University of Southern California Dissertations and Theses

The boundaries of the intentional

(USC Thesis Other)

The boundaries of the intentional

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content THE BOUNDARIES OF THE INTENTIONAL by Brandon Darryl Johns 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) May 2012 Copyright 2012 Brandon Darryl Johns ii DEDICATION Without the undying love and support of my Mother and Father, this dissertation would never have been completed. I dedicate this work to them with love and gratitude. iii ACKNOWLEDGEMENTS Throughout my collegiate career, I have been fortunate to receive instruction from some of the brightest and most distinguished scholars in the field of philosophy. I am most grateful to my advisor, Professor Gideon Yaffe. As a young graduate student, Gideon first introduced me to topics in the philosophy of action via an independent study. This study has had a profound impact on the trajectory of my academic life. Gideon has spent countless hours meeting with me at various locations around the city, reading my work, and providing thorough feedback and direction. He has been patient and encouraging even during periods he easily could not have been. I am forever in his debt. I would also like to thank the members of my dissertation committee for their guidance and feedback: Professor Thomas Lyon of the USC Law School, Professor Stephen Finlay of the USC Philosophy Department, and Professor Gary Watson of the USC Law School and Philosophy Department. Several staff members in the USC Philosophy Department have been immensely helpful throughout the years as well: Mrs. Cynthia Lugo, Mrs. Tomiko Higuchi, Mrs. Valerie Hunt, and Mrs. Barrington Smith- Seetachitt. Without the assistance of these women, graduate school would have been considerably more complicated and less enjoyable. I also thank Professors Mark Schroeder and Eric Schwitzgebel. Mark is perhaps the hardest working individual I’ve ever encountered. In addition to helping me become a better philosopher and teacher, he continually pressured me to complete my iv dissertation; and though this pressure often caused me a great deal of stress, I’m convinced I would not have completed this dissertation had it not been for his efforts. For this I am supremely grateful. I am also supremely grateful to Eric. Eric was my undergraduate mentor. It was in virtue of his encouragement and instruction that I chose to pursue philosophy at the graduate level. Last but not least, I thank my three roommates. Much of my dissertation was written during the “graveyard shift” hours. Two of my roommates—my brother Aaron Johns and Emily Charelian—often adjusted their schedules and routines in order to accommodate and support my unusual schedule. My third roommate, my cat Risotto, also played a crucial role: had it not been for her late night antics I would surely have lost my mind. v TABLE OF CONTENTS Dedication ii Acknowledgements iii Abstract vi Introduction 1 Chapter 1: Trying, Intending, and the Double Phenomenon View 19 The Belief/Desire Model of Intention 22 The Planning Theory of Intention 32 The Double Phenomenon View 45 Chapter 2: The Single Phenomenon View vs. the Double Phenomenon View 52 Two Arguments for SPV 55 Two Better Arguments for SPV 61 Two Arguments against SPV 72 Chapter 3: An Actional Account of Refraining 80 Brand and Yaffe on Refraining 82 A Difficulty for REF 85 Reformulating REF 106 Three Objections to REF* 109 Chapter 4: Non-Intentional Action 115 Gorr and Horgan’s Argument 115 The Simple View Argument 119 Chan’s Argument 121 Meiland’s Argument 125 An Argument from Side-Effect Action 129 An Argument from Lucky Action 142 Conclusion 150 Comprehensive References 164 vi ABSTRACT My dissertation explores the nature and extent of intentional action. I begin by arguing for a necessary connection between intentional action and the state of intention. I then argue that this commonsense view—known as the Single Phenomenon View—is inconsistent with two highly influential theories of intention; thus, both theories require modification. My focus then turns to the possibility of negative intentional action. I defend an account of refraining whereby refrainings are conceived as a type of intentional action. Furthermore, my account of refraining leads to a novel argument for the existence of non-intentional action—action that is neither intentional nor unintentional. This argument involves, among other things, developing an account of unintentional action. One important consequence of my dissertation is that all action—intentional, unintentional and non-intentional—issues from the state of intention. 1 INTRODUCTION The Mind-Body Problem One reason why philosophy of action is important is because of its connection to the mind-body problem. The mind-body problem is a cluster of problems or puzzles surrounding the relation between the mental and the physical. One popular form of the problem traces back to Descartes. Descartes famously held that mind and body were distinct substances. 1 This view, commonly known as substance dualism, claims that mind is essentially non-physical or immaterial, whereas the body is material. Conceiving of mind and body along these lines, however, invites a difficult and well-known problem involving causation. It is apparent that the mental and the physical interact causally. External objects and events exert causal influence on our minds, as in perception. The red apple on the table, for example, causes me to believe that there is a red apple on the table and produces in me a desire to eat it. But interaction also runs the other way round; our minds exert influence over our bodies, as in action. My belief that there is an apple on the table coupled with my desire to eat it results in an intention to eat the apple. This intention compels me to reach for the apple and take a bite. But, how can bi-directional interaction occur between mental and physical if the mental is non-physical? After all, physical effects seem to require physical causes. 1 Descartes (1641); see especially the Second and Sixth Meditations. 2 The trend today is toward a monistic view of substance. More specifically, philosophers, in one form or other, tend to be physicalists or materialists about the mental. By conceiving of mind and body as a single substance, physicalists are perhaps better positioned to address the problem of causation troubling substance dualists. 2 Even still, various puzzles concerning mind and body remain. 3 The form of the mind-body problem that I am concerned with is the challenge of identifying the mental items that contribute to human behavior. This challenge is related to the problem of mental causation above. I am interested in exploring the following complex question: Which states of mind are responsible for action and what is the nature of these states? For instance, many assume that the state of intention is necessary for intentional action. Is this assumption true? If so, what are intentions? And how do they relate to other states like beliefs and desires? The complex question above constitutes a form of the mind-body problem—one that is firmly situated in the philosophy of action. The philosophy of action, therefore, has the potential to solve an important piece of the mind-body problem. Intentional Action and Intention My dissertation, broadly speaking, attempts to further our understanding of the relation between mind and body. It’s an investigation into the nature and extent of intentional 2 Of course, this depends on what kind of physicalist one is. Non-reductive physicalists perhaps still face a problem concerning mental causation. For more on this, see Crane (1995); Davidson (1993); and Jackson (1996). 3 For instance, there is the famous challenge for materialists to explain conscious experience. 3 action: i.e., what it is and where in the world we find it. Chapters 1-3 explore the nature of intentional action; chapters 3-4 its extent. 4 As Al Mele has pointed out, intentional action is of primary importance to the philosophy of action. 5 First, if there were no intentional actions, then arguably there would be no actions whatsoever. Donald Davidson and Jennifer Hornsby, for instance, have argued that all action is intentional under some description. 6 Furthermore, chapter 4 of my dissertation implies that all action necessarily involve doing something intentionally. Second, discussions about free action usually proceed from the intentional. Action theorists working in this important area are primarily interested in identifying conditions under which intentional actions are free. This is not to say that the conditions under which, say, non-intentional acts are free is not itself an interesting issue; but there’s a sense in which it’s not as fundamental as the question of the conditions under which intentional actions are freely performed. Third, discussions of moral responsibility tend to focus on actions that are intentionally performed. This is presumably no accident: in order to develop an account of the conditions under which one is responsible for an unintentional action, one should first have an idea about what to say in paradigmatic cases of actions for which one should be held responsible. These paradigmatic cases are most likely ones in which the relevant actions are performed intentionally. An investigation into intentional action quickly leads to issues involving intention. Intuitively, intentional action is linked to the state of intention. After all, it 4 This is not a typo. Chapter 3 explores both the nature and extent of intentional action. 5 Mele (1997), p. 16. 6 See Davidson (1971) and Hornsby (1993). 4 hardly seems an accident that the term ‘intention’ is present in ‘intentional action’. But in what way exactly are the two linked? One natural proposal goes like this: intentionally performing an act, A, requires intending to do A. In the literature, this thesis is called the Simple View (SV). 7 SV posits the tightest possible fit between intention and action: in order to act intentionally, one must intend that very act. I would not be surprised if the general public were largely sympathetic to SV. And perhaps for good reason: it would seem that most of our intentional actions, especially paradigmatic ones, are intended. When I intentionally go to the library downtown, I usually intend to go; and when I eat dinner at Orochon afterwards, it, too, is exactly what I intend. SV does a great job of capturing the link between intention and action when it comes to many of our actions. But is SV true? Probably not. Michael Bratman has forcefully argued against it. 8 Bratman imagines a scenario in which an agent tries to do each of two things, but cannot do both. Sarah is playing a video game. She tries to hit target 1 and tries to hit target 2, while knowing that she can’t hit both. If she hits target 1 in the way she tries, then she hits the target intentionally. According to SV, Sarah must intend to hit target 1. Bratman argues that if she intends this, then by symmetry she must intend to hit target 2. But if she has both of these intentions, she must be irrational because she knows she can’t hit both targets. Since Sarah is not irrational, Bratman concludes she must not have either intention; and SV fails. 7 SV is endorsed by Adams (1986) and McCann (1991). In a recent paper, Wasserman (forthcoming) defends SV from various objections; however, it’s unclear whether he ultimately endorses the thesis. 8 Bratman (1987), ch. 8. 5 Many have been persuaded by Bratman’s argument. Some, however, aren’t convinced. Hugh McCann has argued that an intention to try to do something entails an intention to succeed. 9 McCann also thinks that Sarah intends to try to hit target 1 and intends to try to hit target 2 in Bratman’s example. If she has these intentions, then she must, contrary to Bratman, intend to hit target 1 and intend to hit target 2. Thus, Bratman’s example does not threaten SV. In chapter 4 (section 2), I argue against McCann’s claim that an intention to try to A entails an intention to A. The argument essentially makes the point that in some cases it is perfectly reasonable for an agent, S, to intend to try to do something, A, yet not reasonable for S to intend to do A. In which case, the latter intention must not follow from the former. Ryan Wasserman objects to Bratman’s argument on different grounds. 10 He argues that Sarah neither hits the target intentionally nor unintentionally, but rather non- intentionally. And since SV is only a claim about intentional action, Wasserman concludes that Bratman’s argument poses no threat to SV. I’m not persuaded by Wasserman’s argument for the following reason. As I argue in chapter 4, non- intentionality comes in two forms: lucky action and side-effect action. Sarah’s hitting the target certainly is not a side-effect—at least on any plausible definition of side-effect action. Is Sarah’s hitting the target lucky? Well, here things are less clear. If her hitting it is lucky, then I agree with Wasserman that her act is non-intentional. So let us grant that in Bratman’s original scenario Sarah gets lucky when she hits target 1. The problem 9 McCann (1991). 10 Wasserman (forthcoming). 6 for Wasserman is that we can alter Bratman’s example in a way that eliminates luck. Suppose Sarah is highly skilled at hitting the targets, but she often doubts her skill. She tries to hit both because she lacks confidence that she will hit either on this occasion, and she reasons that trying to hit both will increase her chances of hitting one. When she hits target 1 in just the way she tries, her hitting it is intentional rather than lucky. Now perhaps Wasserman might insist that this just goes to show that the category of the non-intentional is not exhausted by lucky and side-effect action. That is, Wasserman might stand his ground and insist that even in the altered scenario Sarah’s hitting the target is non-intentional. But it’s difficult to see what would motivate this view. Sarah’s target-hitting behavior is initiated and guided by an intention. Furthermore, she tries to hit the target and (non-luckily) succeeds in just the way she tries. Her hitting the target is straightforwardly intentional. To claim otherwise is implausible. In addition to Bratman’s example, there also seem to be other, simpler counterexamples to SV. It seems that one might intend to do A or B, while maintaining no preference for doing one over the other. I might intend to buy a ticket to see Pulp Fiction or Kill Bill. I slam the money on the counter and demand a ticket to either film. When the theater employee hands me a ticket to Kill Bill, it seems that I buy the ticket intentionally. This, despite not intending (flat out) to buy a ticket to Kill Bill. The implausibility of SV, then, calls for a loosening of the connection between intentional action and intention. We can loosen the connection in the following way: intentionally A-ing requires having an intention. This weaker thesis is known in the 7 literature as the Single Phenomenon View (SPV). 11 On SPV, an agent might A intentionally by intending to do A, but she might only intend to try to A, or intend to A if one can, or intend to A or B, and so on. To use Bratman’s terminology, one’s A-ing is intentional in case A is in the ‘motivational potential’ of one’s intention. 12 SPV, it would seem, does a better job than SV of capturing the relation between intentional action and intention. Chapter 1 SPV is one of the more attractive principles in philosophy of action. One way this is apparent is in our expectations concerning the theory of intention. We expect plausible theories of intention to support SPV. This brings us to chapter 1 of my dissertation. SPV informs us that intentional action necessarily involves intention. This raises the question: what is an intention? There are two basic types of views about this. One view holds that intentions are composed of other, more basic mental states. The other view, simply put, denies this. The former type of view constitutes a reductive approach to intention; the latter view a non-reductive approach. In chapter 1, I argue that one major reductive theory of intention and one major non-reductive theory both fail to support SPV; both are, in other words, incompatible with SPV. For this reason, I conclude that both theories require modification. 11 This term was coined by Bratman (1987), ch. 8. Proponents of SPV include: Bratman (1987); Harman (1976); and Mele (1992). 12 See Bratman (1987), ch. 8. 8 In contemporary philosophy of intention three important theories dominate the scene. Two of these three theories are reductionist in approach. What does each theory reduce intention to? One view holds that intention is a special kind of belief: namely, a belief or expectation that one will act in a particular way. 13 On this view, intentions are usually thought to be motivated by desires of some sort, though desire is not a proper part of intention. According to the second major reductive view, intentions consist of pairs of beliefs and desires. 14 The most influential belief/desire theory places three major conditions on intention. The belief condition is perhaps best known in the literature as the Strong Belief Requirement: intending to do A requires believing that one will (probably) A. 15 The desire condition we’ll call the Strict Desire Requirement on intention: intending to do A requires (intrinsically or extrinsically) desiring to do A. 16 The third condition is essentially causal; let us call it the Causal Requirement: intending to do A requires that one’s belief that one will (probably) A be caused by one’s desire to do A. Michael Bratman is the architect of the third leading theory of intention. 17 Bratman’s theory is easily the most influential and sophisticated non-reductionist account of intention ever developed. According to Bratman, intention differs from other practical 13 Important proponents of this kind of view include Setiya (2007b) and Velleman (1989). 14 Important proponents of this kind of view include Audi (1973) and Davis (1984). 15 Proponents of this requirement include Audi (1973; 1986), Davis (1984); Grice (1971); Harman (1976); Setiya (2007a; 2007b); and Velleman (1989). For a more complete list of proponents, see note 36, ch. 1. 16 Proponents of this requirement include Audi (1973; 1986); Davis (1984); and Goldman (1970). For a more complete list of proponents, see note 37, ch. 1. 17 Bratman (1987). 9 attitudes (e.g., desire) in virtue of its role in inter- and intra-personal planning. Central to Bratman’s planning theory is a trio of norms governing rational intention. First is the Strong Consistency Requirement: intending to A rationally requires not believing that not- A. The Strong Consistency Requirement is, in other words, the demand that one’s beliefs and intentions be consistent. The second norm is the Means-End Coherence Requirement: if one intends to A and believes that B is a necessary means to A, then rationality requires that one intend to B. The third is known as the Agglomerativity Requirement. Formulating this requirement is perhaps more challenging than the first two. Regarding agglomerativity, Bratman himself says, “If I both intend to [A] and intend to [B], there will be rational pressure for me to intend to [A and B].” 18 This remark suggests the following formulation of the Agglomerativity Requirement: if one intends to A and intends to B, then rationality requires that one intend to A and B. 19 This trio of norms is justified on Bratman’s view according to their tendency to promote the aims of planning agency. In chapter 1 I develop an argument that the belief/desire theory and Bratman’s planning theory of intention are both committed to rejecting SPV. 20 As a result, both must embrace what I call the Double Phenomenon View (DPV): intentional action does not require having an intention. The argument runs as follows. I first present a pair of 18 Bratman (1987), p. 134. 19 Yaffe (2006) has argued for an alternative formulation of the Agglomerativity Requirement. See also ch. 2, section 2.1 in Yaffe (2010) for a discussion of all three norms. I discuss Yaffe’s formulation of the Agglomerativity Requirement in the Conclusion. 20 More accurately, I should say the versions of these two theories sketched above are committed to rejecting SPV. To keep things simple, I won’t keep reminding readers that my argument concerns these two particular versions of the belief/desire and planning theory. When I discuss these two theories below, one should assume that I am referring to the versions of them sketched above (unless otherwise indicated). 10 examples in which agents attempt to perform an action. In the first example, I describe an agent that violates the Strong Belief and Strict Desire Requirements with respect to numerous, possible act-motivating intentions. I conclude that on the belief/desire view, trying to act does not require having an intention. In the second example, I utilize Bratman’s three norms above in order to eliminate numerous, possible act-motivating intentions. I conclude that on the planning theory, trying to act, too, does not require having an intention. From here, I argue that DPV follows. If trying does not require having an intention, then intentional action does not require having an intention. This, I suggest, is because either (a) trying itself constitutes intentional action, or (b) trying that results (in the right way) in success constitutes intentional action. Thus, the belief/desire theory and the planning theory of intention are incompatible with SPV; as such, both theories require modification. Chapter 2 The conclusion that the two theories of intention require modification because they are at odds with SPV relies on the assumption that SPV is more plausible than any alternative principle about the relation between intentional action and intention. I suggested above that SPV was preferable to SV. But the question remains: is SPV more plausible than DPV? Like SPV, DPV is a principle about the relation between intentional action and intention; it posits an unnecessary connection between intentional action and intention. Perhaps the fact that these two important theories of intention entail DPV provides a good 11 reason to doubt SPV? The question of whether SPV is preferable to DPV is the topic of chapter 2. I argue here that SPV is, in fact, preferable to DPV. In chapter 2 I present a total of four arguments in favor of SPV. Along the way, I offer what I believe are the strongest possible responses on behalf of those committed to DPV. The first two arguments for SPV—The Linguistic Argument and The X-Phi Argument—are unpersuasive. The Linguistic Argument essentially makes the point that the word ‘intention’ is the stem of the adjective ‘intentional’. This fact, it is claimed, suggests a connection between intentional action and intention. Thus, SPV. The proponent of DPV can respond to this argument by denying that one is entitled to draw conclusions about the deep nature of intentional action based on mere linguistic facts. After all, we do not think the term ‘coffee table’ captures a necessary connection between coffee and the tables for which we apply the term. Still, proponents of DPV should say something about why intentional action is action so-called. Robert Audi, a proponent of DPV, claims (not unreasonably) that intentional action is so-called because paradigmatic intentional actions involve intention. 21 The X-Phi Argument is an argument about ordinary people’s intuitions. It essentially claims that if we were to poll ordinary people, we would very likely find that a high percentage have intuitions that align with SPV. Therefore, SPV is preferable to DPV. I think there are at least a couple things proponents of DPV can say here. First, proponents of DPV might point out that if we attach weight to this sort of argument, we should probably anticipate it counting in favor of SV. It’s not a stretch to suppose that 21 Audi (1986), section 3. 12 the majority of ordinary people would find SV intuitively appealing. But most philosophers would not take this to constitute a forceful consideration favoring SV. Thus, we shouldn’t take The X-Phi Argument to constitute a forceful consideration favoring SPV (over DPV). Second, proponents of DPV might point out that truth is not determined by consensus. In order for The X-Phi Argument to have teeth, one has to relate the (assumed) empirical result in some interesting way to the nature of intentional action. This hasn’t been done. The third and fourth arguments for SPV—The Settling Arguments—carry more force. Both arguments turn on the settling feature of intention. Most philosophers of action would accept the following pair of claims about settling: (1) Settling on a course of action initiates and sustains action (at the appropriate time), and (2) Settling on a course of action is subject to certain norms of rationality. I argue that SPV, but not DPV, preserves claims (1) and (2). The upshot of chapters 1 and 2 is that the belief/desire theory and the planning theory of intention are implausible because each is incompatible with SPV. Indeed, The Settling Arguments of chapter 2 coupled with the argument of chapter 1 entails the falsity of (all versions of) the belief/desire theory of intention. There is, however, no principled reason why the planning theory could not be altered so as to conform to SPV. I briefly explore this issue in the Conclusion. 13 Chapter 3 In chapter 3 I explore the possibility of negative intentional action. Intuitively, refraining from doing something constitutes intentional omission. But, if behavior is intentional, then it must be actional. This presents a puzzle: if action essentially involves doing something, how can omissions (i.e., refrainings) be proper actions? Aren’t they rather inactions? In this chapter I answer this question by developing an actional account of refraining. My account constitutes an improvement over my earlier account because of its sensitivity to non-intentional action. In 2009, I forwarded the following account of refraining: (REF) S refrains from A-ing iff (1) S considers A-ing, (2) S has the ability to A, (3) S Bs intentionally, (4) S lacks the belief that B-ing is compatible with A-ing, and (5) B-ing is incompatible with A-ing. 22 Under REF, refraining from acting always involves doing something (see condition (3)). It is this positive component of refraining that entitles us to the claim that refrainings are proper actions. Thus, REF preserves the sense we have that refrainings are intentional omissions. The problem with REF is that it is insensitive to non-intentional action. Al Mele and Steven Sverdlik have argued for the existence of action that is neither intentional nor 22 Johns (2009). My statement of REF here isn’t quite its statement in (2009). Here I’ve added condition (4) and have also altered some of the language of conditions (2) and (3). 14 unintentional; they call such action ‘non-intentional’. 23 What counts as non-intentional action? Mele and Sverdlik argue that side-effect acts are non-intentional. An example here will help. Consider Gilbert Harman’s famous ‘sniper’ scenario. 24 A sniper wants to shoot a high-ranking enemy soldier. However, he knows that if fires his rifle, he will alert the enemy to his presence. Still, the sniper reasons that shooting the soldier is worth the risk involved in alerting the enemy. So, the sniper shoots the soldier; in so doing, he alerts the enemy to his presence. The sniper’s shooting the soldier is straightforwardly intentional. What about his alerting the enemy? Mele and Sverdlik argue that it is not intentional because it’s not the sniper’s aim or goal to alert the enemy. They also argue that it’s not unintentional on grounds that the sniper knowingly and non-accidentally alerts the enemy. They conclude that the sniper non-intentionally alerts the enemy. I agree with Mele and Sverdlik that side-effect acts like the sniper’s are non- intentional. 25 What I do in this chapter is extend their insight to the category of omissions. That is, just as intentional action can be positive or negative, I suggest that non-intentional action can be positive or negative. The examples of negative non- intentionality I present here are ones in which omissions occur as the result of an intentional action. Since these omissions are not the agent’s aim or goal, they are not intentional failures (i.e., refrainings). Since they are expected and taken into consideration, however, they are not unintentional. Therefore, such omissions are non- intentional. 23 Mele and Svedlik (1996). 24 Harman (1976), p. 433. 25 Though as we’ll see in ch. 4, I take issue with one of Mele and Sverdlik’s premises. 15 The problem with REF is that it is insensitive to non-intentional omissions. REF wrongly classifies non-intentional omissions as intentional. To fix this, I revise REF by adding a sixth necessary condition on refraining: (6) S’s not A-ing is S’s aim or goal. Chapter 4 REF relies on the existence of non-intentional action. Non-intentionality, however, is more controversial than I’ve let on above. In this chapter I supply two arguments in favor of non-intentionality. I begin by discussing four arguments for non-intentionality that have been forwarded in the literature. I argue that none of these four arguments are successful. I then present two arguments for non-intentionality that I believe are more promising. The first is inspired by Mele and Sverdlik’s argument from side-effect action discussed above. The second is an argument from lucky action. The common argumentative strategy deployed by proponents of non-intentional action involves two steps: first, provide conditions for intentional and unintentional action; then identify a particular action (or action type) that fails to satisfy both sets of conditions. I object to four prior arguments for non-intentionality on grounds that they rely on implausible accounts of intentional action. Two of these four arguments take the implausible SV thesis as a premise. Recall that SV is the claim that intentionally performing A requires intending A. Thus, both of these arguments are unconvincing. The third argument relies on an account of intentional action that is inconsistent with the apparent truth that one can intentionally perform an act by executing an intention to try to perform the act. Finally, the fourth argument places a knowledge condition on 16 intentional action: intentionally A-ing requires knowing that one is A-ing (at the relevant time). But as Davidson has persuasively argued, a knowledge condition on intentional action is too strong. 26 I proceed to offer two arguments in favor of non-intentional action that I believe are more promising. The first argument is a variation of Mele and Sverdlik’s argument. Recall that Mele and Sverdlik argue that side-effect acts are non-intentional. They claim that the sniper does not alert the enemy intentionally because it is not his aim or goal to alert the enemy. I disagree, however, with their claim that the sniper’s alerting the enemy is not unintentional simply because it is expected or anticipated. I supply an example in which an agent knowingly performs an action unintentionally. In order to show why the sniper’s alerting the enemy is not unintentional, I develop an account of unintentional action. It is then seen that the sniper’s alerting the enemy violates my conditions on unintentionality. I conclude that the sniper’s act is non-intentional. My second argument involves lucky action. I present an example in which an agent tries to perform an action and succeeds by sheer luck. I argue that because the agent’s success is not based on skill but rather coincidence, it does not count as intentional. Yet because it violates my conditions on unintentionality, it is not unintentional either. Therefore, the agent’s lucky act is non-intentional. 26 Davidson (1978), p. 92. 17 INTRODUCTION REFERENCES Adams, Frederick (1986), “Intention and Intentional Action: The Simple View,” Mind and Language 1, pp. 281-301. Audi, Robert (1986), “Intending, Intentional Action, and Desire,” in The Ways of Desire, ed. Joel Marks (Chicago: Precedent), pp. 17-38. Audi, Robert (1973), “Intending,” Journal of Philosophy 70, pp. 387-402. Bratman, Michael (1987), Intentions, Plans, and Practical Reason (Cambridge, MA: Harvard University Press). All references are to the CSLI Publications (1999) re- issue. Crane, Tim (1995), “The Mental Causation Debate,” Proceedings of the Aristotelian Society, Supplementary Volume 69, pp. 211-36. Davidson, Donald (1993), “Thinking Causes,” in Mental Causation, eds. John Heil and Alfred Mele (Oxford: Oxford University Press), pp. 1-17. Davidson, Donald (1978), “Intending,” in Philosophy of History and Action, ed. Yirmiahu Yovel (Dordrecht, NL: D. Reidel), pp. 41-60. Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 83- 102. References are to Davidson (2001). Davidson, Donald (1971), “Agency,” in Agent, Action, and Reason, eds. Robert Binkley, Richard Bronaugh, and Ausonio Marras (Toronto: University of Toronto Press). Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 43-61. Davis, Wayne (1984), “A Causal Theory of Intending,” American Philosophical Quarterly 21, pp. 43-54. Descartes, Rene (1641), Meditations on First Philosophy, trans. and ed. John Cottingham (Cambridge: Cambridge University Press, 1996). Goldman, Alvin (1970), A Theory of Human Action (Englewood Cliffs, NJ: Prentice- Hall). Grice, H. P. (1971), “Intention and Uncertainty,” Proceedings of the British Academy 57, pp. 263-79. 18 Harman, Gilbert (1976), “Practical Reasoning,” Review of Metaphysics 79, pp. 431-63. Reprinted in Mele (1997), ch. 7. Hornsby, Jennifer (1993), “Agency and Causal Explanation,” in Mental Causation, eds. John Heil and Alfred Mele (Oxford: Oxford University Press), ch. 10. Reprinted in Mele (1997), ch. 12. Jackson, Frank (1996), “Mental Causation,” Mind 105, pp. 377-413. Johns, Brandon (2009), “Refraining and the External,” Ratio 22 (2), pp. 206-15. McCann, Hugh (1991), “Settled Objectives and Rational Constraints,” American Philosophical Quarterly 28, pp. 25-36. Reprinted in (Mele (1997), ch. 9. Mele, Alfred (1997), The Philosophy of Action (Oxford: Oxford University Press). Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). Mele, Alfred and Steven Sverdlik (1996), “Intention, Intentional Action, and Moral Responsibility,” Philosophical Studies 82 (3), pp. 265-87. Setiya, Kieran (2007a), “Cognitivism about Instrumental Reason,” Ethics 117, pp. 649- 73. Setiya, Kieran (2007b), Reasons without Rationalism (Princeton, NJ: Princeton University Press). Velleman, David (1989), Practical Reflection (Princeton, NJ: Princeton University Press). Wasserman, Ryan (forthcoming), “Intentional Action and the Unintentional Fallacy,” Pacific Philosophical Quarterly. Yaffe, Gideon (2010), Attempts (Oxford: Oxford University Press). Yaffe, Gideon (2006), “Trying, Intending and Attempted Crimes,” Philosophical Topics 32, pp. 505-32. 19 CHAPTER 1 TRYING, INTENDING, AND THE DOUBLE PHENOMENON VIEW I might intend now to go to the Lakers game tonight. And should the evening arrive with my intention intact, I will, all things being equal, act upon it: I will go to the game intentionally. This simple example illustrates a connection between intention and intentional action that most of us take for granted. In the philosophy of action, this connection is known as the ‘Single Phenomenon View’: Single Phenomenon View (SPV): Intentional action necessarily involves the state of intention. 27 SPV is an intuitively attractive principle; indeed, it is presumably no accident that the adjective ‘intentional’ embeds the noun ‘intention’. One goal of the theory of intention, then, is to support SPV. In this chapter, I will argue that two important theories of intention fail to meet this goal. Theories of intention can broadly be classified as reductionist or non-reductionist. Reductionist theories, as the label indicates, analyze the state of intention in terms of other, more basic mental states. Arguably the most influential reductive strategy has been the belief/desire model of intention. On the belief/desire model, intentions consist 27 The term ‘Single Phenomenon View’ was coined by Bratman (1987), ch. 8. 20 of belief/desire pairs. 28 One well-known source of motivation for the belief/desire strategy lies in the thought that belief and desire are the two most fundamental states of mind—and that (perhaps) all other states of mind can be explained by them. While belief/desire theorists may disagree about what constitutes necessary and sufficient conditions for intending, most agree that an intention to act necessarily involves both a belief that one will act and a desire to act. Not all philosophers of mind subscribe to reductive approaches to intention. Michael Bratman has developed the most influential non-reductionist theory of intention. 29 On Bratman’s view, intention, like desire, is a practical attitude; however, intention differs from other practical attitudes in virtue of its role in planning for the future. And so the norms of rationality that govern intention promote the aims of planning agency. Whichever kind of theory of intention one might favor—reductionist or non- reductionist—a plausible theory will be one that is compatible with SPV. I think belief/desire theorists and planning theorists are, in fact, committed to denying SPV and therefore endorsing what I call the ‘Double Phenomenon View’: Double Phenomenon View (DPV): Intentional action does not require having an intention. 30,31 28 For two important belief/desire views, see Audi (1973) and Davis (1984). Another important reductive approach involves reducing intention to a kind of belief; see here Setiya (2007b) and Velleman (1989). 29 Bratman (1987). 30 Audi (1986) has argued for this principle. My argument in section 1.1 below resembles Audi’s argument in certain respects. But it is also importantly different. I will not discuss these differences in this paper. For a criticism of Audi’s argument, see Mele (1992), pp. 186-87. 21 DPV is a principle that severs a necessary connection between intentional action and intention. Thus, I will be arguing that the belief/desire model and the planning theory are flawed because they fail to preserve a common-sense feature of intentional action. This is the second of two main points I’ll attempt to make in this chapter. The first point I will make concerns the relation between these two theories of intention and trying to act. I will argue that belief/desire theorists and planning theorists are committed to the following thesis: trying to act does not require having an intention. To be clear, the claim here is not merely that trying to do A does not entail intending to do A. Rather, the thesis is the strong claim that one can try to do something without being motivated by any intention whatsoever. 32 In what follows, I will first argue that the belief/desire view of intention implies that trying does not require having an intention (section 1). I will next argue that Bratman’s planning theory of intention implies that trying does not require having an intention (section 2). I will then use the results of sections 1 and 2 to argue that both theories of intention are committed to rejecting SPV and thus endorsing DPV (section 3). 31 As I noted in the Introduction (note 20), I’ll be arguing that two important versions of these theories are committed to DPV. In what follows, when I speak of the belief/desire theory and the planning theory of intention, it should be assumed that I’m referring to the two relevant versions of these theories unless otherwise indicated. The two relevant versions were briefly mentioned in the Introduction and will be made clearer below. 32 In what follows, I will sometimes use the expression ‘trying without intending’ to refer to this thesis. 22 1 The Belief/Desire Model of Intention According to a long tradition, intending is thought to involve cognitive and motivational elements. 33 Many philosophers in this tradition take the state of belief to constitute the cognitive aspect of intention, and the state of desire to constitute the motivational aspect. There are different ways, though, in which belief and desire might constitute intention. The most influential belief/desire accounts of intention hold that intending to do something necessarily entails (i) the belief that one will act and (ii) the desire so to act. 34 Whether or not (i) and (ii) are sufficient for intending is a matter of debate. Wayne Davis, for instance, has argued forcefully for the necessity of a third, causal condition that grounds one’s (relevant) belief in one’s (relevant) desire. 35 For our purposes, let us understand the belief/desire view as the view that (i) and (ii) are necessary conditions on intending. We can leave open the question of sufficiency since my argument below will apply equally to all belief/desire views that take (i) and (ii) to be necessary constraints on intending. As indicated, (i) and (ii) are metaphysical—as opposed to normative—conditions on intending. It’s not that were an agent to lack (i) or (ii) then her intention to act would be irrational. Rather, one cannot have an intention without satisfying (i) and (ii): intention just is belief and desire (plus?). Let us call condition (i) the ‘Strong Belief Requirement’ on intending and formulate it as follows: 33 See Audi (1973); Austin (1873); Beardsley (1978); Davidson (1978); Davis (1984); Grice (1971); Harman (1976); Kenny (1966); and Salmond (1930). 34 See, e.g., Audi (1973) and Davis (1984). 35 Davis (1984). 23 Strong Belief Requirement (SBR): Intending to A requires believing that one will A. 36 And let us call condition (ii) the ‘Strict Desire Requirement’ on intending and formulate it as follows: Strict Desire Requirement (SDR): Intending to A requires (intrinsically or extrinsically) desiring to A. 37 In section 1.1, I will argue that SBR and SDR commit proponents of the belief/desire view to the thesis that trying to act does not require having an intention. 1.1 Trying without Intending We’ll get to the argument shortly. First, let us briefly consider a “direct” argument for the claim that trying does not require intention. Suppose that I am driving down the road when suddenly a dog shoots out in front of my car. My perception of the dog causes me to apply pressure to the brake pedal. In my haste, my foot misses the brake pedal. In this situation we might think that, first, there is an attempted action; and second, that the action attempted lacks an intention. By ordinary standards, it would seem that I try to step on the brakes. And under the circumstances, it may seem as though I don’t have the time to form a (belief/desire) intention of the sort that would motivate my attempt to step 36 Proponents of this requirement include Audi (1973; 1986); Austin (1873); Beardsley (1978); Chisholm (1970); Davis (1984); Grice (1971); Harman (1976); Kim (1976); Pears (1964; 1980; 1984; 1985); Rankin (1972); Salmond (1930); Sellars (1966); Setiya (2007a; 2007b); and Velleman (1989). 37 Proponents of this requirement include Audi (1973; 1986); Austin (1873); Beardsley (1978); Churchland (1970); Davis (1984; 1986); Goldman (1970); and Nagel (1970). 24 on the brakes. And so on the belief/desire view trying seems not to require having an intention. I think this argument is unconvincing for the following reason. It seems plausible to suppose that there exist two ways in which one might acquire an intention: actively and passively. 38 Actively acquiring an intention—or forming an intention—involves deliberating about a course of action. When I attempt to slam on my brakes in order to avoid the dog, there is presumably not enough time for deliberation; and so I cannot be said to have formed an intention. However, not all intentions are formed: we passively— i.e., non-deliberatively—acquire intentions as well. My perception of the dog (together with various dispositions of mine) can produce an intention, say, to avoid the dog or to slam on the brakes—or at least to try to act in these ways. 39 The distinction between actively and passively acquiring an intention allows for sudden actions (such as trying to brake) to involve suddenly acquired intentions. And so the “direct” argument is, at best, inconclusive. 40 The better argument for trying without intending, I believe, is less direct. Consider the following scenario: 38 This distinction has been drawn by Mele (1992), p. 184. 39 As Mele (1992) points out, other mental states, such as belief and desire, also appear subject to a similar distinction between active and passive acquisition. Sometimes we form beliefs; sometimes, beliefs “happen to us.” For instance, when the dog shoots out across the road, I immediately acquire the belief that there is dog in the road. There is no time, it seems, to form this belief. But few would deny that I have such a belief. The active-passive distinction can explain this. 40 Of course, it’s not my claim that proponents of the belief/desire model must embrace the distinction between active and passive intention. For those that will deny it, I think the example of spontaneous action is plausibly a case of trying without intending. 25 (FOOTBALL) Sam, a star quarterback, has just injured his arm during a football game. Sam is in considerable pain, but it is not yet clear to what extent Sam’s arm is injured. The team’s athletic trainer instructs Sam to try to raise his arm. If Sam can raise his arm, then he is not seriously injured and can immediately return to the game. However, if Sam cannot raise his arm it’s an indication that his injury is more severe; and so Sam would not, in this case, be able to play. Sam is aware of this, and due to his competitive nature, wants to raise his arm and get back in the game. Because of the sharp pain he experiences, Sam does not believe that he will succeed in raising his arm; and so he does not believe that he will get back in the game. However, Sam believes there is a chance he could raise his arm and so believes that he will try. Sam, therefore, struggles in pain to raise his arm. By ordinary standards, Sam’s struggle to raise his arm is his trying to raise his arm. For all we have said, though, Sam’s attempt to raise his arm involves an (belief/desire) intention. What we’re going for, you’ll recall, is the claim that on the belief/desire model, one can try to act without having a motivating intention; to get there, then, I need to show now that Sam’s attempt need not involve an intention. My plan is to utilize SBR and SDR in a way that precludes various intentions. The idea is to dismiss a series of intentions until it becomes reasonable to conclude that Sam’s attempt to raise his arm does not spring, even in part, from intention. One question we can ask ourselves is: what doesn’t Sam believe in our scenario above? Well, to begin, we know that: 26 (1) Sam lacks the belief that he will raise his arm. You’ll recall that Sam is in considerable pain, and so does not believe that he will raise his arm. This fact by itself is not terribly interesting. What’s interesting is the consequence of pairing (1) with SBR. According to SBR, intending requires believing you will; so: no belief, no intention. Sam lacks the belief that he will raise his arm. Given SBR, then, Sam must not intend to raise his arm; for if he did, he would believe that he would raise his arm and (1) would be false. Thus, when Sam tries to raise his arm, he does so without having the intention to raise his arm. Here is something else that is true of Sam: (2) Sam lacks the belief that he will get back in the game. Again, Sam is in pain, and so neither believes that he will raise his arm nor that he will return to the game. Given SBR, it follows that Sam must not intend to get back in the game. Thus, Sam tries to raise his arm without having the intention to get back in the game. Given that Sam lacks the belief that he will raise his arm ((1)) and the belief that he will get back in the game ((2)), it follows that: 27 (3) Sam lacks the belief that he will get back in the game by raising his arm. In virtue of SBR, Sam does not intend to get back in the game by raising his arm. Thus, Sam tries to raise his arm without having the intention to get back in the game by raising his arm. What else could be true of Sam? Well, given that Sam does not believe that he will get back in the game ((2)), it’s possible that: (4) Sam lacks the belief that he will get back in the game by trying to raise his arm. (4) and SBR jointly entail that Sam does not intend to get in the game by trying to raise his arm. Hence, Sam tries to raise his arm without having the intention to get in the game by trying to raise his arm. Let us suppose that at the time of Sam’s injury, the outcome of the football game is undetermined: both teams have a fighting chance to win. Since Sam lacks the belief that he will return to the game ((2)), it follows that: (5) Sam lacks the belief that he will help his team to victory. (5) and SBR jointly entail that Sam does not intend to help his team to victory. And so Sam tries to raise his arm without intending to (help his team) win the game. 28 You might think that trying to do something involves intending the means. If this is true, then Sam must intend the means to raising his arm given that he tries to raise his arm. Let us attribute to Sam the following: (6) Sam lacks a belief about the means to raising his arm. (6) is possible—and hence permissible—given the nature of arm-raising (or arm- movement more generally). Raising one’s arm is a basic action in the following sense: in doing it, one does not (ordinarily) have instrumental beliefs about how it is done; arm- raising is, in other words, something one usually does without having beliefs about the means to doing it. And so we can suppose that Sam does not have a belief about the means to raising his arm. In virtue of SBR, it follows that Sam does not intend the means to raising his arm. Thus, Sam tries to raise his arm without intending the means to raising his arm. 41 You might think that Sam desires to try to raise his arm given that he desires to raise his arm. If this were true, then Sam would be in a position to intend to try to raise his arm given that he believes that he will try to raise his arm. Trying to raise one’s arm is a means to raising it. So the relevant question here is: under what condition is it true that one desires the means given that one desires the end? I think that one desires the 41 Some have suggested that actions such as arm-raisings cannot be attempted because they are basic. (See, e.g., Danto (1973) and Taylor (1966).) However, if our example involving Sam is correct, then agents can in fact try to perform basic actions. (Mele (1992) holds that one can try to raise an arm and has claimed elsewhere that arm-raisings are basic (see Mele and Moser (1994)). O’Shaughnessy (1973) holds that one can try to raise an arm, though I’m not sure whether he thinks arm-raisings are basic.) As Hornsby points out, even though basic actions are performed “directly,” or without instrumental beliefs, it is not that one cannot try to do such things; it is rather that when one attempts something basic, one’s attempt is likewise directly performed. (See Hornsby (1995), p. 532.) 29 means only if one believes that the means is a means to the end. So, does Sam desire to try to raise his arm? He certainly does not desire this intrinsically—for Sam has no interest in trying to raise his arm for its own sake. But Sam also does not extrinsically desire to try to raise his arm. For if he did, he would believe that trying to raise his arm is a means to raising his arm. But we know from (6) above that Sam does not have a belief about the means to raising his arm. And so it follows that: (7) Sam lacks the (intrinsic and extrinsic) desire to try to raise his arm. There is an important consequence of pairing (7) with SDR. According to SDR, intending to do something requires desiring to do it: so, no desire, no intention. Sam lacks the desire to try to raise his arm. Given SDR, then, it follows that Sam does not intend to try to raise his arm. Still, Sam’s desire to raise his arm is compelling, and so Sam tries to raise his arm without intending to try to raise his arm. 42,43 42 Given that Sam does not intend to try to raise his arm, I take it that Sam does not intend to do what he can to raise his arm, or intend to give arm-raising a shot. This is because the latter intentions are identical to the intention to try. Furthermore, if Sam intended to try to get back in the game, then he would intend to try to raise his arm. This is because Sam knows that the only way to get back in the game is by raising his arm. Since Sam does not intend to try to raise his arm, then, it follows that he must not intend to try to get back in the game. 43 Putting aside my claim that trying can be a means to doing, one might have an intuition that wanting to A simply entails wanting to try to A. Such an intuition, however, should be resisted. Suppose that I desire to bench press one-thousand pounds in order to impress a girl. I know that I cannot press even one-quarter this much weight. And if my friend were to ask me, “Do you want to try to press one-thousand pounds?” I think a natural response would be, “Of course not!” Wanting to do something does not appear to entail wanting to try. 30 You might wonder if Sam possesses a conditional intention when he tries: perhaps he intends to raise his arm if he can. If Sam intends this, then by SBR he must believe that he will raise his arm if he can. However, we can attribute to Sam the following: (8) Sam lacks the belief that he will raise his arm if he can. Under what circumstances might Sam not believe that he will raise his arm if he can? Recall that Sam is in considerable pain, and for all he knows the pain could intensify when he tries to raise his arm. Now suppose that prior to his attempt Sam believes there is a certain degree of pain, n, such that if he were to experience n while attempting to raise his arm, he would abort his attempt. The reason Sam would abort an n-attempt is due to his concern for his future as an athlete. That is, an n-attempt, Sam believes, is such that it would cause him to fear for his athletic career because he would fear returning to the game and possibly injuring himself further and more seriously. Moreover, suppose Sam believes that he could successfully execute an n-attempt if he wanted to. Since Sam believes that his attempt to raise his arm could be an n-attempt, Sam does not form the belief that he will raise his arm if he can. Therefore, by SBR, Sam does not intend to raise his arm if he can. 44 Perhaps, then, Sam intends to raise his arm if doing so is not too painful. We can dismiss this intention in virtue of the following possibility: 44 Given that Sam does not intend to raise his arm if he can, it follows that he lacks the intention to raise his arm if it is possible. This is because under the relevant sense of ‘possibility’—which approaches ‘ability’ or ‘capability’—the intention to raise his arm if he can is identical to the intention to raise his arm if it is possible. 31 (9) Sam lacks the belief that he will raise his arm if doing so is not too painful. What makes (9) possible? Sam is agnostic as to the extent of his injury. He knows that his arm is in pain, but is unsure whether or not he has suffered structural damage. That is, for all Sam knows, he may not be able to raise his arm even were he to take, say, an (instantly) effective pain killer. Given the epistemic possibility of structural damage to his arm, we can suppose, therefore, that Sam does not believe that he will raise his arm if doing so is not too painful; for Sam’s arm might (structurally) be such that it cannot be raised. (9) and SBR jointly entail that Sam does not intend to raise his arm if doing so is not too painful. So perhaps Sam believes that he will raise his arm if doing so is not too painful and his arm is not structurally damaged (in the relevant way); and if he believes this conditional, then Sam is in a position to intend it. We can certainly imagine a scenario in which Sam believes, and even intends, this conditional. However, it is possible that: (10) Sam lacks the belief that he will raise his arm if doing so is not too painful and his arm is not structurally damaged. (10) is permissible in case it’s possible; and it’s possible in case its negation is not necessary. Now then, the negation of (10) does not appear necessary; i.e., under the circumstances, Sam need not have the belief that he will raise his arm if doing so is not too painful and his arm is not structurally damaged. But even if there are doubts about this, we can always simply add further details to the example that would render (10) 32 possible. In virtue of SBR, then, Sam can lack the intention to raise his arm if doing so is not too painful and there is no structural damage. To summarize: Sam desires to raise his arm because he desires to return to the game; he believes he can return to the game by raising his arm and so believes he’ll try to raise his arm. Sam tries to raise his arm. In virtue of (1)-(10), SBR and SDR, Sam’s attempt to raise his arm does not involve intention. What’s more, it seems unlikely that there exists a belief/desire pair, p, such that (i) Sam necessarily possesses p, and (ii) p constitutes an intention, i, such that i would motivate Sam’s attempt to raise his arm. We conclude that on the belief/desire model of intention, trying to act does not require intending anything whatsoever. 2 The Planning Theory of Intention Michael Bratman has developed the most sophisticated and influential non-reductive theory of intention. 45 According to Bratman, intention, like desire, is a practical attitude—it plays a role in deliberation and the motivation and guidance of action. But what distinguishes intention from other practical attitudes is its role in planning for the future. Planning is important because it helps us to achieve our goals. If I intend to go to the Lakers game tonight, then I can plan on stopping by my office prior to the game to pick up the book I need; and I can expect to read several chapters of the book tomorrow; and so on. Planning helps us to achieve our ends by coordinating our activities with our future selves. Planning also helps us coordinate our activities with others in order to 45 Bratman (1987). 33 achieve joint goals. For instance, in order to save money on parking, I can plan on carpooling with Dan (and Dan can expect to carpool, etc.). Intentions for Bratman, then, are distinct practical attitudes that are (typically) elements in coordinating plans. Bratman identifies three norms of rationality that govern intentions: Strong Consistency, Means-End Coherence, and Agglomerativity. 46 These norms serve to promote the forms of inter- and intra-personal coordination supported by intentions. The requirement of Strong Consistency can be stated as follows: Strong Consistency Requirement: Intending to A rationally requires not believing that not-A. The Strong Consistency Requirement, in other words, is the demand that one’s intentions and beliefs be consistent. The requirement of Means-End Coherence is this: Means-End Coherence Requirement: If one intends to A and believes that B is a necessary means to A, then rationality requires that one intend to B. Regarding the requirement of Agglomerativity, Bratman says, “there is rational pressure for an agent to put his various intentions together into a larger intention.” He continues, “If I both intend to [A] and intend to [B], there will be rational pressure for me to intend to [A and B].” 47 Let us then formulate the Agglomerativity norm as follows: 46 For a discussion of the first two norms, see Bratman (1987), ch. 3. For a discussion of Agglomerativity, see Bratman (1987), ch. 9. For a helpful discussion of all three norms, see Yaffe (2006). 47 Bratman (1987), p. 134. 34 Agglomerativity Requirement: If one intends to A and intends to B, then rationality requires that one intend to A and B. 48 There is, of course, a lot more to be said about each of these norms. But let us push on. In section 2.1, I will argue that Bratman’s planning theory of intention commits him to the thesis that trying to act does not require having an intention. 2.1 Trying without Intending Before we get to the argument that I think commits Bratman to the thesis about trying, let us briefly revisit the issue of spontaneous action. Recall that in section 1.1, we discussed the possibility of spontaneous trying without intending. There, I suggested that proponents of the belief/desire model could embrace the distinction between active and passive intention acquisition and thereby deny spontaneous trying without intending. Bratman addresses the same issue as it involves his planning theory. 49 He offers the following example: Suppose you unexpectedly throw a ball to me and I spontaneously reach up and catch it. On the one hand, it may seem that I catch it intentionally…On the other hand, it may seem that, given how automatic and unreflective my action is, I may well not have any present-directed intention that I am executing in catching the ball. 50 48 See Yaffe (2006; 2010) for an alternative formulation of this requirement. I discuss Yaffe’s formulation in the Conclusion of my dissertation. 49 Bratman (1987), ch. 8. 50 Bratman (1987), p. 126. 35 If Bratman can catch the ball intentionally, then presumably he can try to catch it. And if in such cases one need not have an intention, then trying does not require intending. Bratman responds by suggesting that spontaneous attempts involve the execution of “long-standing, personal policies,” rather than present-directed intentions. Regarding the (spontaneous) attempt to catch the ball, Bratman says “I do not have a present- directed intention specifically to catch this very ball, but my action still involves an intention, namely: my general intention to protect myself in such circumstances.” 51 I’m not completely convinced by this response. While an appeal to long-standing, general intentions may be plausible in some cases of spontaneous trying (such as the ‘catch’ and ‘dog’ examples), it is questionable in others. Suppose that while walking in Tokyo, I unexpectedly see a childhood friend on the other side of the street. I spontaneously begin waving my arms and calling her name in an attempt to grab her attention. There does not seem to be any obvious long-standing, general intention that we might plausibly appeal to in order to explain my attempted action. (Is it a long-standing intention to connect with old friends? Is it a long-standing intention to get the attention of people I know or have known? What could it be?) Rather than deny that spontaneous trying involves present-directed intention, I think Bratman, like the belief/desire theorists, ought to accept Mele’s distinction between actively and passively acquiring an intention (see section 1.1). Bratman could then claim that in cases of spontaneous trying, agents spontaneously (i.e., non-deliberatively, 51 Bratman (1987), p. 126. 36 passively) acquire an intention. Sometimes our actions are spontaneous—why not our intentions? In order to show that Bratman is committed to trying without intending, we need an example of non-spontaneous action. Here we can use Bratman’s own ‘video game’ example. In “Two Faces of Intention,” Bratman presents a famous example designed to refute a particular version of SPV—what he calls the ‘Simple View’ (SV): intentionally A-ing requires intending to do A. 52 Bratman imagines a video game in which players are rewarded for guiding missiles into targets. The game allows the player to choose between two options of game play: on option 1, the player uses a single controller to shoot at a single target; on option 2, the player uses two controllers—one for each hand— to shoot at the two targets simultaneously. The rules of the game are as follows: regardless of the option selected, the player can only hit a single target. If one plays option 2 and is on the verge of hitting both targets, the game senses this and shuts down immediately. In this case, no targets are hit and the player loses. Neither option is inherently easier or more difficult than the other. A prize is awarded for hitting a target. Now suppose that a player, Sarah, elects to play the game. Sarah is aware of the rules of the game; in particular, she knows that she cannot hit both targets in option 2. She also knows from personal experience that it is quite difficult to hit a target; however, she has successfully hit the target in the past. Sarah wants to win the prize. Since she is ambidextrous, she chooses to play option 2. She reasons that the opportunity to shoot at 52 Bratman (1987), ch. 8. Proponents of SV include Adams (1986) and McCann (1991). Wasserman (forthcoming) recently defends SV from a string of objections, but it is not clear whether he endorses SV. 37 two targets outweighs the risk of shutting the game down. Sarah thus tries to hit target 1 and tries to hit target 2. Bratman points out that if Sarah hits target 1 in the way she tries (and non-luckily, I would add), then she hits it intentionally. According to SV, then, the following must be true: (11a) Sarah intends to hit target 1. But, given the symmetry of the example, if Sarah intends to hit target 1, then it must also be the case that: (11b) Sarah intends to hit target 2. If Sarah has both intentions, then by the Agglomerativity Requirement she is under rational pressure to intend to hit both targets. But if she intends to hit both targets, then she is criticizably irrational because her intention is inconsistent with her belief that she cannot hit both (thus violating the Strong Consistency Requirement). However, Sarah is not criticizably irrational for trying to hit each target: her strategy makes good sense. Therefore, Sarah must neither intend to hit target 1 nor intend to hit target 2—and SV fails. On Bratman’s view, when one tries (or endeavors) to A but does not intend A—as in Sarah’s case—one’s attempt is guided by a desire to A. 53 Sarah has a guiding desire to 53 Bratman (1987), p. 137. 38 hit target 1 and a guiding desire to hit target 2. However, having these guiding desires is not incompatible with Sarah’s having an intention that contributes to her trying behavior. What we’re going for, you’ll recall, is the claim that Bratman is committed to the thesis that trying does not require an intention. So, like our strategy in section 1.1, let us proceed to consider various possible intentions Sarah might possess and that might serve to motivate her pair of attempts. The plan will be to dismiss these intentions (mostly) by utilizing Bratman’s norms of rationality introduced above. We know that Sarah neither intends to hit target 1 nor intends to hit target 2. Let us now consider and dismiss the following pair of intentions: (12a) Sarah intends to try to hit target 1. (12b) Sarah intends to try to hit target 2. According to Bratman, tryings (endeavorings), unlike intentions, are not subject to an agglomeration requirement. 54 If they were, Sarah would be under rational pressure to try to hit both targets given that she tries to each of them. However, Bratman points out that it would be irrational for Sarah to try to hit both targets in light of her knowledge that she cannot hit them both. Since Sarah’s behavior is not irrational, Bratman concludes that tryings are not agglomerative. Now, if trying to hit both targets is irrational and so not something that Sarah does, then (12a) and (12b) appear problematic. Given (12a) and (12b) and the Agglomerativity Requirement, Sarah must intend to try to hit both targets in order to 54 Bratman (1987), p. 134. 39 maintain rationality. But it is difficult to see how rationality could require her to intend to try to hit both targets if it is irrational to try to hit them both. Of course, we might add to the example bizarre details that might make it rational for Sarah to intend to try to hit both targets. But this is beside the point. The point is that the current example does not rationally necessitate that Sarah intend to try to hit both targets; thus, we can eliminate (12a) and (12b). Given that Sarah does not intend to try to hit target 1 or intend to try to hit target 2, we can eliminate the following pair of intentions: (13a) Sarah intends to shoot at target 1. (13b) Sarah intends to shoot at target 2. (13a) and (13b) entail (12a) and (12b). Suppose Derek, a basketball player, shoots 35% from three-point range. Ordinarily, when he shoots a three-pointer, it follows that he’s trying to sink the shot: he wants to sink it and utilizes his three-point shooting skill and knowledge to guide his conduct. He doesn’t merely throw the ball wildly toward the basket or shoot with indifference toward the outcome. Similarly, if Sarah shoots at target 1 and shoots at target 2, then she tries to hit target 1 and tries to hit target 2. Sarah has a guiding desire to hit each target because she wants to win the prize, and her game-playing skill and knowledge guides her conduct. And so an intention to shoot at target 1 entails an intention to try to hit target 1. Thus, if Sarah intends to shoot at target 1 ((13a)) and intends to shoot at target 2 ((13b)), then she intends to try to hit target 1 ((12a)) and 40 intends to try to hit target 2 ((12b)). Since she does not possess the latter pair of intentions, she does not possess the former pair. Perhaps Sarah’s intentions are conditional: (14a) Sarah intends to hit target 1 if she can. (14b) Sarah intends to hit target 2 if she can. We can dismiss (14a) and (14b) by imagining certain conditions that would lead Sarah (not irrationally) to believe that she would not hit target 1 or target 2 even if she could hit either one of them. Suppose the game is such that it sometimes informs players of their chances of hitting a target. For instance, after many hours of play, a player might be informed that she has, say, a 95% chance of success. Now, just prior to playing the game, Sarah forms the belief that she will not hit target 1 if her little sister walks in the room and the belief that she will not hit target 2 if her little sister walks in the room. Perhaps she forms these beliefs because she knows her little sister would be extremely envious of her were she to hit a target; and she’d rather not hit a target than upset her sister. Since she realizes that her sister may walk in at a point when she can hit either target 1 or target 2, Sarah forms the belief that, under certain circumstances, she will not hit target 1 or target 2 even if she can. If Sarah intends to hit target 1 if she can ((14a)) and intends to hit target 2 if she can ((14b)), then she is irrational given her belief that, under certain circumstances, she will not hit either target even if she can. In other words, 41 (14a) and (14b) would violate the Strong Consistency Requirement and would thus render Sarah irrational. But Sarah is not irrational; so we can eliminate (14a) and (14b). 55 Instead of a pair of intentions, perhaps Sarah has a single intention: (15) Sarah intends to hit one target or the other but not both. We can dismiss this intention by adding further details to the scenario. Suppose that successfully hitting a target takes approximately eight hours and Sarah is aware of this. It is twelve o’clock noon and Sarah is preparing to play the game. Before she begins playing she attempts to develop some further plans for the afternoon. Suppose her favorite television program airs at 5 p.m., her friends have invited her to coffee at 6 p.m., and her boyfriend is only available to have dinner with her at 7 p.m. Sarah would like to watch her favorite show, have coffee with her friends and have dinner with her boyfriend. So, what can Sarah plan on doing? Well, if (15) is true, then Sarah plans to hit a target. If she plans to hit a target, then she plans to play the game for eight hours—i.e., she’s committed to playing for eight hours. But if she plans to play the game for eight hours, then she should plan on not doing the three things she would like to do. Imagine how strange and inappropriate it would be if (at noon) Sarah told her friends: “I plan on playing the game for eight hours 55 At this point, one might wonder if perhaps (14a') Sarah intends to hit target 1 if she can and her sister does not walk in, and (14b') Sarah intends to hit target 2 if she can and her sister does not walk in. But to eliminate these two intentions, all we need to do now is add further details to the example. Perhaps in addition to her sister walking in, Sarah believes she will abandon her attempts at hitting the targets if her mother walks in (for whatever reason). The point is, no matter how one tries to bloat these conditional intentions, we can always add further details to the example such that the intentions run inconsistent with Sarah’s beliefs. And since Sarah is not irrational, she will not have these intentions. 42 and meeting for coffee at 6 p.m.” The problem, however, is that Sarah believes that she will probably not hit either target; thus, she believes that she will probably not play the game for eight hours. Rationality, it seems, should not require Sarah to plan on not doing all these things she wants to do when she rightly believes that she will probably not hit a target. In which case, rationality must not require that Sarah intend to hit one target or the other but not both ((15)). In claiming that rationality does not require that Sarah intend (15), there’s a question here of whether I am attributing to Bratman the following principle: If one intends to A, then rationality requires that one not believe that probably not-A. I think there are a couple things to say here. First, this principle very closely resembles Bratman’s own Strong Consistency Requirement; the only difference between the two is the word ‘probably’ in the former. I am inclined to view these two principles are interchangeable. If this is right, we can dismiss (15) by simply stating that it is inconsistent with Sarah’s belief—and therefore is an irrational intention. Secondly, suppose the principle noted above is in fact distinct from the Strong Consistency Requirement. In this case, I don’t think I’m committed to attributing it to Bratman in order to dismiss (15). All I would insist on here is that rationality does not necessitate that Sarah possess the intention contained in (15). Thus, (15)’s dismissal. 43 We can also dismiss the following intention: (16) Sarah intends to hit one of the targets. The intention to hit one target or the other ((15)) entails an intention to one of the targets ((16)). If rationality does not require Sarah to intend the former, then it must not require her to intend the latter. If Sarah is not under rational pressure to intend to hit a target ((16)), then it is not the case that she is under rational pressure to intend the following: (17) Sarah intends to win the prize. Recall the Means-End Coherence Requirement: if one intends to A and believes that B is a necessary means to A, then rationality requires that one intend to B. Sarah believes that hitting a target is a necessary means to winning the prize. By the Means-End Coherence Requirement, then, if Sarah intends to win the prize ((17)), she is under rational pressure to intend to hit one of the targets ((16)). But we know that Sarah is not under rational pressure to intend to hit one of the targets. So, she must not be under rational pressure to intend to win the prize ((17)). If rationality does not require that Sarah intend to win the prize, then we can eliminate (17). 44 Here is another possible intention: (18) Sarah intends to try to hit one of the targets. The problem here is that an intention to try to hit one of the targets is just the wrong intention for Sarah to have under the circumstances. Sarah is trying to hit each of the targets (though not both), not just one of them. If Sarah were to possess (18), then her trying behavior is relatively unintelligible. Since her behavior is not unintelligible, she must not have the intention to try to hit one of the targets ((18)). Sarah wants to hit a target because she wants to win the prize. She realizes that hitting a target is a difficult task, so she chooses to play option 2. She reasons that the opportunity to shoot at two targets outweighs the risk of shutting the game down. Sarah thus tries to hit target 1 (with the guiding desire to hit it) and tries to hit target 2 (with the guiding desire to hit it). By utilizing the norms of Strong Consistency, Means-End Coherence, and Agglomerativity, I’ve argued that Sarah’s attempts need not be motivated by an intention. It is true that I have not considered and dismissed every possible intention that Sarah might have, but I do not anticipate any further intentions derailing our argument. We conclude that Bratman’s theory of intention implies that trying to act does not require intending anything whatsoever. 45 3 The Double Phenomenon View I have argued above that both belief/desire theorists and planning theorists are committed to the view that trying does not require having an intention. In this section, I hope to show that this commitment entails the Double Phenomenon View (DPV). DPV, you’ll recall, is the view that intentional action does not require having an intention. I will argue that there are two routes that take us from the trying without intending thesis to DPV. The success of either route will be sufficient to yield DPV. First Route Let us return to the FOOTBALL scenario. Sam tries to raise his arm. And his trying to raise his arm just is his arm-raising behavior motivated by his desire to raise his arm and his belief that he can return to the game by raising his arm. Sam’s desire and belief constitute a motivating reason. Behaving for a reason entails acting intentionally. Therefore, Sam’s attempt to raise his arm constitutes an intentional action. 56 Moreover, if Sam’s attempt to raise his arm is an intentional action and if his attempt does not involve an intention, then it follows straightaway that intentional action does not require intention. The same goes for Bratman’s ‘video game’ example. Sarah attempts to hit each target for a reason. If her attempts constitute action, then they are intentional (yet 56 According to Cleveland (1997), all attempts are intentional actions. Perhaps Cleveland is right, but I wish to remain neutral about this for now; my argument only requires the weaker claim that an attempt can be an intentional action. 46 without intention). Thus, belief/desire theorists and planning theorists are committed to DPV. Second Route One way to block the first route is to deny that attempts are proper actions. Instead, one might hold that attempts are (merely) necessary for intentional action. 57 On this view, attempts are not intentional actions; and so the first route does not deliver DPV. I do not here wish to engage such a view. If one insists on this view of trying, then there is a second route to DPV. If one tries to do something and (non-luckily) succeeds in the way one was trying, the action one succeeds in performing is intentional (whatever intentionality amounts to). If Sam succeeds in raising his arm because of his attempt, Sam’s arm-raising is intentional: it’s motivated by a reason and is under Sam’s control. The same goes for Sarah. If she (non-luckily) hits target 1 in the way she tries, then her hitting it is intentional. Now, we have seen that Sam and Sarah’s attempted actions do not involve intention. If Sam succeeds in raising his arm, then, his arm-raising is both intentional and without intention. And if Sarah succeeds in hitting target 1, then her hitting it, too, is both intentional and without intention. Thus, the belief/desire theory and the planning theory of intention entail DPV: intentional action does not require intention. To summarize: either attempts can be intentional actions or successful attempts that are not the result of luck can be intentional actions (or both). Add to this disjunction 57 O’Shaughnessy (1973), e.g., holds this view. 47 the thesis that trying does not require intending anything, and it follows that intentional action does not entail intention (DPV). Thus, both belief/desire reductionists and planning theorists are committed to rejecting SPV. 4 Conclusion It is usually taken for granted that trying involves intention. 58 What I have argued is that two highly influential accounts of intention—the belief/desire model and Bratman’s planning theory—are committed to rejecting this assumption. The argumentative strategy, you’ll recall, was to show that for any given intention that might produce an attempt in a given context, a story can be supplied that permits (or necessitates) the elimination of such an intention. Now, critics might mount the following complaint: the arguments in sections 1 and 2 reach an end well before all possible intentions are considered and eliminated. And this fact enables critics to resist the argument(s). Critics might insist that for any point at which the argument actually halts, an intention can be located that (a) can explain the attempt in question and (b) has not yet been eliminated. And it’s this intention that is responsible for the relevant attempt. Thus, they’ll conclude that my argument does not demonstrate that trying does not require having an intention on the two theories in question. This is a fair concern. It’s difficult to know how best to respond to it. I suppose I might just insist that for any intention my opponent might locate here, I can add further details to the example(s) in order to eliminate it. But at this point we appear to reach 58 See, e.g., Ginet (2004). Yaffe (2006) is an exception; he argues that trying to A requires intending to A. 48 something of a tie or stalemate. My opponent will insist that there will always be further intentions that have not yet been excluded; and I, in turn, will insist that further details can always be added in order to eliminate them. I’m not sure who wins here. The issue here is about getting the last word: the one with the last word gets what they need. For now, all I can say is that my opponent has no better claim to the last word than I. In section 3 I suggested that the thesis about trying without intending gives rise to DPV: intentional action does not require having an intention. This is because either attempts themselves can constitute intentional action or successful attempts not involving luck constitute intentional action. Therefore, proponents of the belief/desire model and the planning theory of intention are committed to rejecting SPV. Since plausible theories of intention are only those that are compatible with SPV, we conclude that both theories call for modification. 59 59 I am indebted to Steve Finlay, Daniel Kwon, Tom Lyon, Johannes Schmitt, Gary Watson, George Wilson, and Gideon Yaffe for their valuable feedback in various forms. Versions of this chapter were presented at the 2009 Southern California Philosophy Conference and the 2010 Berkeley-Stanford-Davis Graduate Student Conference in Philosophy. I would like to thank my commentators Ben Wolfson and Masahiro Yamada for their insightful comments and criticisms, and the audiences for their helpful feedback. 49 CHAPTER 1 REFERENCES Adams, Frederick (1986), “Intention and Intentional Action: The Simple View,” Mind and Language 1, pp. 281-301. Anscombe, G. E. M. (1963), Intention (Ithaca, NY: Cornell University Press). Audi, Robert (1986), “Intending, Intentional Action, and Desire,” in The Ways of Desire, ed. Joel Marks (Chicago: Precedent), pp. 17-38. Audi, Robert (1973), “Intending,” Journal of Philosophy 70, pp. 387-402. Austin, John (1873), Lectures on Jurisprudence Vol. 1 (London: John Murray). Beardsley, Monroe (1978), “Intending,” in Values and Morals: Essays in Honor of William Frankena, Charles Stevenson, and Richard Brandt, eds. Alvin Goldman and Jaegwon Kim (Dordrecht, NL: D. Reidel), pp. 163-84. Bratman, Michael (1987), Intentions, Plans, and Practical Reason (Cambridge, MA: Harvard University Press). All references are to the CSLI Publications (1999) re- issue. Chisholm, Roderick (1970), “The Structure of Intention,” Journal of Philosophy 67, pp. 633-47. Churchland, Paul (1970), “The Logical Character of Action Explanations,” The Philosophical Review 79 (2), pp. 214-36. Cleveland, Timothy (1997), Trying without Willing (Brookfield, VT: Ashgate). Danto, Arthur (1973), Analytical Philosophy of Action (Cambridge: Cambridge University Press). Davidson, Donald (1978), “Intending,” in Philosophy of History and Action, ed. Yirmiahu Yovel (Dordrecht, NL: D. Reidel), pp. 41-60. Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 83- 102. Davis, Wayne (1986), “The Two Senses of Desire,” in The Ways of Desire, ed. Joel Marks (Chicago: Precedent), pp. 63-82. Davis, Wayne (1984), “A Causal Theory of Intending,” American Philosophical Quarterly 21, pp. 43-54. 50 Ginet, Carl (2004), “Trying to Act,” in Freedom and Determinism, eds. Joseph Keim Campbell, Michael O’Rourke, and David Shier (Cambridge, MA: MIT Press), pp. 89-102. Goldman, Alvin (1970), A Theory of Human Action (Englewood Cliffs, NJ: Prentice- Hall). Grice, H. P. (1971), “Intention and Uncertainty,” Proceedings of the British Academy 57, pp. 263-79. Harman, Gilbert (1976), “Practical Reasoning,” Review of Metaphysics 79, pp. 431-63. Hornsby, Jennifer (1995), “Reasons for Trying,” Journal of Philosophical Research 20, pp. 525-39. Kenny, Anthony (1966), “Intention and Purpose,” Journal of Philosophy 63, pp. 642-51. Kim, Jaegwon (1976), “Intention and Practical Inference,” in Essays on Explanation and Understanding, eds. Juha Manninen and Raimo Tuomela (Dordrecht, NL: D. Reidel), pp. 249-69. McCann, Hugh (1991), “Settled Objectives and Rational Constraints,” American Philosophical Quarterly 28, pp. 25-36. Mele, Alfred and Paul Moser (1994), “Intentional Action,” Nous 28, pp. 39-68. Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). Nagel, Thomas (1970), The Possibility of Altruism (Oxford: Clarendon Press). O’Shaughnessy, Brian (1980), The Will 2 Vol. (Cambridge: Cambridge University Press). O’Shaughnessy, Brian (1973), “Trying (as the Mental ‘Pineal Gland’),” Journal of Philosophy 70, pp. 365-86. Pears, David (1985), “Intention and Belief,” in Essays on Davidson, eds. Bruce Vermazen and Merrill B. Hintikka (Oxford: Clarendon Press), pp. 75-88. Pears, David (1984), Motivated Irrationality (Oxford: Clarendon Press). Pears, David (1980), “Intentions as Judgements,” in Philosophical Subjects; Essays Presented to P. F. Strawson, ed. Zak van Straaten (Oxford: Clarendon Press), pp. 222-37. 51 Pears, David (1964), “Predicting and Deciding,” Proceedings of the British Academy 50, pp. 193-227. Rankin, K. W. (1972), “The Non-Causal Self-Fulfillment of Intention,” American Philosophical Quarterly 9, pp. 279-89. Salmond, Sir John (1930), Jurisprudence (London: Sweet & Maxwell). Sellars, Wilfrid (1966), “Thought and Action,” in Freedom and Determinism, ed. Keith Lehrer (New York: Random House), pp. 105-39. Setiya, Kieran (2007a), “Cognitivism about Instrumental Reason,” Ethics 117, pp. 649- 73. Setiya, Kieran (2007b), Reasons without Rationalism (Princeton, NJ: Princeton University Press). Taylor, Richard (1966), Action and Purpose (Englewood Cliffs, NJ: Prentice-Hall). Velleman, David (1989), Practical Reflection (Princeton, NJ: Princeton University Press). Wasserman, Ryan (forthcoming), “Intentional Action and the Unintentional Fallacy,” Pacific Philosophical Quarterly. Yaffe, Gideon (2010), Attempts (Oxford: Oxford University Press). Yaffe, Gideon (2006), “Trying, Intending and Attempted Crimes,” Philosophical Topics 32, pp. 505-32. 52 CHAPTER 2 THE SINGLE PHENOMENON VIEW VS. THE DOUBLE PHENOMENON VIEW In contemporary philosophy of intention three important theories tend to dominate the landscape. Two of these three theories explain intentions by appeal to other, more basic mental states. These theories constitute reductive approaches to intention. What does each theory reduce intention to? One view has it that intentions are a special kind of belief: namely, a belief or expectation that one will act in a particular way. 60 The other view, as we saw in chapter 1, sees intentions as consisting of pairs of beliefs and desires, respectively. 61,62 The most influential version of the belief/desire theory places three major conditions on intending. Two of the three conditions are metaphysical. We called these conditions the ‘Strong Belief Requirement’ and ‘Strict Desire Requirement’ in chapter 1. The Strong Belief Requirement (SBR) can be stated as follows: intending to A requires believing that one will (probably) A. The Strict Desire Requirement (SDR) is this: intending to A requires (intrinsically or extrinsically) desiring to do A. The third 60 Two important proponents of this type of view are Setiya (2007) and Velleman (1989). Setiya holds that intentions are motivating, desire-like beliefs that one is will act. Velleman holds that intentions are self- fulfilling expectations of action. For a helpful discussion of Setiya and Velleman’s views, see Paul (2009). 61 Two important proponents of this view are Audi (1973) and Davis (1984). 62 In what follows, I will use the terms ‘desire’ and ‘want’ interchangeably. 53 condition is the Causal Requirement: it specifies that intending to A requires that one’s belief that one will (probably) A be caused by one’s desire to do A. By linking belief and desire in this way, this latter condition guards against potential problems stemming from causally deviant beliefs. Michael Bratman is responsible for the third major theory of intention. 63 Bratman’s theory is the most influential non-reductionist account of intention ever developed. We saw in chapter 1 that on Bratman’s view the state of intention, like desire, is a practical attitude; however, intention differs from other practical attitudes in virtue of its role in planning for the future. Bratman identifies three norms of rationality that govern intention: Strong Consistency, Means-Ends Coherence, and Agglomerativity. 64 According to Bratman, these norms are justified in virtue of their tendency to promote the aims of planning agency. Though philosophers are divided between the three theories of intention crudely sketched above, there is near universal support for the idea that there exists a necessary connection between intentional action and the state of intention. The principle expressing this connection is commonly known as the ‘Single Phenomenon View’: Single Phenomenon View (SPV): Intentional action necessarily involves the state of intention. 63 Bratman (1987). 64 Bratman (1987), p. 134. I will not here define each norm again; for this, see section 2 of ch. 1. 54 SPV is one of the more commonsensical principles in all of philosophy of mind; as such, we should expect theories of intention to support it. SPV should not be confused with what is known as the ‘Simple View’ (SV). On SV, intentionally A-ing entails intending to do A. 65 SV is a strict version of SPV. Proponents of SPV are not committed to SV. In chapter 1 I argued that proponents of two of the three theories of intention above were, in fact, committed to rejecting SPV. More specifically, I argued that the belief/desire model and Bratman’s planning theory of intention entail what I there called the ‘Double Phenomenon View’: Double Phenomenon View (DPV): Intentional action does not require having an intention. The argument went roughly like this. I presented two cases, one for each theory of intention. In the first case (FOOTBALL), I described a situation in which a rational agent, Sam, behaves intentionally yet lacks various beliefs and desires. By lacking these beliefs and desires, we see that Sam fails to satisfy SBR and SDR with regard to numerous, potentially act-motivating intentions. In the second case (the ‘video game’ example), I described a situation in which a rational agent, Sarah, behaves intentionally. I use the norms of Strong Consistency, Means-End Coherence and Agglomerativity in various ways to show that Sarah need not have an act-motivating intention. Thus, I conclude that on the belief/desire model and Bratman’s theory of intention, intentional action does not entail having an intention (DPV). 65 Proponents of this view include Adams (1986) and McCann (1991). 55 This result is problematic for the two theories of intention. By abandoning SPV, both theories come out on the wrong side of commonsense. But besides being commonsensical, is there any other reason to accept SPV? After all, commonsense can be misleading. Indeed, perhaps the fact that these two theories of intention entail DPV provides grounds for doubting SPV. In this chapter, my aim is to defend SPV. I will present various arguments in favor of SPV (sections 1 and 2). I will attempt to explain how proponents of DPV might best respond to them. 66 As we’ll see, I think proponents of DPV can plausibly respond to some—but not all—of these arguments. I’ll close by examining and dismissing a pair of arguments against SPV (section 3). If things go well, then we’ll conclude that SPV is preferable to DPV and thus both the belief/desire model and Bratman’s planning theory of intention are defective. 1 Two Arguments for SPV In this section I present two arguments for SPV. I show how proponents of DPV can adequately respond to them. In section 2, we’ll turn to what I believe are more compelling arguments in favor of SPV. 1.1 The Linguistic Argument It is perhaps surprising to learn that not more than a few arguments for SPV exist in the literature. Many philosophers tend to view the thesis as obvious. I suspect one reason for 66 The term ‘proponents of DPV’ will be used throughout to refer to belief/desire theorists and planning theorists, respectively. 56 this is that many presume something along the lines of what I’ll call ‘The Linguistic Argument’. The Linguistic Argument begins with the following observation: the word ‘intention’ is the stem of the adjective ‘intentional’. This fact naturally suggests that intentional action is action that involves the state of intention. Thus, SPV. Proponents of DPV might respond to The Linguistic Argument more or less charitably. The least charitable response goes something like this. It’s difficult to see why we should take linguistic considerations such as this seriously. After all, we can fasten any label we want to any item we choose. For instance, we call certain tables ‘coffee tables’, but no one thinks that this term captures a necessary connection between coffee and the tables for which we apply the term. Furthermore, we could easily start calling coffee tables something else if we wished to. It would be crazy, then, to think that mere appeal to the term ‘intentional action’ can be used to justify conclusions about the nature of intentional action. Although I sympathize to some extent with this response, I think proponents of DPV can say a bit more about The Linguistic Argument. As noted in chapter 1, Robert Audi rejects SPV. 67 He argues that intentional action can spring from desire as well as intention. He responds to The Linguistic Argument by claiming that paradigmatic intentional actions are “bodily actions under our direct control.” And since bodily actions directly under our control usually involve intention, we apply the adjective ‘intentional’ to all actions directly under our control whether or not they involve intention. 67 Audi (1986), section 3. 57 Audi’s point essentially is that the majority of intentional actions involve intention, and this is why we call such conduct ‘intentional’. I think proponents of DPV can say a bit more yet. After all, critics might point out that other action terms do not appear subject to the sort of explanation Audi provides for the term ‘intentional’. For example, the term ‘immoral action’ does not refer to a category of action that includes non-immoral acts; the term ‘unintentional action’ does not refer to a category that includes behavior that is not unintentional; and the term ‘mental action’ does not refer to a category that includes non-mental acts. Immoral, unintentional and mental actions are (presumably) necessarily immoral, unintentional and mental. Proponents of DPV must admit that the term ‘intentional action’ is somewhat misleading. Broadly speaking, I take ‘intentional action’ to refer to action that is aimed at. According to DPV as I conceive of it, there is more than one way to aim at action: with an intention and with a certain kind of belief. 68 So the term ‘intentional action’ is misleading under DPV. But proponents of DPV might point out that there is more than one way to refer to action of the type that is aimed at. We use the terms ‘intentional action’ and ‘purposive action’ to refer to actions 68 I call this kind of belief ‘practical belief’ below; for a discussion of practical belief, see section 2. Note that there are different ways of interpreting DPV. In what follows, I understand the view as claiming that there are two types of mental state—intention and belief—capable of motivating intentional action. I understand the view along these lines in light of my argument in ch. 1. Audi, on the other hand, holds that intentions and desires can produce intentional action. Some might even hold that there are three or four (or more) state types capable of producing intentional action. I’m inclined to think my conception of DPV is most plausible—but we should keep in mind that other versions might produce different responses to the arguments for SPV presented in this chapter. 58 that are aimed at. 69 This fact perhaps helps proponents of DPV to neutralize The Linguistic Argument. How so? Well, the term ‘purposive action’ better lends itself to DPV’s account of action that is aimed at. Recall that on DPV, agents can aim at action with or without having an intention. The stem of the adjective ‘purposive’ is ‘purpose’. Annette Baier has argued that purposes or goals are not identical to intentions. 70 One can have a goal (or purpose) without having an intention. My goal might be to play “Stairway to Heaven” on the guitar this afternoon; but since I don’t know how to play guitar, I might not intend to play the song this afternoon. However, if I do intend to play it, then it follows that I have playing the song as a goal. Having an intention plausibly entails having a goal, but not vice-versa. Goals (purposes), therefore, constitute a broader category than intentions. The observation that we use the terms ‘intentional action’ and ‘purposive action’ to refer to action that is aimed at allows proponents of DPV to flip The Linguistic Argument around on SPV. Here is The New Linguistic Argument: the word ‘purpose’ is the stem of the adjective ‘purposive’. Having a purpose does not entail having an intention. This suggests that purposive action is action that does not necessarily involve intention. The terms ‘purposive action’ and ‘intentional action’ refer to the same category of action: namely, action that is aimed at. Thus, DPV: intentional action does not entail having an intention. 69 The website <www.dictionary.com> defines intentional as: of or pertaining to intention or purpose. Malle and Knobe (1997) treat the terms ‘intentionally’ and ‘purposefully’ as synonymous in their study of people’s concept of intentional action. 70 Baier (1970), p. 649. 59 The New Linguistic Argument is not a forceful consideration in favor of DPV. 71 The point here is that The Linguistic Argument does little more to justify SPV than The New Linguistic Argument does to justify DPV. In other words, it’s more or less a push. But anything approaching a push in this context is probably good news for DPV. As I suggested above, I think many philosophers are immediately drawn to SPV because of something like The Linguistic Argument. And though I do not doubt that intentional action is so-called because it (necessarily) involves intention, I think that proponents of DPV have the resources to undermine The Linguistic Argument. 1.2 The X-Phi Argument It’s undeniable that DPV does not exactly radiate with intuitive appeal. Indeed I would bet good money that ordinary people’s intuitions tend to align with SPV. This hunch gives rise to an argument for SPV that is inspired by experimental philosophy. Experimental philosophers are yet, to my knowledge, to conduct experiments regarding people’s intuitions about SPV. But certainly we can imagine such an experiment having been conducted. Suppose that we polled hundreds of ordinary people about their intuitions regarding intentional action. And suppose we found that the majority of subjects (say, 80%) reported intuitions that align with SPV. This result, it might be suggested, would count in favor of SPV and against DPV. There are a few things proponents of DPV can say about this argument. First, if we were to give The X-Phi Argument much weight, it’s worth pointing out that we 71 For more on The New Linguistic Argument, see section 3.1 below. 60 should probably think that it would ultimately count in favor of the SV version of SPV. Recall that according to SV, intentionally A-ing entails intending to do A. SV is the strictest possible formulation of SPV. I wouldn’t be surprised that if we polled ordinary people about their concept of intentional action, most would express approval of SV. Betram Malle and Joshua Knobe polled 159 undergraduates in California about their concept of intentional action. They write, “In people’s folk concept of intentionality, performing an action intentionally requires the presence of…an intention to perform the action.” 72 Though we should await further studies before we conclude that the folk approve of SV, Malle and Knobe’s finding is largely what I would expect to be the case. But, few philosophers of action have found SV to be a plausible formulation of SPV. This remark is not itself supposed to constitute a criticism of SV. The point is that if The X-Phi Argument is taken seriously, then it looks as though it will cause trouble for both DPV and (non-SV versions of) SPV. In this sense, The X-Phi Argument causes too much damage. Secondly, and more importantly, proponents of DPV might remind us that truth is not determined by consensus. If it were, we would probably have to dismiss a slew of theories in such fields as physics, chemistry, psychology and philosophy, to name a few. If we polled ordinary people, I bet the majority would not have intuitions favoring various theories in these respective fields. But this would not constitute a compelling reason to abandon any one of them. The fact (assuming it is a fact) that most people have intuitions favoring SPV is not all by itself a deep problem for DPV. In order for The X- 72 Malle and Knobe (1997), p. 111. 61 Phi Argument to be effective, one has to relate the empirical result in some interesting (and non-question-begging) way to the nature of intentional action. It’s not exactly clear how this would be done. Still, I think proponents of DPV should say something about why people’s intuitions are as The X-Phi Argument assumes them to be. I suppose they might do this by returning to a point made above in section 1.1. According to proponents of DPV, the term ‘intentional’ is misleading: it appears to refer to a class of action that necessarily involves intention. On their view, this isn’t true: the term ‘intentional’, instead, refers to action that is aimed at. Aiming does not require intending. Thus, they might say the misleading character of the adjective ‘intentional’ can explain the statistical information contained in The X-Phi Argument. We should conclude that The X-Phi Argument does not pose a serious threat to DPV. 2 Two Better Arguments for SPV Many proponents of SPV would accept the following pair of claims about settling: (1) Settling on a course of action initiates and sustains action (at the appropriate time), and (2) Settling is subject to norms of rationality. These two claims form the bases of a pair of arguments for SPV. In this section I present these arguments and explain how defenders of DPV can most plausibly respond. Unlike the arguments of section 1, I suggest that the arguments in this section present genuine difficulties for DPV. 62 2.1 The Settling Argument In Springs of Action, Al Mele presents an argument for SPV that we will call ‘The Settling Argument’. 73 I will begin by stating Mele’s argument. Then I will explain how defenders of DPV can best respond to it. Finally, I will explain why we should be skeptical of their best response. Mele begins with the familiar claim that if one has an intention to act (at t), then it follows that one is settled upon so acting (at t). There are, furthermore, two types of intention: those directed toward the present (present-directed intentions) and those directed toward the future (future-directed intentions). Present-directed intentions— intentions to act now—involve being settled upon acting now; future-directed intentions involve (currently) being settled upon acting at a later time. Barring disruption, the settling feature of present-directed intention triggers actional mechanisms. Since only intentions have this settling feature, it must be that all intentional action involves present- directed intention. Thus, SPV. There are two ways, I believe, for proponents of DPV to respond to The Settling Argument: (1) deny that settling states are necessary for initiating intentional action or (2) deny that intentions provide the only route to settling on a course of action. Option (1) appears at first glance to be problematic for the following reason. If one denies that settling is necessary for initiating action, then one is committed to the view that certain non-settling states are sufficient for initiating action. What non-settling states might these be? Well, if settling on a course of action involves regarding the question of 73 Mele (1992), ch. 9. 63 whether to act as closed and so being committed to acting, then desires are clearly motivational states that do not incorporate a settling dimension. The problem is that many of our desires—even predominant ones—are not acted upon. But if our desires are not sufficient to initiate action, then presumably no non-settling states are. Thus, states containing a settling dimension appear necessary for triggering action. I suspect the more plausible move for proponents of DPV is to opt for (2). They should agree with Mele that all intentional action springs from settling states; but they should disagree that having an intention is the only path to settledness. Mele says that a predominant desire to A results in an intention to A as long as nothing prevents the formation of the intention. And present-directed intentions result in (intentional) action as long as nothing prevents the triggering of actional mechanisms. 74 Now, if the link between predominant desire and intentional action were ever direct (i.e., skipping intention), then Mele points out that we could easily imagine cases of intentional action without an intention. Defenders of DPV should claim that the examples of chapter 1 involving Sam and Sarah show just this: a “direct link” to intentional action. This direct link is not from predominant desire to intentional action. Rather, the link is from belief (to trying) to intentional action. In other words, what I think proponents of DPV should probably say is that certain beliefs about one’s own behavior—what I will call ‘practical beliefs’—entail being settled on a course of action. Let us see how this works in the examples of Sam and Sarah. 74 Mele (1992), p. 190. 64 In FOOTBALL, Sam the injured football player wants to get back in the game. He believes that he has to raise his arm now in order to return to the game. Sam wants to raise his arm. So, he forms the belief that he will try to raise his arm now (where ‘now’ more or less marks the time immediately following the formation of his belief). Sam’s belief that he will try to raise his arm now, belief/desire theorists should say, constitutes his now being settled on trying to raise his arm. Since Sam lacks the desire to try to raise his arm, belief/desire theorists will say that Sam does not intend to try to raise his arm. So Sam is settled on trying without intending to try. Because Sam is settled and nothing prevents the triggering of actional mechanisms, he tries to raise his arm. However, given that Sam lacks certain other beliefs and desires, we see that Sam lacks an intention on the belief/desire theory. Thus, there is a link from practical belief to trying. Furthermore, if Sam (non-luckily) succeeds in raising his arm in the way he tries, then Sam’s success is intentional. Thus, belief/desire theorists will claim that FOOTBALL is an example of practical belief resulting in intentional action. In the ‘video game’ example, Sarah the ambidextrous video-game player wants to win the prize. She believes that she has to hit a target in order to win the prize. Sarah wants to hit a target, but she knows that hitting one is difficult. So she forms the belief that she will try to hit each target now (again, where ‘now’ more or less marks the time immediately following the formation of her belief). Sarah’s belief that she will try to hit each target now, planning theorists should say, constitutes her now being settled upon trying to hit each target. And because Sarah is settled and nothing prevents the triggering of actional mechanisms, she is moved to try to hit each target. If Sarah (non-luckily) 65 succeeds in hitting a target in the way she tries, then her success is intentional. However, given that Sarah maintains rationality, we see that she lacks an intention under Bratman’s planning theory. Thus, the ‘video game’ example is also an example of practical belief issuing in intentional action. So it seems that proponents of DPV might agree with Mele that settling is necessary for initiating intentional action. They might also agree with Mele that intentions entail being settled. But they might disagree with him that intentions are necessary for being settled. On their view, practical beliefs entail being settled on a course of action. In the cases of Sam and Sarah, both are settled on acting in virtue of their practical beliefs; both, after all, appear to regard the question of whether to act as closed. And so defenders of DPV will claim that The Settling Argument does not vindicate SPV. In proposing that one can settle on a course of action by forming a practical belief, I understand proponents of DPV to suggest that one can be settled by generating a belief of the form ‘I will A at t’, where ‘A’ is an action verb and ‘t’ a specific time or duration of time. Furthermore, I take it these beliefs need to be generated in a certain way. Consider this: Jim visits a psychic who predicts that he will attempt to commit murder on Thursday. Jim is deeply distressed about this prediction, for he is a gentle soul and a law- abiding citizen—he wouldn’t dream of purposely trying to harm anyone. For some reason, Jim believes the psychic’s prediction; he believes, that is, that he will attempt to commit murder on Thursday. It seems wrong to say of Jim that in virtue of his (practical) belief he is settled on attempting to commit murder on Thursday. Rather, he seems 66 resigned to attempting murder. Being resigned and being settled are distinct states— neither entailing the other. Jim’s case shows that proponents of DPV should hold that practical beliefs need to be formed in a particular way: namely, on the basis of one’s pro- attitudes. Jim is not settled on attempting to commit murder because he does not want or hope or judge it best to murder (in fact, he wants not, hopes not, and judges it bad to murder). Had he believed that he would attempt murder because he wanted (or hoped or judged it best) to murder, then his belief would, on the view in question, constitute his being settled on attempting murder. Let us note one final point about practical belief before we turn to an objection. You might think that somewhere we have gone wrong in our development of the notion of practical belief for the following reason. When Sam and Sarah form their respective practical beliefs, it might be supposed that they make a decision about acting. Sam (perhaps) decides to try to raise his arm and Sarah (perhaps) decides to play the video game. But the state of decision plausibly entails intending. So, Sam and Sarah must intend what it is each is settled on doing. But if practical belief entails intending, then the notion of practical belief sketched above is problematic. I don’t think this is a real problem for proponents of DPV. I think the thing to say here is that Sam and Sarah’s practical beliefs are not proper decisions. Proponents of DPV should agree that deciding to do something entails intending to do it (and so being settled on doing it). Since Sam and Sarah lack the relevant intentions, they should, therefore, say that practical beliefs are neither identical to, nor the result of, nor result in, the state of decision. 67 The rough account of practical belief sketched above appears coherent: certain pro-attitude-based beliefs result in an agent’s being settled on a course of action; and when such beliefs are directed toward the present, they trigger (intentional) action. The problem with the notion of practical belief is that it attributes settledness to agents in cases it shouldn’t; in other words, it is not always true that forming a belief that one will act that is based on an appropriate pro-attitude results in one’s being settled on so acting. Consider the following example. 75 On Monday morning, Wes’s colleague, Richard, purposely humiliates Wes by insulting him in front of their boss. Richard, it turns out, has a history of this sort of offensive behavior. As a result, Wes desires to insult and humiliate Richard in return at the mandatory office party later that evening. If there is one thing Wes cannot tolerate it’s being unnecessarily humiliated. Wes knows from past experience that when he desires to humiliate a humiliator, he more often than not succeeds in doing so. And so Wes forms the belief that he will (probably) try to humiliate Richard at the party because he now desires to humiliate him. According to the account of practical belief above, Wes is settled on trying to humiliate Richard at the party. But suppose Wes believes that it is immoral to insult and humiliate others—even in cases of “getting even.” Indeed, suppose Wes wants not to insult and humiliate Richard at the party, and even wants not to want to insult and humiliate Richard at the party. Wes has always had something of a temper and thirst for revenge, and he lately wishes he could change this about himself. Still, Wes believes that he (probably) will try to insult Richard because of his desire to do so. It seems 75 This example is inspired by an example from Mele (1992). 68 implausible to suppose that Wes is settled on trying to insult Richard at the party. Wes is too conflicted about trying to count as being settled. Settling on a course of action characteristically involves viewing the question of whether to act as closed. Although Wes is relatively confident that he will try to insult Richard tonight, he need not (presently) view the question of whether to try to insult him as closed. On the contrary, Wes might well be inclined now to take measures to decrease his chances of insulting Richard at the party. For example, he might plan on meeting with friends for lunch to solicit advice on how to best handle the matter; he might decide (now) not to drink at the party since alcohol tends to reduce his inhibitions; and so on. Moreover, settling tends to influence one’s further deliberation and intentions. If one were settled on A-ing, then (ceteris paribus) one would be irrational if one simultaneously formed intentions aimed at significantly decreasing the likelihood of A-ing. But it is not at all irrational for Wes to form intentions aimed at significantly decreasing the likelihood of his trying to insult Richard at the party: he can rationally intend to meet his friends for lunch in order to get advice, and rationally intend not to drink at the party, etc. Thus, Wes must not be settled on trying to insult Richard at the party. Contrary to proponents of DPV, then, practically believing that one will A does not entail settling on A. Proponents of DPV might try to respond to this objection in the following way. The reason why Wes is not settled on trying to insult Richard is because he’s conflicted about insulting him. If he weren’t conflicted, then his practical belief would result in his being settled on trying to insult Richard. In other words, practically believing that one will A entails settling on A as long as one is not conflicted about A-ing. I don’t think this 69 sort of response is convincing. Regardless of what it means to be “conflicted” here, the response is ad hoc. When one settles on a course of action via intention, one might be (more or less) conflicted about so acting. I might intend to smoke a cigarette after dinner and thus be settled on smoking a cigarette; and this intention might be the product of a desire to smoke after dinner. But I might also have a standing desire not to smoke. Indeed, I might also hope not to smoke and judge it best not to smoke. This inner conflict I experience, however, does not alter the fact that I am currently settled on smoking after dinner. If conflict does not disrupt settling in cases of intending, then we should not expect it to in cases of practical belief. To summarize: the examples of Sam and Sarah show us that on the belief/desire model and the planning theory of intention, one can act intentionally without having an intention (DPV). Since acting intentionally plausibly requires being settled on so acting, belief/desire theorists and planning theorists must explain the sense in which Sam and Sarah are settled on acting. I’ve suggested that the most plausible way to do this is by appeal to the notion of practical belief: believing that one will act at t because of an appropriate pro-attitude(s) entails being settled on acting at t. However, I then presented an example in which practical belief and settledness come apart. We conclude that proponents of DPV cannot account for the settling component of intentional action. Thus, The Settling Argument favors SPV. 70 2.2 The Normative Settling Argument The Settling Argument above focuses on the necessity of settling for intentional action. Another argument for SPV can be formulated that focuses on the normative aspect of settling. Let us call this argument ‘The Normative Settling Argument’. 76 The argument runs like this. The state of being settled on a course of action appears subject to the following rational constraints: (C1) If S settles on A-ing (at t), then S is irrational if S settles on not-A-ing (at t). (C2) If S settles on A-ing (at t), then S is irrational if S believes that S will not-A (at t). (C3) If S settles on A-ing and believes that B-ing a necessary means to A-ing, then S is irrational if S does not settle on B-ing. Notice that C1-C3 can be explained by rational constraints on intention. On the assumption that intending to A entails settling on A: C1 can be explained by the demand that one’s intentions be consistent; C2 can be explained by the demand that one’s intentions and beliefs be consistent (Strong Consistency); and C3 can be explained by Bratman’s principle of Means-End Coherence. Given the ease with which intention can explain these normative constraints on settling, we conclude that there is a necessary connection between settling and intending. And since all intentional action plausibly springs from being settled (see section 2.1), it follows that intentional action necessarily springs from intention. 76 Gideon Yaffe brought this argument to my attention. 71 I don’t think that DPV can adequately account for C1-C3. Recall that DPV is the view that intentional action does not require the state of intention. And we know that acting intentionally plausibly requires settling. Proponents of DPV are thus committed to the view that one can be settled on acting via a state other than intention. Now then, when agents are settled on acting via intending, proponents of DPV can hold that C1-C3 are justified by rational constraints on intending (as noted above). But what about when agents are settled via a state other an intention? Up until the end of section 2.1, I’ve been suggesting that proponents of DPV should hold that practical belief is the other state capable of producing intentional action. But we now know that practical belief does not entail settledness. If it did, proponents of DPV could claim that when agents are settled via practical belief, C1-C3 are justified by rational constraints on belief. To see how this would work, let us return to our three rational constraints on settling. The first is this: (C1) If S settles on A-ing (at t), then S is irrational if S settles on not-A-ing (at t); the second: (C2) If S settles on A-ing (at t), then S is irrational if S believes that S will not-A (at t); and the third: (C3) If S settles on A-ing and believes that B-ing is a necessary means to A-ing, then S is irrational if S does not settle on B-ing. In cases in which settling issues from practical belief, proponents of DPV could claim that C1 and C2 are explained by the demand that one’s beliefs be consistent. C3 is just a version of the Means-End Coherence principle—one that involves belief. Explaining C1-C3 by appeal to rational constraints on belief is a coherent strategy. But the problem with this strategy is by now clear: practical belief does not entail settledness. If practical belief does not entail settledness, as was argued in section 72 2.1, then proponents of DPV must identify some other state (beside intention) that entails settledness. 77 It’s difficult to imagine this could be done. Furthermore, even supposing that such a state type could be identified, proponents of DPV are still left with the task of accounting for C1-C3. Both of these challenges appear daunting. Thus SPV, but not DPV, is capable of accounting for C1-C3. DPV is the view that intentional action can be produced by intention and by (practical) belief. I have argued in this section that DPV faces two serious problems. First, practical belief does not entail being settled on a course of action (2.1). This is problematic because intentional action plausibly requires settling. Thus, practical belief is incapable of producing intentional action. Second, DPV is incapable of accounting for rational constraints on settling (2.2). We conclude that SPV is the more plausible of the two views. 3 Two Arguments against SPV We’ve examined four arguments above in favor of SPV. I have argued that two of these arguments pose genuine difficulties for DPV. In this section I raise two arguments against SPV. I explain why neither argument is persuasive. 77 The state of decision is not an option because it presumably entails intention. 73 3.1 The New Linguistic Argument (Revisited) The New Linguistic Argument, you’ll recall, was introduced in section 1.1. I will restate it here and quickly dismiss it. The more interesting argument is the argument in section 3.2 below. The New Linguistic Argument runs as follows. The terms ‘intentional action’ and ‘purposive action’ are interchangeable. The noun ‘purpose’ is the stem of the adjective ‘purposive’. But having a purpose (or goal) does not entail having an intention. Thus intentional action does not entail intention (DPV). Just like The Linguistic Argument for SPV above (section 1.1), The New Linguistic Argument is weak. Regardless of whether one sympathizes with SPV or DPV, both camps want the same thing: to understand the nature of intentional action. I just don’t think we should place much stock in linguistic considerations such as these. Having said that, I think proponents of SPV ought simply to reject the premise in The New Linguistic Argument that the terms ‘intentional action’ and ‘purposive action’ are interchangeable. Given that acting with a purpose (presumably) does not entail acting from intention, proponents of SPV will hold that purposive action constitutes a broader category than intentional action—intentional action is a subclass of purposive action. Denying the interchangeability of these terms is not unreasonable; so The New Linguistic Argument is blocked. 74 3.2 The Regress Argument The Regress Argument is a more substantive consideration against SPV. The argument is inspired by Gilbert Ryle. In The Concept of Mind, Ryle presents a series of arguments for the elimination of volition. 78 One of Ryle’s arguments focuses on the apparent inability of volition to account for voluntary action—a task that Ryle believes is supposed to ontologically ground volition. I’ve reconstructed Ryle’s argument using the notions of intention and intentional action rather than volition and voluntary action. The result is a regress argument against SPV. We can state it as follows: (1) Intentions are either formed intentionally (call these ‘i-intentions’) or they are passively acquired (call these ‘p-intentions’). (2) Because i-intentions are formed intentionally, they must issue from prior intentions, and those from prior intentions, ad infinitum. (3) P-intentions are such that we “cannot help” but have them. 79 (4) If we cannot help but have p-intentions, then it would be “absurd” to call actions that are caused by them ‘intentional’. (5) Therefore, (p- and i-) intentions are incapable of producing intentional action; and SPV is false. It’s worth noting that The Regress Argument not only does violence to SPV, but to standard versions of DPV as well. Most proponents of DPV will not deny that intentions can cause intentional action; they only deny that intentions are necessary for intentional 78 Ryle (1949), ch. 3; The Regress Argument is based on an argument on p. 67. 79 Double quotation marks in this argument indicate Ryle’s language. 75 action. But if The Regress Argument is right, intentions can never produce intentional action. Proponents of SPV can block the regress of (2) by holding that intentional action, including the formation of i-intentions, ultimately issues from p-intentions. Since the acquisition of p-intentions is non-actional, they need not result from prior intentions. Thus, Ryle’s regress is blocked. But we are not out of the woods yet; Ryle will now direct our attention to premises (3) and (4). It’s Ryle’s claim that p-intentions are incapable of generating intentional action—indeed, he thinks it’s “absurd” to suppose otherwise. If this is right, then employing p-intentions to halt the regress will not work. Let us then turn to the issue of p-intending. Ryle claims in (3) that we “cannot help” but have our p-intentions. From this, he concludes in (4) that it would be “absurd” to call actions ‘intentional’ that result from p- intentions. Ryle is right that there is a sense in which we cannot help having p-intentions. However, there is also a sense in which we can help having them. It is uncontroversial that we can often revise, suspend, and even abandon our intentions. This seems true of our p-intentions no less than our i-intentions. Contrary to Ryle, then, often we can do something about having our p-intentions. In order for Ryle’s (3) to be plausible, it has to be stated as a claim about our coming to have p-intentions: (3*) We cannot help coming to have our p-intentions. But once we read (3) as (3*), (4) becomes problematic. Instead it should read: 76 (4*) If we cannot help coming to have our p-intentions, then it would be absurd to call actions caused by them ‘intentional’. I think proponents of SPV ought to reject (4*). Intentions normally trigger action if they persist until the relevant time. Their persistence is something we usually have a certain amount of control over. In normal circumstances, agents permit or allow their p-intentions to persist until the time for action arrives. When our p-intentions are acquired, we are not forced into retaining them and therefore acting upon them. If we were, then perhaps (4*) would be plausible. But, many of the p-intentions we come to have we do not act upon 80 : again, we often revise, suspend, and abandon them. Our p-intentions set us on course for action rather than locking us in. If our p-intentions move us it’s usually because we allow ourselves to be moved by them. Now, when we allow our p-intentions to move us, the resulting actions are, I submit, rightly deemed intentional. Proponents of SPV should hold that the intentionality of action does not depend upon the way we come to have our intentions. If we permit our p-intentions to be effective, then our resulting conduct is intentional. Contrary to premise (4*), such a position is not “absurd.” Proponents of SPV, therefore, ought to reject (4*) and thus (5). So agents must allow their p-intentions to move them in order for their conduct to be intentional. But proponents of The Regress Argument might press us further. They might ask: is this allowing intentional or not? If it is, then Ryle’s regress is back up and 80 This is probably more common in cases involving future-directed p-intentions. 77 running; all this business about intentions and allowing above will have done nothing more than push the issue back one step. I think what we can say here is that intention- allowing need only occur as the result of one’s wanting or desiring to do (or more broadly: a pro-attitude favoring) what is intended. And so one’s allowing of a p- intention to persist need not itself be intentional in order for one’s resulting conduct to be intentional; rather p-intentions need only be allowed because of a pro-attitude on the part of the agent to act in the way that he or she (p-)intends. Surely there is a lot more that can be said about this issue of allowing and pro- attitudes just introduced. Furthermore, it’s not my claim that this move is the only move available to proponents of SPV. But I think what has been said above is enough to undermine the regress proposed by The Regress Argument. 4 Conclusion In chapter 1 I argued that the belief/desire view and Bratman’s planning theory of intention were equally committed to DPV: acting intentionally does not require having an intention. I claimed that this commitment was problematic on grounds that an adequate theory of intention ought not be incompatible with SPV. This latter claim was based on an appeal to commonsense. In this chapter I have tried to show that SPV is preferable to DPV on philosophical grounds. I presented four arguments for SPV above, and explained how proponents of DPV can best respond to each of them. I argued that their responses to the two settling arguments for SPV (section 2) were unsatisfactory. We then turned to two arguments against SPV; I explained why neither argument succeeds. I 78 conclude that theories of intention ought to be compatible with SPV; and thus the belief/desire theory and the planning theory of intention are defective. 81 81 Thanks to Gideon Yaffe, Daniel Kwon, and Gary Watson. 79 CHAPTER 2 REFERENCES Adams, Frederick (1986), “Intention and Intentional Action: The Simple View,” Mind and Language 1, pp. 281-301. Audi, Robert (1986), “Intending, Intentional Action, and Desire,” in The Ways of Desire, ed. Joel Marks (Chicago: Precedent), pp. 17-38. Audi, Robert (1973), “Intending,” Journal of Philosophy 70, pp. 387-402. Baier, Annette (1970), “Act and Intent,” Journal of Philosophy 67, pp. 648-58. Bratman, Michael (1987), Intentions, Plans, and Practical Reason (Palo Alto, CA: CSLI). Davis, Wayne (1984), “A Causal Theory of Intending,” American Philosophical Quarterly 21, pp. 43-54. McCann, Hugh (1991), “Settled Objectives and Rational Constraints,” American Philosophical Quarterly 28, pp. 25-36. Malle, Bertram and Joshua Knobe (1997), “The Folk Concept of Intentionality,” Journal of Experimental Psychology 33, pp. 101-21. Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). Paul, Sarah (2009), “How We Know What We’re Doing,” Philosophers’ Imprint 9, pp. 1- 24. Ryle, Gilbert (1949), The Concept of Mind (Chicago, IL: The University of Chicago Press). Setiya, Kieran (2007), Reasons without Rationalism (Princeton, NJ: Princeton University Press). Velleman, David (1989), Practical Reflection (Princeton, NJ: Princeton University Press). Yaffe, Gideon (2006), “Trying, Intending and Attempted Crimes,” Philosophical Topics 32, pp. 505-32. 80 CHAPTER 3 AN ACTIONAL ACCOUNT OF REFRAINING We do lots of things. Some of the things we do are intentional: for instance, my turning on the computer and typing these words, and your reading them and scribbling curse words in the margin. Broadly speaking, intentional behavior is behavior that we aim at performing, either as an end or a means to an end. Of course, not everything we do is aimed at. Many things we do unintentionally: we’ve all stepped on someone’s foot or misspelled a word. Unintentional behavior can be characterized as behavior that is accidentally or inadvertently performed. These two categories—the intentional and unintentional—do not, in my opinion, exhaust the ways of doing things. Al Mele and Steven Sverdlik have argued that some of our behavior plausibly runs between the intentional and unintentional. 82 They call such behavior non-intentional. Mele and Sverdlik’s notion of non-intentionality can be understood by way of an example. Consider Gilbert Harman’s famous ‘sniper’ scenario. 83 Suppose a sniper wants to kill a high-ranking, enemy soldier. He gets in position and trains the sight of his rifle on the soldier. But he knows that if he fires, he will alert the enemy to his presence. Still, the sniper thinks that killing the soldier is worth the danger involved in alerting the enemy to his presence. So, he fires his rifle at the soldier; in so doing, he alerts the 82 Mele and Sverdlik (1996), p. 274. 83 Harman (1976), p. 433. 81 enemy to his presence. His firing directly issues from intention and so is undoubtedly intentional. What about his alerting the enemy to his presence? Is that intentional? Harman thinks it is. 84 But alerting the enemy to his presence is not something the sniper aims at doing, either as a means or an end; it is not what he tries to do. It also seems implausible to describe the sniper’s alerting the enemy as unintentional; after all, the sniper knew that he was going to alert the enemy by firing the gun. According to Mele and Sverdlik, behavior that is neither intentional nor unintentional is non-intentional. The sniper non-intentionally alerts the enemy to his presence. 85 So there are several different ways of doing things. There are also several different ways of not doing things. Some of the things we don’t do are intentional. Let us call intentional non-performances ‘refrainings’. Refraining from acting, as we’ll see a bit later, requires considerable work to achieve. Most of the things we do not do we do not refrain from doing. Rather, the majority of our omissions are unintentional. 86 I am not currently levitating or shaking hands with Elvis. But surely I’m not refraining from doing these things. I hadn’t even considered doing them, and I’m pretty sure it’s impossible for me to succeed in doing either. At any given moment there are zillions of things we are not doing—only a handful of which are likely intentional. Furthermore, as in the case of doing, I take it that some of our non-doings are also non-intentional. 84 Bratman (1984) would agree with Harman. 85 I’m inspired by Mele and Sverdlik’s argument, though I’m not completely convinced by it. In ch. 4 I’ll develop an argument for non-intentionality that differs in certain respects from Mele and Sverdlik’s argument. 86 I use the terms ‘omission’, ‘not doing’ and ‘failure’ interchangeably. 82 Doing and not doing, then, can be achieved three ways apiece. The category of non-intentional omission, however, has not, to my knowledge, received attention in the literature on omitting. I think this category is important because of its impact on the theory of refraining. Refraining is the most philosophically interesting type of omission. It is important for understanding the nature of intentionality, as well as morality and responsibility (legal, moral and causal). By paying closer attention to non-intentional omissions, we can decrease the likelihood of mistaking them for refrainings and vice versa. This, in turn, can lead to the development of more plausible theories of refraining. In this chapter I want to develop an actional account of refraining that respects non-intentionality. It is my goal to improve upon my recently offered account of refraining. 87 In sections 1 and 2 I will explain my recent account (including its roots) and present a challenge to it. The challenge will involve cases of non-intentional omission. In sections 3-7, I will attempt to revise my account in light of this challenge. In section 8 I will consider several objections to my revised view. 1 Brand and Yaffe on Refraining We ordinarily use the term ‘refraining’ to denote the intentional. We say things like, “Allison dislikes John and so she refrained from going to his party last night” and “Leonard, a chain smoker, refrained from smoking a cigarette after dinner.” I, too, did not go to John’s party last night, nor did I smoke a cigarette after dinner, but it doesn’t seem that I refrained from doing these things. Why not? Well, part of the answer seems 87 Johns (2009). 83 to be that I unknowingly did nothing with respect to John’s party and smoking: I don’t know John and so I wasn’t aware that he threw a party last night; and I don’t smoke, so it wouldn’t have occurred to me to smoke after dinner. Allison and Leonard, on the other hand, purposely do not do the things in question. Refraining is, by ordinary standards, intentional non-doing. Moreover, if something is intentional, it must be actional. With a little reflection, however, this may seem puzzling. If intentional action necessarily involves doing something, then how can refraining properly constitute action? Isn’t it rather inaction? In his essay “The Language of Not Doing,” Myles Brand presents an account of refraining that helps to preserve our sense that refrainings are actional. 88 According to Brand, refraining from doing something necessarily involves doing something else. We can state Brand’s actional view as follows: (MB) S refrains from A-ing iff (1) S does not A, and (2) S Bs in order that S does not A. On Brand’s view, we would say that Allison refrained from going to John’s party by going to the movies; and Leonard refrained from smoking by taking a nap. Under MB, refrainers are always doing something; and so MB supports the claim that refrainings are actions. I am attracted to Brand’s account of refraining. But even if we accept Brand’s view, there remains much work to be done. MB provides us with an important but broad understanding of what refraining comes to. A complete understanding of refraining 88 Brand (1971), p. 49. 84 requires us to draw out the details that are packed in to MB’s (1) and (2). In his essay “Locke on Refraining, Suspending, and the Freedom to Will,” Gideon Yaffe provides an account of refraining that elaborates MB. 89 Yaffe’s view is this: (GY) S refrains from A-ing iff (1) S considers A-ing, (2) S Bs voluntarily, and (3) B-ing is incompatible with A-ing. I’m sympathetic to GY’s (1)-(3). However, I have argued that (1)-(3) do not constitute sufficient conditions for refraining. 90 It’s my view that refraining from doing something requires having the ability (or power) to do it. I therefore proposed that GY should be modified as follows: (REF) S refrains from A-ing iff (1) S considers A-ing, (2) S has the power to A, (3) S Bs voluntarily, and (4) B-ing is incompatible with A-ing. Though I believe that REF is on the right track, I am no longer inclined to think that it is an adequate account of refraining. As with GY, REF does not supply sufficient conditions for refraining. In the next section, I will argue that REF’s (1)-(4) are insufficient for refraining. This will lead us to an improved account of what it is to refrain from acting. 89 Yaffe (2001), p. 378. I’m not suggesting here that Yaffe constructed GY with an eye to elaborating Brand’s MB; I’m only suggesting that we can understand GY as an expansion of MB. 90 Johns (2009), section 3. 85 2 A Difficulty for REF Before we get to what’s wrong with REF, let us slightly alter REF as follows: (REF) S refrains from A-ing iff (1) S considers A-ing, (2) S has the ability to A, (3) S Bs intentionally, (4) S lacks the belief that B-ing is compatible with A-ing, and (5) B-ing is incompatible with A-ing. The major difference between this version of REF and the older version is the addition of condition (4). From here on, let ‘REF’ refer to this set of conditions on refraining. I will not here attempt to establish the necessity of REF’s five conditions. This is a task for another day. My goal in this chapter is not to convince readers that a true account of refraining must incorporate REF’s (1)-(5). Rather, it is to provide an account of refraining that appeals to sympathizers of the actional approach to refraining. That said, it is worth making a few surface-level remarks about REF’s (1)-(5). Condition (1)—the considering condition—helps explain why most of the things we do not do are not refrainings. What makes it true that five minutes ago my failure to do one- hundred jumping jacks was not a refraining? Well, at least part of the answer seems to be that five minutes ago it had not occurred to me to do jumping jacks. Many non-doings fail to satisfy (1). Many also fail to satisfy condition (2), the ability condition. Although I didn’t do one-hundred jumping jacks several minutes ago, I could have (I think!). Contrast this failing with my failing several minutes ago to shake hands with Elvis. Intuitively, this latter failing falls short of refraining even had I considered shaking hands 86 with Elvis. Condition (2) captures this intuition: my failing to shake Elvis’s hand several minutes ago is not a refraining because I was unable to shake it. Condition (3) states that one must do something intentionally; and condition (5) tells us that what is done intentionally must not involve doing what one refrains from doing. Condition (5) is obviously true if we accept the necessity of condition (3). Why should we accept condition (3)? As noted, condition (3) helps preserve the sense we have that refrainings are bona fide actions. If action necessarily involves doing something, then condition (3) supplies the doing component of refraining. Lastly, I have come to think that condition (4) belongs in an account of refraining in light of the following sort of consideration. Suppose we have an agent that satisfies conditions (1), (2), (3) and (5). Brett considers going to Canada, and he can go if he chooses to. He then intentionally travels to Alaska. Since traveling to Alaska is incompatible with traveling to Canada, Brett refrains from traveling to Canada if these four conditions are sufficient for refraining. But suppose that Brett believes that Alaska is in Canada, and so he believes that going to Alaska involves going to Canada. It seems false to think now that Brett refrains from traveling to Canada. Brett’s belief seemingly prevents him from refraining from doing this. Hence, condition (4). As I mentioned, I think REF is on the right track. However, I think it is inadequate as it stands. Consider the following example: (BABY 1) Two babies are floating down a river at a fast rate of speed. The babies are one-hundred feet apart and are certain to drown if no one intervenes. Jasper is standing downstream directly in between the approaching babies. He realizes that he can only save one of 87 them. There is no reason for Jasper to prefer saving one baby over the other—he is not related to either baby, he has no reason to believe that one will grow into a moral monster, etc. Jasper considers saving the baby to his left (Baby L), and he has the ability to do so. He then flips a coin, prompting him to form the intention to save the baby to his right (Baby R). He thus saves Baby R. Jasper knows that saving Baby R is incompatible with saving Baby L. According to REF, Jasper refrains from saving Baby L. But this seems wrong. If Jasper refrains from saving Baby L, then he intentionally does not save her. If he intentionally does not save her, then not saving her is what he aims at doing, either as an end or a means to an end. But clearly not saving Baby L is not Jasper’s aim or goal. He is not striving or trying not to save her. What he strives to do is save Baby R—this is his goal. Aiming at doing something, X, and knowing that X results in something else, Y, does not entail aiming at Y. Is Jasper’s failure to save Baby L unintentional? It seems unlikely. Jasper knows (as he embarks on saving Baby R) that not saving Baby L is a consequence of saving Baby R; his failure to save her is not accidental. Consider this: as Jasper saves Baby R, he also (unknowingly) fails to save a particular starving child on the other side of the world. This failure of Jasper’s is unintentional—it’s inadvertent. But certainly this failure is unlike his failure to save Baby L. If Jasper’s failure to save Baby L is neither intentional nor unintentional, it must be non-intentional. Jasper intentionally saves Baby R. And he is aware that his saving Baby R carries a consequence: namely, that he will not save Baby L. So Jasper’s failure, like the sniper’s alerting the enemy to his own presence, is a foreseen consequence of an 88 intentional action. The problem with REF is that it fails to appreciate the category of non-intentional omissions. Consider another example: (BABY 2) A baby is floating down a river certain to drown if no one intervenes. Jerry is standing downstream and realizes that he can save the baby. He considers saving the baby and has the ability to do so. But Jerry has reason to believe that the baby will grow up to be a murderer. So Jerry decides not to save the baby, and intentionally stands still as the baby floats by. Standing still is incompatible with saving the baby, and Jerry is aware of this. Furthermore, suppose that if Jerry were to (try to) save the baby, he would get his hair wet. Jerry is aware of this too (he considers getting his hair wet). But Jerry is indifferent about whether or not he gets his hair wet. According to REF, Jerry refrains from saving the baby and refrains from getting his hair wet. But whereas he plausibly refrains from saving the baby, it seems wrong to suppose he refrains from getting his hair wet. The reason is the same as in BABY 1: Jerry does not aim at not getting his hair wet. Just as it wasn’t Jasper’s goal not to save Baby L, it isn’t Jerry’s goal to avoid wetting his hair. Of course, Jerry’s not getting his hair wet is not unintentional either. It’s a foreseen consequence of his intentional failure to save the baby. It is, therefore, non-intentional. BABY 1 and BABY 2 are examples of non-intentional omission. The main difference between the scenarios is that in BABY 1 Jasper wants to do the thing that he non-intentionally omits doing (saving Baby L); whereas in BABY 2, Jerry is indifferent about doing what he non-intentionally omits doing (getting his hair wet). Both scenarios motivate a sixth necessary condition on refraining: 89 (6) S’s not A-ing is S’s aim or goal. Condition (6) guarantees that REF does not mistake non-intentional omissions for intentional ones. But what exactly does it mean for one’s not doing something to be one’s aim or goal? This is a difficult question. Though I certainly cannot thoroughly address it here, we will explore it below. In the next five sections we’ll investigate this question by identifying some of the different ways condition (6) can be satisfied. The payoff will come at the end of section 7 when we arrive at a more precise and informative statement of condition (6). 3 Wanting Commonsense suggests that often people’s aims and goals are linked to what it is they want. 91 Kobe wants to win an NBA championship and so he has winning a championship as a goal; Angelica wants to get an ‘A’ on the exam because her goal is to earn an ‘A’ in the class. Examining the state of wanting is a natural place to begin our investigation into some of the different ways condition (6) might be satisfied. In both of the BABY cases, neither of our agents want to do what they (non-intentionally) fail to do. Recall that in BABY 1, Jasper wants to save both babies—and so wants to save Baby L. But given the difficulty of the circumstances, he (non-intentionally) fails to save Baby L. In BABY 2, Jerry neither wants to get his hair wet nor wants not to get his hair wet. Jerry decides not to save the baby and thus (non-intentionally) fails to get his hair wet. These facts about 91 As in previous chapters, I use the terms ‘want’ and ‘desire’ interchangeably. 90 Jasper’s and Jerry’s wants seem to explain why neither refrains from the relevant actions. Usually our goals are accompanied by our wanting to achieve them. Since neither wants what they’re (non-intentionally) not doing, neither seems to have their relevant omissions as goals. For ease of reference, let us re-state REF (sans condition (6)): (REF) S refrains from A-ing iff (1) S considers A-ing, (2) S has the ability to A, (3) S Bs intentionally, (4) S lacks the belief that B-ing is compatible with A-ing, and (5) B-ing is incompatible with A-ing. Now then, given our discussion of Jasper’s and Jerry’s wants (or lack thereof), the following form of condition (6) naturally emerges. (6a) S wants not to A. Condition (6a) helps REF capture the fact that what one refrains from doing is something that is aimed at. When added to REF, we see that (6a) gets the right results in the BABY cases. In BABY 1, ‘A’ represents the action ‘save Baby L’. And we know that Jasper lacks the want not to save Baby L. So REF and (6a) jointly entail that Jasper does not refrain from saving Baby L. In BABY 2, ‘A’ represents ‘get hair wet’. Since Jerry lacks the want not to get his hair wet, REF and (6a) jointly entail that Jerry does not refrain from getting his hair wet. These are precisely the results we’re after. 91 So condition (6a) gets it right in the examples discussed so far. But we’re yet to demonstrate the necessity of (6a). Indeed, on closer inspection, we’re yet to establish that (6a) plus REF’s (1)-(5) constitute sufficient conditions on refraining. It seems that one’s wanting not to A needs to relate to one’s intentionally B-ing in a certain way in order to guarantee that one’s not A-ing is one’s goal. The reason is this. One can want to do something without having doing it as a goal. I might want to smoke a cigarette, but because I am trying to quit, smoking a cigarette is not my goal; Mel might want to run five miles in order to burn calories, but because running five miles is hard work, she fails to take it up as a goal. What this shows us is that one might satisfy REF plus (6a) and yet still not have not A-ing as a goal—and therefore not properly refrain from A-ing. REF plus (6a) are insufficient. In order to guarantee that one’s not A-ing is one’s goal when one Bs, we need one’s wanting not to A to play an important role in the production of one’s B-ing. So the question is: how might we link one’s wanting not to A with one’s intentionally B-ing? To answer this question, let us re-introduce the Single Phenomenon View (SPV). Recall that according to SPV, intentionally X-ing entails having an intention, though the intention in question need not be an intention to do X. In chapter 2 I defended SPV on grounds that it is more plausible than the alternative Double Phenomenon View (DPV): the view that intentionally X-ing does not entail having an intention. When we couple SPV with condition (3) of REF, it follows that refraining from A-ing requires having an intention that motivates one’s (intentional) B-ing. Now, this intention does not have to be an intention to B, though it could be. Let’s call the intention that is required by (3) plus SPV 92 the intention to B*. On REF plus SPV, it follows, then, that one’s intention to B* produces one’s act of B-ing. With this point in place, we are now in a position to connect one’s wanting not to A ((6a)) with one’s B-ing intentionally ((3)). (7a) S’s want not to A is (part of) S’s reason for intending to B*. It seems to me too strong to require that S’s want not to A be the sole reason for S’s intending to B*. After all, S might intend B* because S both wants not to A and wants to (try to) B. It would still follow that not A-ing is (one of) S’s goal. Hence, the ‘part of’ qualifier. By attaching (6a) and (7a) to REF, we arrive at a set of minimally sufficient conditions for refraining. Let us return to the example involving Mel. Mel considers running five miles and has the ability to do it. She wants to run in order to burn calories; however, she also wants not to run because of the effort it requires. Her want not to run is stronger and eventually wins out, resulting in her forming the intention to take a nap. She takes a nap intentionally. Mel knows that taking a nap is incompatible with running. REF plus (6a) and (7a) entail that Mel refrains from running. And this seems exactly right: not running is (one of) Mel’s goal(s); her not running, therefore, is intentional. So we get sufficiency by attaching (6a) and (7a) to REF. The question now is: is (6a) a necessary condition on refraining? (Of course, whether (7a) is necessary depends on whether (6a) is necessary; if the latter isn’t, then we know the former can’t be.) Condition (6a) captures the aiming feature of refraining. Are there any other conditions 93 that can also capture this feature? To answer this, we need to examine further possible conditions on refraining. If we can identify any others that support the aiming feature, we can conclude that conditions (6a) and (7a) are unnecessary for refraining. So let us temporarily bracket (6a) and (7a) and search for other conditions that, when attached to REF’s (1)-(5), deliver sufficiency. Another wanting condition that comes to mind is: (6b) S wants to B. It should be clear, however, that (6b) cannot account for the aiming feature of refraining. Recall that in BABY 1, Jasper was in possession of this want (‘B’ = ‘save Baby R’). If we couple (6b) with REF’s (1)-(5), it follows that Jasper refrains from saving Baby L. But we know that Jasper does not so refrain. And so attaching (6b) to REF cannot constitute sufficient conditions for refraining. Now just because (6b) and REF are not sufficient for refraining, it does not follow that there is no place in our account of refraining for (6b). It might be that (6b) is necessary for refraining. If it is, however, it is in a sense we’re not particularly interested in here. Here’s why. Condition (6b) cannot by itself explain the aiming feature of refraining; and what’s more, it does not appear necessary to explain it. Therefore, if (6b) is necessary for refraining, it is because wanting to B is entailed by intentionally B-ing (condition (3) of REF). But whether or not (3) entails (6b) is a question we can leave open here. If you are inclined to think B-ing intentionally entails wanting to B, then you 94 can understand REF’s condition (3) as embedding (6b). If, on the other hand, you are adverse to the relevant entailment, then do not read (6b) into (3). We can generate a third wanting condition by combining the two above: (6c) S wants to B and not to A. Condition (6c) entails that one wants not to do the thing they do not do. And we’ve already seen that wanting not to do something can lead to one having not doing it as a goal or aim. Let us introduce an appropriate linking condition: (7b) S’s want to B and not to A is (part of) S’s reason for intending to B*. The pairing of (6c) and (7b) with REF, then, generates another set of sufficient conditions for refraining. Are (6c) and (7b) necessary conditions? I don’t think they are. A conjunctive want entails each of its conjuncts. Condition (6c), therefore, entails that S wants not to A. But as we’ll see just below in our discussion of condition (6e), wanting not to A is plausibly unnecessary for refraining from A-ing. Thus, neither (6c) nor (7b) are necessary in order to refrain. 92 92 So let us bracket (6c) and (7b) and continue to examine other possible conditions on refraining. 95 Here is another possible wanting constraint on refraining: (6d) S wants to B more than S wants to A. Condition (6d) cannot capture the aiming feature of refraining. One might want to do A and B, and yet slightly want more to B. I might want to go both to Miami and Honolulu, yet want slightly more to go to Honolulu. But in (deciding on) going to Honolulu, it is not (other things equal) thereby my goal not to go to Miami. Thus, attaching (6d) to REF will not produce sufficiency. Is (6d) necessary? Again, if it is, it will be in a way that, for our purposes, isn’t particularly interesting. Condition (6d) cannot account for aiming. So if (6d) is necessary for refraining, it is because some other of REF’s conditions entail it (e.g., (1), (3) and (4)). But this is not an issue I here want to investigate. As in the case of (6b), then, let us leave it open as to whether (6d) is entailed by REF. Here is another wanting condition: (6e) S wants to B in order not to A. I used to think that condition (6e) entailed that S wants not to A; so I thought we could connect it to (6a), (7a) and REF and thereby create a new set of sufficient conditions for refraining. But I no longer think that’s right. If Jerry wants to stand still in order not to save the baby, then Jerry might want not to save the baby; but he doesn’t have to want 96 this. Instead, Jerry might not want to save the baby. Not wanting to do something is not wanting not to do it. So (6e) does not entail (6a). Having said that, I think (6e) can capture the aiming feature of refraining. If Jerry wants to stand still in order not to save the baby, and Jerry thereby stands still, then his not saving the baby is straightforwardly one of his goals. We need to create a condition that links (6e) to one’s intention to B*: (7c) S’s want to B in order not to A is (part of) S’s reason for intending to B*. By conjoining (6e) and (7c) and REF, we arrive at a set of minimally sufficient conditions for refraining. Jerry wants to stand still in order not to save the baby. This results in Jerry forming the intention to stand still. Jerry intentionally stands still. And Jerry believes truly that standing still is incompatible with saving the baby. Thus, Jerry refrains from saving the baby. To the question of whether or not (6e) is a necessary condition on refraining, the answer is clearly ‘no’. Conditions (6e) and (7c), as noted, account for the aiming feature of refraining. But we’ve seen that conditions (6a) and (7a) do so as well. And if we add REF to the latter two conditions, we get refraining. We get refraining, that is, without (6e) and (7c). Hence, conditions (6e) and (7c) are not necessary for refraining. 93 93 So again let us temporarily bracket (6e) and (7c) and continue to examine other possible conditions. 97 Lastly, it appears that the following is a plausible condition on refraining: (6f) S wants to try not to A. Suppose that I am a smoker and I’ve just arrived at a party full of smokers. Indeed, there are so many smokers present that I am constantly being handed free cigarettes before I can refuse them. I consider smoking at the party and certainly can if I want. Because I’ve lately been trying to cut back, I want to try not to smoke at the party. Let us now introduce the following linking condition: (7d) S’s want to try not to A is (part of) S’s reason for intending to B*. My want to try not to smoke is the reason I form the intention to snap in half every cigarette that I’m handed. This intention leads me to (intentionally) snap in half the cigarettes I am handed. I know that snapping cigarettes is incompatible with smoking them; so, I do not smoke. Is my failure to smoke a goal of mine? Certainly it is: snapping the cigarettes issues from my desire to try not to smoke at the party. Thus, (6f), (7d) and REF constitute minimally sufficient conditions for refraining. 94 94 It’s fairly obvious that (6f) and (7d) are not necessary for refraining. So as with the other conditions above that, when conjoined to REF, deliver sufficiency, let us temporarily bracket these two conditions. These, and all the bracketed conditions, will resurface when we finalize REF below (section 7). And so we don’t have to keep inserting notes such as this, go ahead and automatically bracket any conditions below that, when paired with REF entail sufficiency, but are not themselves necessary. They will reappear in our final formulation of REF. 98 4 Intending In addition to wanting, intention is an obvious state to mine for further aiming conditions. Without committing to any particular view of intention, it is (normally) true that when one intends to do something, one has doing it as an aim or goal. This point naturally suggests the following condition: (6g) S intends not to A. Not surprisingly, (6g) gets the right results in the BABY cases. To see this, let’s first introduce the appropriate linking condition: (7e) S’s intention not to A is (part of S’s reason for) S’s intending to B*. Note that S’s intention not to A might simply be S’s intention to B*, but it doesn’t have to be. In BABY 1, one reason Jasper does not refrain from saving Baby L (‘A’ = ‘save Baby L’) is that he does not intend not to save Baby L. However, had Jasper intended this and had it resulted in his intentionally saving Baby R, then it would be correct to describe Jasper as refraining from saving Baby L. In BABY 2, Jerry intends (via deciding) not to save the baby (‘A’ = ‘save the baby’), and this results in his intentionally standing still. Thus, Jerry refrains from saving the baby. Furthermore, Jerry does not satisfy (6g) with respect to getting his hair wet (‘A’ = ‘get hair wet’): i.e., he does not 99 intend not to get his hair wet. And so the fact that Jerry does not refrain from getting his hair wet is consistent with (6g), (7e) and REF constituting sufficiency. So (6g), (7e) and REF plausibly constitute minimally sufficient conditions for refraining. Are (6g) and (7e) also necessary? Clearly they are not. We have already encountered other conditions above that, when paired with REF, also produce sufficiency. And it is unlikely that each of these collections of sufficient conditions require or entail conditions (6g) and (7e). For instance, we know that (6f), (7d) and REF’s (1)-(5) are sufficient for refraining. But it is difficult to imagine any of these conditions entailing (6g) and (7e). Let us next consider the following condition: (6h) S intends to B. It is unlikely that (6h) is a condition on refraining. This intention is present in BABY 1: Jasper intends to save Baby R. But we know that Jasper does not refrain from saving Baby L. So (6h) plus REF does not get us sufficiency. Is (6h) necessary for refraining? It is difficult to see why it would be. We take it that SPV is true: intentionally X-ing entails having an intention, though it need not be an intention to X. If (6h) were necessary, then it would have to be because intentional B-ing requires intending to B when the B-ing is part of a refraining from A-ing. I suppose it’s possible to tell a story about why refraining from A-ing entails intending to B, but I have no idea what the 100 details would look like. So for now we conclude that (6h) should not factor into an account of refraining. We can generate a third intending condition by combining the two above: (6i) S intends to B and not A. Condition (6i) is not necessary for refraining. But (6i) easily captures the aiming feature of refraining. Let us introduce a linking condition: (7f) S’s intention to B and not to A is (part of S’s reason for) S’s intending to B*. If Jerry intends to stand still and not save the baby, and he stands still in virtue of this intention, then his failure to save the baby is aimed at. So (6i), (7f) and REF constitute minimally sufficient conditions for refraining. Another possible intending condition is as follows: (6j) S intends to B in order not to A. Intending to B in order not to A plausibly entails intending not to A. Intending not to A, we know, is condition (6g). And since (6g) is not necessary for refraining, (6j) is plausibly unnecessary for refraining. But it is apparent that (6j), with the appropriate linking condition, can support the aiming feature of refraining: 101 (7g) S’s intention to B in order not to A is (part of S’s reason for) S’s intending to B*. Conditions (6j) and (7g) and REF constitute sufficient conditions for refraining. Lastly, the following pair of conditions are plausible constraints on refraining: (6k) S intends to try not to A. (7h) S’s intention to try not to A is (part of S’s reason for) S’s intending to B*. If one can aim at not A-ing by way of wanting to try not to A ((6f)), then certainly one can do the same by way of intending to try not to A ((6k)). So (6k), (7h) and REF constitute minimally sufficient conditions for refraining. But clearly one need not intend to try not to A in order to refrain from A-ing—one could simply intend not to A. Thus, (6k) and (7h) are unnecessary. 5 Preferring Preferring A neither entails wanting A nor intending A. But it is clear that preferences, like wants and intentions, can play a role in the production of aims and goals. Consider: (6l) S prefers not to A. And: 102 (7i) S’s preference not to A is (part of) S’s reason for intending to B*. When we attach (6l) and (7i) to REF, we get it that S’s not A-ing is plausibly one of S’s aims. Thus, (6l) and (7i) and REF constitute minimally sufficient conditions for refraining. In light of our discussion of wants and intentions above, we can, then, be confident that the following pairs of preferences and corresponding linking conditions also constitute sufficient conditions when combined with REF: (6m) S prefers to B and not to A. (7j) S’s preference to B and not to A is (part of) S’s reason for intending to B*. (6n) S prefers to B in order not to A. (7k) S’s preference to B in order not to A is (part of) S’s reason for intending to B*. (6o) S prefers to try not to A. (7l) S’s preference to try not to A is (part of) S’s reason for intending to B*. 6 Hoping, Deciding, Choosing A brief review: we started with an account of refraining, REF, which was incapable of handling cases involving non-intentional omission (the BABY cases). REF’s shortcoming was that it too liberally ascribed refrainings. What REF needed was a 103 further condition—condition (6)—that ensured that what a refrainer failed to do was one of her aims or goals at the time of acting. But there are multiple ways of satisfying (6). So we identified various wants, intentions and preferences that, when connected to REF in the right way, captured the aiming character of refraining. In this section, we turn to three kinds of mental states that involve either wanting or intending. Their respective entailments render them relevant to refraining. 6.1 Hoping Let us start with hopes. As with our discussion of wanting, intending, and intentional action above, I want to remain as neutral as I can about the nature of hoping (as well as deciding and choosing below). The reason for this is that I want REF to be broad enough to be compatible with various accounts of these items. I here make only the following assumption about hope: hoping to do something, A, involves wanting to do A. Once we make this assumption, we see that hoping can figure into refraining in all the ways wanting can. Here are four hoping conditions on refraining: (6a h ) S hopes not to A. (6c h ) S hopes to B and not to A. (6e h ) S hopes to B in order not to A. (6f h ) S hopes to try not to A. 104 Each hope condition entails a wanting condition from section 3: (6a h ) → (6a), (6c h ) → (6c), (6e h ) → (6e), (6f h ) → (6f). This generates four new sets of sufficient conditions for refraining. Let us state each of the four sets. First, if S hopes not to A, then S wants not to A. If S’s want not to A results in S’s intention to B*, and S satisfies REF’s (1)-(5), then S refrains from A-ing. In other words, (6a h ), (6a), (7a) and REF constitute (non-minimally) sufficient conditions on refraining. The remaining sets of sufficient conditions follow the same pattern as the first, so let’s state them more briefly. The second set is: (6c h ), (6c), (7b) and REF. The third set: (6e h ), (6e), (7c) and REF. And the fourth: (6f h ), (6f), (7d) and REF. 6.2 Deciding As I indicated in chapter 1, I take it there are two ways of coming to have an intention: actively and passively. 95 Passively acquiring an intention, i, is non-actional: one doesn’t (intentionally) do anything in coming to have i (i “happens” to the agent). Actively forming an intention, on the other hand, is actional. The active formation of an intention is what we call decision. Thus, deciding to do A entails intending to do A. With this assumption in place, we see that deciding can figure into refraining in all the ways intending can. Here are four decision conditions on refraining: 95 For an account of the distinction between active and passive intention, see Mele (1992), p. 141 and Mele (2003), ch. 9. 105 (6g d ) S decides not to A. (6i d ) S decides to B and not A. (6j d ) S decides to B in order not to A. (6k d ) S decides to try not to A. Each decision condition entails an intending condition from section 4: (6g d ) → (6g), (6i d ) → (6i), (6j d ) → (6j), (6k d ) → (6k). This generates four new sets of sufficient conditions for refraining. First, if S decides not to A, then S intends not to A. If S’s intention not to A results in S’s intention to B*, and S satisfies REF’s (1)-(5), then S refrains from A-ing. In other words, (6g d ), (6g), (7e) and REF constitute sufficient conditions for refraining. The remaining three sets follow the same pattern as the first, so let’s state them more briefly. The second set is: (6i d ), (6i), (7f) and REF. The third set: (6j d ), (6j), (7g) and REF. And the fourth: (6k d ), (6k), (7h) and REF. 6.3 Choosing What about the state of choosing? It might well be that choosing just is deciding. At any rate, I take it that choosing to do something, A, entails intending to do A. This assumption prompts the following four choosing conditions: (6g c ) S chooses not to A. (6i c ) S chooses to B and not to A. 106 (6j c ) S chooses to B in order not to A. (6k c ) S chooses to try not to A. Of course, if one is inclined to identify choosing with deciding, then these four conditions just are the four deciding conditions in 6.2. But if we don’t identify choosing with deciding, then these four conditions generate four new sets of sufficient conditions for refraining. The first set is: (6g c ), (6g), (7e) and REF. The second set: (6i c ), (6i), (7f) and REF. The third set: (6j c ), (6j), (7g) and REF. And the fourth: (6k c ), (6k), (7h) and REF. 7 Reformulating REF We are now in a position to formulate the new and improved version of REF. You’ll recall that originally REF contained five conditions. We saw that cases involving non- intentional omission necessitated a sixth condition: (6) S’s not A-ing is S’s aim or goal. Condition (6) guarantees that REF does not mistake non-intentional omissions for intentional ones. But, (6) can be satisfied in different ways. The bulk of this chapter has been spent trying to identify some of the different ways we might satisfy (6). 107 We can reformulate REF as follows: (REF*) S refrains from A-ing iff (1) S considers A-ing, (2) S has the ability to A, (3.1) S intends to B*, (3.2) S intentionally Bs in virtue of S’s intention to B*, (4) S lacks the belief that B-ing is compatible with A-ing, (5) B-ing is incompatible with A-ing, and (6.1) S wants not to A, or (6.2) S wants to B in order not to A, or (6.3) S wants to try not to A, or (6.4) S intends not to A, or (6.5) S intends to B and not to A, or (6.6) S intends to try not to A, or (6.7) S prefers not to A, or (6.8) S prefers to B in order not to A, or (6.9) S prefers to try not to A, and (7.1) S’s want not to A is (part of) S’s reason for intending to B*, or (7.2) S’s want to B in order not to A is (part of) S’s reason for intending to B*, or (7.3) S’s want to try not to A is (part of) S’s reason for intending to B*, or (7.4) S’s intention not to A is (part of S’s reason for) S’s intending to B*, or (7.5) S’s intention to B and not to A is (part of S’s reason for) S’s intending to B*, or (7.6) S’s intention to try not to A is (part of S’s reason for) S’s intending to B*, or (7.7) S’s preference not to A is (part of) S’s reason for intending to B*, or (7.8) S’s preference to B in order not to A is (part of) S’s reason for intending to B*, or (7.9) S’s preference to try not to A is (part of) S’s reason for intending to B*. As formulated, REF* contains nine sets of minimally sufficient conditions for refraining. This means that REF* does not exhaust all the different ways an agent might refrain from 108 doing something. As we’ve seen, one might also refrain via hoping, deciding, or choosing. I’ve decided to list only minimally sufficient sets in REF* because including more would increase complexity unnecessarily. I do not suspect that REF* captures all possible sets of minimally sufficient conditions for refraining. I wouldn’t be surprised if, for instance, I’ve failed to identify further wanting, intending, and preferring conditions (and their appropriate linking conditions). I mentioned earlier that REF* does not assume any particular account of wanting, intending, preferring, etc.; as a result, it is compatible with various accounts of these items. With this in mind, let us note, then, that some might embrace REF* yet disagree over which sets of conditions are minimally sufficient. For example, some hold that intending entails wanting (desiring). 96 Under such a view, REF* might (other things equal) be thought to contain only six sets of minimally sufficient conditions (namely, the wanting and preferring sets). The intention sets would, on this view, be more like the hoping sets of non-minimally sufficient conditions. Indeed, with some re-working, REF* could even be rendered consistent with DPV (the view that X-ing intentionally does not entail having an intention). The point is that REF* can be tweaked in different ways depending on one’s respective views about various mental and behavioral phenomena. Having formulated REF*, an important point emerges. 97 Conditions (6.1)-(6.9) appear to entail the following preference condition: 96 E.g., Audi (1973) and Davis (1984). 97 Gideon Yaffe brought this to my attention. 109 (6') S prefers not A-ing to A-ing. Likewise, conditions (7.1)-(7.9) appear to entail the following linking condition: (7') S’s preferring not A-ing to A-ing is (part of) S’s reason for intending to B*. We can then combine (6') and (7') into a single condition that captures the aiming feature of refraining: (6*) S prefers not A-ing to A-ing, and this preference is (part of) S’s reason for B-ing (or intending to B*). 98 REF*’s (1)-(5) plus (6*), therefore, constitute necessary and sufficient conditions for refraining from acting. 8 Three Objections to REF* Under REF*, refraining incorporates both a positive and a negative component: it is doing something while failing to do something else. The refrainer aims at both the doing and the failing. The doing component grounds the actional nature of refraining; while the aiming feature is what separates refrainings from (mere) non-intentional omissions. This approach to refraining is subject to several objections. 98 Or alternatively: S has a reason for not A-ing, and this reason is (part of) why S Bs (or intends to B*). 110 8.1 The Linguistic Concern The first objection we’ll call ‘The Linguistic Concern’. It goes like this. It might be suggested that claims of refraining only make reference to what is not done. We say things like ‘Allison refrained from going to John’s party’ and ‘Leonard refrained from smoking a cigarette’. But that our refraining claims only mention failure is a good indication that refraining does not involve a positive component as REF* suggests. Thus, REF* is implausible. This objection is not particularly forceful. Proponents of REF* will deny the premise that all refraining claims only mention what one fails to do. Certainly it is true that some of our refraining claims only mention failure. However, we sometimes mention the positive aspect of refraining—that which is done—as well. When Allison’s friend asks if she went to John’s party last night, Allison might respond, ‘No I didn’t, I went to see a movie instead’. This response, on our view, is (or at least can be) a claim of refraining. Statements such as this explicitly involve both positive and negative behavior. So, why is it that some of our refraining claims only mention the negative? I suspect that sometimes what is done as part of a refraining, r, an agent might have reason not to mention (or at least might not have reason to mention) when expressing r. If John’s friend asks Allison if she went to John’s party, it might be appropriate for her only to mention that she didn’t. After all, she might not know John’s friend well enough to feel comfortable mentioning what she did instead, or she might not want John to know what she did instead. Thus, the positive aspect of her refraining (quite reasonably) goes unstated. 111 8.2 Bach’s Objection Kent Bach has presented an objection to actional views of refraining. 99 Suppose Mary has a pint of ice cream in her freezer. In order not to eat it, she might throw it out. But Bach thinks she doesn’t have to do this, or anything else, to refrain from eating the ice cream. Mary just has to decide not to eat it—and not eat it because of her decision—to count as refraining. Bach claims that it’s Mary’s decision, rather than any action of hers, that keeps her from eating it. Refraining, therefore, does not require doing anything beyond making a decision. Hence, actional views like REF* are false. Bach claims that Mary’s decision not to eat the ice cream is what keeps her from eating it; thus, refraining only requires decision. What actionalists will say here is that, on the assumption that Mary refrains, Mary’s decision not to eat the ice cream results in her performing some action such as standing still or turning and walking out of the kitchen. Bach will be unsatisfied by this response on grounds that it’s question-begging. But his objection begs the question no less than our response. On our view, if one refrains by deciding not to do something, then one’s decision contributes to one’s refraining—it sets one down the path to refraining. But the decision all by itself is not what keeps one from doing what ultimately is not done. It’s the decision not to A plus the intentional B-ing that (intentionally) keeps one from A-ing. Bach’s example fails to mention what it is Mary does instead of eating the ice cream. But failing to mention this does not in any clear sense point up a defect in our view. 99 Bach (2010). 112 Furthermore, let me remind readers of the following point. It seems intuitive to describe Mary as intentionally refraining from eating the ice cream (however we are to ultimately conceive of refraining). And again, behavior that is intentional is actional. Bach, however, cannot capture this intuition because he denies that refrainings are actions; they are, for him, inactions. Suppose Mary refrains from eating the ice cream because she is on a strict diet. Intuitively, her refraining is intentional. Bach’s objection commits him to rejecting this intuition. 8.3 Weakness of Will A third worry about REF* comes from cases involving weakness of will. Suppose that Jeff has a slew of pro-attitudes towards working: he wants to get work done, he intends to work, he believes that it is best that he work, he prefers working to not working, and so on. Furthermore, Jeff considers working and could work. But instead of working, Jeff intentionally goes to the beach. It might be thought that Jeff has refrained from working. If he has, then REF* is false on account of condition (6*)—the aiming condition. Given that it is not Jeff’s aim or goal not to work, proponents of REF* should say that cases like Jeff’s do not involve refraining: Jeff’s failure to work is not intentional. It’s plausibly not unintentional either since Jeff considers working and knows that going to the beach is incompatible with working. Jeff’s failure must therefore be non- intentional. Perhaps what motivates the intuition (if there is such an intuition) that Jeff’s failure is intentional is an urge to hold Jeff accountable for his failure. That is, the intuition that Jeff refrains might stem from an inclination to want to criticize Jeff for not 113 working despite his (pro-)attitudes toward working. But certainly one can be held accountable for behavior that is not intentional. A drunk driver might hit and kill a pedestrian unintentionally, yet still be held (morally and legally) accountable. 100 If the drunk driver can be subject to blame, then Jeff can be as well. 9 Conclusion Refrainings are intuitively intentional. Behavior that is intentional must be actional. Hence, our preference for an actional account of refraining. I’ve attempted here to develop such an account in the form of REF*. I don’t suppose that REF* is complete by any measure—it’s a work in progress. 101 That said, REF* points a way to understanding how to conceive of refrainings as bona fide actions. 102 100 I’ve stolen this example from Mele and Sverdlik (1996), pp. 273-74. 101 In the Conclusion I consider a further possible condition on REF*. 102 Thanks to Gideon Yaffe and Gary Watson for extremely helpful comments. A version of this chapter was presented at the 65th Annual Mountain-Plains Philosophy Conference and the 2012 APA Central Division Meeting. I am indebted to Annaleigh Curtis and Michelle Mason for their instructive comments, and to both audiences for their valuable feedback. 114 CHAPTER 3 REFERENCES Audi, Robert (1973), “Intending,” Journal of Philosophy 70, pp. 387-402. Bach, Kent (2010), “Refraining, Omitting, and Negative Acts,” in A Companion to the Philosophy of Action, eds. Timothy O’Connor and Constantine Sandis (Oxford: Wiley-Blackwell). Brand, Myles (1971), “The Language of Not Doing,” American Philosophical Quarterly 8 (1), pp. 45-53. Bratman, Michael (1984), “Two Faces of Intention,” The Philosophical Review 93 (3), pp. 375-405. Davis, Wayne (1984) Davis, Wayne (1984), “A Causal Theory of Intending,” American Philosophical Quarterly 21, pp. 43-54. Harman, Gilbert (1976), “Practical Reasoning,” Review of Metaphysics 79, pp. 431-63. Johns, Brandon (2009), “Refraining and the External,” Ratio 22 (2), pp. 206-15. Mele, Alfred (2003), Motivation and Agency (Oxford: Oxford University Press). Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). Mele, Alfred and Steven Sverdlik (1996), “Intention, Intentional Action, and Moral Responsibility,” Philosophical Studies 82 (3), pp. 265-87. Yaffe, Gideon (2001), “Locke on Refraining, Suspending, and the Freedom to Will,” History of Philosophy Quarterly 18 (4), pp. 373-91. 115 CHAPTER 4 NON-INTENTIONAL ACTION The existence of intentional and unintentional action is relatively uncontroversial. The question of whether there exists a third type of action is the subject of this chapter. In chapter 3, I forwarded an account of refraining whereby refraining is a kind of intentional action. My account of refraining relied, among other things, on the existence of a third type of action—what is called non-intentional action. I am, therefore, committed to a third category of action. In this chapter I will present a pair of arguments for the existence of non- intentional action. This chapter will contain two parts. The first part is negative in nature. In sections 1-4, I will consider four arguments that some have thought establish the existence of non-intentional action. I will argue in these sections that none of the four arguments succeed. In the second part of the chapter (sections 5 and 6), I will develop a pair of arguments for the existence of non-intentional action. 1 Gorr and Horgan’s Argument The most common argumentative strategy deployed by proponents of non-intentional action involves two steps: first, provide conditions for intentional and unintentional 116 action; then identify a particular action (or action type) that fails to satisfy both sets of conditions. When such an action is located, it gets classified as non-intentional. This is precisely how Michael Gorr and Terence Horgan attempt to establish the existence of non-intentional action. In their essay “Intentional and Unintentional Action,” Gorr and Horgan argue that knowledge is the central feature of intentionality. 103 They provide the following account of intentionality: (I) S’s A-ing at t is intentional under the description ‘A-ing’ iff (i) This event is an act, and (ii) S knows, at t, of this act, that it is an A-ing by him. 104 And they offer the following account of unintentionality: (UI) S’s A-ing at t is unintentional under the description ‘A-ing’ iff (i) This event is an act, and (ii) S does not believe, at t, of this act, either that it is an A-ing by him or that there is a significant chance that it is an A-ing by him. 105 The question now is: does all action fall under I and UI? 103 Gorr and Horgan (1982). 104 Gorr and Horgan (1982), p. 255. It might seem that Gorr and Horgan’s account of intentionality ignores an essential feature of intentional action: the state of intention. If all that is required of an action A to be intentional is that one know that one is A-ing, then it may appear as though one can act intentionally without acting from any intention whatsoever. I think Gorr and Horgan can escape this worry. Gorr and Horgan are providing an account of what it is for an action to be intentional; they are not here providing an account of what it is for conduct to be action. It remains open to Gorr and Horgan, then, to conceive of action as behavior that springs from intention. On this view, behavior that is produced by intention and known about (under a description) is intentional action (under that description); behavior that is produced by intention but not believed (under a description) is unintentional action (under that description). By building intention into action in this way, action is a kind of behavior: all action is behavior, but only some behavior is action. 105 Gorr and Horgan (1982), p. 255. 117 The asymmetry of I and UI lead Gorr and Horgan to conclude that some action cannot be classified as either intentional or unintentional. According to Gorr and Horgan, action that is intentional under a description requires knowledge that one is acting under that description. Unintentional action, on the other hand, not only requires not knowing that one is acting, but not believing that there is a significant chance that one is so acting. But Gorr and Horgan note that it is possible to perform an act A at t, believe at t that there is a significant chance that one is A-ing, and yet fail to know at t that one is A-ing. Consider the following example. Alex wants to assassinate the mayor by rigging his office with dynamite and detonating it from a safe distance. One afternoon while the mayor is out, Alex rigs the mayor’s office with dynamite then flees down the block. He waits several hours to detonate the dynamite. Now suppose that the mayor is in his office 50% of the time and Alex is aware of this fact, but that Alex does not have any beliefs about the mayor’s schedule. Furthermore, Alex is fully rational. As Alex presses the button to detonate the dynamite, he does not believe—and thus does not know—that he is killing the mayor. If it turns out that the mayor is in his office at the time of detonation, then under I, Alex’s action of killing the mayor is not intentional. And given that Alex believes there is a significant chance that he is killing the mayor as he detonates the dynamite, Alex’s killing the mayor is not unintentional either. Therefore, Gorr and Horgan conclude that Alex’s killing the mayor is non-intentional. The problem with Gorr and Horgan’s argument lies in their account of intentionality. As Davidson has pointed out, a knowledge condition on intentional action is too strong. In his essay “Intending,” Davidson presents his famous ‘carbon copy’ 118 example. 106 Mary tries to produce ten carbon copies. Producing ten copies is difficult, and she is not confident that she will succeed. But if she succeeds because she tries, then her success is intentional: she intentionally produces ten carbon copies. And she does this without believing—and thus without knowing—that she is succeeding. So knowledge must not be a necessary condition on intentionality. Now it might be objected that Mary’s success can be deemed lucky; and since lucky action is presumably not intentional, this example fails to harm Gorr and Horgan’s account of intentionality. But we can skirt this objection by slightly modifying Davidson’s example. Suppose that Mary is quite skilled at producing carbon copies; in fact, Mary is one of a handful of people in the world that can produce ten carbon copies at will. Mary succeeds at producing ten copies 97% of the time. What is peculiar about Mary is that she tends to lack self-confidence. Because of her lack of self-confidence, when she attempts to produce ten copies, she does not believe that she will succeed. But if she succeeds in the way she tries and in a way that depends on her copy-making skills, her success is straightforwardly intentional and not lucky. Gorr and Horgan’s account of intentionality is too demanding: knowledge is not a requirement for acting intentionally. This fact undermines Gorr and Horgan’s reason for thinking that cases like Alex’s are examples of non-intentionality. How ought we to classify Alex’s act of killing the mayor? Well, a definitive answer to this question will depend (among other things) on how we conceive of luck, and whether we think Alex’s killing the mayor is lucky. For my part, I suspect that Alex’s killing the mayor is 106 Davidson (1978), p. 92. 119 intentional rather than lucky. Alex tries to kill the mayor and succeeds in just the way he tries and in a way that appears to depend on his relevant skills. 2 The Simple View Argument Intentional action plausibly involves the state of intention. 107 A second argument for non-intentional action can be formulated using a principle that links intentional action and intention—this principle is known as the Simple View (SV). Recall that SV is the principle that intentionally performing an action A requires intending to do A. 108 SV posits the tightest possible connection between intending and intentional action. SV possesses considerable intuitive appeal because of its application in normal cases of intentional conduct. When we intentionally raise our arms, we usually intend to raise our arms; and when we intentionally drink coffee, we usually intend to drink coffee. The first premise in The Simple View Argument, then, is SV. It might then be pointed out that in the process of executing intentions, agents sometimes bring about events that are not intended. Often these consequences are unforeseen: e.g., in executing my intention to raise my arm, I might, to my great surprise, strike a friend who happens to be walking behind me. But unintended consequences can also be foreseen or anticipated. Consider again Harman’s famous ‘sniper’ example 109 : a sniper reasons that killing a high-ranking soldier is worth the danger involved in alerting the enemy to his 107 I argued for this claim in ch. 2. 108 Proponents of SV include Adams (1986) and McCann (1991). In a recent paper Wasserman (forthcoming) defends SV from several objections, but doesn’t explicitly endorse the principle. 109 Harman (1976), p. 433. 120 presence. So, he fires his rifle at the soldier; in so doing, he alerts the enemy to his presence. By SV, proponents of The Simple View Argument hold that the sniper’s alerting the enemy to his presence is not intentional: if it were, it would be intended, which it’s not. But because alerting the enemy is foreseen—and so perhaps not an accident or a mistake when it occurs—it is not unintentional. If an action is neither intentional nor unintentional, it is concluded that it must be non-intentional. Hence, the existence of non-intentional action. While I agree that Harman’s ‘sniper’ scenario is an instance of non-intentional action, I am skeptical that The Simple View Argument establishes this. The clearest defect with the argument is that it relies on SV. As we have seen in chapter 1 (section 2.1), SV is vulnerable to Bratman’s ‘video game’ objection. SV also appears subject to a counterexample of the following sort. It seems that one can intend to try to do something and do it because of this very intention. When one succeeds, one’s success appears straightforwardly to be intentional. Pam wants to move a log. She believes that she will probably not be able to move it because it appears heavy. She forms the intention to try to move the log. If Pam’s attempt to move the log is successful, then (barring luck and causal deviance) her moving it is intentional. Under SV, however, Pam’s success cannot be intentional. Thus, SV is implausible. Now it might be objected here that intending to try to A entails intending to do A 110 ; and so the example above is not a counterexample to SV because Pam intends to move the log if she intends to try to move it. But this entailment is implausible. 110 McCann (1991) and Pears (1985), e.g., hold this view. Bratman (1987), Davis (1984), Mele (1992) and Velleman (1989) oppose it. 121 Certainly it is appropriate for Pam to intend to try to move the log given that she believes that she will probably fail. If the entailment claim above were true, then Pam would also intend to move the log. But attributing this latter intention to Pam is problematic. Presumably, intentions promote intra- and inter-personal planning. If Pam were to intend to move the log, then it would be appropriate for her to make certain further plans based on her intention to move it. She might, for example, plan to collect a soil sample from under the log and to examine the sample in her lab an hour from now, and so on. But, other things being equal, forming these latter plans would be inappropriate given that Pam believes that she probably will not move the log. Given Pam’s circumstances, it is more appropriate for her to form conditional plans: to collect a soil sample and examine it at her lab if she can move the log, etc. Conditional plans like these better coincide with an intention (plan) on Pam’s part (only) to try to move the log. Thus, it is implausible that intending to try entails intending to succeed. 3 Chan’s Argument In his essay “Non-Intentional Actions,” David Chan presents an argument for non- intentional action. 111 Chan argues that intentional action essentially involves acting for a reason. More specifically, he defines intentional action as follows: (I1) If an action is the bringing about of events appropriately caused by the agent’s reason for so acting, the action is intentional. 112 111 Chan (1995). 112 Chan (1995), p. 140. 122 But what is it to perform an act for a reason? Chan says that in order for a reason R to be an agent’s reason for A-ing, R must enter into a process of practical reasoning. The outcome of this process, he claims, is an intention to perform A. Since Chan thinks that intentions are only ever formed or acquired as conclusions of practical reasoning, he goes on to offer a second definition of intentional action: (I2) If the events brought about in acting are caused in the appropriate way by the agent’s intention to so act, the action is intentional. 113 The antecedents of I1 and I2 are interchangeable since acting for a reason necessarily involves acting on an intention and vice versa. Chan proceeds to define unintentional action as: (U) An unintentional action is performed if an intention that an agent acts upon causes an unintended effect and the bringing about of the latter effect counts as an action. 114 Since unintentional action, like intentional action, involves acting on an intention, it follows on Chan’s view that both types of action involve acting for a reason. The question now is: are there any actions that are not performed for a reason? If there are, such actions would constitute non-intentional action. 113 Chan (1995), p. 140. I1 and I2 are stated as sufficiency conditions on intentional action; but I take it Chan holds that I1 and I2 are necessary and sufficient. 114 Chan (1995), p. 141. 123 Chan presents two examples of action he believes to be non-intentional. The first example belongs to the class of behavior called “mannerisms”: A public speaker may have acquired the mannerism of tugging at his ear- lobe once or twice as he is giving a speech. And he may do this fully conscious of the quirky action he is performing. So, on a particular occasion, he knowingly raises his arm in order to tug at his ear. But afterward, he tells us quite sincerely that he was wholly indifferent to whether he tugged at the ear. It is just something that he now and then does. 115 The second example involves habitual action or routine. Some of our habits may have initially been formed on the bases of reasons, but are no longer performed for these reasons. Indeed, they may no longer be performed for any reason at all. One may have started sprinkling salt on her food years ago because she prefers the flavor, but she has since abandoned this preference; she continues to salt her food only out of pure habit. Chan concludes that mannerisms and certain habits are not performed for reasons; they are, therefore, non-intentional. One serious difficulty with Chan’s argument is that it relies on SV. Recall that Chan claims that when one intentionally performs an act A, one’s A-ing must be for a reason. In order for a reason to count as one’s reason for A-ing, it must factor into one’s practical deliberation. The outcome of this deliberation is an intention to do A. So A-ing 115 Chan (1995), p. 141. 124 intentionally necessarily involves intending to do A. 116 This is SV. But we know that SV is implausible (see section 2). Chan’s argument is no more plausible than SV. A second worry is that Chan’s argument relies on the claim that intentions are only ever formed as conclusions of practical reasoning. It seems to me that we can acquire intentions “spontaneously” or without engaging in practical reasoning. Sometimes when I’m at the market I suddenly acquire an intention to touch an item on the shelf as I slowly walk down an aisle. When I touch the item, I do so intentionally. My intention to touch the item, however, wasn’t the result of deliberation: it just “popped up.” Al Mele’s distinction between active and passive intention acquisition is helpful here. 117 Mele points out that some intentions are the result of deliberation; these intentions are actively formed. Some intentions, on the other hand, are plausibly passively acquired. Suppose a dog shoots out in front of my car and I slam on my brakes. My slamming on the brakes is plausibly intentional. On the supposition that intentional action requires intention, I must intend something such that it renders my behavior intentional. Since I don’t have time to deliberate and thereby form an intention, it must be that my intention is (passively) acquired. Since Chan holds that all intentions are actively formed, he must reject Mele’s plausible distinction. One last complaint involves Chan’s account of unintentionality (U). Under U, if in the course of acting on an intention one brings about an unintended effect E (where E 116 Further proof that Chan accepts SV lies in his following remark on p. 145: “[Bratman] argues, contrary to my view, that one can act intentionally without having the intention to so act” (1995). 117 Mele (1992), p. 184. This distinction was discussed briefly in ch. 1, section 2.1, and again in ch. 3, section 6.2. 125 is an action), it follows that E is unintentional. But this account delivers counterintuitive results in cases like Harman’s ‘sniper’. Recall that in ‘sniper’, a sniper intentionally shoots the soldier; in so acting, he knowingly alerts the enemy to his presence. Given that the sniper does not intend to alert the enemy to his presence, it follows on U that the sniper unintentionally alerts the enemy. But intuitively this seems wrong: the sniper’s alerting the enemy does not seem accidental or inadvertent. If unintentional action is, broadly speaking, action that is accidentally performed, then there is no obvious sense in which the sniper’s alerting the enemy can be classified as unintentional. 118 It seems Chan’s U is too liberal. 4 Meiland’s Argument J. W. Meiland creatively argues for the existence of non-intentional behavior. 119 Meiland begins by pointing out that actions have results. Consider the act of buying a loaf of bread. In buying a loaf of bread, one result is that (a) the loaf of bread is in the buyer’s possession (or the possession of the person the buyer is buying it for). But this act could also have the following results: (b) the depletion of the baker’s supply of bread; (c) the feeding of a hungry child; (d) the buyer running out of money; etc. The difference between (a) and (b)-(d), according to Meiland, is that the former must be intended if one is to intentionally perform the act of buying a loaf of bread. Results (b)-(d), on the other hand, need not be intended in order for one to intentionally buy bread. Results that must 118 More on the sniper and unintentionality in section 5. 119 Meiland (1963). Meiland does not have a term for behavior we’re calling ‘non-intentional’; he describes it as behavior in which the terms ‘intentional’ and ‘unintentional’ do not apply. 126 be intended in order for one to perform an act intentionally Meiland calls intrinsic results; results that do not have to be intended in order to perform an act intentionally are called extrinsic results. Put differently: if an agent performs an act A, the intrinsic result of A necessarily is one of the agent’s ends; whereas an extrinsic result of A may or may not be one of the agent’s ends. So one intentionally A-s only if one intends A’s intrinsic result. With this view of intentionality, it is natural to expect that unintentional action will be a matter of acting while failing to intend the act’s intrinsic result. But this is not how Meiland understands unintentionality. According to Meiland, if one fails to intend the intrinsic result of an act A, then one cannot perform A at all—A-ing is, to use Meiland’s expression, “non- existent.” In short, all action is intentional; there are no unintentional actions. So what of non-intentional behavior? Meiland claims that behavior that lacks intrinsic results can neither be classified as intentional nor unintentional. He cites running as an example of such behavior. As Meiland puts it: If one is performing certain movements, then one is running, and no reference is made to any intrinsic result. One need not traverse any distance in order to be running, since one is running when one is “running in place”….And both “He ran intentionally” and “He ran unintentionally” represent improper usage. If someone uttered these expressions, we should not know what he was trying to say. 120 If running is neither intentional nor unintentional behavior, the conclusion is that it is non-intentional. 120 Meiland (1963), p. 381. 127 I think there is a deep problem with Meiland’s argument. To begin, it seems obvious to me that running behavior can be intentional. Consider the following example: Carl is a world-class sprinter. As he approaches the starting-line, Carl forms the intention to run when the gun sounds. The gun sounds and Carl executes his intention. Carl’s running is straightforwardly intentional. And if one were to remark that Carl intentionally ran, I think most competent speakers would take the claim to imply that Carl ran on purpose, that his running was under his control and was produced by an intention, that he probably had certain pro-attitudes (besides intention) toward running, and so on. Contrary to Meiland’s claim, there’s nothing particularly puzzling about such a remark. 121 The problem is that Meiland maintains that running can never be performed intentionally. If it could, then running would have an intrinsic result; but this is precisely what Meiland denies in the passage quoted above. So perhaps Meiland simply selected a poor example of behavior that supposedly lacks an intrinsic result, and that all we need is a better example of such behavior and Meiland’s argument is back up and running. But if Meiland grants that running can be performed intentionally, then what behavior will count as lacking intrinsic results? Perhaps walking or moving an arm? But like running, it’s fairly obvious that one can walk or move an arm intentionally. 121 I also suspect that running can be performed unintentionally. Suppose Cal is a race walker. As he nears the finish-line in a race, he notices that he must speed up or else risk not placing. Cal is fiercely competitive so he speeds up. Cal’s speeding up accidentally results in his running; he is thus disqualified from the race. All else being equal, it seems to me that Cal’s running was unintentional. 128 What is perhaps becoming apparent is that all physical behavior can be performed intentionally. But what about mental behavior? Perhaps thoughts lack intrinsic results. 122 The problem is that if thoughts can count as behavior, then, like the physical behavior discussed above, thoughts can be formed intentionally. I might intend now to think about some philosophical puzzle tonight. When the evening arrives, I execute my intention and begin thinking about the puzzle. If my thinking here is behavioral at all, then it is plausibly intentional. And as in the case of running, it would make sense to most competent speakers if it were claimed that my thinking about the puzzle was intentional. Hence, it appears difficult to locate behavior—physical or mental—that lacks intrinsic results. Meiland’s argument falls short of establishing the existence of non-intentional behavior. If this conclusion were the worst of it, perhaps Meiland wouldn’t be too bad off. But there is a more serious problem: Meiland’s account of intentionality is extremely implausible. On Meiland’s view, one must intend an act’s intrinsic result in order to perform that act intentionally. But why is this true? We have already seen that one can plausibly intend to try to do something and thereby try to do it; and if the individual (non- luckily) succeeds in the way she tried, then her success is plausibly characterized as intentional (see section 2). I might intend to try to buy a loaf of bread, but I don’t intend to have bread in my possession because I suspect the bakery might be out of bread for the day. I place my money on the counter and ask the baker for a loaf of bread. The baker 122 Hugh McCann (1974) famously argues that thoughts lack results and are, therefore, basic actions. McCann’s notion of a result, however, is not identical to Meiland’s notion of an intrinsic result. 129 comes back with the very last loaf and hands it to me. I intentionally buy the bread without intending that the bread be in my possession. Meiland’s account of intentionality is no more plausible than SV. In fact, it’s less plausible. At least it is true under SV that one can intend an act and thereby perform that act intentionally. On Meiland’s view, this apparently is not the case. If I buy a loaf of bread because I intend to buy a loaf of bread, I have not intentionally bought a loaf of bread according to Meiland. This is because I did not intend that the loaf of bread be in my possession. But surely one can intentionally A by intending A. Any view that is inconsistent with this claim is almost certainly false. Now perhaps Meiland would respond by claiming that intending to do A necessarily entails intending A’s intrinsic result. But it’s difficult to imagine why this would be true. Meiland’s argument for non- intentionality is unpersuasive. 5 An Argument from Side-Effect Action Now let us switch gears. In this section I will argue that side-effect acts are non- intentional. I will attempt to improve upon an argument initially developed by Al Mele and Steven Sverdlik in their essay “Intention, Intentional Action, and Moral Responsibility.” 123 Though the argument developed here will not be conclusive, I hope to persuade readers that it is defensible. Harman’s ‘sniper’ scenario will be helpful here: the sniper knows that if he fires at the soldier, he will alert the enemy to his presence. Still, he thinks that killing the 123 Mele and Sverdlik (1996). 130 soldier is worth the risk. So, he opens fire; in so doing, he alerts the enemy to his presence. The sniper’s firing on the soldier is straightforwardly intentional. What has been less clear to philosophers is the status of the sniper’s side-effect act of alerting the enemy. Harman and Bratman, for instance, hold that the sniper’s alerting the enemy is intentional. Chan, on the other hand, is committed to classifying the sniper’s act as unintentional in virtue of his respective account of unintentionality. Mele and Sverdlik and Wasserman have argued that both camps are mistaken: the sniper non-intentionally alerts the enemy. 5.1 Alerting the Enemy is Not Intentional Mele and Sverdlik begin by pointing out that intentional action is, broadly speaking, action that an agent aims at performing either as an end or a means to an end. But, there is no plausible sense in which the sniper aims at alerting the enemy. Notice that the sniper does not try or strive to alert the enemy; nor does he intend, want, or prefer to alert the enemy and thereby so act. And so Mele and Sverdlik rightly conclude that the sniper does not intentionally alert the enemy. 124 5.2 Alerting the Enemy is Not Unintentional Why isn’t alerting the enemy unintentional? Mele and Sverdlik claim that it is not unintentional on grounds that the sniper knowingly alerts the enemy. The idea here seems to be that if an agent intentionally performs an act A, and A-ing leads to B-ing, and the 124 I’m aware that this piece of the argument is hasty; we’ll return to it in section 5.3 below. 131 agent knows at the time of A-ing that she will thereby B, then the agent’s B-ing is not accidental or unintentional. Here I part company with Mele and Sverdlik. Though I agree that the sniper does not unintentionally alert the enemy, it is not because he knows or is aware that he will alert the enemy when he fires. Knowing that one will (or is currently performing) B, I contend, is compatible with one’s B-ing unintentionally. Consider this: the sniper decides to shoot the soldier and pulls the trigger as a result of his decision. While in the process of pulling the trigger, the sniper becomes aware of a child standing near the soldier. The sniper is an expert marksman, and so the child is not in danger of being physically harmed. But the sniper realizes that by shooting the soldier he (probably) will scare the child. The sniper does not desire to scare the child. Because he is in the act of pulling the trigger, he does not have time to re-think or re-consider his decision so to act. As soon as he realizes that he (probably) will scare the child, he completes his act of pulling the trigger. The sniper intentionally fires on the soldier. In so doing, he scares the child. It seems that his scaring the child is most naturally construed as accidental and unintentional. This, despite the fact the sniper knows that he (probably) will scare the child by shooting the soldier. 125 Thus, one can expect to do something and yet do it unintentionally. 125 One might wonder whether the sniper could know that he (probably) will scare the child given that it only takes a fraction of a second to pull a trigger. In such a brief amount of time it might be thought that the sniper cannot possibly process the relevant information; and so he can’t be said to know that he (probably) will scare the child. This concern shouldn’t bother us. If one is worried about the time it normally takes to pull a trigger, then imagine that the sniper has an abnormal trigger on his rifle. His trigger takes a little longer to fire than normal triggers. How long? However long it takes for the example above to be realistic. Of course, we don’t even need the example of the sniper to make our point about the compatibility of knowledge with unintentional action. All we need is for it to be possible that while in the process of performing an act A, an agent comes to realize or expect that she will perform B, but does not have time to re-think or re-consider her A-ing. In such cases, I contend the agent’s B-ing is accidental and thus unintentional. 132 So I disagree with Mele and Sverdlik that knowledge or expectation is what renders the sniper’s alerting the enemy non-intentional rather than unintentional. In order to see precisely why the sniper’s alerting the enemy is not unintentional, we need to develop an account of unintentional action. To generate an account of unintentionality, let us begin by examining paradigmatic cases of unintentional action. We can then try to derive necessary and sufficient conditions from these cases. Consider the following three cases: (ARM-RAISE) A teacher poses a question to the class. Shawn, a student, raises his arm in order to answer the question (A). While raising his arm, he happens to strike a classmate walking behind him (B). Shawn had no intention or desire to strike his classmate. Shawn’s striking his classmate is unintentional. 126 (DRIVE) Sally is driving to work. She realizes that her off-ramp is rapidly approaching, so she quickly decides to turn right. Sally makes a right turn (A). In so doing, she cuts off a driver to her right (B). Sally had no intention or desire to cut off anyone. Sally’s cutting off the driver is unintentional. (MISSPELL) Sue is typing an essay (A). She begins to type the word ‘theory’, but she misspells it (B). Her misspelling the word is unintentional. In each case, I take it to be obvious that the agent performs action B unintentionally. Now let us examine pertinent facts about each case in order to identify conditions on unintentional action. 126 The variables ‘A’ and ‘B’ in each example mark distinct actions. 133 First, in each case the agent is acting on an intention. 127 Of course, exactly which intentions each agent acts upon is unclear given that the scenarios are under-described. But, minimally, we know that each agent intends to do A: Shawn intends to raise his arm, Sally intends to turn right, and Sue intends to type her essay. And so we know that each agent is intentionally A-ing. Moreover, each agent’s unintentional action occurs in the course of, or as a result of, his or her A-ing: Shawn strikes his classmate as a result of raising his arm, Sally cuts off the driver while turning right, and Sue misspells the word while typing her essay. These facts suggest two initial conditions under which an agent, S, unintentionally performs B: (1) S A-s or is in the course of A-ing intentionally. And (2) S B-s in the course of, or as a result of, A-ing. Understood as necessary conditions, (1) and (2) essentially inform us that unintentional action is always performed in the course of, or as a result of, doing something else intentionally. They do not imply that an unintentional action cannot directly result from another unintentional action; (1) and (2) only imply that any given unintentional act is ultimately traceable back to an intentional act. One might wonder if the following sort of example threatens the necessity of conditions (1) and (2): Sal is being examined by his doctor. The doctor firmly taps Sal’s 127 The view that unintentional action involves acting on an intention, and thus doing something intentionally, is held by Davidson (1971) and Chan (1995), among others. 134 knee with a rubber mallet. Sal’s leg kicks out in response to the tap. Sal kicks the doctor. Here it might be thought that Sal’s kicking the doctor is unintentional; yet his kicking does not occur in the course of intentionally doing anything else. And so unintentionally B-ing does not require intentionally A-ing—and conditions (1) and (2) are unnecessary. I don’t think this example succeeds in showing that (1) and (2) are unnecessary. First, it might well be that Sal’s kicking is not a proper action; in which case, it isn’t an unintentional action. Second, even on the supposition that Sal’s kicking is an unintentional action, it is not the case that it does not occur in the course of doing something else intentionally. When Sal’s leg kicks, he’s in the process of intentionally relaxing his leg. 128 After all, the doctor’s examination of Sal can only succeed if Sal’s leg is relaxed (not tensed). Thus, even if we grant that Sal’s kicking is an unintentional action, this example fails to undermine the necessity of conditions (1) and (2). Another fact in each of the three cases is that the agent does not expect to do what it is he or she does unintentionally. Shawn does not expect to strike his classmate; Sally does not expect to cut off the driver; and Sue does not expect to misspell the word. This might suggest the following negative expectation condition on unintentionally B-ing: S does not expect to B. I have argued above, however, that one can expect to do something and yet do it unintentionally. Thus, any such negative expectation condition on unintentionality is false. An important fact in each of the above cases involves what information enters into the agent’s practical reasoning. In deliberating about A-ing, none of the agents 128 We might also describe Sal’s behavior here as refraining from moving his leg. I argued in ch. 3 that refraining is a type of intentional action. 135 consider performing B. Shawn does not consider striking his classmate when deciding to raise his arm; Sally does not consider cutting off the driver when deciding to turn right; and Sue does not consider misspelling ‘theory’ when typing her essay. This suggests the following condition: (3) S does not consider B-ing (in S’s deliberation that leads to A-ing). Condition (3) supplies a sense in which unintentional action is accidental. If an agent does not consider performing an action when deliberating about acting, then its performance is an accident. By not considering performing B when deciding to A, each of the agents above can be characterized as accidentally or inadvertently B-ing. Though condition (3) provides one sense in which unintentional action is accidental or inadvertent, I suspect that (3) is not necessary for unintentionally B-ing. Consider the following case: (SOCCER) Silvia is a professional soccer player. During a match, she deliberates about whether to pass the ball back to her team’s goalie. It occurs to her that if she kicks the ball past the goalie and into the net, she could make a substantial amount of money (say, on a crooked bet). But scoring on her own goal could jeopardize her team’s chances of winning the match. Silvia decides against scoring on her own goal; as a result, she attempts to pass the ball to the goalie. As it turns out, Silvia kicks the ball too hard and winds up scoring on her own goal. 136 All else being equal, Silvia’s scoring on her own goal is straightforwardly accidental and unintentional. But Silvia considered doing just this when she deliberated about what to do with the ball. Hence, condition (3) is unnecessary for acting unintentionally. In SOCCER, notice that after Silvia considers scoring on her own goal, she decides against it. Her decision not to score the goal (together, perhaps, with the fact that this decision motivates her attempt to pass the ball to the goalie) is what appears to make her scoring accidental and unintentional. SOCCER renders condition (3) unnecessary, though not irrelevant. Let us then label condition (3) as (3a) below and re-cast it as part of a disjunction. S unintentionally B-s only if either: (3a) S does not consider B-ing (in S’s deliberation that leads to A-ing), or (3b) S considers B-ing but decides not to B. We can combine conditions (3a) and (3b) into the following, necessary condition on unintentional action: (3ab) If S considers B-ing (in S’s deliberation that leads to A-ing), then S decides not to B. Conditions (3a) and (3b) mark two important ways in which an agent’s B-ing can be accidental. Condition (3ab) is, therefore, an improvement over condition (3). 137 Condition (3ab) is an improvement, but it is not yet adequate. For it seems that one can consider doing something and decide to do it, yet still act unintentionally. Consider the following example 129 : (ENEMY) Ralph gets in his car with the intention of going to the market. As he starts his car, he notices his archenemy, Dez, standing behind his car. Ralph considers backing over Dez, and decides to do it. Just before he begins to back out, Ralph looks in his rear-view mirror and no longer sees Dez standing behind his car. Believing that Dez has snuck away, Ralph carries on with his original plan to go to the market. Ralph backs his car out. As it happens, Dez was crouched down behind Ralph’s car. Ralph backs over Dez. Intuitively, Ralph’s backing over Dez was unintentional despite his considering and deciding to do it. Condition (3ab), therefore, needs altering. What seems important about Ralph’s act of backing over Dez is that it is not the result of his decision to back over him. Ralph acts on his decision (plan, intention) to go to the market. This is what seemingly renders his backing over Dez accidental. This point recommends the following modification to condition (3ab): (3ab*) If S considers B-ing, then either (a) S decides not to B, or (b) S decides to (try to) B, but S’s B-ing is not brought about (in the right way) by S’s decision to (try to) B. 130 Condition (3ab*) introduces a third sense in which action can be accidental. I believe that (3ab*) is a necessary condition for unintentionally B-ing. 129 Thanks to Gideon Yaffe for bringing this example to my attention. 130 The ‘in the right way’ clause is meant to elude problems involving deviant causation. 138 Let us now state our account of unintentional action. Where ‘S’ is an agent and ‘A’ and ‘B’ are distinct behaviors: (UA) S unintentionally performs B iff (i) S A-s or is in the course of A-ing intentionally, (ii) S B-s in the course of, or as a result of, A-ing, and (iii) If S considers B-ing, then either (a) S decides not to B, or (b) S decides to (try to) B, but S’s B-ing is not brought about (in the right way) by S’s decision to (try to) B. UA correctly captures each of the five clear-cut cases of unintentionality discussed in this section. The agents in ARM-RAISE, DRIVE, and MISSPELL each act unintentionally because they fail to consider doing what they plainly do unintentionally. In SOCCER, Silvia acts unintentionally under UA because she decides not to score the goal, then (purely accidentally) does so anyway. In ENEMY, Ralph’s backing over Dez is not the result of his decision to back over him, and thus is unintentional under UA. Now let us return to the example involving the sniper. We are now in a position to see why the sniper’s alerting the enemy is not unintentional. The reason it is not unintentional is because it fails to satisfy all three conditions of UA. It’s true that the sniper’s alerting the enemy to his presence satisfies conditions (i) and (ii) of UA: the sniper shoots the soldier intentionally, and alerts the enemy in the course of shooting the soldier. However, his alerting the enemy violates condition (iii): though he takes into consideration alerting the enemy when deliberating about whether to shoot the soldier, he neither decides not to alert the enemy nor does he decide to (try to) alert the enemy. If the sniper decided not to alert the enemy, then his decision to shoot the soldier would 139 render him irrational given his knowledge that shooting the soldier will alert the enemy to his presence. But the sniper is not (or at least, need not be) irrational in our scenario. If the sniper decided to (try to) alert the enemy, then it would follow that he intends to (try to) alert the enemy—which clearly he does not. The sniper’s violation of UA’s condition (iii) helps explain our sense that the sniper’s alerting the enemy is not accidental or inadvertent. Thus, the sniper’s alerting the enemy is not unintentional. If an act is neither intentional nor unintentional, then it must be non-intentional. Thus, the sniper non-intentionally alerts the enemy to his presence. 5.3 Is Alerting the Enemy Intentional? As I indicated above in section 5.1, I believe that Mele and Sverdlik are correct to claim that the sniper’s alerting the enemy is not intentional because it is not the sniper’s aim or goal. The problem is that it is difficult to specify the sense in which intentional action is, necessarily, action that one aims at performing. And in the absence of such a specification, it remains open to philosophers like Bratman and Harman to deny a necessary connection between aiming and acting intentionally. Rather than attempt to ground the aiming feature of intentional action, let us instead discuss two considerations that might lead some to suppose that side-effect acts are intentional. I will explain why neither consideration should be afforded much weight. Then I’ll briefly discuss a pair of differences between paradigmatic intentional acts and side-effect acts like the sniper’s alerting of the enemy. These differences will help suggest a difference between intentional acts and the sniper’s act. Alas, our discussion 140 will fall short of establishing that side-effect action is not intentional. But I hope that it renders the position viable. One reason one might suppose that side-effect acts are intentional involves an aversion to SV. Recall SV is the principle that intentionally A-ing requires intending to do A. Proponents of SV are committed to denying that side-effect acts, such as the sniper’s alerting the enemy, are intentional: the sniper does not intend to alert the enemy, and so does not alert the enemy intentionally. Opponents of SV, then, might feel some pressure to affirm that such acts are intentional on account of their more relaxed view of the connection between intention and intentional action. However, an aversion to SV should not motivate the view that side-effect acts are intentional. The view that side- effect acts are non-intentional is fully compatible with a rejection of SV. A second reason one might feel pressure to affirm that side-effect acts are intentional rather than non-intentional might involve a concern about committing to a third category of action. If one holds that side-effect acts are non-intentional, then one is committed to the view that there exists a third type of action. And perhaps the prospect of this commitment is enough to make some theorists a bit nervous. In section 6 below I will present an independent argument in favor of non-intentionality—one that I believe is decisive. If I am right, then any such “fear of commitment” should not count in favor of classifying side-effect action as intentional. There is (at least) a pair of differences between paradigmatic intentional acts and side-effect acts such as the sniper’s alerting the enemy. First, paradigmatic intentional acts are always welcomed by the agent. The sniper’s shooting the soldier is not only 141 obviously intentional, but it is also something the sniper welcomes doing. When Shawn intentionally raises his arm in order to answer the teacher’s question, he welcomes the raising of his arm; and when he intentionally answers the question, he welcomes that too. Contrast these cases with the sniper’s alerting the enemy. Clearly the sniper does not welcome alerting the enemy to his presence. Indeed, if it were up to him, he’d surely shoot the soldier without alerting the enemy. One difference, then, between paradigmatic intentional action and the sniper’s alerting the enemy is that the former is welcomed whereas the latter is not. The second difference between paradigmatic intentional action and the sniper’s alerting the enemy is this: the former involves a disposition to adjust one’s behavior in order so to act. Suppose that as the sniper is preparing to shoot the soldier, he becomes aware of a peculiarity about his rifle. Perhaps in order to fire his rifle (once), he’ll have to pull the trigger three times. So, the sniper pulls the trigger three times and (intentionally) shoots the soldier. The sniper makes the adjustments that he believes are necessary for shooting the soldier. Or suppose I (intentionally) walk to the market. As I’m walking, I realize that the street I usually take to get to the market is closed off. What to do? The answer is easy: I simply select an alternate street to take to the market. In cases of paradigmatic intentional action, agents are disposed to make certain adjustments in order to successfully act. This, however, is not true of the sniper’s alerting the enemy. If the sniper comes to learn that in order to alert the enemy he’ll have to fire his first shot straight up in the air, the sniper will simply shoot the soldier without alerting the enemy. Or suppose that as the sniper prepares to shoot the soldier, the soldier 142 abruptly changes his position (suppose he runs one-hundred yards up the road). In order to shoot the soldier, the sniper must change his own position. All else being equal, the sniper will make the necessary change to his position. But if while changing his position it occurs to the sniper that when he shoots the soldier he (probably) will no longer alert the enemy, the sniper will not be inclined to pursue measures to alert the enemy. Indeed, doing so would be irrational. The sniper’s alerting the enemy is not unintentional because it violates UA. It’s not intentional because alerting the enemy is not the sniper’s aim or goal. Therefore, it must be non-intentional. 6 An Argument from Lucky Action In this section, I will present a second argument for non-intentionality; one that I believe is more promising than all the preceding arguments. Using my account of unintentional action developed in the last section, I will argue that cases of lucky action are plausibly non-intentional. 6.1 Lucky Action is Not Intentional Consider the following scenario: (BASKET) Derek, an amateur basketball player, has the ball at half-court. His team is behind by one point as the game clock is expiring. Derek shoots the ball from half-court in an attempt to make a basket. Lo and behold, the shot goes in. 143 Derek’s making the basket is straightforwardly lucky. The question is: what is the status of his lucky shot? Many philosophers have denied that lucky acts like Derek’s are intentional. 131 Intuitively, Derek’s making the basket is too coincidental to count as intentional. Examples like BASKET illustrate the need for a skill condition on intentionality. Let us state such a condition as follows: (SK) S intentionally performs B at t only if S’s B-ing at t manifests a suitably reliable skill of S’s in B-ing in the way S B-s at t. 132 However we are to understand the concept of skill, it is clear that Derek violates SK: Derek, like most people, is not skilled in making half-court shots. Thus, Derek’s making the shot is not intentional. Is SK too strong? It might be thought that one need not be skilled at doing something in order to do it intentionally. Instead, perhaps intentionality merely requires that one possess evidence that one will (probably) succeed in acting. Let us add to BASKET the following fact: God tells Derek to shoot the ball when the clock reaches one second because the conditions at that time will be nearly perfect for making the shot. Sure enough, when the clock reaches one second, Derek shoots and makes the basket. It might be thought here that Derek makes the basket intentionally, despite being unskilled at making half-court shots. If so, SK is too strong. 131 See, e.g., Harman (1976), Mele and Moser (1994), O’Shaughnessy (1980), and Wasserman (forthcoming). 132 For an enlightening discussion of skill and intentionality, see Mele and Moser (1994). 144 In response to this, we might say that in this case Derek is skilled at making half- court shots. That is, in situations in which God assures Derek that he will (probably) make a half-court shot if he shoots it, it’s the case that Derek possesses the skill of making half-court shots. I’m not sure if this is the right thing to say; fortunately, we don’t need to commit to saying it here. We can go ahead and simply modify SK to include an evidence disjunct: (SK*) S intentionally performs B at t only if either (i) S’s B-ing at t manifests a suitably reliable skill of S’s in B-ing in the way S B-s at t, or (ii) S possesses evidence at t that S will (probably) B at t. 133 Whether SK* is preferable to SK is a question we can leave open for now. I take it that the disjunction of SK and SK* is extremely plausible. What’s important for our purposes is to recognize that Derek (in the original BASKET scenario) violates both SK and SK*. We’ve already seen that he violates SK. He violates SK* because (in BASKET) he lacks evidence that he will (probably) make the shot. So, Derek violates both SK and SK*. And on the assumption that either SK or SK* is true, it follows that Derek does not intentionally make the half-court shot. 6.2 Lucky Action is Not Unintentional Why isn’t Derek’s making the shot unintentional? To answer this, we need to return to UA. Derek luckily makes the half-court shot. Derek satisfies UA’s (i) and (ii): he shoots 133 For a defense of a condition along the lines of SK*, see section 4 of Mele and Moser (1994). 145 the ball intentionally, and makes the basket as a result of shooting the ball. However, Derek violates UA’s condition (iii): he considers making the shot and decides to try to make it, but his making it is brought about in the right way by his decision to try to make it. In other words, Derek decides to try to make the shot and thereby tries as a result of his decision; furthermore, he makes the shot in just the way he tries to make it. Thus, Derek’s making the shot is not unintentional. We’ve seen above (section 6.1) that Derek’s lucky shot is not intentional because it violates a skill condition (or skill-evidence condition, if you like) on intentionality. Neither is the shot unintentional because it violates condition (iii) of UA. If an action is neither intentional nor unintentional, it must be non-intentional. We conclude that Derek non-intentionally sinks the half-court shot. 7 Conclusion In this chapter I have argued that both lucky and side-effect action are best conceived as non-intentional. But there are (at least) two ways of interpreting this claim. The weaker interpretation goes something like this. Our ordinary notions of intentionality and unintentionality are not currently equipped to handle lucky and side-effect action. It’s also true that there is no ordinary notion of non-intentionality. And so in saying that lucky acts and side-effect acts are non-intentional, we’re essentially pointing out a shortcoming in our ordinary notions of intentionality and unintentionality. The idea is that we need to modify these two notions in a way that captures lucky and side-effect action. Until we do this, we’ll categorize lucky and side-effect action as “non- 146 intentional.” But let’s not fool ourselves: the suggestion is not that there is a real, third category of action; rather we’ll temporarily call these acts “non-intentional” until we (i.e., the linguistic community) determine whether they should be classified as intentional or unintentional. On this weak interpretation, labeling an act “non-intentional” is merely an expression of agnosticism. Let me be clear: in claiming that there exists non-intentional action, I mean to say something stronger than what the weak interpretation has in mind. It is my view that action that is rightly classified as non-intentional can never later be rightly classified as anything to the contrary. Just as an act that is correctly deemed to be intentional or unintentional can never later be correctly deemed to have been otherwise, so it is for non- intentional action. In other words, I believe that the category of non-intentionality has the same ontological status as the categories of intentionality and unintentionality. Non- intentionality is real. If I am right about lucky action and side-effect action both being non-intentional, an interesting fact about non-intentional action emerges. Broadly speaking, we can characterize intentional action as action that is aimed at and unintentional action as action that is not aimed at. We’ve seen that some non-intentional action is aimed at (luck) and some is not (side-effects). There is a sense, therefore, in which non-intentional acts are nearer or farther from intentionality and unintentionality: lucky acts are “closer” to being intentional than unintentional; side-effect acts are “closer” to being unintentional than intentional. 147 The category of non-intentionality occupies space in between the intentional and unintentional. I have argued that lucky and side-effect acts are non-intentional. I wouldn’t be surprised if the category of non-intentionality were larger yet. I noted earlier that the existence of non-intentionality could potentially have important implications for moral and legal responsibility. I look forward to the challenge of tracking down these implications. 134 134 Thanks to Al Mele, Mark Schroeder, and Gary Watson for helpful comments and suggestions for improvement. I am especially grateful to Gideon Yaffe for helpful comments and many hours of discussion on the topic of non-intentionality. A version of this chapter will be presented at the 2012 APA Pacific Division Meeting. I thank Preston Greene for his helpful comments. 148 CHAPTER 4 REFERENCES Adams, Frederick (1986), “Intention and Intentional Action: The Simple View,” Mind and Language 1, pp. 281-301. Bratman, Michael (1987), Intentions, Plans, and Practical Reason (Cambridge, MA: Harvard University Press). All references are to the CSLI Publications (1999) re- issue. Chan, David K. (1995), “Non-Intentional Actions,” American Philosophical Quarterly 32 (2), pp. 139-51. Davidson, Donald (1978), “Intending,” in Philosophy of History and Action, ed. Yirmiahu Yovel (Dordrecht, NL: D. Reidel), pp. 41-60. Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 83- 102. References are to Davidson (2001). Davidson, Donald (1971), “Agency,” in Agent, Action, and Reason, eds. Robert Binkley, Richard Bronaugh, and Ausonio Marras (Toronto: University of Toronto Press). Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 43-61. References are to Davidson (2001). Davis, Wayne (1984), “A Causal Theory of Intending,” American Philosophical Quarterly 21, pp. 43-54. Gorr, Michael and Terrence Horgan (1982), “Intentional and Unintentional Actions,” Philosophical Studies 41, pp. 251-62. Harman, Gilbert (1976), “Practical Reasoning,” Review of Metaphysics 79, pp. 431-63. McCann, Hugh (1991), “Settled Objectives and Rational Constraints,” American Philosophical Quarterly 28, pp. 25-36. McCann, Hugh (1974), “Volition and Basic Action,” The Philosophical Review 83, pp. 451-73. Meiland, J. W. (1963), “Are There Unintentional Actions?” The Philosophical Review 72 (3), pp. 377-81. Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). Mele, Alfred R. and Paul K. Moser (1994), “Intentional Action,” Nous 28, pp. 39-68. 149 Mele, Alfred and Steven Sverdlik (1996), “Intention, Intentional Action, and Moral Responsibility,” Philosophical Studies 82 (3), pp. 265-87. O’Shaughnessy, Brian (1980), The Will (Cambridge: Cambridge University Press). Pears, David (1985), “Intention and Belief,” in Essays on Davidson, eds. Bruce Vermazen and Merrill B. Hintikka (Oxford: Clarendon Press), pp. 75-88. Velleman, David (1989), Practical Reflection (Princeton, NJ: Princeton University Press). Wasserman, Ryan (forthcoming), “Intentional Action and the Unintentional Fallacy,” Pacific Philosophical Quarterly. Yaffe, Gideon (2006), “Trying, Intending and Attempted Crimes,” Philosophical Topics 32, pp. 505-32. 150 CONCLUSION In my dissertation I have argued for the following six major claims: (C1) The belief/desire theory and the planning theory of intention each entail DPV. (C2) DPV is implausible. So, (C3) Both theories of intention require modification. (C4) Negative intentional action exists in the form of refraining. (C5) Non-intentional action exists in the form of lucky and side-effect acts. (C6) Negative non-intentional action exists in certain cases involving expected omissions. In this conclusion I will do two things: I will briefly remind readers how we arrived at claims C1-C6 and discuss several interesting consequences and issues related to some of these claims. C1, C2 and C3 Let us begin with the closely related claims C1, C2 and C3. In chapter 1 I presented an argument for C1. The argument had the following structure: first, show that both theories of intention—the belief/desire theory and the planning theory—were committed to the view that trying to act does not require having an intention whatsoever. Then show that 151 this view about trying leads to DPV: intentional action does not require having an intention whatsoever. I suggested at the end of chapter 1 that both theories of intention required fixing in light of each theory’s commitment to DPV. But this raised the question: why should we think these two theories require fixing? Perhaps DPV is more plausible than SPV— the view that intentional action requires intention? In order to make good on my suggestion at the end of chapter 1, then, I needed an argument for C2. C2 was defended in chapter 2. There I argued that the following pair of claims about settling was preserved by SPV but not DPV: (1) Settling on a course of action initiates and sustains action, and (2) Settling is subject to certain norms of rationality. If this is right, then a plausible theory of intention will be one that is compatible with SPV. Thus, C3: the belief/desire theory and the planning theory require fixing. Let us say a little more about C3 here. C3 is slightly misleading because it is a bit understated. If my arguments for C1 and C2 are plausible, then the belief/desire theory of intention can’t be true. This is because the belief/desire theory cannot capture the settling feature of intention. Presumably only intentions involve being settled on a course of action. Recall that in chapter 2 it was argued that the belief that one will perform A does not entail being settled on performing A. And of course, even if we add to one’s belief the desire to do A we still do not have settledness. Thus, claim C3 is more accurately stated as the following claim: (C3') The belief/desire theory is false and the planning theory requires modification. 152 Now then, how might we alter the planning theory to bring it into conformity with SPV? One thing we might do is permit exceptions to the Strong Consistency Requirement. 135 Recall that the Strong Consistency Requirement was the demand that one’s intentions and beliefs be consistent. If we permit exceptions to this demand, then we can attribute to Sarah an intention in my scenario in chapter 1. Of course, one consequence of permitting exceptions to the Strong Consistency Requirement is that Bratman’s ‘video game’ objection to SV—the view that intentionally A-ing requires intending A—seems to lose its force. If it is possible to rationally intend to A while believing that one will not A, then proponents of SV can presumably cook up a story about why Bratman’s game player rationally intends to hit the two targets all the while believing she cannot hit both. Another option is to alter the Agglomerativity Requirement. Recall that in chapter 1 I formulated the requirement as follows: if one intends to A and intends to B, then rationality requires that one intend to A and B. Let us call this formulation of the requirement ‘Agglomerativity 1’. Gideon Yaffe has argued convincingly that Agglomerativity 1 is implausible. 136 He points out that Agglomerativity 1 is too demanding; it requires mental work of agents even when there is no apparent pay off. Suppose I intend to do my laundry in five minutes and I intend to visit Hong Kong in twenty-five years. According to Agglomerativity 1, I am under rational pressure to pull my two intentions together into a single, conjunctive intention. But why would 135 Bratman (1987) himself seems to endorse this position. He says: “To coordinate my activities over time a plan should be, other things equal, internally consistent…So, assuming my beliefs are consistent, such a plan should be consistent with my beliefs, other things equal” (p. 31; emphasis mine). 136 See Yaffe (2006). 153 rationality require me to do this even when there is no sense in which I’m better off for having done it? Indeed, to extend Yaffe’s point, suppose at any given moment I am such that I’m in possession of approximately one-thousand intentions. And suppose my set of intentions shifts regularly; some intentions get executed, others abandoned, still others are newly acquired, etc. It would seem virtually impossible to form massive conjunctive intentions at any given moment. It would take too much time to form them—my set of intentions would shift before I could completely form most of them. Under Agglomerativity 1, then, I would perpetually be irrational. But this seems wrong. Instead of Agglomerativity 1, Yaffe prefers the following: Agglomerativity 2: It is irrational to intend to A and intend to B if the intention to do both A and B would itself be irrational according to some other principle of rational intention. 137 The question for our purposes is: does Agglomerativity 2 render the planning theory compatible with SPV? Recall that I argued in chapter 1 that Sarah does not intend to try to hit target 1 nor intend to try to hit target 2 because if she did, then by Agglomerativity 1 she must intend to try to hit both targets. But since Bratman holds that it is irrational for Sarah to try to hit both given her knowledge she can’t hit both, I suggested that rationality could not require her to intend to try to hit them both. Thus, we could eliminate the pair of intentions to try. But, what if we insert Agglomerativity 2 in place of Agglomerativity 1? Does anything change? Perhaps. We might hold that it is not irrational for Sarah to intend to try to hit both targets in this situation on grounds that it 137 Yaffe (2006); see also Yaffe (2010), p. 55. 154 will best help her to achieve the outcome she most desires (i.e., hitting a target/winning the prize). And if this conjunctive intention is not irrational, then (by Agglomerativity 2) it follows that neither is the intention to try to hit target 1 nor the intention to try to hit target 2 irrational. In which case, we can attribute these intentions to Sarah and cite them in the production of her intentional act of hitting one of the targets. Thus, we have intentional action with an intention—and SPV is preserved. But this move raises the question: if it’s not irrational for Sarah to intend to try to hit both targets, why would it be irrational for her to try to hit them both? Here, perhaps we should simply part company with Bratman on this issue and hold that it is not irrational for Sarah to try to hit both targets. But I’m not certain about this; there are questions here about the nature of trying that I am not prepared to engage (e.g., can one try to do something if one believes that success is impossible? If not, then Sarah cannot even try to hit both targets.) I need to think more carefully about Agglomerativity 2 and how it relates to the argument of chapter 1. For now, let us move on to claims C4-C6. C4 In chapter 3 I defended C4: Negative intentional action exists in the form of refraining. Suppose Velma has the opportunity and the ability to insult her friend, but she refrains from doing so. I have suggested it is intuitive to describe Velma’s refraining as intentional. If her refraining is intentional, then it must be actional. But if actions necessarily are events, then how can Velma’s refraining be an action? “Pure” omissions are not events—they’re non-occurences. I respond to this difficult challenge in chapter 3 155 by developing an account of refraining that preserves the intuition that refrainings are proper intentional actions. Key to my conception of refraining is the condition that in order to refrain from doing something (A), an agent must thereby do something else (B) intentionally (where A is incompatible with B). In other words, I propose that refraining involves both a positive and a negative component. The positive component entitles us to conceive of refrainings as negative intentional action. C5 Actions can be performed intentionally and unintentionally. Question: is there a third way to perform an action? In chapter 4 I argued there was—that some of our acts are performed non-intentionally. I first argued that side-effect acts, such as the sniper’s alerting the enemy to his presence, were non-intentional. My argument for this conclusion was inspired by Mele and Sverdlik’s famous argument. I then argued that lucky acts, such as Derek’s sinking the half-court shot, were also non-intentional. Hence, C5: Lucky and side-effect acts constitute non-intentional action. In the course of constructing the argument for C5, I also developed an account of unintentional action. Two important consequences of my argument for C5 are worth noting here. First, it follows on my view that all action involves intentional action. In other words, one cannot act intentionally, non-intentionally or unintentionally without acting intentionally. This point reinforces Mele’s claim about the primacy of intentional action mentioned in the Introduction. If I am right, all action depends on intentional action. The second consequence is this: all action is produced by the state of intention. If whenever one acts 156 one acts intentionally, then given the truth of SPV, it must be that all three types of action spring from the state of intention. Let us now turn to a potential difficulty facing C5. C5 is a conjunctive claim. One of the conjuncts is this: Non-intentional action exists in the form of side-effect acts. A potential obstacle for this conjunct comes from an important study by Joshua Knobe on ordinary ascriptions of intentionality. 138 Subjects in Knobe’s study read one of two vignettes. One of the vignettes involved a harmful side-effect action: (HARM) The vice-president of a company went to the chairman of the board and said, ‘We are thinking of starting a new program. It will help us increase profits, but it will also harm the environment.’ The chairman of the board answered, ‘I don’t care at all about harming the environment, I just want to make as much profit as I can. Let’s start the new program.’ They started the new program. Sure enough, the environment was harmed. 139 The subjects were then asked whether the chairman intentionally brought harm to the environment. A large majority of the subjects (82%) reported that the chairman’s harming of the environment was intentional. 140 138 Knobe (2003). 139 Knobe (2003), p. 191. 140 The other vignette that was read by Knobe’s subjects was essentially identical to HARM except that the word ‘harm’ was replaced by the word ‘help’. When asked whether the chairman’s helping the environment was intentional, a large majority of subjects (77%) answered ‘no’. This asymmetry in subjects’ ascriptions of intentionality has become known as ‘The Knobe Effect’: in cases involving morally bad side-effects, people are inclined to attribute intentionality; in cases involving morally good side-effects, people are not similarly inclined. 157 Part of C5 is the claim that non-intentional action exists as side-effect action. The potential difficulty here is that HARM is a case of side-effect action in which people (apparently) widely believe is intentionally performed. I think there are a few things to say in response to Knobe’s study and the potential threat it poses to C5. First, it’s unclear, of course, whether people’s intuitions in Knobe’s study are accessing the true nature of intentionality (more on this below). 141 But let us suppose for a moment that the folk are correct about the chairman: he intentionally harms the environment. What would this show? To begin, notice that it would not undermine my most important point from chapter 4 that non-intentional action exists. Even if it turns out that side-effect acts are not non-intentional, my Argument from Lucky Action remains intact. But we can say more than this. Even if the folk are right, it would not require me to abandon my view that non-intentional action exists in the form of side-effect acts. It would only require me to qualify my view roughly in the following way: non-intentional action exists in the form of non-moral side-effect acts. This would be compatible with it being true that morally good and bad side-effect acts are not non-intentional. In fact, in Knobe’s ‘help’ vignette, subjects only report that the chairman does not intentionally help the environment. But ‘not intentionally helping the environment’ is compatible with ‘non-intentionally helping the environment’. So perhaps for now I need, at most, only concede that morally bad side-effects do not constitute non-intentional action. Still, this concession is consistent with the truth of C5. 141 Adams and Steadman (2004), e.g., argue that the asymmetry in ascriptions of intentionality in Knobe’s study is (merely) pragmatic. They argue that the subjects’ ascriptions of intentionality are expressions of praise and blame for the chairman’s side-effect action. 158 I think the remarks above, however, are too charitable. We were granting that subjects’ intuitions were accessing the true nature of intentionality; that the moral status of side-effect acts determine whether or not they are performed intentionally. But perhaps we ought not grant this point. Dean Pettit and Knobe have recently conducted a study on ordinary people’s ascriptions of decision in HARM. 142 Subjects read HARM and were then asked whether they agree or disagree with the following statement: (S1) The chairman decided to harm the environment. 143 Ratings were recorded on a scale of 1 (‘disagree’) to 7 (‘agree’). The subjects tended to agree with S1; the mean rating was 4.6. This suggests that ordinary people tend to believe that the chairman (in HARM) decided to harm the environment. Question: should we think that people’s intuitions here are supplying information about the true nature of the state of deciding? I have my doubts. As we’ve already discussed, deciding to do something plausibly entails intending to do it. 144 If the chairman decides to harm to the environment, then it follows that he intends to harm the environment. But it seems unlikely that he intends to harm the environment; rather what he intends to do is start the new program (or increase profits)—this is his aim. He 142 Pettit and Knobe (2009). 143 Some subjects read the ‘help’ version of the story. They too were asked whether they agree or disagree with this statement (substituting ‘harm’ for ‘help’). This part of the study we can ignore for now. 144 See Mele (1992), p. 141 and Mele (2003), ch. 9. I discussed this point in chs. 1 (section 1.1) and 3 (section 6.2). 159 doesn’t care one way or the other about harming the environment. If he doesn’t intend to harm the environment, then he doesn’t decide to harm it. And so Pettit and Knobe’s subjects are likely mistaken in their attributions of decision. And if their intuitions are wrong about decision, then perhaps we ought to be suspicious about Knobe’s subjects’ intuitions about intentionality. A similar point applies to the state of desire. Tannenbaum, Ditto and Pizarro conducted a study on people’s intuitions about desire in HARM. 145 Subjects were asked whether they agree or disagree with the following statement: (S2) The chairman desired to harm the environment. As with S1, ratings were recorded on a scale of 1 (‘disagree’) to 7 (‘agree’). In this case, subjects seemed conflicted; the mean rating was 3.4. This rating indicates that the group of respondents neither agrees nor disagrees with S2. But, it’s fairly clear that the chairman does not desire to harm the environment—he’s indifferent. Now, it’s possible that some people think that the chairman has an extrinsic desire to harm the environment: he wants to harm it in order to increase profits. But of course, this would be a misreading of HARM; the harm that comes to the environment is purely a side-effect rather than a means for increasing profits. It’s not through harming the environment the new program will increase profits. If this misreading is (even partially) responsible for the results related to S2, then perhaps a similar misreading is occurring in Knobe’s original study 145 Tannenbaum, Ditto and Pizarro (unpublished). 160 involving people’s ascriptions of intentionality. Either way, I think the results pertaining to S1 and S2 motivate a certain degree of skepticism about ordinary people’s intuitions concerning intentionality. Knobe’s studies about folk attributions are intriguing and certainly deserve a more careful treatment than I’ve provided here. Nothing I’ve said is intended to be conclusive. I have merely been concerned with discussing how claim C5 relates to some of Knobe’s important work on intentionality. C6 I argued in chapter 3 that there are two basic types of intentional action: positive and negative. I argued in chapter 4 that positive non-intentional action exists in the form of lucky and side-effect action. If this is right, then we should expect there to exist negative non-intentional action. In chapter 3 I defended C6: Negative non-intentional action exists in certain cases involving expected omissions (e.g., the BABY cases). The BABY cases were essentially the negative analogue of positive side-effect acts, such as the sniper’s alerting the enemy. Recall that these negative non-intentional acts motivated a sixth necessary condition on refraining: (REF) S refrains from A-ing iff (1) S considers A-ing, (2) S has the ability to A, (3) S Bs intentionally, (4) S lacks the belief that B-ing is compatible with A-ing, (5) B-ing is incompatible with A-ing, and (6) S’s not A-ing is S’s aim or goal. 161 One question that I did not address in chapter 3 is this: are there other types of negative non-intentional action besides side-effect omissions? I’m guessing there probably is at least one other type: lucky omissions. If lucky acts are non-intentional, then we should expect lucky omissions (assuming they exist) to constitute negative non- intentionality. Suppose I am walking down the side walk and I spot a bug up ahead. I decide to refrain from stepping on him. So, I step wide right, believing that in so stepping I will avoid squashing him. Sure enough, my belief is true: I avoid stepping on him. Other things being equal, I seem to refrain from stepping on the bug. But suppose that, as it turns out, my failure to step on the bug was rather lucky: there was a 99.5% chance that I would squash the bug stepping exactly where I stepped. Have I now refrained from stepping on the bug? I’m guessing that I haven’t. If luckily doing something does not constitute intentional action, then why should luckily failing to do something be any different? This consideration generates a seventh necessary condition on refraining: (7) S’s not A-ing is not lucky. Now, I certainly admit that the ‘bug’ example is not obvious: it doesn’t “scream out” non-intentionality. But this might (at least partially) be explained by the fact that we are not usually aware of whether our omissions are a matter of luck or not. When Derek the basketball player hits the half-court shot, it’s apparent to everyone that he was lucky. We are often aware of whether our positive acts are luckily performed. When I, on the 162 other hand, fail to step on the bug, I am simply unaware how lucky I was. For this reason, perhaps classifying lucky omissions as non-intentional is something of a non- obvious affair. However, if lucky positive action is non-intentional, and we have reason to believe there exists negative non-intentional action (we do: expected omissions), then we should probably view certain lucky omissions as negative non-intentional acts as well. 163 CONCLUSION REFERENCES Adams, Fred and Annie Steadman (2004), “Intentional Action in Ordinary Language: Core Concept or Pragmatic Understanding?” Analysis 64, pp. 173-81. Bratman, Michael (1987), Intentions, Plans, and Practical Reason (Cambridge, MA: Harvard University Press). All references are to the CSLI Publications (1999) re- issue. Knobe, Joshua (2003), “Intentional Actions and Side-Effect Acts,” Analysis 63, pp. 190- 94. Mele, Alfred (2003), Motivation and Agency (Oxford: Oxford University Press). Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). Pettit, Dean and Joshua Knobe (2009), “The Pervasive Impact of Moral Judgment,” Mind & Language 24 (5), pp. 586-604. Tannenbaum, David, Peter Ditto and David Pizarro (unpublished), “Different Moral Values Produce Different Judgments of Intentional Action.” Yaffe, Gideon (2010), Attempts (Oxford: Oxford University Press). Yaffe, Gideon (2006), “Trying, Intending and Attempted Crimes,” Philosophical Topics 32, pp. 505-32. 164 COMPREHENSIVE REFERENCES Adams, Frederick (1986), “Intention and Intentional Action: The Simple View,” Mind and Language 1, pp. 281-301. Adams, Fred and Annie Steadman (2004), “Intentional Action in Ordinary Language: Core Concept or Pragmatic Understanding?” Analysis 64, pp. 173-81. Anscombe, G. E. M. (1963), Intention (Ithaca, NY: Cornell University Press). Audi, Robert (1986), “Intending, Intentional Action, and Desire,” in The Ways of Desire, ed. Joel Marks (Chicago: Precedent), pp. 17-38. Audi, Robert (1973), “Intending,” Journal of Philosophy 70, pp. 387-402. Austin, John (1873), Lectures on Jurisprudence Vol. 1 (London: John Murray). Bach, Kent (2010), “Refraining, Omitting, and Negative Acts,” in A Companion to the Philosophy of Action, eds. Timothy O’Connor and Constantine Sandis (Oxford: Wiley-Blackwell). Baier, Annette (1970), “Act and Intent,” Journal of Philosophy 67, pp. 648-58. Beardsley, Monroe (1978), “Intending,” in Values and Morals: Essays in Honor of William Frankena, Charles Stevenson, and Richard Brandt, eds. Alvin Goldman and Jaegwon Kim (Dordrecht, NL: D. Reidel), pp. 163-84. Brand, Myles (1971), “The Language of Not Doing,” American Philosophical Quarterly 8 (1), pp. 45-53. Bratman, Michael (1987), Intentions, Plans, and Practical Reason (Cambridge, MA: Harvard University Press). All references are to the CSLI Publications (1999) re- issue. Bratman, Michael (1984), “Two Faces of Intention,” The Philosophical Review 93 (3), pp. 375-405. Reprinted in Mele (1997), ch. 8. Chan, David K. (1995), “Non-Intentional Actions,” American Philosophical Quarterly 32 (2), pp. 139-51. Chisholm, Roderick (1970), “The Structure of Intention,” Journal of Philosophy 67, pp. 633-47. 165 Churchland, Paul (1970), “The Logical Character of Action Explanations,” The Philosophical Review 79 (2), pp. 214-36. Cleveland, Timothy (1997), Trying without Willing (Brookfield, VT: Ashgate). Crane, Tim (1995), “The Mental Causation Debate,” Proceedings of the Aristotelian Society, Supplementary Volume 69, pp. 211-36. Danto, Arthur (1973), Analytical Philosophy of Action (Cambridge: Cambridge University Press). Davidson, Donald (1993), “Thinking Causes,” in Mental Causation, eds. John Heil and Alfred Mele (Oxford: Oxford University Press), pp. 1-17. Davidson, Donald (1978), “Intending,” in Philosophy of History and Action, ed. Yirmiahu Yovel (Dordrecht, NL: D. Reidel), pp. 41-60. Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 83- 102. References are to Davidson (2001). Davidson, Donald (1971), “Agency,” in Agent, Action, and Reason, eds. Robert Binkley, Richard Bronaugh, and Ausonio Marras (Toronto: University of Toronto Press). Reprinted in Davidson (2001), Essays on Actions and Events (Oxford: Oxford University Press), pp. 43-61. Davis, Wayne (1986), “The Two Senses of Desire,” in The Ways of Desire, ed. Joel Marks (Chicago: Precedent), pp. 63-82. Davis, Wayne (1984), “A Causal Theory of Intending,” American Philosophical Quarterly 21, pp. 43-54. Reprinted in Mele (1997), ch. 6. Descartes, Rene (1641), Meditations on First Philosophy, trans. and ed. John Cottingham (Cambridge: Cambridge University Press, 1996). Ginet, Carl (2004), “Trying to Act,” in Freedom and Determinism, eds. Joseph Keim Campbell, Michael O’Rourke, and David Shier (Cambridge, MA: MIT Press), pp. 89-102. Goldman, Alvin (1970), A Theory of Human Action (Englewood Cliffs, NJ: Prentice- Hall). Gorr, Michael and Terrence Horgan (1982), “Intentional and Unintentional Actions,” Philosophical Studies 41, pp. 251-62. 166 Grice, H. P. (1971), “Intention and Uncertainty,” Proceedings of the British Academy 57, pp. 263-79. Harman, Gilbert (1976), “Practical Reasoning,” Review of Metaphysics 79, pp. 431-63. Reprinted in Mele (1997), ch. 7. Hornsby, Jennifer (1995), “Reasons for Trying,” Journal of Philosophical Research 20, pp. 525-39. Hornsby, Jennifer (1993), “Agency and Causal Explanation,” in Mental Causation, eds. John Heil and Alfred Mele (Oxford: Oxford University Press), ch. 10. Reprinted in Mele (1997), ch. 12. Jackson, Frank (1996), “Mental Causation,” Mind 105, pp. 377-413. Johns, Brandon (2009), “Refraining and the External,” Ratio 22 (2), pp. 206-15. Kenny, Anthony (1966), “Intention and Purpose,” Journal of Philosophy 63, pp. 642-51. Kim, Jaegwon (1976), “Intention and Practical Inference,” in Essays on Explanation and Understanding, eds. Juha Manninen and Raimo Tuomela (Dordrecht, NL: D. Reidel), pp. 249-69. Knobe, Joshua (2003), “Intentional Actions and Side-Effect Acts,” Analysis 63, pp. 190- 94. Malle, Bertram and Joshua Knobe (1997), “The Folk Concept of Intentionality,” Journal of Experimental Psychology 33, pp. 101-21. McCann, Hugh (1991), “Settled Objectives and Rational Constraints,” American Philosophical Quarterly 28, pp. 25-36. Reprinted in Mele (1997), ch. 9. McCann, Hugh (1974), “Volition and Basic Action,” The Philosophical Review 83, pp. 451-73. Meiland, J. W. (1963), “Are There Unintentional Actions?” The Philosophical Review 72 (3), pp. 377-81. Mele, Alfred (2003), Motivation and Agency (Oxford: Oxford University Press). Mele, Alfred (1997), The Philosophy of Action (Oxford: Oxford University Press). Mele, Alfred (1992), Springs of Action (Oxford: Oxford University Press). 167 Mele, Alfred and Paul Moser (1994), “Intentional Action,” Nous 28, pp. 39-68. Reprinted in Mele (1997), ch. 10. Mele, Alfred and Steven Sverdlik (1996), “Intention, Intentional Action, and Moral Responsibility,” Philosophical Studies 82 (3), pp. 265-87. Nagel, Thomas (1970), The Possibility of Altruism (Oxford: Clarendon Press). O’Shaughnessy, Brian (1980), The Will (Cambridge: Cambridge University Press). O’Shaughnessy, Brian (1973), “Trying (as the Mental ‘Pineal Gland’),” Journal of Philosophy 70, pp. 365-86. Reprinted in Mele (1997), ch. 3. Paul, Sarah (2009), “How We Know What We’re Doing,” Philosophers’ Imprint 9, pp. 1- 24. Pears, David (1985), “Intention and Belief,” in Essays on Davidson, eds. Bruce Vermazen and Merrill B. Hintikka (Oxford: Clarendon Press), pp. 75-88. Pears, David (1984), Motivated Irrationality (Oxford: Clarendon Press). Pears, David (1980), “Intentions as Judgements,” in Philosophical Subjects; Essays Presented to P. F. Strawson, ed. Zak van Straaten (Oxford: Clarendon Press), pp. 222-37. Pears, David (1964), “Predicting and Deciding,” Proceedings of the British Academy 50, pp. 193-227. Pettit, Dean and Joshua Knobe (2009), “The Pervasive Impact of Moral Judgment,” Mind & Language 24 (5), pp. 586-604. Rankin, K. W. (1972), “The Non-Causal Self-Fulfillment of Intention,” American Philosophical Quarterly 9, pp. 279-89. Ryle, Gilbert (1949), The Concept of Mind (Chicago, IL: The University of Chicago Press). Salmond, Sir John (1930), Jurisprudence (London: Sweet & Maxwell). Sellars, Wilfrid (1966), “Thought and Action,” in Freedom and Determinism, ed. Keith Lehrer (New York: Random House), pp. 105-39. Setiya, Kieran (2007a), “Cognitivism about Instrumental Reason,” Ethics 117, pp. 649- 73. 168 Setiya, Kieran (2007b), Reasons without Rationalism (Princeton, NJ: Princeton University Press). Tannenbaum, David, Peter Ditto and David Pizarro (unpublished), “Different Moral Values Produce Different Judgments of Intentional Action.” Taylor, Richard (1966), Action and Purpose (Englewood Cliffs, NJ: Prentice-Hall). Velleman, David (1989), Practical Reflection (Princeton, NJ: Princeton University Press). Wasserman, Ryan (forthcoming), “Intentional Action and the Unintentional Fallacy,” Pacific Philosophical Quarterly. Yaffe, Gideon (2010), Attempts (Oxford: Oxford University Press). Yaffe, Gideon (2006), “Trying, Intending and Attempted Crimes,” Philosophical Topics 32, pp. 505-32. Yaffe, Gideon (2001), “Locke on Refraining, Suspending, and the Freedom to Will,” History of Philosophy Quarterly 18 (4), pp. 373-91.

Abstract (if available)

Abstract My dissertation explores the nature and extent of intentional action. I begin by arguing for a necessary connection between intentional action and the state of intention. I then argue that this commonsense view—known as the Single Phenomenon View—is inconsistent with two highly influential theories of intention

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The mental states first theory of promising

PDF

Language and the structure of Berkeley's world

PDF

Describing authority

PDF

The minimal approval view of attributional-responsibility

PDF

Vagueness, legal content, and legal interpretation

PDF

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

PDF

Thomas Reid: motives and the anatomy of the mind

PDF

Crime and punishment: examining proportionality issues within criminal law

PDF

Contrastive reasons

PDF

Units of agency in ethics

PDF

The constitution of action

PDF

The case for moral skepticism

PDF

Thomas Reid on singular thought

PDF

Role of emotion and empathy in moral judgments of real-life situations

PDF

Reasoning with uncertainty and epistemic modals

PDF

The dynamics of reasonable doubt

PDF

Intrinsic and extrinsic values of distributional equality

PDF

Reasons, obligations, and the structure of good reasoning

PDF

Contract, from promise to commodity

PDF

Just imagining things: Hume's conception-based account of cognition

Asset Metadata

Creator Johns, Brandon Darryl (author)

Core Title The boundaries of the intentional

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Philosophy

Publication Date 03/27/2012

Defense Date 12/07/2011

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

Tag action theory,intention,intentional action,negative action,non-intentional action,OAI-PMH Harvest,philosophy of mind,refraining,unintentional action

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Yaffe, Gideon (committee chair), Finlay, Stephen (committee member), Lyon, Thomas D. (committee member), Watson, Gary L. (committee member)

Creator Email bjohns43@gmail.com,brandodj@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-1068

Unique identifier UC11288154

Identifier usctheses-c3-1068 (legacy record id)

Legacy Identifier etd-JohnsBrand-554.pdf

Dmrecord 1068

Document Type Dissertation

Rights Johns, Brandon Darryl

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