are a struggling actor who has promised your roommate that you will buy milk on your way home tonight. While en route home, your agent calls offering you a once-in-a-lifetime chance to audition for a major director, so long as you come to the audition immediately. If you go to the audition, you will not be able to get milk before the store closes. It seems clear that you do not behave immorally all-things-considered if you go to the audition instead of buying the milk. This is because your massively strong prudential obligation to pursue your most important career goals trumps the weak moral obligation to keep an unimportant and low-stakes promise. Furthermore, it is not the case that only a more important conflicting obligation can outweigh a promise. For it is plausible that there could be a very important reason for an agent to refrain from Φing that does not rise to the level of obligation, but that nevertheless outweighs an unimportant promise to Φ. For example, suppose Anthony lives only half a mile from the airport. He doesn’t like to walk home dragging his suitcase, but it’s not overly burdensome for him to do so. Betsy, who lives five miles !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 207 De Officiis, Book 1, Section 32. ! 223! from the airport, has promised Anthony that she will pick him up. Shortly before Anthony’s flight arrives, President Obama makes a surprise visit to town, causing massive road closures and traffic jams. The roads are so backed up that it would take Betsy an hour to drive to the airport, whereas it would take Anthony only 10 or 15 minutes to walk home. Avoiding the unexpected cost of wasting time and gas by sitting in traffic for an hour is not easily construed as an obligation. But it seems to be a strong enough reason to function as an escape condition on Betsy’s promise to pick Anthony up, especially since it is so easy for Anthony to walk home. Finally, there might be multiple conflicting reasons or obligations in play, all of which must be balanced to determine whether the promise must be kept on the whole. Consider the following delightful example from Cicero: Suppose that a millionaire is making some wise man his heir and leaving him in his will a hundred million sesterces; and suppose that he has asked the wise man, before he enters upon his inheritance, to dance publicly in broad daylight in the forum; and suppose that the wise man has given his promise to do so, because the rich man would not leave him his fortune on any other condition; should he keep his promise or not? I wish he had made no such promise; that, I think, would have been in keeping with his dignity. But seeing that he has made it, it will be morally better for him, if he believes it morally wrong to dance in the forum, to break his promise and refuse to accept his inheritance rather than to keep his promise and accept it— unless, perhaps, he contributes the money to the state to meet some grave crisis. In that case, to promote thereby the interests of one's country, it would not be morally wrong even to dance, if you please, in the forum. 208 There are three conflicting obligations at stake: (1) a moderately strong obligation to keep the promise to the millionaire, (2) a very strong moral obligation to refrain from dancing publicly in the forum, and (3) a very strong moral obligation to promote the interest’s of one’s country by giving a large amount of money to the state. If only the first and second obligations were present, the second would outweigh the first. But the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 208 ibid, Book 1, Section 93. ! 224! weight of the third obligation combined with the weight of the first outweighs the second, making it the case that the best thing to do all-things-considered is dance in the forum, receive the money, and donate it. These cases show us that sufficiently more important conflicting reasons or obligations (MICROs, for short) of any kind that unexpectedly arise after a promise has been made can function as escape conditions, making it the case that the promisor is not all-things-considered obligated to act as promised. Since the original promise was valid, the promisor owes a directed, pro tanto obligation to the promisee, which she fails to satisfy. Because there is an escape condition, this failure is morally permissible, all-things-considered. The promisor still owes reparatory duties of some sort; you should buy milk for your roommate tomorrow morning, and Betsy must do something like warn Anthony that she cannot pick him up, and offer to give him a ride the next time he needs one. 2.1. The importance threshold How important must a MICRO be in order to count as an escape condition? The answer to this question will vary from one promise to another. We can best see how this works by considering an example. Suppose you promise your friend Christa that you will fly across the country to attend her wedding next summer. When you make the promise, you have a good job and can easily afford to purchase a plane ticket. However, you lose your job several months before Christa’s wedding. Purchasing the flight would now be a major financial burden for you. Whether this burden counts as an escape condition will depend on the circumstances of your promise. ! 225! Suppose Christa is a college friend you have not spoken to in years. You were never extremely close, and you do not have many mutual friends. There will be 200 people at the wedding, and you will not be missed very much by anyone if you do not attend. In this version of the case, it seems like your unexpected financial burden is indeed an escape condition. Now suppose instead that Christa has been your best friend since elementary school. She has been telling you for months that she is looking forward to seeing you at the wedding. You have many mutual friends, all of whom are looking forward to seeing you, as well. Christa can afford to invite only 50 people to her wedding reception, and you know that she has offended some of her distant relatives by not inviting them in order to invite close friends like you. In this version of the case, it seems like your unexpected financial burden is not an escape condition. The importance threshold of a promise is a term of art that specifies how important a MICRO must be in order to count as an escape condition on that promise. In the second version of the wedding case, the importance threshold is much higher than it is in the first version of the case. This is why the same financial burden counts as an escape condition in the first version but not in the second. To figure out what accounts for this difference, we need to better understand what features are responsible for a promise’s importance threshold. What causes one promise to have a higher importance threshold than another? The most obvious feature that affects the importance threshold is how important the promise is to the promisee. In general, the more the promise matters to the promisee, the more important a MICRO will have to be to count as an escape condition. How much a promise matters to the promisee is determined by a number of factors, such as: to what extent is the promisee relying on the promisor to perform the promised ! 226! action? Has the promisee sunk any costs in expectation of the promise being fulfilled? If the promise is part of a reciprocal bargain, has the promisee already performed her share? To what degree, if any, does the promisee care whether the promise is kept apart from considerations about cost and reliance? How close or morally significant is the relationship between the promisor and the promisee? etc. A promise that someone is relying on will have a higher importance threshold than a promise that would have little to no impact on the promisee’s life if it were broken; this is why the importance threshold is lower for your promise to buy milk for your roommate than it would be for a promise to drop your roommate’s time-sensitive rent check in the mail. If the promisee has already performed her part of a reciprocal bargain—if Betsy is scheduled to pick up Anthony from the airport as part of a bargain they made to trade airport rides, and Anthony has already given her a ride—then the importance threshold will be higher than if Betsy’s promise to pick Anthony up were entirely gratuitous. Promises to your closest loved ones will usually have higher importance thresholds than promises to casual acquaintances or to strangers, because your relationship with the former is more valuable than with the latter. It is often obvious from context how important a promise is to the promisee. But since the promisor cannot always or infallibly know how much the promise matters to the promisee, the determination of the importance threshold may have to involve the promisor’s best educated guess. If for all you can be reasonably expected to know, it does not matter very much to your roommate that you bring home milk, the importance threshold will be fairly low—even if in fact, and contrary to all reasonable expectations, your roommate cares a great deal about you bringing home milk. ! 227! Letting the promisor’s interests fix the importance threshold instead would have problematic implications. If you have merely intended to attend Christa’s wedding, how much you feel like going will affect whether you ought to go if an unexpected financial burden arises. But promising Christa that you will go changes the situation: whether you dislike attending weddings does not matter for determining whether an unexpected financial burden counts as an escape condition. Rather, what matters for the importance threshold is how important the promise is to Christa—which is fixed by factors like how much Christa cares about your presence there, how close the two of you are, etc. The normative weights of promises would be highly unstable if they fluctuated with the promisor’s interests in the way that the weights of mere intentions do. As the wedding draws nearer, the guests might find themselves faced with better offers or conflicting obligations. If whether each guest was morally obligated to attend permissibly varied on the basis of how much he or she happened to value attending the wedding at any particular moment, it would be very easy for guests to create escape conditions for themselves and find themselves off the moral hook for attending. Fixing the importance threshold in terms of how much Christa values each guest’s attendance avoids this difficulty. This is not to say that the promisor’s interests never play any role whatsoever in determining whether a given conflicting reason counts as an escape condition. For the promisor’s interests can fix the strength of the conflicting reason that competes with the promise. If a promisor highly values her career, a career based MICRO will be very strong; if a promisor does not care at all about her career, a career based MICRO will have little weight. This changes the strength of the conflicting reason. But it does not ! 228! affect how strong a potential conflicting reason must be in order to count as an escape condition—which is to say that it does not affect the importance threshold. 2.2. Implicit conditionality Another factor that can affect the importance threshold is the implicit conditionality of a promise. I have been arguing that all promises are explicitly conditional on the acceptance of the promisee. Promises might be explicitly conditional on other factors, as well; I might promise you that I will go jogging with you tomorrow morning, only on the condition that it doesn’t rain—in which case, I form a resolution that is conditional both on your acceptance and on dry weather. Many promises also have implicit conditions built into them that help determine what kinds of conflicting reasons can cross the importance threshold. For example, suppose that I promise to attend your band’s show on Friday night. I have a tendency to go on dates on Friday nights, and have skipped many of your Friday night gigs in the past because of this. One reason I promised you that I would go to the show (rather than just telling you that I planned to be there) was to assure you that I would not bail on you for a date as I have in the past. If a highly desirable partner asks me out on a date for Friday night, this will not be an important enough MICRO to count as an escape condition. This is because it was implicit in my making the promise that having a Friday night date would not excuse me from going to your show. In this context, my having a date is a conflicting reason of the wrong sort to count as an escape condition, even if my attendance at the show doesn’t matter much to you. But if a MICRO of a truly unexpected sort arose—of a sort that was not implicit in our mutual understanding of what keeping my promise entails—then an escape ! 229! condition might be in place. Suppose that I typically attend all of your band’s shows, and I promise you that I will go to your show this Friday. Lately, I have been struggling to recover from a difficult breakup. You have been relentlessly encouraging me to go out on dates whenever possible, but I have been very reluctant. A highly desirable partner asks me out on a date for Friday night, and for the first time in months I am interested in going. In this case, it is implicitly assumed in my making of the promise that my wanting to go on a date can override my promissory obligation to you and excuse me from going to your show. So my having a date is a reason of the right sort to count as an escape condition in this scenario, even if my attendance at the show would otherwise be very important to you. The implicit conditionality involved in a particular promise need not be explicitly considered in advance by the parties to the promise. Common social understanding about what obligations of different sorts typically involve can do a lot of work in fixing what counts as an implicit condition on a promise. For example, suppose you promise me that you will attend a departmental party on Friday. While I would enjoy your company, I will have other friends at the party, and would enjoy myself even if you did not attend. Your car stops working on Friday afternoon, and the only way you can get to the party is by hiring a cab for $50. In this situation, it seems that an escape condition is in place, and it is permissible for you to skip the party. But what if your car is in working order, and on Friday afternoon you are presented with an unexpected opportunity to earn $50 by working during the party? In this version of the case, it seems that an escape condition is not in place, and it is not permissible for you to ! 230! skip the party unless you obtain my release. 209 General shared assumptions about what social obligations entail account for this difference. When someone promises to be somewhere, we typically assume that she will be able to travel there in her typical way. A circumstance that flouts that assumption—say, a car breaking down, or transit workers going on strike—is an unexpected complicating factor that must be taken into account as a possible escape condition. But when someone promises to be somewhere, we do not typically assume that she will go only if she has no better options (such as picking up extra money at work.) And so that sort of circumstance is not the kind of thing we must take into account as a possible escape condition. 210 I have not offered a precise metric for assessing a promise’s importance threshold that will yield a clear and determinate answer in all cases. But this is what we should expect. Morality is often complicated, messy, and highly dependent on individual circumstances. Promises are no exception to this. Whether a particular MICRO counts as an escape condition is a difficult question that will need to be tackled on a case-by-case basis. Up until this point in the discussion, I have been assuming that the initial promise is such that the agent ought to keep it if there are no MICROs in place. But there are many circumstances in which a valid promise is itself flawed in some way that !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 209 Thanks to Shyam Nair for suggesting this pair of cases, and for helpful discussion of them. 210 Conditions that are specific to particular situations could override these general shared assumptions. If the party you are going to is your grandmother’s 100 th birthday party, and it would break your grandmother’s heart if one grandchild wasn’t there, having to pay $50 for a cab would not be an important enough conflicting reason to license escape.! ! 231! entails that it does not yield an all-things-considered obligation, apart from any conflicts with other reasons or obligations. I turn next to such cases. 3. Impossibility Suppose Diego promises his mother Emily that he will bring her donuts from her favorite shop the next time he goes to visit her. The night before Diego’s visit, there is a fire in the donut shop, and it is forced to close. So Diego is unable to get the donuts for his mother, even if he wanted to. Has Diego done anything morally wrong in breaking his promise to Emily? It seems obvious that he has not, and it would be unfair to blame him for failing to buy donuts that are impossible to get. This does not change the fact that Diego owes a reparatory duty of some sort to Emily when he cannot keep his promise, such as bringing her donuts from another shop. For the promise to bring donuts was a valid promise that yields a pro tanto obligation, even though the fact that the promise has become impossible to keep makes it the case that Diego is not all- things-considered obligated to keep it. This example shows us that impossibility is an escape condition on promises. If you make a promise and then discover that keeping it is impossible—either because it was impossible all along and you just realized this, or because the situation has changed unexpectedly such that keeping the promise has become impossible since you first made it—you are not morally required to keep it, all-things-considered. Claiming that impossibility is an escape condition on promising requires claiming that it is possible to be pro tanto obligated to do impossible things; we can assume that ought implies can for all-things-considered obligations. It is independently plausible that we can be pro tanto obligated to do the impossible. By their nature, pro ! 232! tanto obligations are such that they can conflict, and we are unable to keep all of them all of the time: you can be pro tanto obligated to do A, and pro tanto obligated to do ~A, when doing both A and ~A is impossible. It is for you possible to do A, and possible for you to do ~A. But if you do A, then ~A becomes impossible. And it seems that doing A does not thereby negate your pro tanto obligation to do ~A, or cause it to disappear—even though ~A is now impossible. 211 Furthermore, if pro tanto obligations aggregate—if a pro tanto obligation to do A and a pro tanto obligation to do ~A yields a pro tanto obligation to do (A & ~A)—then we are frequently faced with pro tanto obligations to do the impossible. If we cannot be pro tanto obligated to do the impossible, then impossibility will be an invalidating condition rather than an escape condition. However, this will not undermine ability of MSF to explain the escape conditions on both promises and resolutions, so long as impossibility is an invalidating condition on resolutions, as well. It is important to be cautious about how we constrain what an agent “can” do. We do not want agents to be able to bring about an escape condition on a promise by deliberately making it the case that a promise they have made becomes impossible to keep. For example, suppose Diego promises his mother in the evening that he will buy her donuts from her favorite shop the following morning. Overnight, Diego lights the donut shop on fire. When he wakes up, it is impossible for him to get the donuts, because the shop is burned down. Diego is not off the moral hook, since this impossibility is self-imposed—had he not burned the donut shop down, he would have been able to buy the donuts. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 211 Julia Driver (2011) offers a helpful and more detailed argument for this point. ! 233! In the sense of “can” that tracks moral blameworthiness, Diego could have gotten the donuts. Instead, he made a deliberate decision that ruled this out. So it is fair to hold him fully morally responsible for his failure. Impossibility only functions as an escape condition if keeping the promise is made impossible due to factors beyond the agent’s control and for which she is not blameworthy. Similarly, an agent who is secretly drugged and who is therefore incapacitated while driving home is not morally blameworthy for an accident she causes. But an agent who makes an informed decision to incapacitate herself by taking drugs and then driving home is responsible, for this is something she could have easily avoided. What about cases in which an agent knowingly makes a promise that is impossible to keep? Making a promise that you take to be impossible from the get-go, you violate the aptness condition on promising that we discussed in Chapter 3; it is inappropriate to promise to do something that you do not believe you can do. Are such inapt promises subject to an escape condition? Suppose Diego calls the donut shop, and learns that they are closed for renovations. He then promises Emily that he will bring her donuts from that shop, even though he knows that this is impossible. Is Diego morally obligated to keep his promise, all-things-considered? It might be tempting to claim that he is. After all, Diego seems blameworthy for failing to get the donuts in this case, in a way that he does not seem to be when the impossibility arises unexpectedly. But how could Diego be obligated, if all-things-considered obligation implies can, and he can’t? We can more clearly draw out our intuitions about this by considering a pair of cases. In Case 1, Diego insincerely promises Emily that he will bring her donuts, while fully aware that this is impossible. Diego clearly commits a moral wrong when he makes this promise, for he is deliberately deceiving Emily and ! 234! cultivating false expectations in her. As with unforeseen impossibility, Diego owes reparatory duties; he is morally obligated to give Emily sufficient warning that the promise cannot be kept, and to attempt to make it up to her in some other way. He commits a moral wrong if he does not do this. Does Diego commit another moral wrong when he actually fails to buy the donuts? He does not. To see why, compare this to Case 2, in which Diego calls the donut shop, hears the voicemail state that the shop is closed for repairs, and then insincerely promises Emily that he will bring her donuts anyway. Diego calls the shop again, and an employee picks up the phone and informs him that the voicemail is outdated and the shop has now re-opened. Diego fails to buy Emily the donuts anyway, even though doing so is now possible. Diego’s actions in Case 2 seem worse in a different way than his actions in Case 1. But promising insincerely is equally bad in each case. What explains the difference in our intuitions is that Diego commits a second moral wrong in Case 2, but not in Case 1. In Case 2, Diego is given the opportunity to drastically mitigate the badness of making the insincere promise by buying Emily the donuts after all. If he did this, his initial wrong would seem quite minor; he erred in making an insincere promise, but Emily was not harmed by it because he kept the promise after all. If Diego doesn’t do this, he commits a second wrong. This shows us that initial impossibility functions as an escape condition even in a case where the agent knowingly makes an impossible promise. In Case 1, we have the intuition that Diego is blameworthy because his promise is inapt and because he made a deceptive and insincere promise, and not because he is blameworthy for failing to do the impossible. ! 235! 4.! Initial Immorality Suppose that Claudio has gravely wronged Beatrice’s dear friend Hero. As proof of his loyalty to her, Beatrice asks Benedick to promise her that he will kill Claudio in revenge, and he agrees. Clearly, Benedick is not morally obligated to follow through with this promise. Agents are not morally obligated all-things-considered to keep immoral promises, and it would be absurd if promising to murder someone sufficed to make murdering someone morally required. This shows us that immorality is an escape condition on promises. This is so whether the promise is immoral in itself—like promising to murder someone—or whether the promise is immoral because it is an otherwise permissible action that conflicts with a prior moral obligation. If Benedick has already, say, agreed to meet Claudio for lunch at noon, and then promises to meet Beatrice at noon tomorrow, one of these promises is immoral to keep, and an escape condition is in place. 212 (Recall that in Chapter 1 I argued that, contrary to what some have claimed, immoral promises are valid promises that yield pro tanto obligations, which are overridden by the immorality of the act.) In the lunch version of the case, Benedick owes reparatory duties to Beatrice; he should apologize for missing their lunch and offer to meet her at another time, or something of the sort. But in the first version of the case, he owes no such reparatory duties. For if a promisee does wrong in soliciting or accepting an immoral promise, the reparatory duties that the promisor would otherwise owe to her are negated. It is !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 212 We usually assume that if two conflicting and equally important promises are made—e.g., each breach would have equally bad consequences, the same sort of relationship holds between promisor and promisee, etc.—we ought to keep whichever promise was made first. It’s surprisingly difficult to articulate exactly why this is so; see Liberto (ms). This is a topic I plan to explore in future work. Here, all I need to claim is that there is an escape condition on one of conflicting promises. ! 236! wrong for Beatrice to ask Benedick to kill Claudio as a test of his loyalty. So when he fails to do so, he does not owe Beatrice anything. Promises that involve significant promisee wrongdoing are the exception (alluded to in Chapter 1) to the rule that reparatory duties always accompany the failure to keep a valid promise. 213 4.1. Promises that are immoral to make (but not to keep) We have seen that promises that are immoral to keep, either because they conflict with prior obligations or are promises to do independently immoral things, are subject to escape conditions. Are promises that are immoral to make subject to escape conditions, as well? Recall the example from Chapter 1, in which your current spouse asks you to promise that you will not make any promises to your ex-partner, because he does not want you to display the sort of special concern for your ex that making a promise displays. You agree to this promise. You then promise your ex that you will attend only the Central APA this year if he will attend only the Pacific, to ensure that you do not run into each other. This promise was immoral for you to make, because in doing so you broke your promise to your spouse. But this promise would not be immoral to keep. To the contrary, coordinating your APA attendance with your ex is a good idea, and is a gesture that your current spouse would appreciate. Is an escape condition in place for this promise? !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 213 There are cases in which an agent makes an initially immoral promise, but in the circumstances it would be even more immoral for the agent not to keep it. In such cases, the agent is required to act as promised. But that is because this is the lesser of two evils, and is therefore the morally preferable option. For example, Holly Smith (1997, p. 155) discusses a case of a university dean who makes a nepotistic promise to hire her relative for a new faculty position. The new hire is then advertised, and a great number of graduate students enroll in the university solely for the purpose of working with the dean’s relative. Were the dean to fairly hire a new candidate now, these students would be left in the lurch. And so all-things- considered, the dean is morally required to keep her initially immoral promise. ! 237! It might seem so, because you have done wrong in making the promise in the first place. You ought not make any more promises to your ex, and you ought to apologize to your spouse for the promise you already have made. But it would not be fair to your ex (who has not wronged you in any way that is related to the making of your promise) if this counts as an escape condition on your promise, especially since your ex is relying on the promise. You were obligated not to promise, and you violated that obligation. You owe reparatory duties to your spouse as a result of this violation. But now the promise has been made, and is not itself immoral or otherwise problematic. And so the promise is not subject to an escape condition. Immorality as an escape condition has only to do with the immorality of keeping the promise, and not any distinctive immorality involved in making it. 5.! Initial Imprudence Suppose Frida’s friend George wants to take a luxury cruise, and is trying to persuade Frida to join him. The cruise costs much more than Frida can afford, and she would have to go into debt to pay for it. But Frida tends to make rash decisions, and is not very good with money. So without thinking it through very carefully, she promises George that she will join him on the cruise. Is Frida morally required all-things- considered to go on the cruise, even though it means financial disaster? It seems plausible that the imprudence of Frida’s promise functions as an escape condition, for it seems plausible in general that we are not morally required to behave entirely irrationally. At very least, it is plausible that this is true of our promising convention. Promising would not be a very popular or widely accepted convention if it required us to act in extraordinarily imprudent ways. This does not mean that the agent who makes a deeply imprudent promise has no obligation; Frida still has a directed, pro ! 238! tanto obligation towards George, and she must make reparation of some sort to him when she does not keep her promise (say, by apologizing for making such an imprudent promise, and offering to plan a more affordable vacation together instead, or pledging to join George on a cruise someday if she ever has money to do so.) How imprudent must an agent’s promise be in order to count as an escape condition? As with MICROs, this will vary in accordance with the importance threshold of a promise. A significantly imprudent promise might count as an escape condition on a fairly unimportant promise, while the level of imprudence would have to be quite great to count as an escape condition on a more important promise. And lesser degrees of imprudence do not excuse. Agents are never required to do the immoral simply because they promised to do so. But an agent may very well be required to act imprudently, simply because she promised to do so. For example, suppose Frida has a stack of papers to finish grading by tomorrow. George calls her in the morning and asks if she wants to go see a play with him that night, and Frida promises to go. This is an imprudent decision, for going to the play will take up several hours of her time, which would be better spent grading. Going to the play does not make grading impossible; Frida will be able to finish grading even if she goes to the play. But she will have to stay up late to do so, when she needs to wake up early the next day. The imprudence of Frida’s promise in this case does not seem to function as an escape condition. She should not have made the promise in the first place. But given that she made it anyway, she is stuck with a late night. Sidgwick argues that this is common sense, writing that “it is scarcely thought that a promise is not binding because it was a foolish one, and will entail an amount of pain or burden on ! 239! the promiser out of proportion to the good done to the promisee.” 214 The difference between the play case and the vacation case is one of degree. Having to stay up late to grade is inconvenient. Taking on a huge amount of debt in order to pay for a vacation is insane. In both versions of Frida’s case, she is negligent in making her promise. While she would recognize upon reflection that going on the cruise or going to the play does not make much sense from her own perspective—and while this is the sort of reflection that she ought to be engaging in before making a promise—Frida does not clear- sightedly take the imprudence of her action into full account when she makes the promise. I want to leave it open that agents might be able to willfully promise to perform transparently imprudent actions, because they deliberately want to bind themselves to do something for another person at great cost to themselves. For example, a knight errant might promise his lady that he will perform a series of arduous and pointless tasks, the doing of which is greatly imprudent, as a proof of his devotion to her. So long as what the knight promises to do is possible and morally permissible, I don’t wish to rule out that it is possible for the knight to so bind himself. We can use the following story about a famous imprudent promise to illustrate the various ways in which imprudence may or may not count as an escape condition: Once when Jacob was cooking a stew, Esau came in from the field, and he was famished. Esau said to Jacob, “Let me eat some of that red stuff, for I am famished!” . . . Jacob said, “First sell me your birthright.” Esau said, “I am about to die; of what use is a birthright to me?” Jacob said, “Swear to me first.” So he swore to him, and sold his birthright to Jacob. Then Jacob gave Esau bread and lentil stew, and he ate and drank, and rose and went his way. Thus Esau despised his birthright. 215 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 214 ibid, p. 308 215 Genesis 24:29 – 34, New Revised Standard Version of the Bible ! 240! The birthright that Esau gives away entitles him to receive the largest portion of the family inheritance, as well as a special blessing from his father. Promising to give all of this away in exchange for a bowl of lentil stew is massively imprudent. This is obviously not a promise that Esau ought to have made in the first place. But since he has made it, is he obligated all-things-considered to keep it? The answer to this question depends on how we cash out the particular details of the case. We assume that Esau is engaging in hyperbole when he states that he is dying. But suppose he has been out in the fields for two days with nothing to eat, and is famished. In this case, his hunger is likely a coercive factor that undermines Esau’s moral responsibility, in which case his “promise” is invalid. (If you find it difficult to imagine hunger exerting such coercive force, imagine Esau as a heroin addict in withdrawal, who “promises” Jacob his share of the family inheritance in exchange for the hit of heroin that Jacob is cooking.) Suppose Esau’s hunger is not so extreme as to be coercive. We may still have evidence that Esau’s responsibility is undermined. For Esau seems to care so little about his future self that he would exchange a bowl of soup now for a massive fortune, the goodwill of his family, and the blessings of God in the future. Someone who displays such bizarre patterns of thought may not be a functional agent. That is, the fact that Esau cares so little about his future self may be evidence that he is in the grip of depression or some other mental illness, or that he is incapable of seeing himself as a unified agent extended through time. And if this is the case, then perhaps Esau’s moral responsibility is undermined so greatly as to make his “promise” to Jacob invalid. ! 241! Finally, it could be that Esau makes a deliberate, clear-sighted decision to trade his birthright for some soup—a decision that, although imprudent, is consistent with his other values. The text states that Esau “despised” his birthright. Perhaps this means that Esau has unusual preferences, such that he hates money and family blessings but instead cares exclusively about seizing the day, living in the moment, and satisfying his whims as they occur. Even though it would objectively be in Esau’s best interests to hang on to his birthright, when his whim for soup comes upon him he values satisfying this whim more than securing his long-term future prospects and honoring his family. It might seem that imprudence does not function as an escape condition in such a case, and that Esau is morally obligated to keep his foolish promise. Perhaps he is like the knight errant and unlike Frida: he willingly made a bad choice, and since he fully endorsed this decision he is now stuck with the consequences. We have seen that promises that are initially flawed because they are impossible or immoral are subject to escape conditions. The cases discussed above show us that initial imprudence can count as an escape condition on a promise, as well. This happens when the imprudence is great enough, and when the imprudent promise has been made rashly or negligently rather than clear-sightedly or for a deliberate purpose. 6.! Undermining False Belief Suppose that Harry and Irene are planning to go to a movie tonight, and the movie theater website states that it is cash only. Harry has only $5 cash on him, and the movie costs $10. Irene promises to buy him a ticket, because they both believe that he will be unable to purchase one himself; were this not the case, Irene would not have offered to buy Harry a ticket. However, when Irene arrives at the theater she discovers ! 242! that the website was mistaken, and the theater takes credit cards after all. Is she nevertheless morally obligated to buy Harry a ticket? It seems that she is not, as the only reason Irene promised to buy his ticket is that they both thought he would be unable to buy one himself. It would not be fair to hold Irene to performing an action that was prefaced entirely on a misconception. Since it will actually be quite easy for Harry to buy his own ticket, he no longer needs to rely on Irene to purchase one for him. Of course, Irene cannot leave Harry in the lurch; she must warn him that he will have to purchase the ticket himself, and she should buy the ticket herself if he cannot be warned in time—which is to say that Irene owes Harry reparatory duties. And even if Harry could be warned in time, it would still be permissible for Irene to buy Harry’s ticket. But doing so would be supererogatory, rather than morally required on the basis of Irene’s promise. Sidgwick is right when he notes that “if the promise was understood to be conditional on the truth of a statement which is found to be false, it is of course not binding.” 216 (At least, Sidgwick is right if we interpret this as the claim that such promises are not binding all-things-considered; as I will argue below, even promises that were grounded in an undermining false belief are valid and generate pro tanto obligations.) A promise is conditional on the truth of a given belief when that belief played an essential role in the making of the promise. That is, the belief must have been in some way necessary for the formation of the promise. If the belief turns out to be false, a necessary reason for which the promise was made is undermined. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 216 ibid, p. 305 – 06 ! 243! We can counterfactually test whether Irene’s false belief is undermining by asking whether she would have promised to buy a ticket for Harry had she known that the movie theater accepted credit cards. Since she would not have, her belief that the theater is cash only is an undermining false belief. If Irene would have promised to buy Harry’s ticket even if she had known the theater accepted credit cards, then the false belief will not be undermining, and will not yield an escape condition. For the falsity of a belief that was merely one of multiple sufficient reasons for the making of the promise does not yield an escape condition. Suppose Irene promised to buy Harry’s ticket because he has no cash and Irene believes the theater to be cash only, and because she owes him from the last time they went to the movies. The undermining of one of these reasons does not license escape from the promise, since there remains another sufficient reason for which Irene would have made the promise anyway. We have to be careful how we define when a false belief counts as undermining in this way, lest we overgenerate escape conditions. For there are cases in which what might first appear to be an undermining false belief does not also appear to be an escape condition. 217 For example, suppose Irene promises Harry that she will take him to visit Chicago. Irene does this because she desires to see Jacques-Louis David’s painting The Death of Socrates, and believes that the painting is at the Art Institute of Chicago. They then discover that they were mistaken about the location of the painting, which is in fact at the Metropolitan Museum of Art in New York City. Is an escape condition in place? The answer to this question will depend on the reasons for which Harry accepted the promise. Had Harry cared about going to Chicago only because he wanted to see !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 217 Thanks to Jake Ross for pressing me about an example with the structure of the painting example I discuss here, and suggesting that it matters whether the reasons that are undermined belong to the promisee or the promisor. ! 244! The Death of Socrates, the fact that the painting is not in Chicago would count as an escape condition, since Harry’s necessary reason for going to Chicago would be undermined. So Irene would not be morally required to take Harry to Chicago, although she would owe him reparatory duties (such as offering to take him to New York instead, or expressing sincere regret that they will not be going to Chicago.) But what if Harry accepted the promise because he wants to see the painting of Socrates, and because he has an equally strong desire to view Grant Wood’s American Gothic, which he (correctly) believes is also at the Art Institute? In this case, there does not seem to be an escape condition in place. For even if Irene would not have made the promise had she known that The Death of Socrates was not at the Art Institute, Harry would have still accepted the promise. That is, the fact that The Death of Socrates is in New York and not Chicago undermines only one of Harry’s two sufficient reasons for going to Chicago. And a false belief that undermines one of multiple sufficient reasons does not license escape. This example shows us that in order to count as an escape condition, a false belief must undermine the necessary reasons of the promisee. We saw in Section 2.1 that the importance threshold of a promise is fixed by the interests of the promisee rather than by the interests of the promisor; it does not matter how much you feel like going to Christa’s wedding, but how much it matters to Christa whether you are there. The same sort of phenomenon is at work here. 6.1. Deceptive promises Intuitively, it seems that promises grounded in deception are subject to escape conditions, as well. For example, suppose Harry knows that the movie theater accepts ! 245! credit cards, but lies to Irene and tells her that it only accepts cash in a tricky attempt to avoid having to buy his own ticket. This case is clearly subject to an escape condition; when Irene finds out about Harry’s lie, she will not be morally obligated to keep her promise to him. This is another exception to the rule that promise-breaking leads to reparatory duties. Harry’s wrongdoing in deceiving Irene entails that the reparatory duties she would otherwise owe him are negated; she is not obligated to warn him in advance about her non-performance in the way that she would otherwise be. Similarly, an escape condition would seem to be in place if Irene were deceived by a third party. If Irene discovered that her friend who likes to play jokes on people deliberately misinformed her about the theater’s cash only policy, Irene would not be morally required to buy Harry’s ticket. So long as Harry was not actively involved in the deception, Irene would owe him standard reparatory duties—although these duties would be negated if Harry actively solicited the third party deception, or knew that the promise was essentially grounded in a falsehood but accepted it anyway without informing Irene. 218 However, deception is not a distinct category of escape condition, but rather a particular kind of undermining false belief. Deceptive promises are subject to escape conditions only when and because the false belief resulting from the deception !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 218 There might be cases of promisee deception that are justified all-things-considered on paternalistic grounds. For example, suppose I am insisting that it is a good idea to go for a midnight swim in treacherous waters tonight. I won’t listen to all of your excellent reasons why this is a bad idea. So you deceive me, telling me that the beach is closed and asking me to promise to go to the pub with you tonight instead. (Thanks to Jake Ross for suggesting this example.) An escape condition will be in place if I discover your deception, for a necessary reason for which I made (and you accepted) the promise to go to the pub with you has been undermined. However, if your deception was justified on paternalistic grounds, it is plausible that I will owe you reparatory duties anyway, for you have not engaged in wrongdoing of the sort the negates such duties. ! 246! undermines the promisee’s necessary reasons for accepting the promise in the first place. The deception itself plays no essential role in licensing escape. We can illustrate this by considering cases of unsuccessful deception. Suppose Irene is fully aware that the movie theater takes credit cards, and that Harry is lying to her when he claims that it is cash only. But Irene also knows that Harry is a compulsive liar, who is undergoing therapy in an effort to rid himself of his frequent urges to fib. Irene values maintaining a good relationship with Harry, and believes that calling him out for his compulsive lying will damage their relationship. So she makes a clear-sighted decision to ignore the deception, and promises Harry that she will buy him a ticket. Such a case does not seem to be subject to an escape condition, which shows us that the bare existence of deceptive behavior is not what grounds the escape condition. We want to leave it open that agents might make valid and binding deceptive promises for independent reasons, just as they might make valid and binding imprudent promises for independent reasons. The knight errant might successfully bind himself to his lady by promising her that he will perform arduous and pointless tasks, even though this imprudence would count as an escape condition had the knight entered into the promise without realizing what the tasks required. Similarly, Irene may successfully bind herself to Harry by promising him that she will buy him a ticket, even though this promise was prefaced on a deception that would count as an escape condition had Irene entered into the promise without realizing that Harry was attempting to deceive her. People might have a variety of relationship-based reasons for wanting to make and follow through with unsuccessful deceptive promises, and we want to be able to accommodate this. ! 247! I have granted that in cases where deception is successful, there is an escape condition stemming from the undermining false belief caused by the deception. When the promisee is implicated in the deception, reparatory duties are negated, in much the same way that reparatory duties are negated when the promisee is guilty of soliciting or knowingly accepting an immoral promise. One might worry that it makes more sense to claim instead that there are no reparatory duties in cases with guilty promisees because such promises are invalid, and no pro tanto obligation has been created in the first place. David Owens takes this approach, arguing that promises involving wrongdoing by the promisee (whether due to immorality or deception) are invalid. 219 However, this is not a helpful or natural understanding of deceptive promises. Reparatory duties are obviously owed when an agent fails to keep an immoral or deceptive promise involving an innocent promisee. It would not carve up the phenomena very neatly to claim that there exist both valid immoral promises and invalid immoral promises, as well as valid deceptive promises and invalid deceptive promises, with the only difference between the valid and the invalid cases being promisee wrongdoing. And it would be bizarre if an agent needed to determine whether the promisee had done wrong in soliciting or accepting the promise in order to determine whether a valid promise could be made in the first place. In general, then, lack of promisee wrongdoing is not a plausible validity condition. Rather, it is more natural to claim that the existence of reparatory duties can come apart from a promise’s validity, and is fixed by the promisee’s behavior. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 219 2007 ! 248! 7.! Counterfactual promisee release? We saw in Chapter 5 that promisee release is an invalidating condition: it immediately causes the promise to cease to bind in even a pro tanto way, such that there is no moral residue or need to fulfill reparatory duties. One might wonder whether there is another kind of release that counts as an escape condition. Consider a promisor who reasonably believes that the promisee would release her, but the promisee cannot get in touch with her in time. Call such cases counterfactual promisee release—cases in which the promisor reasonably believes that the promisee would not want her to keep the promise anymore, but is not able to successfully seek release. For example, suppose that we are coworkers who take turns bringing each other lunches every Friday. It’s my turn this week. On Thursday morning, I promise to bring you Chik-Fil-A on Friday, which I know to be your favorite fast food. On Thursday evening, a news story breaks about the president of Chik-Fil-A donating a significant portion of his company’s profits to anti-gay organizations. I know that you are a gay rights activist, and that as a policy you never support companies that are opposed to gay rights. I cannot reach you before Friday morning, so you do not have the opportunity to actually release me from my promise to get you Chik-Fil-A. But if you could reach me, I figure that you would say something like “You’d better not support that homophobic company by buying me a sandwich there!” So I make the entirely reasonable assumption that you no longer want me to carry out my promise, and I bring you a burrito from Chipotle instead. This seems to be morally permissible, which makes it seem like your counterfactual release is an escape condition. ! 249! However, if we look a bit closer we see that it is not actually your counterfactual release that does the work in this situation. For there is an escape condition in place even if you would not have released me. Suppose that when I hand you the burrito, you say, “Hey, I don’t want this! You promised to bring me Chik-Fil-A!” I tell you about the company president’s anti-gay stance. You say “I know and I don’t care; I love chicken sandwiches and will continue to eat them even if they support a cause that I hate.” This does not change the fact that I acted permissibly in bringing you Chipotle instead of Chik-Fil-A. There is an independent escape condition in place on this promise, because certain implicit conditions involved in the promise have been violated: given your history as a gay rights activist, it is mutually assumed between us that I ought not ever support homophobic businesses on your behalf. Promises that are explicitly conditional might appear on the face of it to be cases of counterfactual promisee release, as well. Suppose that a stockbroker promises her client that she will purchase 1000 shares of a certain stock so long as that trade is profitable. If the stock unexpectedly drops 20%, the broker is no longer obligated to purchase the stock. This might seem to be because the client would not want her to do so anymore. But the underlying explanation is that the explicit condition of the promise is no longer met. What appears to be an instance of counterfactual promisee release is actually an escape condition of a more typical sort. In all plausible cases of counterfactual promisee release, the promisee would be counterfactually likely to release the promisor because of a conflicting obligation, initial flaw, or unmet implicit or explicit condition. In such cases, it is the underlying obligation, flaw, or condition that makes it the case that the promise need not be kept, and not the promisee’s counterfactual release. Cases of bare counterfactual release— ! 250! that is, cases in which there is no conflicting obligation, initial flaw, or unmet condition—do not plausibly generate escape conditions. We would overgenerate escape conditions if we allowed them to do so. For example, suppose your brother is extremely accommodating. You know that if you were to ask him to release you from any promise you have made him, he would accommodate you, even if you have no good excuse and he has been relying on the promise greatly. The fact that your brother would counterfactually release you is not in itself any grounds for escape. In this chapter, we have seen that escape conditions of various kinds can make it the case that all else is not equal, and that a pro tanto promissory obligation is overridden or undermined and therefore does not yield an all-things-considered obligation. These escape conditions include: a more important conflicting reason or obligation, determined by the promise’s importance threshold fixed by the promisee’s interests; keeping the promise being impossible, immoral, or deeply imprudent in a way that is not fully endorsed by the promisor; and the discovery of a false belief undermining a reason that was necessary for the promisee’s accepting the promise. In the next chapter, I argue that the same general sorts of escape conditions apply to resolutions, and that MSF gives us a good explanation of this striking similarity. ! 251! CHAPTER 8: THE ESCAPE CONDITIONS ON RESOLUTIONS If promises express conditional resolutions, we should expect the escape conditions on promises and on resolutions to be similar. In this chapter, I will argue that they are: it is rationally permissible to abandon a resolution in roughly the same types of situations in which it is morally permissible to break a promise (with one important modification, as we shall see in Section 8.) 1.! Validity Conditions on Resolutions We saw in Chapters 1 and 5 that in order to yield a valid promise, a promissory offer must be made intentionally by a morally responsible agent who is not acting under coercion, and then accepted by a promisee who does not subsequently release the promisor. It is possible for a resolution to be invalid if the resolver lacks basic rational responsibility, just as a promise is invalid if the promisor is not morally responsible. You cannot make a genuine, binding resolution if you are permanently or temporarily rationally incapacitated or irresponsible; an agent who makes a “resolution” while drunk to the point of incoherence is not bound by it. Moreover, resolutions must also be uncoerced, for similar reasons as promises must be free of coercion. It’s not obvious that there are plausible cases of coerced “resolutions.” 220 But if there are, they are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 220 The following might be a plausible example of a coerced resolution. Suppose that a teacher very much wants an at-risk student to begin doing the readings for class. The teacher explains how important having the right attitude towards one’s schoolwork is, and encourages the student to resolve to do the readings. (The teacher does not care about whether student actually does readings, but wants the student to care about doing them, for the teacher is primarily concerned with making the student a better student.) When the student is resistant, the teacher threatens that unless student resolves to do the readings, the teacher will stop providing extra help. The student forms a resolution to do the readings, and as a result of the threat acquires an instrumental desire not to abandon the resolution. Assuming that the threat of revoking help is ! 252! invalid. Resolutions are valuable in large part because they make agents likely to act in accordance with their stable values and endorsed preferences over the long term. Coerced “resolutions” do not do this, because they line up with the goals and values of the coercer rather than of the agent. Since there is no second party with resolutions, there is no need for acceptance, and no possibility of release. It would not do to allow that resolvers could release themselves from a resolution in the manner that promisees may release—that is, at any time and for any (or no) reason. For resolutions that were so easy to be self-released from would not serve their function: if a vegetarian struggling with the difficulty of resisting steak could permissibly release herself from her resolution to avoid eating meat whenever she felt like it, her resolution would not be very effective. 221 2. More Important Conflicting Reasons or Obligations Suppose that Zoe wants to run a marathon, because she believes training for a marathon will be good for her cardiovascular health. She recognizes that training for a marathon is a long and difficult process, and that in the moment when she is waking up early to go running she might not actually want to do so. So she resolves to diligently follow a strict training schedule. One month into training, Zoe is promoted at work. The new job requires extremely long hours, leaving Zoe with only a fraction of the free time that she had when she initially resolved to train. If Zoe were to follow a rigorous training program with her new work schedule, she would have to give up socializing !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! coercive, this resolution will not yield a pro tanto rational obligation for the student, since it did not come about voluntarily. 221 See Liberman (ms) for an argument that resolutions cannot be subject to self-release, in the context of arguing that resolutions are not plausibly construed as promises to the self. ! 253! after work. She decides that having a social life is more important to her than training for a marathon, and so she abandons her resolution. Intuitively, Zoe does not behave irrationally. To the contrary, it would seem irrational for her to continue training in spite of the newfound difficulty of doing so. This is because Zoe’s resolution to train for a marathon is outweighed by a more important conflicting reason. Her commitment to train every day no longer makes sense now that she does not have enough time to do so. This shows us that sufficiently more important conflicting reasons function as escape conditions on resolutions, just as they do on promises. As with promises, a MICRO need not rise to the level of obligation to count as an escape condition; Zoe’s desiring a social life is most plausibly construed as a strong reason to stop training rather than an obligation to do so. And as with promises, a MICRO need not be a reason of a particular normative sort to serve as an escape condition on a resolution. For example, suppose that Yoshi resolves to stop eating dessert for one year because he wants to lose weight. Yoshi’s elderly and arthritic grandmother painstakingly prepares him her signature chocolate mint layer cake, which was his favorite dessert as a child. It would hurt his grandmother’s feelings greatly if Yoshi did not eat any of it, especially because (in common grandmother fashion) she thinks that Yoshi is too skinny, and would have a hard time understanding why he does not want to eat dessert. Intuitively, it does not seem in any way irrational for Yoshi to have a slice of his grandmother’s layer cake. This is because Yoshi’s moral obligation to avoid hurting his grandmother’s feelings conflicts with and is more important than keeping his resolution to avoid all desserts. This shows us that sufficiently more important conflicting moral obligations (or reasons) ! 254! can be escape conditions on resolutions, as well. The same will hold of conflicting reasons or obligations of other normative sorts. As already noted, when abandoning a resolution there are no reparatory duties, either to yourself or others. However, an agent who abandons a resolution should briefly revisit any other plans and resolutions that are connected to the abandoned resolution to see whether they still make sense, and must take whatever action is necessary to ensure that they do or to abandon them if they do not. For example, suppose Zoe has registered for several 10K races in upcoming months, not for their own sake but solely as part of her marathon training. Once she gives up her resolution to run a marathon, running the shorter races no longer makes sense for her. So she is rationally required to change her plans and not run in the 10k races. 2.1. The importance threshold Like promises, resolutions have a variable importance threshold that determines how strong a MICRO must be to count as an escape condition. We can illustrate this with a pair of cases. In the first case, suppose that you have to finish filling out a lengthy application for a job you very much want, which is due tomorrow morning with a strict deadline. You resolve to complete the application tonight, come what may; since you will not be able to apply for the job unless you finish the application tonight, you want no alternative option that might arise to prevent you from working it, no matter how appealing the option. In the second case, you resolve to work on the application tonight because you have had an unproductive week and are now behind schedule. But the application is not due until the end of the week, and you will be able ! 255! to complete it in time even if you don’t work on it tonight, although you would have to skip a party at the end of the week that you had hoped to attend. Suppose that you receive a call from an old acquaintance that you have had a romantic crush on for years. You have repeatedly asked him out on a date, and he has repeatedly refused. However, he has had a change of heart; he is in town for only one night, and would like to go out with you. This seems to suffice as an escape condition in the second case. That is, it seems perfectly rational to abandon your resolution to complete the job application and to go on the date instead. It makes sense to go out tonight, then scramble to get the essays done and skip the party at the end of the week. In the first case, though, you do seem irrational if you abandon the resolution and go out tonight instead of finishing your job application. After all, you were trying to ensure that you completed the application no matter what better option came up. 2.2. Implicit conditionality Like promises, many resolutions have an implicit conditionality that can affect their importance threshold. When an agent resolves to Φ, she often does so because she wants to protect herself from the influence of some but not all reasons against Φing—to bolster herself against certain forms of temptation and not others. When you form a resolution, you often have some implicit sense of which sort(s) of conflicting reason would justify changing your mind, and which won’t. The agent in our first case presupposes that almost nothing will cause her to change her resolution, while the agent in our second case grants than an especially good alternative (like a great date) would be sufficient, although lesser reasons (like watching cooking competition shows on TV) would not be. The implicit conditions under which an agent wishes to remain ! 256! resolute might be quite specific. Suppose I am in the habit of calling my ex when I feel lonely, and I resolve to refrain from doing this. No amount of loneliness will count as an escape condition, no matter how great. But even weak reasons of different kinds might count as escape conditions—say, calling my ex to notify her that there is crime alert for the neighborhood where she lives. We saw in Chapter 7 that the importance threshold for a promise varies according to the interests of the promisee: the more important the promise is to the promisee, the higher the threshold will be. The importance threshold for resolutions varies according to the interests of the resolving agent. But, crucially, resolutions vary according to the interests that the resolving agent had at the time the resolution was made. We typically make resolutions regarding how to act at some later time L when we expect that we are able to make better decisions now (at time N) than we will be able to make at L. We expect that temptation or laziness or some other barrier to action will arise between N and L, so we form a resolution to counter this. The importance threshold is fixed by the agent’s interests at N, rather than by her interests at L. Whether a date is a strong enough MICRO to license abandoning a resolution to stay home and complete a job application depends on how much staying home and completing the application mattered to the agent at N, when she made the resolution. If it were fixed instead in terms of the agent’s interests at L, resolutions would be much too variable. At the time that you make a resolution, you care very much about going for a morning jog or refraining from eating meat. In the moment of temptation, when you are sleepy or craving a cheeseburger, your desires are likely to shift such that you do not actively care about exercise or vegetarianism. The resolution will be effective only if we fix the importance threshold in terms of the former interests, rather than the ! 257! latter. To fail to do so would be to allow something like problematic judgment shift: not an inappropriate change in what the agent has resolved to do, but an inappropriate change in how important the resolution is, as fixed by the resolver’s interests. David Gauthier cashes this idea out in terms of a difference between what he calls proximate and vanishing-point preferences. 222 Proximate preferences are those short- term preferences that we have at the moment of choice (that is, at time L). Vanishing- point preferences are more stable, and are those that we maintain across time (e.g., from time N to L.) It is not rational to act against one’s vanishing-point preferences by giving in to one’s proximate preferences. Doing so is like succumbing to inappropriate judgment shift in Holton’s sense. Gauthier suggests that agents who act on their stable vanishing-point preferences do better at attaining all of their preferences overall. So we have reason to act resolutely in accordance with our vanishing-point preferences, which are more likely to be our preferences at N when we are making resolutions than they are to be our preferences at L when we are facing temptation and deciding how to act. What happens if an agent undergoes a shift in vanishing-point preferences? Does it make sense for her to stubbornly maintain her resolution anyway? In the next section, I argue that if you undergo such a massive or fundamental preference shift that you no longer identify with or endorse the time N values of the resolving self, then there is an escape condition in place and the resolution no longer has force. 3.! Massive Value or Stable Preference Shift Suppose Xena is a radical anti-natalist environmentalist who believes that humans are a scourge on the planet; we are destroying the Earth, and it would be best if !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 222 1997 ! 258! we stopped reproducing and slowly died out completely, allowing the earth to return to its natural state. In light of these views, Xena resolves never to have children. She recognizes that she might someday be tempted to have children, because she is aware that many humans have strong biological urges to reproduce. Xena fears that these biological urges might get the better of her more reasoned view that having children is immoral. So she resolves to stay childless, in order to combat such temptations. Over the next several years, Xena slowly undergoes an ideological change. After much study and careful consideration, she comes to believe that human life is uniquely valuable, and that continued human existence is compatible with the overall wellbeing of the planet. It seems clear that Xena’s resolution to avoid having children no longer binds her, and that an escape condition on this resolution is in place. 223 Why is this so, given that having children violates Xena’s interests from time N, when she first made her resolution? The difference between this case and the cases discussed in the previous section is that Xena’s beliefs, preferences, and values have undergone a shift so great that her resolution no longer makes sense from her own standpoint as a whole. This is not a temporary lapse in judgment of the sort caused by temptation, but rather a fundamental change in values. Xena’s time N self should not hold sway over her time L self because these selves have significantly different value sets. This is not the case with most resolutions; my wide awake self at 8:00 PM resolving to go to the gym tomorrow !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 223 It is important that we not be led astray by our intuitions about which course of action is objectively best. Consider the converse of the example: Xena resolves to have many children, because she thinks having a large family is the key to a good life. (She must resolve rather than intend because she fears the pain of pregnancy and childbirth, and this fear makes her likely to revise her decision to have children.) After careful study and consideration, Xena becomes an anti-natalist environmentalist. In such a case, there indeed appears to be an escape condition on Xena’s resolution, and it is rationally permissible for her to refrain from having children. ! 259! morning is basically the same as my tired self at 8:00 AM not wanting to wake up. As Gauthier notes, an “agent, at any given time, has no interest in preferences or concerns that she once embraced but holds no longer.” 224 And so those preferences and concerns should not hold rational sway over her. An important constraint on whether massive shifts in values or stable preferences count as escape conditions is that these shifts must be rational. In the first version of the case, Xena undergoes a gradual and coherent shift in her values that is compatible with her evidence and the rest of her value set. But suppose instead that Xena joins a UFO cult that preaches that in 30 years, the Earth will be destroyed and its entire population will be picked up on comets and taken by benevolent aliens to other planets to colonize the galaxy. You need a lot of humans to successfully colonize the galaxy, so Xena believes that it is her duty to have as many children as possible. She accordingly abandons her resolution to refrain from having children. Assume that it is not rational for Xena to believe the claims of the UFO cult (e.g., they have no good evidence for their views, and there is lots of good evidence to the contrary.) Xena’s abandoning of her resolution does not seem rationally licensed in this case, because the judgment shift that led her to abandon the resolution was not itself licensed. If a massive shift is irrational—perhaps because the agent is reasoning poorly or is misinterpreting the evidence—it is not likely to match up with the agent’s endorsed values. Such unendorsed shifts are like the judgment shift that standardly follows from temptation. If we let them count as escape conditions, resolutions would not serve their !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 224 ibid, p. 18 ! 260! purpose. We would be allowing irrationality at one stage (an irrational shift) to infect a later stage (inappropriate reconsideration), which is not something we want to do. 225 4. Impossibility Suppose William resolves to climb Mount Whitney this summer. In late spring, he breaks his leg, and is put into a cast that he will need to remain in for several months. It is not possible for him to climb a mountain in his cast, so William abandons his resolution. Clearly, this is rational, which shows us that impossibility is an escape condition on resolutions, as it is on promises. It is quite plausible to think that it is not rational to intend to do what you believe to be impossible. 226 The same holds for resolutions. Since William believes it to be impossible for him to climb Whitney, it is irrational for him to maintain his resolution to do so. All that is required for my claim is that it is not rational to maintain a resolution to Φ if you believe that it is impossible for you to Φ. This is compatible with a very weak !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 225 There are interesting cases in which someone expects that she will undergo a fundamental shift in values, but does not want to undergo that transformation, and so makes a resolution designed to prevent such a shift. For example, suppose a sheltered person from a small, homogenous town is devoutly religious. She is preparing to go to university and leave home for the first time. She knows that many in her situation have lost their faith while away at university. She predicts that she is likely to undergo such a shift, as well, but in her current state finds the thought of losing her faith appalling. So she resolves to go to church every week whether she thinks church-going matters or not, in hopes that her continuing to go to church will halt or reverse any fundamental value shifts that may occur. While away at college, her values indeed shift as she expected them to, as a rational response to new evidence that she encounters. Is there an escape condition on her resolution to go to church weekly, come what may? It seems to me that there is: the person who has lost her faith “has no interest in preferences or concerns that she once embraced but holds no longer,” even though she formerly embraced those concerns in engaged in a deliberate effort to prevent herself from ceasing to embrace them in the future. 226 See footnote 93 of Chapter 3 for a survey of the ways in which different theories about intention agree about this claim. ! 261! constraint under which it can be rational to resolve to Φ so long as you believe that you might Φ, even if you believe that Φing is highly unlikely. For example, I might resolve to win a race even though I am not the best runner there, and I doubt that I will actually win. I also wish to allow a great deal of flexibility about what is “possible.” Suppose you resolve to work on a paper all day tomorrow. However, you know from past experience that you are not psychologically well-suited to working for an entire day on the same task, so you don’t think it very likely at all that you will actually work on the paper all day. But so long as you recognize that you could work on the paper if you really set your mind to it, it is not irrational to resolve to work on the paper. Initial impossibility that the agent is aware of when she makes the resolution functions as an escape condition on resolutions, as well, just as it does for promises. Suppose William breaks his leg, and resolves to climb Mount Whitney while he is in the hospital having his cast put on. Since this is impossible, William is rationally required to abandon this resolution, just as he would be if the impossibility arose suddenly. The difference between initial and unexpected impossibility is that agents are rationally blameworthy for making initially impossible resolutions, whereas agents who face unforeseen impossibility make no initial rational error. 5. Immorality Some accounts of the relationship between morality and rationality will not accommodate the possibility that an action might be both immoral and all-things- considered rational. This position arises if you understand an action’s being immoral to mean that there is most reason overall not to do it, and understand an action’s being rational to mean that there is most reason overall to do it. Since the same action cannot ! 262! be such that we have most reason overall to do it and most reason overall not to do it, we will never have any cases in which it is rational to do something immoral. So the question of whether initial immorality is an escape condition on resolutions will not matter, because an immoral act will always also be irrational, and as we shall see in the next section, irrationality is always an escape condition on resolutions. What about views under which an action’s being immoral means that there is conclusive moral reason not to do it, but it is an open question whether there is conclusive overall reason not to do it? For example, suppose Vinnie the mob boss resolves to kill his nephew, who has betrayed the family by acting as a police informant. (Vinnie must resolve rather than merely intend to kill the nephew because he fears he may otherwise be too emotional and weak to pull the trigger.) Resolving to kill the nephew is obviously immoral. But suppose that killing the nephew coheres with Vinnie’s overall goals, values, and intentions, and that Vinnie decides to kill his nephew entirely clear-sightedly and continues to endorse this decision. In this case, initial immorality does not seem to be a rational escape condition on Vinnie’s resolution. For he would be irrational if he abandoned his resolution to kill the nephew; his endorsed values would be frustrated, and his plans would be undermined. Vinnie in this case is like Esau in the analogous promising case. Esau clear- sightedly promises to do something imprudent, and continues to endorse this decision. He is morally required to do something foolish as a result of this commitment. Likewise, Vinnie resolves to do something terrible, and continues to endorse this decision. And he is rationally required to do something morally bad as a result of this commitment. Keep in mind that the notion of rationality that we are working with is subjective. We are assessing whether an agent’s action is rational given his own set of ! 263! attitudes—e.g., because Vinnie doesn’t care about morality, it doesn’t make sense for him to let moral demands structure his planning—and not from some objective standpoint—e.g., because morality is in fact important, everyone ought to let moral demands structure their planning. Vinnie’s perspective is wicked, and so following through on wicked actions makes sense from it. Thankfully, most agents are not like Vinnie the mob boss. They care about morality, just like most agents are unlike Esau and care about the well-being of their future selves. Suppose a less callous agent—a newly recruited mob underling—has resolved to kill the snitch. Upon further consideration, he realizes that this is immoral, and that he should not have made such a resolution. In this sort of case, immorality is clearly an escape condition on the underling’s resolution, just as recognizing that promising to take a vacation she cannot afford is imprudent is an escape condition on Frida’s promise to go on a cruise with George. Recognizing or coming to newly appreciate the extent to which a resolution you have made is immoral will license abandoning it, so long as avoiding immorality is something you care about. 6. Imprudence Suppose Ursula works as a pastry chef, a job that includes sampling the pastries to ensure that they are high quality. Ursula wants to reduce her sugar intake, so she resolves to eat no sweets for a year. This resolution is imprudent from the start, for she cannot both keep it and do her job well. She should not have made the resolution in the first place. But when Ursula realizes that she needs to taste test one of the éclairs she has just made, it is not rational for her to stubbornly hold on to her resolution to avoid sweets anyway. One rational mistake must not license a second. Rather, Ursula should ! 264! abandon the resolution. This shows us that irrationality or imprudence is an escape condition on resolutions. Imprudence will function as an escape condition regardless of whether the resolver makes an imprudent resolution deliberately and clear-sightedly, or whether she makes it negligently and only later comes to fully appreciate the extent to which it is imprudent. This is because it makes no sense for rationality to require that an agent act irrationally. The person who makes an imprudent resolution fully knowingly is more rationally blameworthy for doing so than is the person who does so through culpable negligence. And someone who makes an imprudent resolution through non-culpable negligence or unknowingly is not rationally blameworthy at all. But this initial rational blameworthiness does not affect whether it is rational to follow through with the resolution. This is just how the immorality escape condition works for promises. People who make immoral promises are morally blameworthy; the more knowingly they make such promises, the more blameworthy they are. But it makes no sense for morality to require that agents act immorally. And so immorality is always an escape condition on promises. 7. Undermining False Belief Suppose Tim resolves to stop eating carbs because he believes that this will be an effective way for him to lose weight. During a routine checkup, Tim’s doctor discovers that Tim has a rare genetic condition that makes him gain weight if he cuts carbs from his diet, and lose weight if he eats a lot of carbs. So Tim abandons his resolution. This is not irrational, for the grounds on which the resolution was based have been undermined. Since avoiding carbs will not in fact help Tim achieve his goal of losing ! 265! weight, the resolution to do so no longer makes sense. As with promises, then, if your resolution to Φ was prefaced on faulty information that you later discover to be untrue, you are not rationally required to retain and carry out that resolution. And as with promises, the belief that is discovered to be false must have been in some way necessary for the formation of the resolution. We can apply the same counterfactual test that we used with promises to determine whether the belief on which a resolution as based is necessary. Had Tim known that carbs cause him to gain weight, he would not have resolved to avoid them. The undermining of one of multiple merely sufficient reasons will not suffice. If Tim resolved to avoid carbs because he believed this will help him lose weight and because he thought that cutting carbs would give him more energy, the discovery that carbs will not help him lose weight will not undermine his resolution, so long as he still believes that avoiding carbs will give him more energy. Nor will finding out the falsity of information that was unrelated to the making of your resolution count as an escape condition. Suppose Tim falsely believed that beer contained no carbs. He then discovers that alas, beer is full of carbs, and he will have to give beer up as well. If he would have resolved to give up carbs anyway had he known this, discovering that beer contains carbs will not license escape. For promises, the reasons that are undermined by a false belief must be those of the promisee, just as the importance threshold is fixed by the interests of the promisee. For resolutions, the reasons that are undermined by a false belief must be those of the agent at time N when she is making the resolution, just as the importance threshold is fixed by the interests of the resolver at N (rather than at L when the time for action arrives.) The undermined reasons must be those of the resolving agent because that is ! 266! how undermining false belief works as an escape condition: you discover that one of beliefs that led you to form the resolution in the first place was false. 8. Discrepancies in Thresholds for Revision So far, I have been explaining the ways in which the escape conditions on promises are the same as those on resolutions. However, there is a significant difference between the two types of escape condition that has yet to be discussed. The broad categories of escape condition are the same for promises and resolutions— MICRO, impossibility, initial immorality, initial imprudence, and undermining false belief. But there are particular cases in which a situation that would count as an escape condition on a resolution would not count as an escape condition had that resolution been a promise. That is, there are cases in which an agent A resolves to Φ, and condition C makes it the case that A is rationally permitted to abandon her resolution. But had A promised B that she would Φ, C would not make it the case that A is morally permitted to break her promise. 227 Call this a variation in the threshold for revision. Recall the example from Chapter 6 about being in the habit of eating pizza for lunch every day, and resolving to go to a Thai restaurant for lunch because you want to expand your culinary horizons. Suppose you are on your way to the Thai restaurant when you see a flyer announcing that your local pizzeria is giving out free slices today as part of a customer appreciation day. Free pizza is a good enough MICRO for you to abandon his resolution. But suppose you had promised your gluten-intolerant friend who cannot eat pizza that you would meet her for lunch at the Thai restaurant. The prospect of free pizza does not seem to be a legitimate escape condition on your !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 227 Thanks to Jake Ross for first pressing me on cases like these. ! 267! promise to eat Thai food with your friend, even though it is an escape condition on your mere resolution to eat Thai food. Such discrepancies can arise with other types of escape condition, too. Suppose you resolve to go to the Thai restaurant in order to expand your culinary horizons. But this is not a prudent choice; the restaurant is always extremely busy at lunchtime, it is difficult to find parking in the area, and the food is expensive. Coming to fully appreciate these factors seems sufficient to override your resolution. But had you promised to meet a friend at the Thai restaurant, these inconveniences would not license breaking your promise. Or suppose you resolve to go to the Thai restaurant for lunch because you believe it serves Thai iced tea, which is your favorite drink. (You must resolve to go rather than merely intend, because you are such a creature of habit that you are likely to eat pizza for lunch unless you resolve to do otherwise, thereby missing out on a drink that you know you would enjoy.) Before you leave for your lunch break, you learn that the restaurant no longer serves Thai iced tea. It seems perfectly rational for you to abandon your resolution once you discover that the restaurant does not serve your favorite drink after all. But you would not be morally permitted to break a promise to meet your friend at the Thai restaurant simply because you discover that they do not serve Thai iced tea. Are such variations in the threshold for revision a problem for my claim that the escape conditions on resolutions and promises are basically the same? They are not— and to the contrary, we should expect differences of these sorts. In the rest of this chapter, I will explain why. We shall see that the threshold for revision varies across ! 268! different sorts of resolutions, on the basis of how important the resolution is, whether it is conditional, and what sorts of temptations it is designed to overcome. Promissory resolutions are a natural part of this variation, at one end of a continuous spectrum. 9. How MSF Explains the Escape Conditions on Promises and Resolutions MSF gives us a simple and straightforward account of why resolutions and promises have the roughly same escape conditions. We have already established that promises inherit the further commitment norm on resolutions. This norm states that an agent who resolves to Φ is rationally committed to Φing, all else being equal. The escape conditions on resolutions that I’ve been outlining in this chapter are the conditions in which all else is not equal. These escape conditions are therefore an essential component of the further commitment norm on resolutions. When promises inherit the further commitment norm on resolutions, they inherit these escape conditions as a component of that norm. MICROs, impossibility, immorality, irrationality, and undermining false beliefs are the conditions that license breaking promises because these are the conditions that license abandoning the resolutions that promises express. As we saw in Chapter 3 with the SAMS pattern, this holds for both sincere and insincere promises. What then explains the differences we have seen between the escape conditions on promises and those on resolutions? I will first address how MSF can accommodate the particular differences we have observed for different kinds of escape conditions. I will then explain the general way in which the threshold for revision for different kinds of resolutions (including promissory resolutions) varies along a spectrum, and why it is therefore unproblematic to claim that the escape conditions on promises are derived from those on resolutions, in spite of seeming variations in the threshold for revision. ! 269! 9.1. Differences in MICROs When you promise to meet a friend at a Thai restaurant, receiving an offer of free pizza does not count as an escape condition. But when you’ve merely resolved to eat at a Thai restaurant, receiving an offer of free pizza does count. What explains this difference? In general, the more important a promise or resolution, the stronger a MICRO must be to license escape. Usually, a promissory resolution that involves the acceptance of another person will be more important than a resolution to Φ that involves only your own interests. And so a better excuse will typically be needed to break a promise than to break an ordinary resolution. Free pizza is a good enough excuse for abandoning a fairly unimportant ordinary resolution. But it is not a good enough excuse for breaking a significantly more important promissory resolution. Moreover, the importance of a resolution can fluctuate on the basis of changes to the resolving agent’s interests in a way that the importance of a promise cannot. We saw in Chapter 7 that the importance threshold of a promise is fixed by the interests of the promisee, including the extent to which the promisee relies on or forms expectations about the promise. 228 The promisee’s interest in meeting you for Thai food will not be affected or altered if you are offered free pizza while en route to the Thai restaurant. So the offer of free pizza will not lower the promise’s importance threshold. However, we saw in Section 2 of this chapter that the importance threshold of a resolution is fixed by the interests of the resolving agent had at the time she made the resolution. And these interests can be affected by an unexpected offer of free pizza; the attractiveness of getting a free lunch might appropriately cause the resolving agent to cease to be !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 228 Because there is only one agent involved in a resolution, abandoning a resolution will never lead to frustrated epistemic expectations; this is another reason why promises typically have higher importance thresholds than do resolutions. ! 270! interested in expanding her culinary horizons during this particular meal. So the prospect of free pizza can lower a resolution’s importance threshold, in which case a MICRO that would otherwise not be strong enough to count as an escape condition might count as one. 9.2.! Differences in undermining false belief For standard resolutions, an escape condition is in place if you discover that a belief that was essential to your making of the resolution is false; recall the case where the discovery that the Thai restaurant no longer sells Thai iced tea counts as an escape condition on your resolution. For promises, though, the false belief that does the undermining must be a belief not of the promisor’s but of the promisee’s, which was essential not to the making of but to the accepting of the promise. Suppose that you would not have promised to meet your friend at the restaurant had you known that they didn’t sell Thai tea. But when your friend accepted this promise, she was fully aware that they didn’t serve Thai tea, and accepted the promise anyway because she wanted to eat a green papaya salad. Discovering that they no longer sell Thai tea is not an undermining false belief escape condition; discovering that they do not serve green papaya salad would be. In general, the discovery that one of the beliefs on which you based your resolution is false counts as an escape condition when it undermines the aim of your resolution in some way. If I resolve to quit smoking because I care about my health and I believe smoking is bad for me, the discovery that smoking is actually healthy would be an escape condition on my resolution. Had I not been concerned with my health at all but resolved to quit because I thought my spouse hated the smell of cigarettes, the ! 271! discovery that smoking is healthy would not be an escape condition, while the discovery that my spouse actually loves the smell of smoke would be. Undermining false beliefs for promissory resolutions work in the same way: they count as escape conditions when the aim of the promise is in some way undermined. The aim of promising usually includes—but needn’t exclusively consist in— acting as the promisee wants, by taking the promisee’s acceptance and interests into account and ruling out the promisor’s minor, tangential, or not-shared-by-the-promisee interests. For example, you might be motivated to go to the Thai restaurant because you want Thai iced tea. But once your friend accepts the promise, one of the primary aims of your promissory resolution is not your being able to have Thai tea, but rather your acting as your friend wants you to. False beliefs that undermine the promisee’s acceptance will count as escape conditions, because they undermine the promisee- accommodating aim of promissory resolutions. 9.3.! Differences in imprudence and immorality counting as an escape condition Promises and resolutions also differ as to whether imprudence is always or only sometimes an escape condition. Recall that initial imprudence is always an escape condition on ordinary resolutions. If I resolve to Φ because I care about achieving X but Φing does not in fact help me attain X, I am not rationally obligated to Φ. Doing so would be pointless, as it would not enable me to reach my goals. People who knowingly make irrational resolutions are not rationally obligated to keep them just because they have resolved to do so. What X is—what goal the resolution aims at— varies greatly. It could be anything from a single narrow goal (e.g., resolving to learn Hungarian because you want to be able to get by while traveling in Budapest) to a ! 272! holistic set of all of your endorsed aims and values (e.g., resolving to exercise because regular exercise will make your quality of life better, all-things-considered.) If X is broad, then a wide array of circumstances will count as imprudent escape conditions (e.g., exercising is bad for your joints, or takes up so much time that your work suffers, or stresses you out, etc.) If X is narrow, fewer sorts of circumstances will count as imprudent escape conditions (e.g., almost everyone in Budapest speaks English.) As noted above, the aim of promising typically includes satisfying the promisee’s interests in the promise, as evidenced by her acceptance of it. The set of circumstances preventing a promisor from reaching that goal will have only to do with the promisee’s interests, rather than with the interests of the promisor. Discovering that everyone in Budapest speaks English will not undermine a promise to my elderly Hungarian grandfather to learn Hungarian, because the goal of this promise is not my being able to navigate Budapest, but satisfying my grandfather’s interest in my learning Hungarian. It follows from this that many sorts of initial imprudence do not get a promisor off the hook; someone who knowingly makes a foolish promise is usually obligated to keep it, simply because she has promised to do so. Granted, massive initial imprudence can count as an escape condition on a promise, as with Frida’s promising George that she will take a luxury cruise that she cannot afford. But promisors are stuck with their slightly imprudent promises. Promises and resolutions also differ about whether immorality is always or only sometimes an escape condition. Immorality is always an escape condition on a promise; agents cannot be morally required all-things-considered to perform an immoral action, simply because they have promised to do so. But immorality is an escape condition on a resolution only if the resolving agent cares about morality; the ! 273! immorality of killing the snitch is an escape condition for Vinnie the mob boss, but not for the more morally sensitive underling. The explanation for this discrepancy has two parts. First, we must explain why immorality is always an escape condition on promises. The explanation is simple: when a promise is not subject to an escape condition, it yields all-things-considered moral obligation. So if an immoral promise were not subject to an escape condition, it would yield an all-things-considered moral obligation. But the promise is immoral by hypothesis. And if an action is immoral, you cannot be morally required to do it, all- things-considered. So immoral promises cannot yield all-things-considered moral obligations, which means that they must be subject to escape conditions. Second, we must explain why immorality is only sometimes an escape condition on resolutions. We have already seen in our discussion of imprudence that resolutions are subject to escape conditions whenever keeping the resolution would undermine one of the aims that the resolution was designed to achieve. Similarly, immorality counts as an escape condition on resolutions when acting morally is one of the goals at which the resolution aims. An agent who cares about morality will in general have acting morally included in her holistic set of goals and values. If such an agent were to act immorally, she would not attain all of her ends. She would be Φing in service of X when Φing turned out to be anti-X, just as the person who resolves to exercise in the service of a better quality of life fails to attain her end if exercising turns out to decrease her quality of life. It would be imprudent (in the sense of being inconsistent with one’s own goals) for such an agent to behave immorally. Immorality as an escape condition on resolutions is therefore a special instance of imprudence as an escape condition. ! 274! 9.4. How promises and resolutions are part of a spectrum There is not a bright line between ordinary resolutions on the one hand and promissory resolutions on the other. Rather, ordinary unconditional resolutions are at one low end of a spectrum that includes paradigmatic promises at the other high end. The closer two cases are to each other on this spectrum, the more similar their thresholds for revision will be—which is to say, the more similar their escape conditions will be. One way in which resolutions vary along the spectrum is whether they concern your interests alone, or the interests of other people, as well. For example, my resolution to finally wash my dirty car this weekend will be at the low end of the spectrum—it would be nice to have a clean car, but it’s not very important, and there are drought-based reasons weighing against it. My resolution to wash my elderly neighbor’s dirty car will be slightly higher up on the spectrum, since it involves not just my interests but her interests, as well. Another way in which resolutions vary along the spectrum is the importance of the goal in question. My efforts to motivate myself to finish my dissertation by resolving to refrain from watching the new season of Orange is the New Black until I turn the dissertation in is much higher on the spectrum than the resolution to wash my car, because finishing my dissertation is a lot more important. Consider the following non-promissory resolution: Mountain Resolution: Sam’s sister Rita is a mountain climber planning to climb Everest. She becomes chronically ill and can no longer climb. Sam thinks she would be happy if he climbed Everest in her honor, and placed a flag from her alma mater at the top of the mountain. He resolves to do so, so long as this would make her happy, and he doesn’t want to let his own interests get in the way of Rita’s happiness. Sam informs Rita of his resolution, and she confirms that his climbing Everest would make her very happy. A resolution like this one is near the high end of the spectrum, and will accordingly have a high threshold for revision. Realizing how time-consuming, expensive, or ! 275! physically arduous training for Everest will be will not license escape, in the way that such realizations might license escape had Sam resolved to climb Everest for his own personal satisfaction alone. To count as an escape condition on Mountain Resolution, a MICRO will have to quite strong indeed. For the whole point of the resolution is ruling out such considerations, and acting to attain Rita’s happiness. Mountain Resolution seems to have escape conditions of similar strength and type as the following promise: Mountain Promise: Rita wishes she could climb Everest, but a chronic illness prevents her from doing so. She wants to place a USC flag at the top of the mountain. She asks Sam to promise her that he will make it to the top of Everest someday, and put the flag there. He makes this promise, resolving to climb Everst so long as Rita continues to accept the promise. Sam does not want to let his own interests get in the way of keeping the promise. Mountain Promise is near the high end of the spectrum, as well. In both of these cases, only very important MICROs will license escape, and considerations having to do with Sam’s interests alone will not count as escape conditions. 229 There are very strong similarities between the two cases. In both cases, Sam resolves to climb Everest. And in both cases, this resolution is explicitly conditional on Rita’s continuing to hold some kind of pro-attitude towards the resolution: continuing to be happy about it in the resolution case, and continuing to accept it by refraining from releasing Sam in the promise case. Both cases therefore involve resolutions that !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 229 Because Sam publicly declares his resolution to Rita, Rita is expecting him to act. If this weren’t so—if Sam’s resolution to climb Everest remained private—there would be an obvious discrepancy between the two cases. Rita’s expectations in the promising case would function as a strong independent source of additional moral importance, thus making the threshold for revision higher. This explains why the revision threshold for private resolutions is usually much lower than that for promises. ! 276! are explicitly conditional on Rita’s interests, and exclusive of Sam’s interests. 230 This shows us that the resolution that Sam expresses when he makes the Mountain Promise—the resolution he must actually have if he is to be sincere—is very like, if not identical to, the Mountain Resolution. That is, the Mountain Resolution is the kind of mental state that Sam expresses when he makes the promise to Rita. The revision threshold for the Mountain Promise is so robust because the revision threshold for the resolution expressed by that promise is so robust. In general, the escape conditions on most resolutions are less stringent than those on promises. But the escape conditions on the sorts of resolutions we express when we make promises are a lot more stringent than those on most resolutions, which explains why promises seem to have higher thresholds for revision than do most resolutions. MSF derives the norms on promises from those on the mental state promises express. This allows us to explain why we are pro tanto morally obligated to keep all of our promises, even in the most marginal of cases. In these last two chapters, I have argued that MSF has even more explanatory power than it might first appear. For it also explains both when and why these pro tanto obligations are overridden or undermined, such that we are not morally obligated to keep our promises all-things- considered. The escape conditions are part of the norm that gets inherited. MSF shows !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 230 Should this lead us to think that the two cases are actually at the same point on the spectrum, such that Mountain Resolution must be understood not as a mere resolution but as some kind of implicit promise? Perhaps Sam’s declaration of a resolution counts as a promissory offer, and Rita’s telling Sam that his climbing Everest would make her happy counts as acceptance, such that there is no difference in the escape conditions between the two cases because the cases are both promises. However, we can still distinguish between the Mountain Promise and the Mountain Resolution, for the reasons I offered in Section 5 of Chapter 5: (1) the fact that promising involves a recognition that you are promising, which is lacking in Mountain Resolution, and (2) Sam’s declaration that he has a resolution in Mountain Resolution does not count as expressing a resolution, whereas he expresses a resolution in Mountain Promise. ! 277! us that it is permissible to break your promise in all and only those cases in which it is permissible to abandon the conditional resolution your promise expresses. This explanatory power is another point in favor of MSF as a theory of promissory obligation. ! 278! WORKS CITED Allemand, Mathias, Irina Amberg, Daniel Zimprich, and Frank D. Fincham. (2007). “The Role of Trait Forgiveness and Relationship Satisfaction in Episodic Forgiveness.” Journal of Social and Clinical Psychology 26(2): 199 – 217. Altham, J. E. J. (1985). “Wicked Promises.” In Exercises in Analysis, ed. I. Hacking. Cambridge, MA: Cambridge: University Press. Anderson, Scott. (2014). “Coercion.” The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta <http://plato.stanford.edu/archives/spr2014/entries/coercion/>. Anscombe, Elizabeth. (1963). Intention. Oxford: Blackwell. Aristotle. Nicomachean Ethics. Ed. Jonathon Barnes. Princeton: Princeton University Press, 1984. Armstrong, David. (1971). “Meaning and Communication.” The Philosophical Review 80 (4): 427 – 447. Audi, Robert. (1973). “Intending.” Journal of Philosophy 70: 387 – 403. Bedke, Matthew S. (2009). “The Iffiest Oughts: A Guise of Reasons Account of End- Given Conditionals.” Ethics 119 (4): 672 – 698. Bigelow, John, Dodds, Susan M., and Pargetter, Robert. (1990). “Temptation and the Will.” American Philosophical Quarterly 27 (1): 39 – 49. Bradley, Richard. (1999). “Conditional Desirability.” Theory and Decision 43 (1): 23 – 55. Brandom, Robert. (1983). “Asserting.” Noûs 17: 637–650 Bratman, Michael. (1987). Intention, Plans, and Practical Reason. Harvard University Press. -----. (2009). “Intention, Practical Rationality, and Self Governance.” Ethics 119 (3): 411— 443. -----. (Forthcoming). “Intention, Belief, Practical, Theoretical.” Spheres of Reason: New Essays in the Philosophy of Normativity. Ed Simon Robertson. Oxford University Press. Broome, John. (1999). “Normative Requirements.” Ratio 12(4): 398-419. -----. (2007). “Requirements.” In Hommage à Wlodek. Philosophical Papers Dedicated to Wlodek Rabinowicz. Ed. T. Rønnow-Rasmussen, B. Petersson, J. Josefsson & D. Egonsson. <www.fil.lu.se/hommageawlodek.> Brunero, John. (2010). “The Scope of Rational Requirements.” Philosophical Quarterly 60 (238): 28-49. ! 279! Butler, Joseph. (1846). Fifteen Sermons Preached at Rolls Chapel, in The Works of the Right Reverend Father in God, Joseph Butler, D.C.L., Late Bishop of Durham. Ed. Samuel Halifax. Carter. Carter, W. R. (1973). “On Promising the Unwanted.” Analysis 33 (3): 88 – 92. Cholbi, Michael. (2002). “A Contractualist Account of Promising.” Southern Journal of Philosophy 40: 475 – 491. Chwang, Eric. (2011). “On Coerced Promises,” in Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. Cicero, Marcus Tullius. De Officiis. Trans. Walter Miller. Project Gutenberg Ebook Edition. <http://www.gutenberg.org/files/47001/47001-h/47001-h.htm> Cohen, Daniel and Handfield, Toby. (2010). “Rational Capacities, Resolve, and Weakness of Will.” Mind 119 (476): 907 – 932. Cohon, Rachel. (2006). “Hume on Promises and the Peculiar Act of the Mind.” Journal of the History of Philosophy, 44.1: 25 – 45. Dannenberg, Jorah. (Forthcoming). “Promising Ourselves, Promising Others.” The Journal of Ethics. D’Arms, Justin and Daniel Jacobson. (2000). “The Moralistic Fallacy: On the ‘Appropriateness’ of Emotions.” Philosophy and Phenomenological Research 61 (1): 65 – 90. Darwall, Stephen. (1983). Impartial Reason. Ithaca, NY: Cornell University Press. Davidson, Donald. (1978). “Intending.” Reprinted in Essays on Actions and Events. Oxford: Oxford University Press, 1980: 83–102. Davis, Wayne. (1984). “A Causal Theory of Intending,” reprinted in The Philosophy of Action, ed. Alfred Mele. Oxford: Oxford University Press, 1997: 131 – 148. ------. (2008). “Replies to Green, Szabo, Jeshion, and Siebel.” Philosophical Studies 137: 427 – 445. DeRose, Keith. (2002). “Assertion, Knowledge and Context.” The Philosophical Review 111: 167–203. Douven, Igor. (2006). “Assertion, Knowledge, and Rational Credibility.” Philosophical Review 115: 449–485. Dressel, Ashley. (2014). “Directed Obligations and the Trouble with Deathbed Promises.” Ethical Theory and Moral Practice: 1 – 13. Driver, Julia (2011). “Promising Too Much.” in Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. ! 280! Dworkin, Ronald. (2011). Justice for Hedgehogs. Cambridge, MA: Belknap Press of Harvard University Press. Engel, Beverly. “How To Give A Meaningful Apology.” Family Business Center, University of Massachusetts, Amherst. http://www.umass.edu/fambiz/articles /resolving_conflict/meaningful_apology.html Eriksson, John. (2011). “Straight Talk: Conceptions of Sincerity in Speech.” Philosophical Studies 153 (2): 213 – 234. Feinberg, Joel. (1987). The Moral Limits of the Criminal Law, Volume I: Harm to Others. Oxford: Oxford University Press. Fennel, D. L. (1993). “Characteristics of Long-Term First Marriages.” Journal of Mental Health Counseling 15: 446 – 460. Fincham, Frank D. and Steven R. H. Beach. (2002). “Forgiveness in Marriage: Implications for Psychological Aggression and Constructive Communication.” Personal Relationships 9: 239 – 251. Finnis, John. (1980). Natural Law and Natural Rights. Oxford: Oxford University Press. Foot, Philippa. (2001). Natural Goodness. Oxford: Oxford University Press. Friedrich, Daniel and Nicholas Southwood. (2009). “Promises Beyond Assurance.” Philosophical Studies 144 (2): 261 – 280. ------. (2011). “Promises and Trust.” In Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. Gauthier, David. (1986). Morals by Agreement. Oxford: Clarendon Press. ------. (1997). “Resolute Choice and Rational Deliberation: A Critique and a Defense.” Noûs 31 (1): 1 – 25. Gibbs, Raymond and Suzanne Delaney. (1987). “Pragmatic Factors in Making and Understanding Promises.” Discourse Processes 8: 107 – 126. Gilbert, Margaret. (1993). “Agreements, Coercion, and Obligation.” Ethics 103 (4): 679 – 706. ------. Gilbert, Margaret. (2006). A Theory of Political Obligation: Membership, Commitment, and the Bonds of Society. Oxford: Oxford University Press. Green, Mitchell. (2007). “How Do Speech Acts Express Psychological States?” in John Searle's Philosophy of Language: Force, Meaning and Mind, ed. S. L. Tsohatzidis. Cambridge, MA: Cambridge University Press: 267-84. ------ (2009). “Speech Acts.” The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta. http://plato.stanford.edu/archives/spr2009/entries/speech-acts ! 281! Griswold, Charles L. (2007). Forgiveness: A Philosophical Exploration. Cambridge, MA: Cambridge University Press. Habib, Allen. (2009). “Promises to the Self.” Canadian Journal of Philosophy 39 (4): 537 – 557. Harman, Gilbert. (1976). “Practical Reasoning.” Reprinted in The Philosophy of Action ed. Alfred Mele. Oxford: Oxford University Press, 1997: 149–177. Haupt, Angela. (29 August 2012). “How to Forgive, and Why You Should.” US News and World Report. <http://health.usnews.com/health- news/articles/2012/08/29/how-to-forgive-and-why-you-should> Hawthorne, John. (2002). Knowledge and Lotteries. Oxford: Oxford University Press. Heuer, Ulrike. (2012a). “Promising—Part 1.” Philosophy Compass 7 (12): 831-841. -------. (2012b). “Promising—Part 2.” Philosophy Compass 7 (12): 842-851. Hill, Krista M. (2014). When Are Apologies Effective? An Investigation of the Components That Increase an Apology's Efficacy. Order No. AAI3561221, Dissertation Abstracts International: Section B: The Sciences and Engineering. Hills, Alison. (2003). “Duties and Duties to the Self.” American Philosophical Quarterly 40: 131 – 42. Holmes, Janet. (1995). Women, Men, and Politeness. New York: Longman. Holton, Richard. (2004). “Rational Resolve.” Philosophical Review. 113 (4): 507 – 535. -----. (2009). Willing, Wanting, Waiting. Oxford: Oxford University Press. Hooker, Brad. (2011). “Promises and Rule-Consequentialism.” In Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. Hughes, Paul M. (2011). “Forgiveness.” The Stanford Encyclopedia of Philosophy. ed. Edward N. Zalta. <http://plato.stanford.edu/archives/win2011/entries/ forgiveness>Hooker, Brad. (2011). “Promises and Rule-Consequentialism.” In Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. Hume, David. (1741). A Treatise of Human Nature. Ed. L.A. Selby-Bigge and P. H. Nidditch, 2nd Edition. Oxford: Clarendon Press, 1978. Jucker, Andreas H. (2009). “Speech Act Research Between Armchair, Field and Laboratory: The Case of Compliments.” Journal of Pragmatics: 1611–1635. Kagan, Shelly. (1989). The Limits of Morality. Oxford: Clarendon Press. Kant, Immanuel. (1785). Foundations of the Metaphysics of Morals. Trans. Lewis White Beck. New York: Macmillan, 1990. ! 282! Kolodny, Niko. (2005). “Why Be Rational?” Mind 114(455): 509-563. ------. (2007). “How Does Coherence Matter?” Proceedings of the Aristotelian Society 107 (1.3): 229 – 263. -----. (Forthcoming). “Instrumental Reasons.” In The Oxford Handbook of Reasons and Normativity, ed. Daniel Star. Oxford: Oxford University Press. Kolodny, Niko and Jay Wallace. (2003). “Promises and Practices Revisited.” Philosophy & Public Affairs 31(2): 119–154. Korn, Fred and Shulamit R. Decktor Korn (1983). “Where People Don’t Promise.” Ethics 93.3: 445-450. Kovach, Vanya and Fiztpatrick, John. (1999.) “Resolutions.” Australasian Journal of Philosophy. 77 (2): 161 – 173. Lackey, Jennifer. (2006). “Norms of Assertion.” Nous 41: 594–626. Lawler, Kathleen A., J.W. Younger, R.L. Piferi, R. L. Jobe, K.A. Edmondson, and W.H. Jones. (2005). “The Unique Effects of Forgiveness on Health: An Exploration of Pathways.” Journal of Behavioral Medicine 28 (20): 157 – 67. Lawler-Row, Kathleen A., and Johan C. Karremans, Cynthia Scott, Meirav Edlis- Matityahou, Laura Edwards. (2008). “Forgiveness, Physiological Activity, and Health: The Role of Anger.” International Journal of Psychophysiology 68 (1): 51– 58. Lennertz, Ben. (2015). “Quantificational Credences.” Philosopher’s Imprint 15 (9): 1-24. Liberman, Alida and Mark Schroeder. (forthcoming). “Commitment: Worth the Weight.” Forthcoming in Weighing Reasons, ed. Errol Lord and Barry Maguire. Oxford: Oxford University Press. Liberman, Alida. (ms). “Resolutions are Not Promises to the Self.” Liberto, Hallie. (ms). “Promissory Obligation: Against a Unified Account.” Presented at the Arizona Workshop in Normative Ethics, 2015. Lord, Errol. (2011). “Violating Requirements, Exiting from Requirements, and the Scope of Rationality.” Philosophical Quarterly 61 (3): 392-399. Ludwig, Kirk. (1992). “Impossible Doings.” Philosophical Studies 65 (3): 257 – 281. Luper, Steven (2007). “Moral Harm.” Philosophical Quarterly 57.227: 239–251. Lyons, David. (1965.) Forms and Limits of Utilitarianism. Oxford: Clarendon Press. MacFarlane, John. (2005). “Making Sense of Relative Truth.” Proceedings of the Aristotelian Society 105: 321–339. ! 283! ------. (2011). “What is Assertion?” in Assertion, ed. Jessica Brown and Herman Cappelen. Oxford: Oxford University Press: 79-96. Mackie, John. (1977). Ethics: Inventing Right and Wrong. London: Penguin Books. Macnamara, Colleen. (2013). “‘Screw You!’ & ‘Thank You.’” Philosophical Studies 165: 893 – 914. Maier, John. (2010). “Review of Willing, Wanting, Waiting.” Australasian Journal of Philosophy, 89 (2): 361 – 364. Mason, Elinor. (2005). “We Make No Promises.” Philosophical Studies 123: 33-46. Marmor, Andrei. (2009). Social Conventions: From Language to Law. Princeton: Princeton University Press. Marušić, Berislav. (2012). “Belief and Difficult Action.” Philosopher’s Index. 12 (18): 1 – 30. ------. (2013). “Promising Against the Evidence.” Ethics 123.2: 292 – 317. McClennen, Edward F. (1990). Rationality and Dynamic Choice: Foundational Explorations. Cambridge, MA: Cambridge University Press. McCullough, Michael E., Shelly D. Kilpatrick, Robert A. Emmons, and David B. Larson. (2001). “Is Gratitude a Moral Affect?” Psychological Bulletin 127: 249 – 266. McGuire, M.C. (1961). “Decisions, Resolutions, and Moral Conduct.” Philosophical Quarterly. 11 (42): 61 – 67. Mele, Alfred. (1987). Irrationality. Oxford: Oxford University Press. Mihai, Mihaela. (2013). “Apology.” Internet Encyclopedia of Philosophy. http://www.iep.utm.edu/apology/#H1. Nagel, Thomas. (1970). The Possibility of Altruism. Oxford: Clarendon Press. Norcross, Alastair. (2011). “Act-Utilitarianism and Promissory Obligation.” In Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. Owens, David. (2006). “A Simple Theory of Promising.” Philosophical Review 115 (1): 51 – 77. ------. (2007.) “Duress, Deception, and the Validity of a Promise.” Mind 116: 293–315. ------. (2008). “Promising Without Intending.” Journal of Philosophy 105 (12): 737 – 755. ------. (2012). Shaping the Normative Landscape. Oxford: Oxford University Press. Pagin, Peter. (2014). “Assertion.” The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta. http://plato.stanford.edu/archives/spr2014/entries/assertion/ ! 284! Paul, Sarah. (2011). “Review of Willing, Wanting, Waiting.” Mind 120 (479): 889 – 892. Peetz, Vera. (1977). “Promises and Threats.” Mind 86 (344): 578 – 581. Peirce, C. S. (1934). Belief and Judgement. Cambridge, MA: Harvard University Press. Pink, Thomas. (2009). “Promising and Obligation.” Philosophical Perspectives 23 (1): 389 – 420. Pitcher, George (1984). “The Misfortunes of the Dead.” American Philosophical Quarterly 21.2: 183-188. Pratt, Michael G. (2007). “Promises, Contracts, and Voluntary Obligations.” Law and Philosophy 26.6: 531 – 574. Rawls, John. (1971). A Theory of Justice. Cambridge, MA: Harvard University Press. Raz, Joseph (I972). “Voluntary Obligations and Normative Powers II.” Proceedings of the Aristotelian Society, Supplementary Volume 46: 79-101. ------. (1977). “Promises and Obligations.” In Law, Morality, and Society: Essays in Honor of H.L.A. Hart, Ed. P.M.S. Hacker and Joseph Raz. Oxford University Press: 210-228. ------. (1982). “Promises in Morality and Law.” Harvard Law Review: 916 – 938. ------. (1983). The Authority of Law. Oxford: Oxford University Press. ------. (1990). Practical Reason and Norms. Princeton: Princeton University Press. ------. (2005). “The Myth of Instrumental Rationality.” Journal of Ethics and Social Philosophy 1: 1 – 28. Ridge, Michael. (2006). “Sincerity and Expressivism.” Philosophical Studies 131: 487 – 510 Romano, Hara Estroff. (1 March 2004). “The Art of the Compliment.” Psychology Today. <http://www.psychologytoday.com/articles/200403/the-art-the-compliment>. Rorty, Amelie. (1980). “Where Does the Akratic Break Take Place?” Australasian Journal of Philosophy 58: 333 – 46. Rosati, Connie. (2001). “The Importance of Self-Promises.” in Promises and Agreements ed. Hanoch Sheinman. Oxford: Oxford University Press. Ross, Jake. (2009). “How to be a Cognitivist about Practical Reason.” Oxford Studies in Metaethics 4: 243–281. Ross, W.D. (1930). The Right and the Good. Oxford: Clarendon Press. Scanlon, T. M. (1990). “Promises and Practices.” Philosophy & Public Affairs 19: 199–226. ! 285! -----. (1998). What We Owe to Each Other. Cambridge, MA: Harvard University Press. Schroeder, Mark. (2009). “Means End Coherence, Stringency, and Subjective Reasons.” Philosophical Studies 143 (2): 223 – 248. Searle, John. (1969). “The Structure of Illocutionary Acts.” Chapter 3 of Speech Acts, Cambridge, MA: Cambridge University Press: 54 – 71. ------. (1975). “Indirect Speech Acts.” Syntax and Semantics 3: 59 – 82. ------. (2001). Rationality in Action. Cambridge, MA: MIT Press. Seybold, Kevin S., and Peter C. Hill, Joseph K. Neumann, and David S. Chi. (2001). “Physiological and Psychological Correlates of Forgiveness.” Journal of Psychology and Christianity 20: 250 – 259. Sheinman, Hanoch. (2008). “Promise as Practice Reason.” Acta Analytica 23(4): 287-318. Shiffrin, Seana. (2008). “Promising, Intimate Relationships, and Conventionalism.” Philosophical Review 117 (4): 481-524. -----. (2011). “Immoral, Conflicting, and Redundant Promises.” In Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon, ed. R. Jay Wallace, Rahul Kumar, and Samuel Freeman. Oxford: Oxford University Press: 155–78. Shpall, Sam. (2013). “Wide and Narrow Scope.” Philosophical Studies 163 (3): 717 – 736. ------. (forthcoming). “Moral and Rational Commitments.” Forthcoming in Philosophy and Phenomenological Research. Sidgwick, Henry (1907). The Methods of Ethics: Seventh Edition. Indianapolis: Hackett Publishing Company. Singer, Marcus. (1959). “On Duties to Oneself.” Ethics 69 (3): 202 – 205. Smith, Adam. (1759). The Theory of Moral Sentiments. Project Gutenberg Ebook. <http://ebooks.gutenberg.us/WorldeBookLibrary.com/moralsent.htm> Smith, Holly. (1997). “A Paradox of Promising.” Philosophical Review 106 (2): 153 – 196. Smith, Michael. (2011). “The Value of Making and Keeping Promises.” In Promises and Agreements, ed. Hanoch Sheinman. Oxford: Oxford University Press. Tappolet, Christine. (Forthcoming.) “Weakness of Will.” International Encyclopedia of Ethics. Ed. Hugh LaFolette. Wiley-Blackwell. Taylor, Erin. (2013). “A New Conventionalist Theory of Promising.” Australasian Joural of Philosophy 91 (4): 667 – 682. Thomson, Judith Jarvis. (1990). The Realm of Rights. Cambridge, MA: Harvard University Press. ! 286! Tognazzini, Neil. (2007). “The Hybrid Nature of Promissory Obligation.” Philosophy and Public Affairs 35.3: 203-232. Velleman, David. (1989). Practical Reflection. Princeton, NJ: Princeton University Press. Wallace, R. Jay. (2006). Normativity and the Will. Selected Essays on Moral Psychology and Practical Reason. Oxford: Clarendon Press. Watson, Gary. (2004). “Asserting and Promising.” Philosophical Studies 117: 57 – 77. ------. (2009). “Promises, Reasons, and Normative Powers.” In Reasons for Action, ed. David Sobel and Steven Wall. Cambridge, MA: Cambridge University Press. Way, Jonathan. (2010). “Defending the Wide Scope Approach to Instrumental Reason.” Philosophical Studies 147 (2): 213 – 233. -----. (2012). “Transmission and the Wrong Kind of Reason.” Ethics 122 (3): 489 – 515. Weiner, Matthew. (2005). “Must We Know What We Say?” Philosophical Review 114: 227 – 251. Wiggins, David. (1978). “Weakness of Will, Commensurability, and the Objects of Deliberation and Desire.” In Needs, Values, Truth: Essays in the Philosophy of Values. Basil Blackwell, 1987: 239 – 67. Williamson, Timothy. (2000). Knowledge and its Limits. Oxford: Oxford University Press Wright, Crispin. (1992). Truth and Objectivity. Cambridge, MA: Harvard University Press. Zeelenberg, Marcel. (1999). “The Use of Crying Over Spilled Milk: A Note on the Rationality and Functionality of Regret.” Philosophical Psychology 12.3: 325 – 340.

Abstract (if available)

Abstract My dissertation, The Mental States First Theory of Promising, argues that we can explain the norms on promises by appealing to the norms on resolutions, which are the underlying mental state expressed by promises. This approach allows us to explain promissory obligation in a wider variety of cases than typical theories can accommodate, as well as explaining when and why promise breaking is permissible. Moreover, I argue that the norms on a wide variety of speech acts can be explained by appeal to the mental states the acts express in a similar way, which lends support to my theory and suggests a promising avenue for further research.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Contract, from promise to commodity

PDF

Reasoning with uncertainty and epistemic modals

PDF

The boundaries of the intentional

PDF

Units of agency in ethics

PDF

A perceptual model of evaluative knowledge

PDF

The case for moral skepticism

PDF

The minimal approval view of attributional-responsibility

PDF

Process-oriented rationality

PDF

On being bound: law, authority, and the politics of obligation

PDF

Rationality and the primacy of the occurrent

PDF

Contrastive reasons

PDF

Describing authority

PDF

Reasons, obligations, and the structure of good reasoning

PDF

Meaningfulness, rules, and use-conditional semantics

PDF

Aggregating complaints

PDF

Iffy confidence

PDF

Responding to harm

PDF

Constraining assertion: an account of context-sensitivity

PDF

Beliefs that wrong

PDF

Distributive equality and labor burdens

Asset Metadata

Creator Liberman, Alida (author)

Core Title The mental states first theory of promising

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Philosophy

Publication Date 07/10/2015

Defense Date 05/13/2015

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

Tag ethics,OAI-PMH Harvest,Promises,resolutions

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Schroeder, Mark A. (committee chair), Finlay, Stephen (committee member), Gruner, Wolf (committee member), Ross, Jacob (committee member), Watson, Gary L. (committee member)

Creator Email aliberma@usc.edu,alida.liberman@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-591683

Unique identifier UC11300327

Identifier etd-LibermanAl-3585.pdf (filename),usctheses-c3-591683 (legacy record id)

Legacy Identifier etd-LibermanAl-3585.pdf

Dmrecord 591683

Document Type Dissertation

Format application/pdf (imt)

Rights Liberman, Alida

Type texts

Source 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 a...

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