J Ethics (2012) 16:175–209 DOI 10.1007/s10892-012-9124-8
Reason, Responsibility, and Free Will: Reply to My Critics Ishtiyaque Haji
Received: 17 November 2011 / Accepted: 17 November 2011 / Published online: 13 May 2012 Springer Science+Business Media B.V. 2012
Abstract This paper highlights and discusses some key positions on free will and moral responsibility that I have defended. I begin with reflections on a Strawsonian analysis of moral responsibility. Then I take up objections to the view that there is an asymmetry in freedom requirements for moral responsibility and moral obligation: obligation but not responsibility requires that we could have done otherwise. I follow with some thoughts on the viability of different sorts of semi-compatibilism. Next, I turn to defending the ‘‘luck objection’’ to a popular libertarian account of the control that responsibility requires. This is, roughly, the objection that when our decisions are indeterministically caused, their occurrence is a matter of responsibility-undermining luck. Finally, I comment on Frankfurt examples. Keywords Frankfurt-examples Luck objection Peter Strawson Semi-compatibilism
I begin with a summary of some of the key themes pertaining to moral responsibility and moral obligation that I have developed and defended. To be morally responsible for an action or a choice is to be morally blameworthy or praiseworthy, whatever the case may be, for it. According to the analysis of moral blameworthiness I favor, an agent is morally to blame for an action provided it is free, it is performed in light of the nonculpable belief that it is wrong, and it causally issues from antecedents, such as desires, beliefs, and values with respect to which its agent is autonomous.1 Briefly, when a person is blameworthy for something, she willingly and
1
I use ‘wrong’ and ‘impermissible’ interchangeably as I do ‘right’ and ‘permissible.’
I. Haji (&) Calgary, Canada e-mail:
[email protected]
123
176
I. Haji
autonomously does what she nonculpably takes to be wrong. Similarly, an agent is praiseworthy for an action provided that she freely does it on the basis of the nonculpable belief that she is doing something morally permissible or obligatory, and its relevant causal antecedents are ‘‘truly her own’’ (Haji 1998). This analysis suggests three conditions of moral responsibility: a metaphysical one having to do with freedom, an epistemic one having to do with beliefs and intentions, and an authenticity one having to do with when pertinent causal precursors of one’s actions are truly one’s own or authentic. Regarding the first, determinism is the view that, at any instant, there is exactly one physically possible future (van Inwagen 1983, p. 3). To have free will with respect to an action is to have the ability both to do it and to refrain from doing it. Semi-compatibilism is the doctrine that responsibility is compatible with determinism even if determinism is incompatible with our having free will. I am partial to a reasons-responsiveness account of control which is consistent with semi-compatibilism. Concerning the epistemic condition, I have defended the view that moral blameworthiness requires nonculpable belief in what is morally wrong and not what is in fact morally wrong (Haji 2002, 2009a, b). Analogously, praiseworthiness requires nonculpable belief in what is right or obligatory and not what is in fact right or obligatory. As for the authenticity condition, I have argued for ‘‘historicist’’ constraints which imply that how one’s springs of action are acquired makes a difference to moral responsibility. Regarding this condition, I have underscored the view that it is important to distinguish between two stages in moral development, roughly, the stage before we are fully formed free agents and the post childhood stage when we are such agents. The historical constraints must be sensitive to agential differences in these stages (Haji and Cuypers 2008). I have, in addition, defended the view that although there is good reason to believe that the truth of judgments of other varieties of moral appraisal are also conceptually tied to freedom, there is no prima facie reason to think that their freedom-relevant conditions mirror the one for moral responsibility (Haji 2012). In particular, I have argued that, unlike moral responsibility, there is an alternative possibilities requirement for moral and prudential obligation: nothing, for example, can be morally obligatory for a person unless she has free will with respect to it. Lately, I have defended the view that the truth of various judgments of intrinsic value, such as the judgment that Ish takes intrinsic attitudinal pleasure in reading Al’s, John’s, Derk’s, and Michael’s work, also presupposes that we have free will, and so does our having various moral sentiments (2009a, b, 2012). Finally, I continue to maintain that the indeterministic causation of our decisions imperils responsibility for them when these decisions themselves and not, for instance, some of their immediate causal antecedents that are not actions, are caused indeterministically. In what follows, I begin with reflections on a rival Strawsonian analysis of moral responsibility. Then I take up various objections to the view that there is an asymmetry in freedom requirements for moral responsibility and moral obligation: whereas responsibility does not require that we could have done otherwise, obligation does presuppose our having alternatives. I follow with some thoughts on the viability of different sorts of semi-compatibilism. Next, I turn to defending the ‘‘luck objection’’ to a popular libertarian account of the control that responsibility
123
Reason, Responsibility, and Free Will
177
requires. This is, roughly, the objection that when our decisions are indeterministically caused, their occurrence is a matter of responsibility-undermining luck. Finally, I comment briefly on Frankfurt examples.
1 On the Strawsonian Analysis The analysis of responsibility I endorse falls into the ‘‘ledger’’ as opposed to the ‘‘Strawsonian’’ camp. On the ledger view, to be morally responsible is to be such that one’s moral standing or record as a person is affected by some episode in, or aspect of, one’s life.2 The variation of this view I support implies that when one is praiseworthy, one’s moral standing has been enhanced owing to one’s freely doing what one takes to be right or obligatory; and when blameworthy, one’s moral standing has been diminished because one freely and willingly does what one takes to be wrong. The competing view of Peter Strawson has it that to be morally responsible just is to be the appropriate object of one or more of the reactive attitudes such as gratitude, resentment, indignation, forgiveness, and the like (Strawson 1962, 1–25).3 On one construal, the Strawsonian analysis amounts to the following: Strawsonian Analysis-1: S is morally responsible for A = df. it is fitting to adopt some reactive attitude toward S in respect of A.4 There is a question about whether this analysis correctly represents Strawson’s view of the nature of responsibility. I set this query aside and turn to evaluative concerns. Michael Zimmerman argues that the analysis is susceptible to the following sort of objection frequently raised against buck-passing accounts of intrinsic value of this genre: x is intrinsically good = df. it would be fitting for anyone who were to contemplate x to favor x for its own sake (Zimmerman 2010). Suppose a powerful demon threatens that if you do not express anger toward Ish for performing some act, the demon will torture you. It seems fitting for you to express anger toward Ish for performing this act. But contrary to what Strawsonian Analysis-1 implies, Ish need not be morally responsible for the act. To evade this problem, one may want to distinguish ‘‘fitting’’ versions from ‘‘having reasons’’ versions or from ‘‘being deserving of’’ versions of the analysis: Strawsonian Analysis-2: S is morally responsible for A = df. there is a reason to adopt some reactive attitude toward S in respect of A. Strawsonian Analysis-3: S is morally responsible for A = df. S deserves some reactive attitude in respect of A. The example involving the demon, however, seems to undermine Strawsonian Analysis-2 as well (but not Strawsonian Analysis-3). There is a second concern with Strawsonian Analysis-2: unlike the variation of the ledger view I defend, it cannot 2
See, for e.g., Glover (1970), Feinberg (1970), Morris (1976), Zimmerman (1988), and Haji (1998).
3
There is interesting debate on just which sentiments, emotions, or attitudes constitute the reactive attitudes. See, for instance, Wallace (1994).
4
See, also, Wallace (1994), Copp (1997), Fischer and Ravizza (1998).
123
178
I. Haji
underpin semi-compatibilism. To develop this problem let us start by differentiating different sorts of reasons. Davidsonian reasons are, typically, complexes of desires and beliefs (Davidson 1963). One may think of such reasons as ‘‘internal’’ because they are constituted by ‘‘internal’’ mental states (or, perhaps, their physical realizers) of the agent whose Davidsonian reasons they are. In contrast, objective reasons (I use ‘objective’ and ‘external’ interchangeably) are, roughly, facts dissociated from the agent’s desires or other intrinsically motivating attitudes. More precisely, I use ‘motivating desire’ to refer to attitudinal states of mind (or their neural realizers) that are, constitute, or include motivation (Mele 2003, pp. 14–15).5 If R is an objective reason for S to do something, R is not a motivating desire of S, does not have a motivating desire of S as a constituent, and is not, even in part, a fact, truth, proposition, or the like about any actual motivating desire of S (Mele 2003, p. 77). Something’s being intrinsically good or someone’s recognizing that something is intrinsically good, for instance, is an objective reason for one to desire or pursue it. Pro tanto reasons are reasons that other reasons can outweigh, as opposed to allthings considered reasons (or ‘‘oughts’’), which cannot be outweighed. If the term ‘reason’ means pro tanto reason, each reason has a certain weight. Suppose you have several different alternatives, and your pro tanto reasons to act in some way are stronger than your reasons to act in any other way. Then you have most reason to act in this way; you reasons-wise ought to act in this way. Suppose you have sufficient pro tanto reason to act in two or more ways, and no better reason to act in any other way. Then it is reasons-wise permissible for you to act in either of these ways. Finally, suppose you have most pro tanto reason not to act in a certain way. Then acting in this way is reasons-wise forbidden for you.6 Objective pro tanto reasons contrast with such subjective reasons: an agent has a subjective pro tanto reason to do something if and only if she believes that she has an objective pro tanto reason to do it. Some Davidsonian reasons are normative (as opposed to being) merely explanatory reasons, partly, because they have apt subjective reasons as their belief component. For example, suppose you nonculpably believe that you have most reason—you have external pro tanto reasons—to administer medicine A to a patient because you nonculpably believe, on excellent authority, that giving A will save this patient. You believe, too, that you morally ought to administer A. However, you have decisive external pro tanto reasons against giving A because doing so will kill the patient. Suppose, further, that you desire to cure the patient, and you give A partly on the basis of the nonculpable beliefs that you can cure by giving A, and that you morally ought to cure. This desire, together with these beliefs, constitute a normative Davidsonian or internal reason to give A. I have argued that there is an alternative possibilities requirement for our having objective pro tanto reasons. It will be helpful to outline this argument. (Hereafter in this section, unless otherwise specified ‘reasons’ refers to objective pro tanto reasons.) Just as there is an association between the ‘‘ought’’ of morality and ‘‘can,’’ there is a similar association between the ‘‘ought’’ of (objective) reason and ‘‘can.’’ 5
Mele uses ‘motivation-encompassing attitude’ to refer to such states.
6
An instructive paper on, among other things, pro tanto reasons is Broome (2004).
123
Reason, Responsibility, and Free Will
179
Indeed, the moral ‘‘ought’’ implies ‘‘can’’ principle (Kant’s Law) is just a more restricted version of the following general principle: Reasons-Wise ‘‘Ought’’ Implies ‘‘Can’’ (Reason Ought/Can): If one has most reason to do something, A, and, thus, if one reasons-wise ought to do A, then one can do A. If reasons-wise ‘‘ought’’ implies ‘‘can,’’ I see no reason to deny that reasons-wise ‘‘wrong’’ (and reasons-wise ‘‘right’’) imply ‘‘can’’ as well. Here, I merely adumbrate two reasons why ‘‘wrong’’ and ‘‘right’’ imply ‘‘can.’’ First, starting with a comparison, moral praiseworthiness and blameworthiness require freedom (or control). An essential element of the freedom requirement of these responsibility appraisals is captured by the principle that one is morally praiseworthy or blameworthy for doing something only if one could have done it. It would be implausible to suppose that while praiseworthiness requires control, blameworthiness does not (or vice versa). This principle highlights a link that holds of conceptual necessity between moral responsibility and freedom. Again, the link is simply that praiseworthiness and blameworthiness both require control. Moreover, if we think that praiseworthiness requires a certain variety of control, then without good reason to believe the contrary, blameworthiness, too, requires this variety of control. Similarly, obligation, wrong, and right require control. As with responsibility appraisals, it is implausible to suppose, for instance, that while obligation requires control, wrong or right do not. Furthermore, it would seem that the control requirements of moral obligation, unless we have strong reason to think otherwise, should also be the very ones of moral wrong and right. Kant’s Law can be conceptualized as a control principle for moral obligation: if you ought to perform an action, you have obligation–relevant control in performing it. If Kant’s Law expresses just one more incarnation of the association between morality and freedom, then, again, in the absence of special reason to believe otherwise, it should also be the case that the principles that ‘‘wrong’’ implies ‘‘can’’ and ‘‘right’’ implies ‘‘can’’ express two other instances of this association. In sum, if obligation— whether it is reasons-wise obligation or moral obligation—requires control, so do right and wrong; and if obligation requires a certain sort of basic control—that we can do what we are obligated to do—right and wrong require this sort of basic control as well. We may now argue as follows. If it is reasons-wise wrong for one to do something, one reasons-wise ought not to do it. If one reasons-wise ought not to do something, one can refrain from doing it. Hence, if it is reasons-wise wrong for one to do something, one can refrain from doing it. But it is also true that if it is reasonswise wrong for one to do something, one can do it. So, there is an alternative possibilities requirement for reasons-wise ‘‘wrong.’’ Regarding reasons-wise obligation, if one reasons-wise ought to refrain from doing something, it is reasons-wise wrong for one to do it. Furthermore, if it is reasons-wise wrong for one to do something, one can do it (from reasons-wise ‘‘wrong’’ implies ‘‘can.’’) Therefore, if one reasons-wise ought to refrain from doing something, one can do it. But it is also true that if one reasons-wise ought to refrain from doing something, one can refrain from doing it. In other words, just as there is a requirement of
123
180
I. Haji
alternative possibilities for reasons-wise wrongness, so there is such a requirement for reasons-wise obligation. If reasons-wise wrongness and reasons-wise obligation require alternative possibilities, I see little reason to deny that reasons-wise rightness, too, requires alternatives. We may conclude that there is an alternative possibilities requirement for the truth of judgments of objective pro tanto reasons. Second, a powerful analysis of the concept of moral obligation validates the symmetry I propose in the freedom requirements of obligation, wrong, and right (Feldman 1986; Zimmerman 1996, 2008). The analysis provides a plausible treatment of a wide array of deontic puzzles, sometimes partly in virtue of implying that ‘‘wrong’’ implies ‘‘can.’’ The account builds on the idea that at each time of moral choice, there are several possible worlds accessible to a person as of that time: there are, at the time, various ways in which a person might live out her life. For each of these complete ‘‘life histories,’’ there is a possible world—the one that would exist if she were to live out her life in that way. A possible world is accessible to a person at a time if and only if it is still possible, at that time, for the person to see to it that the world is actual.7 Making use of the notion of accessibility, one can say that a state of affairs is possible for a person as of a time if and only if it occurs in some world still accessible to the person at that time. On this analysis, actions are morally judged on the basis of the values of the accessible possible worlds in which they are performed. Worlds may be ranked in accordance to a value-relation; each world is as good as, or better than, or worse than, each other world. A world is best if no world is better than it is. For purposes of ‘‘value-wise’’ ranking worlds, one can supply one’s favorite axiology. I simply label the relevant value ‘‘deontic value.’’ The analysis can now be stated in this way: (MO): A person, S, ought, as of t, to see to the occurrence of a state of affairs, p, if and only if p occurs in some world, w, accessible to S at t, and it’s not the case that *p occurs in any accessible world deontically as good as or deontically better than w.8 More intuitively, and simplifying, according to MO, as of some time, an act is morally obligatory for you if and only if you can do it and it occurs in all the best worlds accessible to you at this time. As of some time, an act is morally permissible for you if and only if you can do it and it occurs in some but not all the best worlds accessible to you at this time. And, as of some time, an act is wrong for you if and only if you can do it and it does not occur in any of the best worlds accessible to you at this time. MO verifies a version of Kant’s Law. If, at some time, you ought to do something that occurs in all the best worlds accessible to you at this time, and accessible-to-you worlds are worlds you can make actual, then you can do that thing. Similarly, given MO, if, at some time, p is wrong for you, p occurs in some world that is accessible to
7
See Feldman (1986, pp. 16–25) for more on accessibility.
8
See, for e.g., Feldman (1986, p. 37). Zimmerman constructs and defends an analysis similar to Feldman’s in his 1996, chap. 2. In his 2008 Zimmerman advances a different analysis but one which still validates the ‘‘wrong’’ implies ‘‘can’’ thesis.
123
Reason, Responsibility, and Free Will
181
you but not in any of the best worlds accessible to you. So, on MO, ‘‘wrong’’ implies ‘‘can’’ as well.9 Reverting, now, to reasons-wise obligation, the reasons-wise ‘‘ought’’ implies ‘‘can’’ principle, just like Kant’s Law expresses the control that ‘‘obligations’’ of reason require. With respect to their control requirements, there is no reason to believe that the control requirements of the moral ‘‘ought’’ differ from those of the reasons ‘‘ought.’’ As I previously ventured, Kant’s Law is just a special case of the general principle that reasons-wise ‘‘ought’’ implies ‘‘can.’’ Again, precluding compelling reasons to think otherwise, if reasons-wise ‘‘ought’’ requires a species of control, reasons-wise ‘‘right’’ and reasons-wise ‘‘wrong’’ require this very species of control as well. We may now argue as follows: If it is reasons-wise wrong for one to do something, one reasons-wise ought not to do it. If one reasons-wise ought not to do something, one can refrain from doing it. Hence, if it is reasons-wise wrong for one to do something, one can refrain from doing it. But it is also true that if it is wrong for one to do something, one can do it. So, there is an alternative possibilities requirement for ‘‘wrong.’’ Consider, next, reasons-wise obligation. If one reasons-wise ought to refrain from doing something, it is reasons-wise wrong for one to do it (from reasons-wise ‘‘ought not’’ is equivalent to reasons-wise ‘‘wrong’’). Moreover, if it is reasons-wise wrong for one to do something, one can do it (from reasons-wise ‘‘wrong’’ implies ‘‘can.’’) Therefore, if one reasons-wise ought to refrain from doing something, one can do it. But it is also true that if one reasons-wise ought to refrain from doing something, one can refrain from doing it. In other words, just as there is a requirement of alternative possibilities for reasons-wise wrongness, so there is such a requirement for reasons-wise obligation. If reasons-wise wrongness and reasons-wise obligation require alternative possibilities, I see little reason to deny that reasons-why rightness, too, requires alternatives. We may conclude that there is an alternative possibilities requirement for our having objective pro tanto reasons. We may now revert to Strawsonian Analysis-2: Strawsonian Analysis-2: S is morally responsible for A = df. there is a reason to adopt some reactive attitude toward S in respect of A. There is nothing to preclude the relevant reason in a case such as the demon’s case, for example, from being an objective pro tanto reason. Although you may not have any relevant desires and beliefs that are constituents of a Davidsonian reason of yours to express anger toward Ish, there is a reason for you to express anger. Then, however, what may be dubbed ‘‘Strawsonian semi-compatibilism’’ is endangered because there is an alternative possibilities requirement for our having objective pro tanto reasons. What of the desert version, Strawsonian Analysis-3? According to this version, moral responsibility is to be understood in terms of being deserving of some reactive attitude. The ground of responsibility is desert. It is someone’s being deserving of a 9
I consider and respond to objections to the view that ‘‘wrong’’ implies ‘‘can’’ elsewhere (Haji 2012, chap. 2). I do not have the space to reproduce that discussion here.
123
182
I. Haji
reactive attitude that grounds moral responsibility; the desert comes first, as it were. But it is not clear, as Zimmerman cautions, that this is true (2010, pp. 103–15). It might reasonably be claimed that it is moral responsibility that comes first: it is one’s responsibility that grounds desert of the reactive attitudes. That is, it is because an agent is morally responsible for some action that this agent deserves some reactive attitude with respect to this action. I want to flag another concern with the desert version. The fundamental thought is that, necessarily, if someone deserves something, there is a reason to see to it that he or she receives or gets this thing. If Tammy is born with a serious congenital impairment, Tammy deserves sympathy, or help, or comfort. If, in a fit of rage, Jack destroyed the windshield of Bennie’s truck, Bennie deserves compensation. If Pam’s young daughter unexpectedly died in a car accident, Pam deserves compassion. In each of these examples involving desert, there is a reason someone has to see to see to it that the person who deserves some thing receives this thing. There is a reason for Tammy to receive sympathy or help; there is a reason for Bennie to receive compensation; there is a reason for Pam to receive compassion. We may safely assume that in many such cases in which there is a reason deriving from considerations of desert—a ‘‘desert-based’’ reason—the reason is an objective pro tanto one. Even if Jack, or anyone else for that matter, has no Davidsonian reason to compensate Bennie, at least someone has a reason to compensate Bennie, and this reason, plausibly, is an objective pro tanto one. That there are desert-based reasons threatens Strawsonian semi-compatibilism on desert versions of the Strawsonian analysis. If, in at least some cases, desert-based reasons are objective pro tanto, then, because there is an alternative possibilities requirement for such reasons, analyses such as Strawsonian Analysis-3 will be committed to our having alternatives.
2 On Determinism’s Undermining Obligation It is now relatively straightforward to discern that moral right, wrong, and obligation require alternatives. There are different pathways to this conclusion. Here is one: obligations are tied to reasons in this way: Obligation-Reasons: If an agent has a moral obligation to do something, A, then the agent has a pro tanto reason to do A. I do not know what sort of argument may be advanced to support this principle because it seems as basic as the principle that obligation and responsibility require control. It may simply be that obligations of any kind are conceptually linked to reasons of the given kind, moral obligations to moral reasons, for instance (Vranas 2007, pp. 172–73). Furthermore, the reasons in question are objective pro-tanto and not subjective or Davidsonian. This is because the view (roughly) that some things are morally wrong or obligatory for an agent irrespective of what desires or beliefs the agent has is compelling.10 You may believe, on the evidence available to you, 10 I realize that this claim would be rejected by those people who think that pro tanto reasons in some way depend on desires. See, for e.g., Schroeder (2007) and Williams (1981).
123
Reason, Responsibility, and Free Will
183
that giving a certain medicine to a sick patient will cure the patient. But if giving this medicine will in fact kill the patient, you do wrong in giving it. You do wrong despite your subjective reason: you believe, let us assume, that you have an objective pro tanto reason to give this medicine, whereas in fact you have no such reason. Indeed, you have a decisive objective pro tanto reason not to give it. Similarly, you do wrong despite your pertinent Davidsonian reasons: you desire to cure the patient and you believe that you can cure the patient by administering this medicine; your having of this desire and belief, in conjunction with other pertinent antecedents of action causally and nondeviantly issues in your giving the medicine. None of this, however, need tell against your act not being wrong for you.11 Obligation-Reason says that if one has a moral obligation to do something, one has an objective pro tanto reason to do it. If one has such a reason to do something, however, one could have done otherwise. Hence, if one has a moral obligation to do something, one could have done otherwise. We may easily derive an analogous result for moral wrong and right. Another, perhaps, simpler route to the conclusion that moral obligation requires free will mimics the route for the conclusion that (objective) reasons-wise ‘‘ought’’ requires free will: If it is morally wrong for one to do something, then one ought not to do it; if one ought not to do something, then one can refrain from doing it. So, if it is wrong for one to do something, then one can refrain from doing it. As ‘‘wrong’’ implies ‘‘can,’’ too, wrong requires alternatives. Regarding obligation, if one ought not to do something, one can refrain from doing it. But if one ought not to do something, then it is also true that it is wrong for one to do it, and, hence (given that ‘‘wrong’’ implies ‘‘can’’), one can do it. Strong alternatives are alternatives incompatible with determinism. Loosely, one has a strong alternative to doing something provided that, given the same past and the natural laws, one could have done otherwise. Weak alternatives are alternatives we can have even if determinism obtains. If obligation requires strong alternatives, then as determinism effaces such alternatives, determinism is incompatible with obligation. Many have advanced instructive criticism of this sort of argument. I start with some of Nick Trakakis’ observations (2008). To set the stage, Derk Pereboom defends hard incompatibilism, the view that as long as no event is agent-caused, regardless of whether determinism obtains or fails to obtain, no one is ever responsible for any event. Pereboom (2001) has taken issue with my view that determinism is incompatible with obligation (provided the alternatives are strong). I responded to these criticisms in Deontic Morality and Control. Trakakis usefully adjudicates this dispute. He considers what he takes to be three of Pereboom’s arguments against my relevant views on hard incompatibilism and obligation, and my responses to them. He concludes that the first of Pereboom’s arguments succeeds ‘‘in providing us with some reason to doubt the claim that if hard 11 Some may, of course, say that if a doctor gives a medicine to a patient that the doctor sincerely and responsibly believes will cure the patient, what the doctor does is not wrong even if it turns out that the medicine unexpectedly kills the patient. Intuitions about these sorts of cases can conflict. I am inclined to claim that the doctor does objective wrong (but is not blameworthy); and that the doctor fulfills his subjective obligation: she does what she believes she has an objective obligation to do.
123
184
I. Haji
determinism were true, none of our actions would be morally wrong’’ (Trakakis 2008, p. 257). Regarding this first argument Pereboom and Trakakis both favor this ‘‘intuitively true principle’’ that connects ‘‘moral goodness with moral rightness’’: GR: Sometimes, actions that bring about the greatest good overall in worlds accessible to S are right for S (Trakakis 2008, p. 53). Trakakis notes that the line of reasoning I sketched above for the view that determinism threatens obligation has nothing explicitly to say about moral goodness and moral badness. So, he judges, I have not given reasons to think that determinism is incompatible with moral goodness, although I have advanced some considerations to think that determinism undermines moral obligation, right, and wrong. Trakakis writes: This, however, commits Haji to an invidious demarcation between judgements about moral goodness and judgements about moral rightness. In other words, if we assume the truth of D [D stands for the proposition that if hard determinism were true, then no one could do otherwise]…and if we follow Haji in accepting (K [Kant’s Law]) and OW [the principle that ‘‘ought not’’ is equivalent to ‘‘wrong’’], then we must also accept the following unintuitive result: …GR is false. But, Pereboom adds, it is far from clear that Haji’s position, as laid out in (I) below, (I) Accept: D [if determinism were true, then no one could do otherwise] & K [Kant’s Law] & OW [‘‘ought not’’ amounts to ‘‘wrong’’] & goodness & blameworthiness and therefore reject: obligation & rightness/wrongness & GR is at all superior to a view (like Pereboom’s) that bids us to (II) Accept: D & K & rightness/wrongness & goodness & GR and therefore reject: blameworthiness & obligation & OW. Each position has unintuitive results, but it is not obvious that position (I) is in better shape than (II). And so, at the very least, we are faced with a kind of de´tente. (Trakakis 2008, pp. 53–54) Suppose, initially, that determinism is no threat to (intrinsic) goodness, something I believe is less than obvious. If GR is the ‘‘difference maker’’ between (I) and (II), i.e., setting aside the other elements of (I) and (II), if whether one is to be preferred to the other turns pivotally on whether either of these positions is consistent with GR (or its basis), then (I) mischaracterizes my view, as I now explain.12 I agree with Trakakis and Pereboom that what I take to the primary basis for affirming GR, namely, that there is a conceptual connection between the right and the good, is intuitively true. What precisely, however, is this connection? I propose that the intuitive attractiveness of GR derives from the worldly analysis of the concept of obligation previously adumbrated: as of some time, S ought to do A if and only if S can do A, and any accessible world in which S refrains from doing A is
12
Rejecting OW—the principle that ‘‘ought not’’ is equivalent to ‘‘wrong’’ strikes me as implausible.
123
Reason, Responsibility, and Free Will
185
deontically inferior to some accessible world in which S does A.13 Thus, I can accept something very close to GR: any action which is such that as of some time, S can do it, and it occurs in some but not all the best worlds accessible to S at this time is permissible (i.e., right) for S. In addition, there are grounds to believe that various judgments of intrinsic goodness also presuppose our having alternatives. Here, I call attention to only one. Use ‘favor’ as an umbrella term to cover any sort of pro-attitude toward whatever it is that is favored. If something is intrinsically good, then anyone who were to contemplate it would have a reason to favor it for its own sake. Again, I think that these reasons are objective pro tanto. An objective pro tanto reason to favor something is not relevantly different from such reason to do something in that, with a reason of this sort for either sort of thing, it is not possible to have a reason to favor something or to do something if one cannot favor it or do it. Without justification to believe otherwise, it is implausible to suppose that this asymmetry obtains: whereas one cannot have an objective pro tanto reason to do something unless one can do it, one can have such a reason to favor something despite one’s not being able to favor it. So, assume that the principles reasons-wise ‘‘ought’’ implies ‘‘can,’’ reasons-wise ‘‘ought not’’ implies ‘‘can refrain from,’’ and reasonswise ‘‘ought not’’ is equivalent to reasons-wise ‘‘wrong’’ ‘‘apply’’ even when it is favoring that is of concern.14 Then it will be true that if you have a reason to favor something, you can refrain from favoring it. We may now invoke this argument to show that the truth of (certain sorts of) judgments of intrinsic value requires our having alternatives: If something is intrinsically good, one has an objective pro tanto reason to favor it; if one has such a reason to favor something, then one could have refrained from favoring it; so, if something is intrinsically good, one had relevant alternatives.
3 On ‘‘Ought’’ Implies ‘‘Can’’ My argument that there is an alternative possibilities requirement for obligation invokes Kant’s Law. Arpaly (2006) has recently launched an interesting challenge against this law. Her reasoning rests primarily on certain putative similarities between the moral ‘‘ought’’ and what may be termed the ‘‘epistemic ‘ought.’’’ Regarding the latter, Arpaly claims that although epistemology is normative, we have no freedom over belief formation—we cannot ‘‘believe otherwise’’ than we do (Arpaly 2006, p. 92). She proposes, however, that it is still ‘‘possible for normative truths to apply to things over which we have no freedom’’ (Arpaly 2006, p. 92). As one illustration of this point, she gives an example in which a student draws the unwarranted conclusion that Homer lived in the first century from having been told that Homer was the first Greek poet. Given various facts about this student, Arpaly 13 There are good reasons, that need not detain us, to amend the analysis in this way: as of some time, S ought to do A if and only if S can do A, S can refrain from doing A, and any accessible world in which S refrains from doing A is deontically inferior to some accessible world in which S does A. 14
See, however, the brief discussion on fittingness or requirement in the next section.
123
186
I. Haji
plausibly claims that the student ought to have known better; she made a judgment that she ought not to have made. Arpaly adds that the student had good reasons for not believing that Homer lived in the first century and that in making her inference ‘‘she failed to respond to epistemic reasons’’ (Arpaly 2006, p. 92). So, unlike cell metabolism, the ‘‘development of beliefs is generally a reason-responsive process (though not universally so, of course)’’ (Arpaly 2006, p. 93). Arpaly emphasizes that ‘‘unless one denies the existence of content efficacy, one cannot assume that choice or control or freedom are necessary for reason responsiveness to take place; that argument can be made, mutatis mutandis, for epistemic reason responsiveness as readily as for practical reason responsiveness’’ (Arpaly 2006, p. 93). She concludes: I am perfectly willing to think of freedom in the traditional sense of selecting among alternative possibilities, and perfectly willing to accept that if determinism is true, this is something that we do not have either with respect to our beliefs or with respect to our actions. In order to think that norms for both practical and theoretical rationality apply to us, we need to believe that human beings are capable of responding to moral and epistemic reasons; we do not necessarily need freedom, however otherwise desirable it might be. (Arpaly 2006, p. 94) To appreciate better Arpaly’s position, the following passage is illuminating: There are genuine epistemic norms: people who infer from P and P ? Q and the absence of strong evidence that not Q to the conclusion that Q come to believe what they ought to believe in such circumstances, and are thereby epistemically rational; to have believed otherwise would have been to have believed something irrational, to have believed something they ought not to have believed. In spite of these normative truths, however, there are no corresponding epistemic abilities. People lack the ability to decide or choose what to believe, and so a fortiori lack the ability to decide or choose to believe what they ought to believe. Nonetheless, they ought to believe it. ‘Ought’, then, does not imply ‘can’ in the strong sense relevant to a debate between compatibilists and incompatibilists, the sense in which an obligation requires a power to bring it about that the obligation is fulfilled. Epistemic ‘ought’ implies epistemic ‘can’ in at most the weak sense that for it to be true that one ought to believe that P, one must have the conceptual resources necessary to formulate the thought that P: P must be the sort of thing that, if it occurred to one with sufficient force, one might come to believe it. This being the case, I see no reason to conclude from general principles that practical ‘oughts’ imply practical ‘cans’ in any more dramatic sense. ‘Ought’ does not in general imply ‘can.’ (Arpaly 2006, p. 106). 1.
Arpaly’s argument against Kant’s Law may be summarized as follows. The epistemic ‘‘ought’’ does not imply ‘‘can.’’ If, for instance, there is good evidence for some proposition, P, and no evidence against it, then one epistemically ought to believe P. (One may, of course, add other conditions for rational belief. One might, for instance, propose that the relevant evidence should be accessible to one in order to believe rationally.
123
Reason, Responsibility, and Free Will
2.
3. 4.
187
These additional details, however, can safely be ignored for purposes of formulating and assessing Arpaly’s argument.) In addition, one epistemically ought to believe P even though one might lack the ability to believe P. We may condense Arpaly’s relevant point in this way: given the overall evidence, it can be appropriate or fitting to believe P despite one’s not being able to believe P. The moral ‘‘ought’’ is relevantly analogous to the epistemic ‘‘ought’’: just as one epistemically ought to believe on the basis of theoretical reason—we are capable of responding to epistemic reasons—so, one morally ought to do something on the basis of practical reason—we are capable of responding to practical reason. Alternatively it may be averred that that just as it can be appropriate or fitting to believe P on the basis of relevant overall evidence despite one’s not being able to believe P, so it can be appropriate or fitting that one morally ought to do something even though one cannot do it, as long as one is suitably responsive to practical reasons. If (1) and (2), then the moral ‘‘ought’’ does not imply ‘‘can.’’ Therefore, The moral ‘‘ought’’ does not imply ‘‘can.’’
To assess this argument, as an initial remark, what might Arpaly argue in response to mundane examples meant to lend support to Kant’s Law such as the following? ‘‘Sid couldn’t have had an obligation to save Snell because Sid was tethered to the pole; in his circumstances, he simply could not have done anything for poor Snell.’’ Arpaly writes that with ‘‘regard to actions, I favor the standard compatibilist substitute for the thought that ‘ought’ implies ‘can’: one ought to take a course of action only if it is true that if one wanted to take this course of action, one could do so’’ (Arpaly 2006, p. 92). Imagine that in Modified Sid, Sid has no desire to help Snell; indeed, assume that in his situation, he could not have acquired a desire to aid Snell. According to Arpaly’s compatibilist reading of Kant’s Law— Compatibilist-Kant’s Law—if one morally ought to perform some action, A, then if one wanted to do A, one could do A. Since it is false that Sid wanted to save Snell, the conditional, if Sid wanted to save Snell, Sid could save Snell, is true. Assuming all other conditions of moral obligation are met, since Compatibilist-Kant’s Law is satisfied in Sid’s case, it follows that Sid morally ought to save Snell. But this is something that ought to be resisted. Perhaps Compatibilist-Kant’s Law should be interpreted in this way: Compatibilist*-Kant’s Law: If one morally ought to do A, then if it were the case that one wanted to do A, one could do A. Suppose Sam suffers from aqua phobia; she simply cannot enter the ocean to save Snell. Imagine, furthermore, that partly because Snell is a hated foe of hers Sam cannot, in her circumstances, acquire the desire to save Snell. Consider the counterfactual (CF): if it were the case that Sam wanted to save Snell, Sam could save Snell. Worlds in which it is true that Sam wanted to save Snell—worlds in which she acquires a desire to save Snell—are far distant from the actual world. There is little reason to believe that in these worlds, Sam suffers from aqua phobia and that she
123
188
I. Haji
loathes Snell, and so little reason to believe that in these worlds Sam could not save Snell. So there is no implausibility to the suggestion that CF is true. We may assume that in the actual world all other conditions for it to be so that Sam morally ought to save Snell are met. It would once again follow, given Compatibilist*-Kant’s Law, that Sam morally ought to save Snell, something hard to accept. Paying, now, close attention to premise (2) of Arpaly’s argument, A.C. Ewing theorized that there are differences between the ‘‘ought’’ of fittingness and the ‘‘ought’’ of moral obligation. Concerning the latter, Ewing proposed that certain value notions, such as being intrinsically good, can be analyzed in terms of the ‘‘ought’’ of fittingness along, roughly, these lines: for something to be intrinsically good is for it to have properties that make it the case that it ought (in the fittingness sense of ‘‘ought’’) to be (dis)favored. This sort of ‘‘ought,’’ Ewing proposed, is not a moral notion and, moreover, it does not imply ‘‘can.’’ However, he insisted that the moral ‘‘ought’’ does imply ‘‘can’’ (Ewing 1939, 1947). Consider, next, some of Zimmerman’s views on the ‘‘ought’’ of fittingness (or, as he prefers to say, the ‘‘‘ought’ of requirement’’). Unlike Ewing, Zimmerman believes the ‘‘ought’’ of requirement is a moral notion but he agrees that this sort of ‘‘ought’’ does not imply ‘‘can.’’ He introduces an example to flesh out the ‘‘ought’’ of fittingness: Suppose that you have gratuitously insulted Bert and that you must now make amends. Perhaps it will be most appropriate if you apologize to Bert in person. Or perhaps it will be equally appropriate if you send Bert flowers with a card expressing your remorse. It may be less appropriate, but nonetheless suitable, if you leave a brief message of apology on Bert’s answering machine. However, you will be overdoing it if you buy Bert a new TV, and it may even be positively inappropriate if you buy Bert a new car. (Zimmerman 2001, p. 92) Zimmerman then proposes that the sort of appropriateness or requirement at issue is more closely related to what is sometimes called the ‘‘ought-to-be’’ (or to what is ideal) as opposed to the ‘‘ought-to-do’’ (or the ‘‘ought’’ of obligation; Zimmerman 2001, p. 93). And he claims that whereas the ‘‘ought’’ of obligation implies ‘‘can,’’ the ideal ‘‘ought’’ does not: I don’t think that ‘‘require’’ or ‘‘appropriate,’’ as they are to be understood in the present context…imply ‘‘can.’’ For example, it would remain true that it would be most appropriate for you to send Bert flowers by way of an apology, even if there were no flowers available. (That is, it’s not just that this would be the, or a, most appropriate thing for you to do if you could; it would be the most appropriate thing period.) Similarly, it would be inappropriate for you to buy Bert a new car, even if you lacked the means to do so. (Zimmerman 2001, p. 94) This slight diversion on the ‘‘ought’’ of fittingness bears on premise (2) in this way: presumably, this ‘‘ought’’ just like the epistemic one or the ‘‘ought’’ of obligation is associated with being responsive to reasons. In principle, one can, on the basis of reasons, figure out what it is fitting to do. Both Ewing and Zimmerman, however, insist that while the ‘‘ought’’ of fittingness does not imply ‘‘can,’’ the moral ‘‘ought’’ does. All told, whether one agrees that the fittingness ‘‘ought’’ does not imply ‘‘can’’ is not directly relevant to this evaluative point. Rather, what is
123
Reason, Responsibility, and Free Will
189
pertinent is it is coherent to suppose that some sorts of ‘‘oughts’’ that are responsive to reasons imply ‘‘can,’’ but others do not. It is pretty clear that the ‘‘ought’’ to be does not imply ‘‘can’’ but it would be inappropriate to generalize from this truth to the further claim that the ‘‘ought’’ of moral obligation, too, does not imply ‘‘can.’’ There would be additional reason to think that such a generalization would be uncalled for if there were independent reason to believe that Kant’s Law is true. I suggested earlier that because the powerful worldly analysis of moral obligation previously introduced validates the ‘‘ought’’ implies ‘‘can’’ principle, there is indeed independent reason to accept Kant’s Law. In addition, it is surely plausible to grant that if one morally ought to do something, then one has a reason to do it, as it is to affirm that if one epistemically ought to believe something, then one has a reason to believe it. But it is also highly credible that one cannot have a practical reason to do something that one cannot do. For example, you cannot now have a reason to see to it that a fact about the past, such as Stefaan puffed on a Havana yesterday, is not a fact about the past. Reflect, next, on these arguments: Argument (A): (1A) If one epistemically ought to believe P, then one has an epistemic reason to believe P. (2A) If one has an epistemic reason to believe P, then one can believe P. Therefore, (3A) If one epistemically ought to believe P, then one can believe P. Argument (B): (1B) If one morally ought to do A, then one has a moral reason to do A. (2B) If one has a moral reason to do A (a reason that one has by virtue of having a moral obligation to do A), then one can do A. Therefore, (3B) If one morally ought to do A, then one can do A. Assuming (1B) is not in question (because moral obligation is conceptually tied to reasons), and that one cannot have a reason (an objective pro tanto one in the case of moral obligation) to do something that one cannot do, (3B), contrary to Arpaly, is credible. Regarding Argument A, again assuming that (1A) is acceptable, Arpaly would reject (2A). What is the relevant difference between (2A) and (2B)? I set aside (2A) and focus on (2B). If (as of some time) one morally ought to do A, then one does A in all the best worlds accessible to one at this time. So, whenever one has a moral reason to do A by virtue of having a moral obligation to do A, there is a world accessible to one in which one does A, and not just some inaccessible possible world in which one A-s. There is a sense, then, in which A is ‘‘realizable’’ by one if one has a moral reason to do A by virtue of being morally obligated to do A. But if A is, in this sense, realizable by one, then one can do A.
4 Semi-Compatibilism The discussion in the previous two sections brings to the fore the issue of how far the semi-compatibilist agenda can be extended. We may broaden our
123
190
I. Haji
conceptualization of semi-compatibilism in this way: the semi-compatibilist seeks to show that the truth of various normative judgments, such as judgments of moral responsibility, moral obligation, intrinsic value, or desert can be true independently of whether we have access to alternatives of any sort. However, the link between freedom and objective reasons does not bode well for several varieties of semicompatibilism, as I now briefly explain. Frankfurt examples have been largely instrumental to fueling semi-compatibilism regarding responsibility. These examples purport to show that a person can, for instance, be blameworthy for doing something despite not being able to do otherwise, as long as the conditions that render her unable to do otherwise play no role in bringing about her action. Many, however, who wish to endorse this brand of semi-compatibilism also accept the principle that blameworthiness requires wrongness (i.e., impermissibility): no one is morally blameworthy for something unless it is morally wrong for one to do it; and they accept, too, the principle that praiseworthiness requires permissibility or obligation: no one is morally praiseworthy for something unless it is permissible or obligatory for one to do it.15 Since there is an alternative possibilities requirement for obligation by virtue of the necessary connection between obligation and objective pro tanto reasons, this brand of semi-compatibilism is jeopardized. A way out of this quandry is to jettison the view that moral blameworthiness and praiseworthiness are tied to impermissibility and obligation (or permissibility) respectively. An alternative, as I have argued, is that these appraisals are tied to relevant nonculpable beliefs regarding what is obligatory, permissible, or impermissible. Semi-compatibilism concerning obligation, however, is another matter. Since the truth of judgments of permissibility, impermissibility, and obligation are essentially associated with our having objective pro tanto reasons, this species of semicompatibilism is imperiled. Similarly, semi-compatibilism concerning various judgments of intrinsic value is also cast into doubt. Finally, some analyses of responsibility will not support semi-compatibilism. I have already mentioned something about this regarding various proposed Strawsonian analyses. It seems that Thomas Scanlon’s interesting somewhat Strawsonianlike analysis will not, in this respect, fare any better either. Like the Strawsonian view, Scanlon’s analysis appeals to persons’ attitudes and their relations with others. The analysis is summarized in this passage: The account of blame that I offer is like Strawson’s in seeing human relationships as the foundations of blame. But it differs from his view in placing emphasis on the expectations, intentions, and other attitudes that constitute these relationships rather than on moral emotions such as resentment and indignation. Briefly put, my proposal is this: to claim that a person is blameworthy for an action is to claim that the action shows something about the agent’s attitudes toward others that impairs the relations that others can have with him or her. To blame a person is to judge him or her to be blameworthy and to take your relationship with him or her to be modified 15
See, for e.g., Fischer (1999, 2003), McKenna (2012).
123
Reason, Responsibility, and Free Will
191
in a way that this judgment of impaired relations holds to be appropriate. (Scanlon 2008, pp. 128–129, note omitted) The general idea of this ‘‘impaired relations’’ analysis is fairly easy to grasp. Suppose two people stand in some sort of relation to one another; maybe they are friends. Among other things, various attitudes, intentions, feelings, expectations, and dispositions of the two constitute this relationship. We may think of friendship as governed by, or subject to, certain norms or standards of behavior. Suppose I, as one of the parties, intentionally does something that violates one of these standards; I break trust, for instance. Such conduct expresses a certain attitude of mine—not being sufficiently sensitive about things that you have disclosed to me in confidence—that may well blight our friendship. You may not want to confide in me anymore or you may be much more reserved about what you decide to confide in me. Scanlon’s analysis (together with the germane facts of the example) entails that I am blameworthy for my conduct in virtue of its expressing attitudes toward you that impairs my relationship of friendship with you. With friendship, it is relatively clear what the ‘‘ground relationship,’’ as Scanlon states, is which provides the standards relative to which the attitudes that an agent’s action reveals constitute an impairment (Scanlon 2008, p. 138). In the general moral case, however, what is this relationship? Scanlon offers the following: [M]orality requires that we hold certain attitudes toward one another simply in virtue of the fact that we stand in the relation of ‘‘fellow rational beings.’’ It requires us to take care not to behave in ways that will harm those to whom we stand in this relation, to help them when we can easily do so, not to lie to them or mislead them, and so on. A morally good person will have standing intentions to regulate his or her behavior in these ways. These intentions concern our behavior toward people in general, not simply toward specific individuals whom we are aware of or could specify. They concern behavior toward people, whoever they may be, whom we happen to interact with in various ways, such as people who may be injured by our driving, or who ask us for directions, or who need our help in other ways….Beyond these intentions, good moral relations with others involve being disposed to have certain other attitudes. These include, in general, being disposed to be pleased when we hear things are going well for other people. (Scanlon 2008, p. 140) This time around, imagine that you and I are strangers, but I do something that violates one of the standards that govern the moral relationship—I fail to help you when I could easily do so with little cost to myself. In so failing, my conduct reveals attitudes of insufficient care that makes it appropriate for you to have attitudes toward me ‘‘different from those that constitute the default moral relation’’ (Scanlon 2008, p. 141). Again, the impaired relations view (in conjunction with apt facts) implies that I am blameworthy for my conduct owing to its disclosing attitudes toward you that impairs my moral relationship with you. Scanlon proposes that determinism poses no general threat to persons’ revealing or expressing in their conduct attitudes that mar the relations that others can have with them. In addition, it appears that the impaired relations view does not appeal to
123
192
I. Haji
our having alternatives to account for blameworthiness. As long as persons can have the relevant attitudes, and these attitudes can be revealing of whether what they do is damaging or potentially so to the pertinent relationship, they can be blameworthy for their conduct. As Scanlon, however, argues, having these attitudes and their being so revealing seems not to require that we have the capacity and opportunity to do otherwise (Scanlon 2008, pp. 180–98). So, it would appear, determinism does not imperil blameworthiness. Determinism may well be incompatible with our having free will, but the Scanlonian impaired relations view implies that determinism is not incompatible with blameworthiness. This view, then, supposedly qualifies as or supports a version of semi-compatibilism. To assess the view, let us begin with a formulation of it that the passages from Scanlon’s work I quoted in this section suggest: Impaired Relations View-1: Agent, S, is blameworthy for doing A if and only if A reveals or expresses something about S’s attitude toward others that impairs the relations that others can have with S. But Impaired Relations View-1 cannot be right. Sid’s behavior may make it abundantly clear to Sandy that he does not want to get romantically involved with her. His behavior expresses attitudes that hinder Sandy from becoming intimately involved with him. But Sid need not be blameworthy for his relevant conduct. Perhaps one thing amiss with this formulation is that it does not presuppose that the parties already are in a particular relationship, in this case, being in love with each other. When Scanlon discusses examples of friendship, or when he speaks of the moral relationship, the relevant persons are already in a particular relationship that is constituted, partly, by various standards or norms of behavior. This suggests the following: Impaired Relations View-2: Agent, S, is blameworthy for doing A if and only if S has some relationship, R (such as the moral relationship or one of friendship), with others, and A reveals or expresses something about S’s attitude toward these others that impairs this relation (R) that others can have with S. Suppose, again, that you and I are friends. I break trust—I say something about you to somebody else that I should not have said. But suppose, further, that this is a one-time occurrence, and you never discover the fault. It would seem that although I may well be blameworthy for breaking trust, the attitude of mine this behavior reveals need not impair my relation of friendship with you. Or imagine that I take myself to be doing something morally wrong; I think I am disclosing something to another person that you revealed to me in confidence. But I am not doing any such wrong. I am simply misremembering. You did in fact tell me about this thing but certainly not in confidence. Perhaps what I revealed about you to this other person may indeed benefit you. My germane behavior expresses attitudes of mine that need not impair my relation of friendship with you, but I may well be blameworthy for this behavior. Or imagine cases of undiscovered cheating. I habitually cheat on my income tax returns, but I am a careful cheater. I know how to manipulate the figures without any
123
Reason, Responsibility, and Free Will
193
real risk of discovery. My attitudes revealed in such behavior need not impair the moral relations that others can have with me. Once again, however, I may be blameworthy for my behavior. Each of the three preceding examples strongly suggests that Impaired Relations View-2 is defective, and that it is not the view Scanlon defends. In each of these cases, the attitudes revealed in my behavior do not, given the circumstances, impair the pertinent relation that others can have with me. However, it may plausibly be thought that these attitudes do give others a reason to revise or modify their relations with me. So, consider: Impaired Relations View-3: Agent, S, is blameworthy for doing A if and only if there is some relationship, R (such as the moral relationship or one of friendship), that S has with others, and A reveals or expresses something about S’s attitude toward these others that makes it appropriate, or fitting, or gives them reason to revise, or modify, or reconsider being in relation R with S. I may have misremembered that what you told me was meant to be strictly confidential. But the fact that I was willing to break trust reveals something about my attitudes toward you that gives you reason to pause about our friendship or at least question its depth. Imagine now, however, that I have come to hold myself to extremely high moral standards. I expect you, as a friend, to attempt to live up to these standards as well; my behavior toward you reveals attitudes of mine that make this clear. You are increasingly annoyed about my demands. My attitudes expressed in relevant behavior gives you reason to revise your relationship with me. But it seems that none of this need entail that I am blameworthy for the germane behavior. We might entertain a slightly different variation of this case. Despite my demands, you merrily carry on in your usual happy-go-lucky fashion. You believe that I have an overinflated sense of what morality requires. We remain very close friends. If Impaired Relations View-3 is true, then my attitudes expressed in relevant behavior still, it seems, give you, or can give you, some reason to modify or revise or review your relationship with me; you just do not take this reason to be important enough to compromise the friendship. According to Impaired Relations View-3, I am blameworthy (to a certain degree) for this behavior. But I need not be blameworthy for this behavior. So this view ought to be rejected. It will not do to modify the view in this way: Impaired Relations View-4: Agent, S, is blameworthy for doing A if and only if there is some relationship, R (such as the moral relationship or one of friendship), that S has with others, A reveals or expresses something about S’s attitude toward these others that makes it appropriate, or fitting, or gives them reason to revise, or modify, or reconsider being in relation R with S, and this reason is also a reason to evaluate S negatively, where this evaluation is a moral one. One obvious problem with this view is with its second clause. There are several different types of moral evaluations, deontic, aretaic, axiological, and so forth. Perhaps you have some reason to think that my character is flawed because I am so
123
194
I. Haji
demanding. But why should this entail that I am blameworthy for the relevant behavior? Moreover, the negative evaluation cannot be one of blameworthiness on pain of circularity. Maybe some will urge that we consider this view: Impaired Relations View-5: Agent, S, is blameworthy for doing A if and only if there is some relationship, R (such as the moral relationship or one of friendship), that S has with others, and A reveals or expresses something about S’s attitude toward these others that (i) makes it appropriate, or fitting, or gives them reason to revise, or modify, or reconsider being in relation R with S, and (ii) gives them reason to adopt or elicit some negative reactive attitude (such as indignation) toward S. But this view would not be Scanlon’s view. He thinks that negative feelings or attitudes are not essential to blameworthiness. He explains: If I ‘‘write someone off’’ as a person I am going to have nothing to do with, then I am blaming him, even if this is accompanied by no hostile feelings, perhaps because I regard him as not worth being angry at. And when someone has been ‘‘written off’’ in this way, even without anger or resentment, there is certainly something for forgiveness to do. (Scanlon 2008, p. 160, note omitted.) In addition, the first version of the case in which I am too demanding casts doubt on Impaired Relations View-5. In this case, I am a sort of perfectionist. I do not take myself to be doing any moral wrong in my demands, and, perhaps, in making my demands I am doing no moral wrong. Nonetheless, my pertinent behavior annoys you. My attitudes that this behavior expresses give you reason to be annoyed and to revise your relationship. Still, I do not see how this entails that I am blameworthy for this behavior. I will not attempt yet further revisions of the analysis. Instead, I want to make a point about versions of the analysis that appeal to reasons. This will shed light on whether the analysis is indeed consistent with semi-compatibilism. Imagine that you get wind of the fact that at a party last week, some people were talking about you, and making some cruel jokes at your expense. You further ‘‘learn that…[your] close friend Joe was at the party, and that rather than coming to…[your] defense or adopting a stony silence, he was laughing heartily and even contributed a few barbs, revealing some embarrassing facts about…[you] that…[you] had told him in confidence’’ (Scanlon 2008, p. 129). The attitudes Joe expresses in his behavior would, presumably, give you reason to modify your relation with him. You might, for example, revise your intentions to confide in him and to encourage him to confide in you (Scanlon 2008, p. 130). The reason or reasons in question may well be moral. Joe has violated a moral obligation (we may assume) by breaching trust, and this is an objective pro tanto reason for you to react in some appropriate way. The general lesson is that in many cases in which it would be appropriate to revise one’s relation with others because of the attitudes toward one that these others reveal in germane behavior, the reasons one would have to alter one’s relations are objective pro tanto. Because there is, however, an alternative possibilities requirement for our having such reasons, Scanlonian semi-compatibilism is endangered.
123
Reason, Responsibility, and Free Will
195
5 On Luck and Antecedent Proximal Control Next, I turn to some thoughts on the luck objection. Libertarians maintain that free action and moral responsibility are incompatible with determinism and that some human beings perform at least some actions, mental or otherwise, that are free and for which they are morally responsible. It is a version of event-causal modest action-centered libertarianism (henceforth ‘‘modest libertarianism’’) that is of concern in this section. Modest libertarian accounts, like their most promising compatibilist competitors on free action, require that to choose or act freely, an agent must have the capacity to engage in practical reasoning and to guide her behavior in light of the reasons that she has.16 The following distinction will be helpful in understanding these accounts: One event deterministically causes a second if and only if the first causes the second, and with the laws of nature and the past as they are, there is no chance that the first occurs without causing the second. An event indeterministically causes another if and only if the former causes the latter, and it is consistent with the laws of nature and the past that the former event occurs and not have caused the latter. Now we may add that, like their compatibilist challengers, modest libertarian accounts dictate that a free decision (or overt action) be made for reasons, and its being made for reasons consists, partially, in its being indeterministically caused appropriately and nondeviantly by the agent’s having those reasons. Libertarian views allow that an indirectly free action—an action whose freedom derives from the freedom of other actions to which this action is suitably related— may be determined by its immediate causal precursors. A directly free action is free independently of deriving its freedom from the freedom of other events. Modest libertarian theories differ from compatibilist ones in that they imply that even the immediate causal antecedents of a directly free action do not determine that action: given these antecedents, and the natural laws, there is some chance that that action not occur.17 On action-centered modest libertarian accounts, the event that is directly free and indeterministically caused is a mental action—the making of a decision (Clarke 2000, p. 23). Accounts of acting for a reason generally require that the connection between the agent’s having the reason and her action comprise, partly, the exercise of a certain degree of control by the agent. To generate a version of modest libertarianism, start with our best compatibilist view of freedom, and then to add to this ‘‘host’’ the requirement that free decisions themselves are indeterministically caused. The resulting libertarianism specifies that an agent’s control—‘‘proximal’’ or ‘‘active’’ control—in making a decision consists in apt agent-involving events causing nondeviantly that decision. The degree of active control the agent exercises depends on which agent-involving events actually cause the decision and on their etiologies. On such a libertarian view, the factors that constitute an agent’s active control in making a free decision are the very ones
16
Such accounts have been defended or discussed by Dennett (1978), Fischer (1995, 2011), Mele (1995), Kane (1996), Clarke (2000, 2003), van Inwagen (2000).
17
A recent defense of this sort of view is in Kane (1996).
123
196
I. Haji
shared by this view and its compatibilist host: roughly, deliberative processes with appropriate causal histories causing nondeviantly the decision. If modest libertarianism does not vary from its best compatibilist rival with respect to active control (as may initially be presumed), one may well wonder why some have thought that this sort of libertarianism undermines free action, responsibility, or freedom-relevant control, especially on the assumption that its best compatibilist competitor does not share these faults. Speaking to this concern regarding control, I quote from a previous work: Suppose Peg is mulling over whether to keep a promise to visit Al. She judges that, all things considered, she ought to keep the promise, though reasons of self-interest tempt her to refrain. She decides to keep the promise, and her having certain reasons to do so, including her making the all things considered judgment that she ought, on this occasion, to keep the promise,…[indeterministically] causes her to make this decision. On an action-centered libertarian view, since Peg’s decision to keep the promise is…[indeterministically] caused, there was a chance that her deliberative process would terminate in a decision not to keep the promise. Had Peg made this other decision, it would have been…[indeterministically] caused by her having reasons of self-interest. Everything prior to the decision that Peg actually makes, including every feature of Peg, might have been just the same, and yet she could have made the alternative decision instead. To underscore this point, consider the nearest possible world with the same past as the past in Peg’s world. This world will have a past in which Peg’s prior deliberations have resulted in the best judgment that the promise ought to be honored but Peg (or if we want, one of Peg’s counterparts, Peg*) decides not to keep the promise. In so deciding, Peg* acquires an intention not to keep the promise. The acquisition of this intention—the making of the decision to refrain from keeping the promise—is seemingly not explained by anything. At least, Peg*’s prior deliberations do not explain why she makes this decision. This is because these deliberations exactly mirror those of Peg’s but Peg’s deliberations…[indeterministically] give rise to the opposed decision to keep the promise. As Alfred Mele has commented, if one agent does one thing and another refrains from doing that thing, ‘‘and there is nothing about the agents’ powers, capacities, states of mind, moral character, and the like that explains this difference in outcome, then the difference really just is a matter of luck.’’18 Luck of this sort seems incompatible with free action or moral responsibility. (Haji 2005, pp. 323–324) If we add that the sort of luck at issue is a matter of something’s not being in one’s control (this need not be accepted but it is plausible), it is not farfetched to suppose that Peg’s decision is not free or she is not morally responsible for it because she lacks freedom-level control in making it. But what sort of control, if any, does Peg lack? Not active, it may preliminarily be supposed (we will return to this presumption later), but what I have previously referred to as ‘‘antecedent 18
Mele (1999, p. 280).
123
Reason, Responsibility, and Free Will
197
proximal control’’ (Haji 2002, pp. 110–111). With fixed pasts, the difference in outcome in Peg’s and Peg*’s cases appears to be merely a function of the indeterminacy in actional pathways leading to choice. I have claimed, however, that it would seem that no agent could exert active or any other sort of control over such indeterminacy to ensure a particular outcome. Amplifying somewhat, if t1 is the time at which Peg makes whatever decision she makes at t1 if she is still alive at t1 (it is not indeterministic, for instance, whether Peg will suffer a fatal stroke at t1), unlike her otherwise similar deterministic counterpart Peggy, Peg does not have the ability or power to ensure that at t1 she decides in accordance with her decisive best judgment about what to do. She may judge that it is best for her to keep the promise, may muster all the powers of self-control that would ordinarily suffice for her deciding in accordance with this best judgment, but still fail to so decide. In Peg’s case, her prior actional antecedents seem not to contribute sufficiently to control. Christopher Franklin (2011) attributes to me what he calls the ‘‘ensurance formulation’’ of the luck argument. He states that its core idea is that if an agent performs an undetermined action u at t, then she could not have ensured or guaranteed that she u-ed rather than w-ed at t; indeterminism prevents agents from having the power to guarantee a particular outcome: agents try their best, but they cannot ensure what they will do. He summarizes this version of the luck argument in this way: (1) Undetermined actions are not ensured, and (2) if an action is not ensured, then the action is a matter of luck (Franklin 2011, p. 210). He adds that according to Haji ‘‘an agent can ensure that he performs an action u only if he has antecedent control over u’’ (Franklin 2011, p. 2010). He suspects that I have something like the following in mind regarding such control: ‘‘An agent S has antecedent control over an action u at t2 just in case S has the power at t1 to w at t1 and if he w-ed at t1, then w would deterministically bring about u at t2’’ (Franklin 2011, p. 210). Franklin explains that although it is possible for agents to possess antecedent control over undetermined actions, it is not possible for them to exercise this control over such actions ‘‘for exercising antecedent control over an action requires that one deterministically bring it about’’ (Franklin 2011, p. 211). He proposes that premise (1) be interpreted as a claim about exercising antecedent causal control: ‘‘undetermined actions are unensured because we cannot, qua undetermined actions, exercise antecedent control over them’’ (Franklin 2011, p. 211). Franklin writes: Under this reading of the ensurance formulation premise, (1) is true, but only at the expense of rendering premise (2) implausible. Under the present reading, (2) asserts that an action u cannot be free unless we perform some earlier action w that deterministically brings about u (i.e., unless we exercise antecedent control over u). A dilemma arises for Haji at this point: either w must itself be free or not. If w must be free, then…a vicious regress is initiated, thus making free action impossible. If w does not need to be free, then Haji is saddled with the implausible claim that free action is (or can be) deterministically brought about by unfree action. (Franklin 2011, p. 211) My response to Franklin turns on clarifying how I construe antecedent causal control, and in explaining why I think such control bears on the luck argument. The discussion should make clear that I do not subscribe to the sort of view regarding
123
198
I. Haji
antecedent causal control that Franklin attributes to me. It is to Franklin’s credit, however, to solicit illumination on just how antecedent causal control is to be understood, and its influence, if any, on free action and responsibility. I start by noting a few things about the Peg/Peg* story. Franklin writes that ‘‘van Inwagen, as did Haji, mistakenly locates indeterminism after choice and for this reason sees indeterminism as a threat to freedom and responsibility’’ (Franklin 2011, p. 217). First, as the sketch of action-centered modest libertarianism should make clear, contrary to Franklin’s assertion, I locate indeterminism at the time of choice and not after choice (I use ‘a choice’ and ‘a decision’ interchangeably). Second, the sort of modest libertarianism I have assumed is consistent with Franklin’s version according to which it is the exercising of one’s ability to choose (or decide) that is undetermined. Some explanation is in order. In a passage in which he addresses some of van Inwagen’s views on luck, Franklin discusses an example in which the probability of Alice’s telling the truth is 50 %. Franklin writes: ‘‘What is undetermined is Alice’s exercising her ability to choose to tell the truth. It is perfectly consistent with event-causal libertarianism that the exercising of the ability deterministically brings about her telling the truth’’ (Franklin 2011, p. 218). Franklin, it appears, conceives of the event that is Alice’s exercising the relevant ability as not itself an action, but as a relation between ‘‘an agent, or agent-involving mental events, and some other event’’ (Franklin 2011, p. 227). In sum, focusing on Alice’s decision itself, Franklin’s view appears to be that what is undetermined is Alice’s exercising her ability to decide to tell the truth, and this event deterministically brings about a second, Alice’s deciding to tell the truth. Further, these two events, her exercising her ability to tell the truth and her deciding occur simultaneously, the former causing the latter. Reverting to the Peg/Peg* scenario, regarding my claim that Peg does not have the ability or power to ensure that at t1 she decides in accordance with her decisive best judgment about what to do—she lacks antecedent control with respect to the decision she makes at t1—the pivotal idea here is this: Assume that an agent is selfcontrolled insofar as she is not akratic. In a deterministic world, suppose that on the basis of her reasons this agent forms the all things considered best judgment that she ought, at t0, to decide to A at t1. Then barring unusual circumstances, such as unexpected death, and in the absence of any information that may influence her verdict about what is all things considered best for her, she can ensure at t0 or at some time prior to t1 that, at t1, she decides in accordance with her best judgment. Her being able to ensure that she so decides simply amounts to her being able to so decide provided nothing goes awry in the causal trajectory of her decision. Wanting to remain faithful to Franklin’s brand of modest libertarianism, one may assume that in the proximal causal history of her decision, there is an event—Peg’s exercising her ability to do A—which causes another event—Peg’s deciding to A—and that these events occur simultaneously. Why should this power or ability—the ‘‘ensuring power’’—as Franklin wonders, be thought to be relevant to free action and moral responsibility? After all, we do act akratically, and strict akratic action is free, intentional action that is contrary to a consciously held best (or better) judgment. Furthermore, we can be morally responsible for our akratic choices or actions. To understand the significance of the
123
Reason, Responsibility, and Free Will
199
ensuring power to freedom and responsibility, assume that deterministic Peggy is perfectly self-controlled. An agent is perfectly self-controlled if (1) this agent does not succumb to akratic or other irrational influences;19 and, (2) barring unusual circumstances, such as sudden death or the occurrence of events over which she lacks any control and which would prevent her from deciding in accordance with her best (or better) judgment, and in the absence of new information, further deliberation, etc., she decides, at least in deterministic worlds, in accordance with such a judgment. I introduce a term of art: we can say that the etiology or causal trajectory, or a segment of such a trajectory, of an action, mental or otherwise, is smooth provided it is free of responsibility-undermining factors, such as, for instance, the influence of manipulation of the sort that vitiates responsibility. Some may claim that deterministic causation is itself responsibility-subversive. To avoid begging questions against either compatibilists or incompatibilists, assume that neither deterministic causation per se nor indeterministic causation per se qualify as responsibility-undermining. Both compatibilists and incompatibilists can provisionally agree on what factors, other than the variety of causation at issue, are at least intuitively and plausibly responsibility-undermining. This assumption is not unwarranted on the constraint that the modest libertarian differs from her compatibilist rival only in that the former insists that directly free actions are not deterministically caused. Imagine that perfectly self-controlled Peggy judges that of her two salient options, A and B, it is best for her to A rather than to B. Assuming that the causal pathway to her decision to A is smooth, she enjoys a certain power in her deterministic world. This is the power to ensure that she decides in accordance with her best judgment, and, again, the ensuring power amounts to this: If she exercises this power and the causal trajectory to her decision to A that ‘‘commences,’’ roughly, with her deliberations about whether to A or to B is smooth, she will decide to A in accordance with her best judgment. Provided this segment of the trajectory is smooth, if as a self-controlled agent, Peggy fails to decide in accordance with her best judgment and decides to B rather than to A, she would not be morally responsible for so deciding. For, if she is perfectly self-controlled and she decides to B, something, somewhere along the causal pathway to her decision, has gone very wrong: no new considerations come to her mind prior to or after her having judged that it is best for her to A, she is uninfluenced by akratic considerations, she is not the victim of surreptitious manipulation, and so on. Her failure to ensure that she decides in accordance with her best judgment is responsibility-subversive. Some may conjecture that there has been a breakdown of agency in a situation of this sort.20 This may be so; if it is, it reinforces the judgment that Peggy is not morally responsible for deciding to B if she decides to B. What, however, of agents like us who, unlike Peggy, are not perfectly selfcontrolled? Suppose that Percy, who sometimes succumbs to akratic influences, judges that of her two options, C and D, it is best for her to D. Suppose the causal pathway or its segment that ‘‘commences,’’ roughly, with her deliberations about 19
I do not deny that an agent may perform an akratic action that it is rational for her to perform.
20
See Mele (2006), pp. 60–61 and pp. 125–129, on breakdowns of agency.
123
200
I. Haji
whether to C or D, to her subsequent decision to D is smooth: she musters sufficient self-control in deciding to D, she is not the victim of secretive manipulation, she is not deliberating under the influence of mind-altering drugs, etc. Like Peggy, in her deterministic world Percy possesses the power to ensure that she decides in accordance with her best judgment. Provided nothing is askew in the segment of the causal pathway to her D-ing, this power can be exercised to achieve Percy’s proximal goal: deciding to D. If on this occasion, had Percy decided to C in the absence of akratic influences or new information that sways her reasoning, unsolicited manipulation, and the like, she would presumably not be morally responsible for deciding to C. We may summarize the general, relevant point about antecedent causal control in this way: Perfectly self-controlled agents, and agents who, though not perfectly selfcontrolled, but who exhibit self-control on various occasions of choice can exercise the power to ensure that they decide in accordance with their best judgments in deterministic worlds. Should they fail to so decide, then in the absence of germane explanatory factors for this failure, they would not be morally responsible for the pertinent decisions they make. These factors should illuminate what it is about the etiology of the decisions in question that render their agents not responsible for these decisions. Revert to our original Peg/Peg* indeterministic case. Assume, first, that Peg is perfectly self-controlled. Judging that it is best for her to A, she decides in accordance with her best judgment. The segment of the causal pathway from her deliberations about whether to A to her decision to A is smooth. The modest libertarian claims that, given exactly the same past up to or just prior to Peg’s decision to A, and the same laws, if this decision is free, Peg was able to make some other decision or refrain from making any decision. So, consider the scenario in which holding fixed the relevant past and laws, one of Peg’s counterpart, Peg* (really, Peg) decides to B. If this decision is not free, then the libertarian has lost the battle because Peg* would not, in this scenario, be morally responsible for deciding to B. Assume, then, that Peg*’s decision to B is a free, intentional mental action. Reflect on what may be suggested are three broad possibilities concerning Peg*’s deciding to B: First, Peg*’s reasons to B causally generate her continent decision to B. This option, however, is inconsistent with the assumption that the past is fixed. Recall, Peg judged that it is best for her to A, so Peg* must have judged similarly. If Peg* decides to B, then her so deciding is akratic: her deciding to B is a free, mental action that is contrary to what we may assume is a consciously held best judgment of Peg*’s. Second, Peg*’s decision to B is akratic. One concern with this option is that it violates the assumption that Peg is perfectly self-controlled. Even if this concern is set aside, there is a problem understanding the etiology of Peg*’s deciding to B. To explain, in customary accounts of akratic action, when an agent performs a strict akratic action, there is a misalignment between the motivational strength of the desire from which her act causally derives (the motivationally strongest desire) and her best judgment.21 If we accept these typical accounts, Peg*’s best judgment that 21
See, for e.g., Mele (1995 chap. 2), Watson (1975, pp. 205–220).
123
Reason, Responsibility, and Free Will
201
she ought to keep the promise should stand opposed to her stronger desire to break the promise. With Peg*’s libertarian supposedly free decision, however, we see no such misalignment because Peg* shares the relevant past with Peg. Given her past, Peg*’s desire to keep the promise does not differ in motivational strength from this desire of Peg’s. Regarding Peg, however, we may assume that her desire to keep the promise has greater motivational clout than her competing desire and, moreover, there is no misalignment between this stronger desire and her judgment that it is better for her to keep the promise. So we have a pretty obvious problem: how are we to explain Peg*’s akratic decision? As I have argued elsewhere (Haji 2005, Sect. 2), I doubt whether there is any plausible way to do so. Regarding this second option, finally, it is implausible that a modest libertarian be committed to the view that in Peg/Peg*-like cases in which Peg acts continently, Peg could have freely done otherwise only if she had akratically done otherwise. On the third option, if Peg* decides to B in opposition to her consciously held best (or better) judgment, she has suffered a breakdown in agency. If so, she is not morally responsible for deciding to B. In sum, the event causal libertarian says that, consistent with the past and the laws being what they are, at t Peg can either freely decide to keep the promise, or at t she can decide to break the promise, and whatever decision she makes at t, there is an action explanation of that decision. Roughly, apt reasons, it is claimed, indeterministically cause the decision that the agent makes. It seems that this view is not quite on target, as Peg*’s scenario illustrates. Active control is a function of agent-involving springs of action appropriately causing one’s actions. It appears that an agent who decides akratically exercises less active control in deciding as she does than an otherwise similar agent who decides continently. So it would be too quick to claim, without further explanation, that indeterministic Peg*, who decides akratically, exercises the same degree of active control in making the decision that she does as deterministic Peggy would in making a type- or near type-identical decision when Peggy is perfectly self-controlled. Still, one might wonder about what precisely is the connection between lack of an action explanation of Peg*’s akratic action and responsibility-level control. I take action explanations to be causal. Kane, an event causal libertarian, explains that an agent’s decision is free only if that agent exercised plural voluntary control in making that decision. Plural voluntary control presupposes that the agent had genuine alternatives; consistent with the past and the laws remaining fixed, the agent (like Peg*) could have made an alternative decision. Assuming that the agent had these genuine options, she had plural voluntary control over these options only if she was able to bring about whichever of the options she willed (or desired) when she willed to do so, for the reasons she willed to do so, on purpose, rather than accidentally or by mistake, without being coerced or compelled in doing so or in willing to do so, or otherwise controlled in doing or in willing to do so by other agents or mechanisms (Kane 2005, p. 138).22 If Peg*, then, has plural voluntary control over her akratic decision, her reason states must suitably and indeterministically cause that decision. If, however, there is no apt causal explanation of her 22
These conditions pertain to self-forming decisions.
123
202
I. Haji
akratic decision in terms of her prior reasons, then we have reason to doubt that she exercises such control in making this decision. To collect results, in Peg/Peg*-like cases, we begin with the assumption that in accordance with her best judgment Peg continently A-s at a certain time, t, in the actual world. In relevant non-actual worlds with the same past and the laws as the actual world, Peg (Peg*, if we want) B-s at at this time. (We may take Peg’s A-ing and Peg*’s B-ing to be Peg’s deciding to A and Peg*’s deciding to B respectively.) In such cases, Peg* lacks the power to ensure that she decides in accordance with her best or better judgment although, having formed this judgment, she does not persist in deliberation, no new information or considerations come to her mind prior to her deciding to B, etc., and there are no reasons to believe that the segment of the causal etiology of her deciding to B from the time she forms the best judgment that she ought to A to her decision to B is not smooth. Abbreviating, Peg* lacks antecedent proximal control in deciding as she does despite the assumption that the relevant segment of the causal etiology of her deciding to B is smooth. If, having formed the best judgment that she ought to A, Peg* does not deliberate any further, no new considerations or information come to her mind prior to her deciding to B, etc., and she lacks proximal antecedent control despite smoothness of the relevant segment of the causal etiology of her deciding as she does, then it does not seem that her reason states appropriately cause her decision to B. Each of the pertinent options—Peg* decides to B continently, she decides to B akratically, or she suffers a breakdown of agency—are untenable (given that the past and the laws are fixed). If Peg*’s prior reason states do not appropriately cause her decision to B, however, then she lacks responsibility- or freedom-level control in deciding as she does. Recall, this control is largely causal. Provided we think that an agent’s active control in making a decision consists in apt agent-involving events causing nondeviantly that decision, we have grounds to challenge the claim that Peg* exercises active control in making the decision that she does. Therefore, Peg* lacks the sort of causal control in deciding as she does that modest libertarianism requires of one to perform free actions. I believe the luck objection can be modified to show that indeterminism may well threaten the truth of judgments of objective and Davidsonian reasons, something I hope to pursue in other work. Prior to leaving the luck objection, it’s worth mentioning that Coffman (2010) has recently advanced an interesting response to this objection. He proposes that appropriately constructed Frankfurt examples involving indeterministically caused directly free actions (which we may take to be putatively free decisions) overturn the objection. Specifically, he states that such examples undermine the presumption of the luck objection that if ‘‘an act (1) was significant for you but (2) lacked a complete explanation, then you are not morally responsible for that act’’ (Coffman 2010, p. 158). Here is an outline of Alfred Mele’s construction of such an example: Black initiates a certain internally deterministic Process P in Bob’s brain at t1 with the intention of thereby causing Bob to decide at t2…to steal Ann’s car. The process, which is screened off from Bob’s consciousness, will culminate in Bob’s deciding at t2 to steal Ann’s car unless he decides on his own at t2 to steal it or is incapable at t2 of making a decision (because, for example, he is dead by t2)….As it happens, at t2 Bob decides on his own to steal the car, on
123
Reason, Responsibility, and Free Will
203
the basis of his own indeterministic deliberation about whether to steal it, and his decision is not deterministically caused. But if he had not just then decided on his own to steal it, P would have issued, at t2, in his deciding to steal it. Rest assured that P in no way influences the indeterministic decision-making process that actually issues in Bob’s decision….Given further details…Bob, in the actual world, W, would appear to be morally responsible for deciding to t2 to steal Ann’s car. (Haji 2006, p. 88, note omitted)23 In reply, it is not evident why a proponent of the luck objection should accept the view that an agent, such as Bob, in an ‘‘indeterministic’’ Frankfurt case, is morally responsible for his decision (Haji 1999, 2003, pp. 271–274, 2006, sect. III). After all, it is open to the proponent to argue that absent the fail-safe mechanism in Mele’s case, Bob is not morally responsible for his decision to steal Ann’s car because the decision is a matter of luck. ‘‘Addition’’ to the case of the fail-safe mechanism should not undermine this verdict unless one has an explanation to the contrary.
6 Some Thoughts on Frankfurt Examples I end by focusing on various putative concerns with Frankfurt examples. To start with a more detailed template of such an example, think of this sort of example as unfolding in two stages. In Stage 1, an agent, Ken, decides to do something, x, and intentionally x-s. For instance, he decides to steal a pear. We are to assume that whether you are a libertarian or a compatibilist, on your account of free action and moral responsibility, Stage 1 Ken is morally responsible for deciding to x (and for xing).24 In Stage 2, the scenario is developed in a way in which something supposedly precludes Ken (Stage 2 Ken) from deciding to do other than x without in any way interfering in Ken’s deciding to x. We are meant to conclude that because Stage 1 Ken is morally responsible—morally blameworthy in this case—for deciding to x, and because Stage 2 Ken does not differ relevantly from Stage 1 Ken regarding deciding to x, Stage 2 Ken is also morally responsible for deciding to x even though he could not have refrained from deciding to x (Frankfurt 1969). Embellish this template in the following two respects. First, assume that in Stage 1 Ken’s deciding to steal and his subsequent stealing are overall and not just prima facie impermissible for him. Thus, in Stage 1 the blameworthiness requires wrongness principle: Blameworthiness Requires Wrongness (BO): An agent is morally blameworthy for doing something only if she overall ought not to do it. can be satisfied. If this is so, then it would seem that in Stage 2, too, there is no bar to this principle’s being satisfied. After all, in Stage 2 Ken acts in the absence of any 23
See, also, Mele and Robb (1998, 2003).
24
A more cautious manner of arguing would be to assume only that it is not demonstrated that the agent is not morally responsible. See, for e.g., Fischer (1999) and Haji (2004) and McKenna (2006). For present purposes, though, we can work with the stronger assumption.
123
204
I. Haji
intervention in just the way in which he did in Stage 1. So, it may be thought, if his deciding to steal and his stealing are wrong in Stage 1, they should also be wrong in stage 2. A proponent of the blameworthiness requires wrongness principle may claim that this should facilitate defending the view that Ken is just as blameworthy in Stage 2 as he is in Stage 1. Second, as I have suggested, and many concur, the control that moral responsibility requires consists largely in one’s decisions being nondeviantly caused by one’s reason states (or their neural realizers).25 Presumably, Ken decides to steal for reasons in Stage 1. He should therefore exercise responsibility-level control in deciding to steal in this stage. It appears that these very sorts of reasons also lead to his decision to steal in Stage 2. So, he should also exercise responsibility-level control in deciding as he does in Stage 2. This further bolsters the judgment that if Ken is blameworthy in Stage 1, he should be blameworthy in Stage 2 as well. The first embellishment generates the following problem. Theorists who accept the blameworthiness requires wrongness principle should renounce the view that Ken’s deciding to steal and his stealing are wrong in Stage 2. For, among other reasons, an argument parallel to the one that shows that there is a requirement of alternative possibilities for reasons-wise right, wrong, and obligation also shows that there is such a requirement for moral right, wrong, and obligation. John Martin Fischer writes that he has ‘‘sought to argue that causal determinism is compatible with moral responsibility [partly by invoking Frankfurt examples]…This result would be considerably less interesting if causal determinism were nevertheless incompatible with the central judgments of deontic morality’’ [i.e., judgments of moral right, wrong, and obligation] (Fischer 2006, pp. 118–216). There is, however, an alternative possibilities requirement for moral right, wrong, and obligation. If, on the one hand, these alternatives are strong, then determinism is ‘‘incompatible with central judgments of deontic morality.’’ If, on the other hand, these alternatives are weak, then semi-compatibilism with respect to moral obligation cannot be sustained. Furthermore, if one insists on the blameworthiness requires wrongness principle, semi-compatibilism with respect to blameworthiness is imperiled. For, if blameworthiness requires wrongness, and wrongness requires alternatives, then blameworthiness requires alternatives as well. Renouncing the blameworthiness requires wrongness principle evades the first concern with Frankfurt examples. Indeed, there are good reasons to give up this principle. First, it has putative counterexamples. Yoshi believes that he can prevent a bomb from going off if and only if he presses a button at 7:00 p.m. Suppose, further, that he inexcusably forgets to press the button at this time. Here, we may assume that he is blameworthy for failing to press the button. Suppose, in addition, that the button is broken: even if he had pressed it, the bomb would still have gone off. Arguably, Yoshi’s not pressing the button is not wrong but he is, nonetheless, blameworthy for not pressing it. Second, blameworthy suberogation presents a serious problem for BO. Common sense morality recognizes, among other normative statuses, the supererogatory, 25
Non-causalists about control, such as Ginet (1990) and McCann (1998), will not accept this view.
123
Reason, Responsibility, and Free Will
205
roughly, permissible action that is beyond the call of duty; and the suberogatory, again, loosely, permissible suboptimality. It also sanctions ascriptions of moral praiseworthiness and moral blameworthiness, respectively, to those who supererogate and to those who suberogate. Elaborating somewhat on the suberogatory, suberogatory actions, the roughly symmetric counterparts of supererogatory ones, are morally optional, not morally indifferent, and are offensive or fall short of decency. In one of Julia Driver’s putative examples of a suberogatory act, a man who takes a seat on a train, thereby intentionally preventing two other people from sitting together despite availability of another seat, does no wrong but falls short of decency (Driver 1992, pp. 286–287). Roderick Chisholm and Ernest Sosa propose that minor discourtesy, such as taking too long in a restaurant when others are waiting, may also be suberogatory (Chisholm and Sosa 1966, p. 326). Assume that even though he mistakenly believes he is doing wrong by preventing the lovers from sitting together despite availability of another seat, Grumpy fails to give up his seat. Here, Grumpy may well be blameworthy for not giving up his seat even though, let us assume, his not giving up his seat is suberogatory and hence morally optional. After all, from his perspective, he is doing wrong: he acts in light of the belief that he is doing wrong. But then Grumpy’s case challenges the principle BO that one is to blame for an action only if one ought not to perform that action. More fundamentally, we accept the view that one can do moral wrong but not be morally blameworthy for the wrongdoing; we acknowledge that one can have a legitimate excuse. In such cases, the negative deontic assessment of the act is not matched by a negative responsibility appraisal of the agent; agent and act appraisals come apart. But if this view is accepted, there is little reason to deny that it is possible that a negative responsibility appraisal of the agent not be matched by a negative deontic appraisal of some pertinent act of that agent. Even if we give up on the blameworthiness requires wrongness principle, the second embellishment having to do with reasons—if Ken decides for reasons in Stage 1, he decides for these same sorts of reasons in Stage 2 as well—gives rise to another concern with Frankfurt examples. If the reasons in question are objective pro tanto, then in Stage 2, Ken’s decision to steal is not caused by such reasons simply because there is an alternative possibilities requirement for having such reasons. We have assumed, however, that Ken could not have refrained from deciding to steal in Stage 2. The moral here is that proponents of Frankfurt cases should not presume that the examples are equally effective no matter what their conception of reasons. Partisans of Frankfurt may insist that although responsibility judgments are conceptually tied to reasons, the reasons are not objective pro tanto reasons. Rather, they are a species of Davidsonian reason, which this example introduced earlier, seems to confirm: you nonculpably believe that you have objective pro tanto reasons to give A to a patient because you nonculpably believe that giving A will cure this patient. You also believe that you morally ought to give A. But you have decisive objective pro tanto reasons against giving A because A will kill the patient. You give A partly on the basis of the nonculpable belief that you morally ought to do so. Are you not morally praiseworthy for giving A although you have no
123
206
I. Haji
objective pro tanto reasons to give A? Moreover, it appears that the sorts of reasons for which you give A are normative Davidsonian reasons. Suppose, again, that you desire to save the patient. You believe that you have decisive objective pro tanto reasons to give A; and you believe, as well, that you morally ought to give A. Assume that your desire to save the patient, together with these beliefs, nondeviantly and causally issues in your forming an intention to give A, which you then execute. Your normative Davidsonian reasons to give A include the belief that you have decisive reason to give A that is false. Still, you give A for reasons that you have even though these reasons are not objective pro tanto reasons. I find this rejoinder plausible, although there is a remaining worry. The proposal that there is no alternative possibilities requirement for Davidsonian reasons is not as forthright as one might initially believe. Suppose you are an ‘‘internalist’’ or an ‘‘externalist’’ about reasons and you accept as a constraint on practical reasons independently plausible principles objective reasons-wise ‘‘ought’’ [‘‘ought not’’] implies ‘‘can’’ [‘‘can refrain from’’], and objective reasons-wise ‘‘ought not’’ is equivalent to objective reasons-wise ‘‘wrong.’’ Then you will be committed to the view that there is a requirement of alternative possibilities for the truth of judgments of internal reasons because these principles imply that your having such reasons requires that you have alternatives. Consequently, you will be able to hold on to the view that there is no requirement of alternative possibilities for the truth of ‘‘internal’’ Davidsonian reasons only by paying a theoretical cost many may well deem too high: regardless of whether you are an ‘‘internalist’’ or ‘‘externalist’’ about reasons, you have to renounce these principles.26 So, is there this symmetry concerning an alternative possibilities requirement with respect to normative Davidsonian and objective reasons? Do both sorts of reason presuppose that if one has a reason of either sort to do something, then one could have done otherwise? Symmetry would, of course, threaten the view that Frankfurt cases undermine the principle of alternative possibilities concerning responsibility—persons are morally responsible for having done something only if they could have done otherwise. There are, I think, reasonable grounds to deny this symmetry in a fashion congenial to the Frankfurt proponent: even if responsibility requires normative Davidsonian reasons, there is no requirement of alternative possibilities for having such reasons. For one thing, lack of alternatives does not preclude one from having desires and beliefs, the fundamental constituents of typical Davidsonian reasons. For another thing, suppose you have an objective reason to give A to a patient because you morally ought to give A to this patient. I find no reason to deny that (objective) reasons-wise ‘‘ought’’ implies ‘‘can’’ and (objective) reasons-wise ‘‘ought not’’ implies ‘‘can refrain from.’’ Crucially, these principles, in combination with others, lead to the conclusion that there is an 26
Williams (1981, pp. 102–103) has proposed that an agent has an ‘‘internal’’ reason to do something— A—only if she would arrive at a motivation (or desire—here, I use the terms interchangeably) to A were she to deliberate rationally from her current ‘‘motivational set,’’ where this set has been corrected to exclude false beliefs and include all relevant true beliefs. Also see Williams (1995a, p. 36). In addition, see Smith (1994, p. 156), Parfit (1997, p. 100). Williams, though, also seems to endorse a ‘‘reason implies can’’ principle: if you have an (internal) reason to A, you must be capable of A-ing (Williams 1995b, pp. 189–190).
123
Reason, Responsibility, and Free Will
207
alternative possibilities requirement for reasons-wise right, wrong, and obligation when the reasons are objective. Suppose, now, you believe falsely that, on a certain occasion, you morally ought to give A to save the patient. This belief, together with a desire of yours to do what you morally ought to do to save the patient on this occasion, is a normative Davidsonian reason you have to save the patient. But one’s believing one ought to give A—a subjective reason—unlike its being true that one ought to give A, does not entail that one can give A. Since ‘‘ought’’ implies ‘‘can’’ does not ‘‘apply,’’ as we may say, to such subjective reasons, these being constituents of pertinent Davidsonian reasons, one might plausibly deny that there is an alternative possibilities requirement for having normative Davidsonian reasons. In brief, one cannot persuasively argue from the premise that pertinent subjective reasons that are constituents of normative Davidsonian reasons, such as the reason that one believes one morally ought to give A, imply ‘‘can,’’ in consort with other principles, such as ‘‘ought not’’ is equivalent to ‘‘wrong,’’ entail that there is an alternative possibilities requirement for such subjective reasons, and, hence, by generalization, for Davidsonian reasons. The argument falls at its first premise. To conclude on a somewhat disturbing note, I find semi-compatibilism regarding moral responsibility attractive. I have deeply admired and learnt a great deal from John Fischer’s fascinating agenda to defend this species of semi-compatibilism as I have done from Michael McKenna’s relevant, insightful contributions. Owing, however, to reasons debt to freedom—the view that having objective pro tanto reasons requires that we have free will—and owing to the essential connection between the truth of various normative judgments, such as those of moral or prudential obligation, and objective pro tanto reasons, there seem to be strict limits to extending the semi-compatibilist’s program. Acknowledgments Many thanks, indeed, to John Fischer and Michael McKenna for discussion on many issues that bear on what I have addressed in this paper. I am deeply grateful to Angelo Corlett for giving me the opportunity to express my views. This paper was written during my tenure of a 2008-2012 Social Sciences and Humanities Research Council of Canada (SSHRC) grant. I am most grateful to this granting agency for its support.
References Arpaly, Nomy. 2006. Merit, meaning, and human bondage. Princeton: Princeton University Press. Broome, J. Reasons. 2004. In Reason and value: Essays on the moral philosophy of Joseph Raz, ed. P. Pettit, S. Scheffler, M. Smith, and R.J. Wallace, 28–55. Oxford: Oxford University Press. Chisholm, Rodrick, and Ernest Sosa. 1966. Intrinsic preferability and the problem of supererogation. Synthese 16: 321–331. Clarke, Randy. 2000. Modest libertarianism. Philosophical Perspectives 14: 21–45. Clarke, Randy. 2003. Libertarian accounts of free will. Oxford: Oxford University Press. Coffman, E.J. 2010. How (not) to attack the luck argument. Philosophical Explorations 13: 157–166. Copp, David. 1997. Defending the principle of alternative possibilities: Blameworthiness and moral responsibility. Nous 31: 441–456. Davidson, Donald. 1963. Actions, reasons, and causes. The Journal of Philosophy 60: 685–700. Dennett, Dennett. 1978. Brainstorms: Philosophical essays on mind and psychology. Montgomery: Bradford. Driver, Julia. 1992. Suberogation. Australasian Journal of Philosophy 70: 286–295. Ewing, A.C. 1939. A suggested non-naturalistic analysis of good. Mind 39: 1–22. Ewing, A.C. 1947. The definition of good. London: Routledge and Kegan Paul.
123
208
I. Haji
Feinberg, Joel. 1970. Doing and deserving. Princeton: Princeton University Press. Feldman, Fred. 1986. Doing the best we can. Dordrecht: D. Reidel Publishing Company. Fischer, John Martin. 1995. Libertarianism and avoidability: A reply to Widerker. Faith and Philosophy 12: 119–125. Fischer, John Martin. 1999. Recent work on moral responsibility. Ethics 110: 93–139. Fischer, John Martin. 2003. ‘Ought-implies-can’, causal determinism, and moral responsibility. Analysis 63: 244–250. Fischer, John Martin. 2006. My way: Essays on moral responsibility. Oxford: Oxford University Press. Fischer, John Martin. 2011. Indeterminism and control: An approach to the problem of luck. In Law and neuroscience, ed. Michael Freeman, 41–60. Oxford: Oxford University Press. Fischer, John Martin, and Mark Ravizza. 1998. Responsibility and control: A theory of moral responsibility. Cambridge: Cambridge University Press. Frankfurt, Harry. 1969. Alternate possibilities and moral responsibility. The Journal of Philosophy 66: 829–839. Franklin, Christopher. 2011. Farewell to the luck (and Mind) argument. Philosophical Studies 156: 199–230. Ginet, Carl. 1990. On action. New York: Cambridge University Press. Glover, Jonathan. 1970. Responsibility. New York: Humanities. Haji, Ishtiyaque. 1998. Moral appraisability: Puzzles, proposals, and perplexities. New York: Oxford University Press. Haji, Ishtiyaque. 1999. Indeterminism and Frankfurt-type examples. Philosophical Explorations 1: 42–58. Haji, Ishtiyaque. 2002. Deontic morality and control. Cambridge: Cambridge University Press. Haji, Ishtiyaque. 2003. Alternative possibilities, luck, and moral responsibility. The Journal of Ethics 7: 253–275. Haji, Ishtiyaque. 2005. Libertarianism, luck, and action explanation. The Journal of Philosophical Research 30: 321–340. Haji, Ishtiyaque. 2006. The principle of alternate possibilities and a defeated dilemma. Philosophical Explorations 9: 179–202. Haji, Ishtiyaque. 2009a. Incompatibilism’s allure: Principal arguments for incompatibilism. Peterborough, ON: Broadview Press. Haji, Ishtiyaque. 2009b. Freedom and value: Freedom’s influence on welfare and worldly value. Dordrecht: Springer. Haji, Ishtiyaque. 2012. Reason’s debt to freedom: Normative appraisals, reasons, and free will. New York: Oxford University Press. Haji, Ishtiyaque, and Michael McKenna. 2004. Dialectical delicacies in the debate about freedom and alternative possibilities. Journal of Philosophy 101: 299–314. Haji, Ishtiyaque, and Michael McKenna. 2006. Defending Frankfurt’s argument in deterministic contexts: A reply to Palmer. Journal of Philosophy 103: 363–372. Haji, Ishtiyaque, and Stefaan Cuypers. 2008. Moral responsibility, authenticity, and education. New York: Routledge. Kane, Robert. 1996. The significance of free will. New York: Oxford University Press. Kane, Robert. 2005. A contemporary introduction to free will. New York: Oxford University Press. McCann, Hugh. 1998. The works of agency: On human action, will and freedom. Ithaca: Cornell University Press. McKenna, Michael. 2012. Agent meaning and conversation: A theory of moral responsibility. New York: Oxford University Press. Mele, Alfred. 1995. Autonomous agents. Oxford: Oxford University Press. Mele, Alfred. 1999. Ultimate responsibility and dumb luck. Social Philosophy and Policy 16: 274–293. Mele, Alfred. 2003. Motivation and agency. Oxford: Oxford University Press. Mele, Alfred. 2006. Free will and luck. Oxford: Oxford University Press. Mele, Alfred, and David Robb. 1998. Rescuing Frankfurt-style cases. The Philosophical Review 107: 97–112. Mele, Alfred and D. Robb. 2003. BBs, magnets and seesaws: The metaphysics of Frankfurt-style cases. In Freedom, responsibility, and agency: Essays on the importance of alternative possibilities, ed. David Widerker and Michael McKenna, 127–138. Aldershot: Ashgate. Morris, Herbert. 1976. On guilt and innocence. Berkeley: University of California. Parfit, Derek. 1997. Reasons and motivation. Proceedings of the Aristotelian Society 71: 99–130.
123
Reason, Responsibility, and Free Will
209
Pereboom, Derk. 2001. Living without free will. Cambridge: Cambridge University Press. Scanlon, T.M. 2008. Moral dimensions. Princeton: Princeton University Press. Schroeder, Mark. 2007. Reasons and agent-neutrality. Philosophical Studies 135: 279–306. Smith, Michael. 1994. The moral problem. Oxford: Blackwell. Strawson, Peter F. 1962. Freedom and resentment. Proceedings of the British Academy 48: 1–25. Trakakis, Nick. 2008. Wither morality in a hard determinist world. In Essays on free will and moral responsibility, ed. N. Trakakis and D. Cohen, 34–70. New Castle upon Tyne: Cambridge Scholars Publishing. van Inwagen, Peter. 1983. An essay on free will. Oxford: Clarendon Press. van Inwagen, Peter. 2000. Free will remains a mystery. Philosophical Perspectives 14: 1–19. Vranas, Peter. 2007. I ought, therefore I can. Philosophical Studies 136: 167–216. Wallace, R.J. 1994. Responsibility and the moral sentiments. Cambridge: Harvard University Press. Watson, Gary. 1975. Free agency. The Journal of Philosophy 72: 205–220. Williams, Bernard. 1981. Moral luck. Cambridge: Cambridge University Press. Williams, Bernard. 1995a. Internal reasons and the obscurity of blame. In Making sense of humanity, ed. B. Williams, 35–44. Cambridge: Cambridge University Press. Williams, Bernard. 1995b. Replies. In World, mind, and ethics, ed. J.E.J. Altham and R. Harrison, 185–224. Cambridge: Cambridge University Press. Zimmerman, Michael J. 1988. An essay on moral responsibility. Totowa: Rowman & Littlefield. Zimmerman, Michael J. 1996. The concept of moral obligation. Cambridge: Cambridge University Press. Zimmerman, Michael J. 2001. The nature of intrinsic value. Lanham: Rowman and Littlefield. Zimmerman, Michael J. 2008. Living with uncertainty. Cambridge: Cambridge University Press. Zimmerman, Michael J. 2010. Responsibility, reaction, and value. The Journal of Ethics 14: 103–115.
123