For us there is only the trying, the rest is not our business. (T.S. Eliot, Four Quartets)

The Principle of Alternative Possibilities (PAP) states that an agent is morally responsible for her action only if she could have done otherwise: we should not praise or blame someone for doing that to which there is no alternative. PAP in conjunction with Causal Determinism (CD)- the doctrine that every occurrence is necessitated by prior events- undermines commonsense assignments of moral responsibility. For every action would be the only possible outcome of preceding events were CD true, rendering unjustified, according to PAP, the apportioning of praise/blame to any agent. In a seminal essay, "Alternate Possibilities and Moral Responsibility," Harry Frankfurt meets this skeptical challenge by providing a counterexample to PAP.

Below, I respond to four recent attempts to show that "Frankfurt cases," as examples structurally identical to Frankfurt’s have come to be known, fail of their purpose. In the first, I expose a misconception of what it is to be disadvantaged. My challenge to the second stems from its reliance upon the notion of "moral luck." The third, while conceding that Franfurt cases do falsify PAP in regards to "complex" actions like casting a ballot, argues that they can not be used to do the same vis-a`-vis "simple" mental acts such as taking a decision. By appealing to the connection between assessments of character and judgments of moral responsibility, I intend to show that this dichotomy does not hold. I close with a discussion of a libertarian alternative to PAP.

Here is a typical Frankfurt case. Black, a partisan neurosurgeon, wants his patient and friend Jones to vote Republican in the upcoming presidential election. Jones is known by Black to be leaning toward the GOP candidate. Black, however, does not wish to take any chances. Thus, after convincing Jones that he needs to undergo exploratory surgery, he implants a device in his brain that will create the irresistible urge to vote Republican should it detect the intention to do otherwise.

It seems, then, that Jones has no alternative to voting Republican. If he decides in favor of the GOP candidate, she will receive his vote; if he decides against her, the outcome will be no different. Yet, when Jones votes Republican 'on his own', carrying out the decision that he had made, it appears to be something for which he is morally responsible- even though it is the only course of action open to him. After all, it is Jones’ resolving to vote Republican that ‘rules out’ the alternatives not the operation of Black’s device. Hence, according to Frankfurt, PAP is false.1

Joseph Campbell challenges the notion that Jones in the above case is unable to avoid voting Republican. On his view, Jones is able to keep from committing that action since there are "accessible worlds" in which he does otherwise.2 Campbell states that another possible world is accessible to an agent just in case it is governed by the laws of nature that hold in that agent's world and bestows upon her all and only the abilities she possesses in her world.3 In other words, A of w can access w’ iff w’ is physically possible relative to w and A would have therein the same capacities that she can exercise in w and no more. On his view, being able to do otherwise is a matter of having access, in virtue of having a free or unobstructed will, to possible worlds in which one acts counterfactually. The will, according to Campbell, is the "set of powers" or "faculty" used in practical reasoning, normative judgment, and the forming and carrying out of one's intentions.4 He lists compulsion, coercion, certain drugs, and mental illness as impairments of the will.5

While I largely concur with this conception of the will (and will later make use of it myself), I reject the idea that Jones can access worlds in which he does not vote Republican. Campbell contends that such worlds are open to Jones despite the fact that in each one he is unencumbered of Black's device, that is, is not limited in terms of what he can do by its presence inside his skull. His being rid of Black’s device, according to Campbell, would not "advantage" Jones, that is, enable him to do anything that the device prevents him from doing, given that it was not activated: "accessible worlds need not include factors which are causally irrelevant to the performance of actions."6

By the same token, though, we would have to say that removing a race horse's blinders would not increase her abilities, unless they are the cause of her looking straight ahead during its run. But, even if she does not glance sideways because of being fixated by that the part of the field that is occupying the center of her field of vision, it is still true that the blinders do not allow her to do something that she otherwise could do- and to that extent they are a disadvantage. Campbell, though, is committed to the view that a thing is disadvantaged by one of its features F only if F actually causes (should figure in the explanation of) a failure on its part. Only proximate causes of untoward circumstances, in other words, count for him as disadvantages. But a thing is also hindered by a condition the lacking of which would allow it to do things it cannot currently accomplish- even if those things remain only potentially undone because they are never actually attempted. Thus, Campbell can not be right in maintaining that Jones can access a world in which he is rid of Black’s device, as therein he would be capable of doing things that he can not do as long as he is encumbered by that mechanism.

Here it is necessary to distinguish between a disadvantage and its manifestations in failures of effort. In Frankfurt's case, the absence of the latter should not lead us to think of Jones as being unhindered by Black's device. To the extent to which he can not attempt to carry out an intention that he is capable of forming, a part of his will (what Campbell, following Thomas Reid, calls the "active powers") is impaired. Campbell’s list of impediments to the will, thus, ought to be expanded to include the presence of counterfactual interveners such as Black’s device. Having an intact will is, thus, the advantage that Jones would gain in a world in which that mechanism is not implanted in his brain, making that world inaccessible to him.

But since his impairment did not play a role in determining how he acted, his action is still something for which he should be held accountable. Thinking otherwise would be like a switch-hitter, who would have been batting right-handed in any event, blaming his striking out on an injury that keeps him from hitting left-handed. Jones’ will is impaired so as to render him incapable of doing otherwise- in that regard he is unfree (he lacks what John Martin Fischer and Mark Ravizza call "regulative control" over his conduct7). It is not impeded, however, in such a way as to mitigate his responsibility for what he does: enough of his will remains intact to allow him to evaluate the various candidates and carry out the decision that actually issues from his deliberations, viz., to vote Republican. Frankfurt's counterexample stands.

Jones in the above example performs a "morally neutral" action: voting Republican is neither obligatory nor something from which he ought to refrain. Let us consider instead a case in which one of the options he is considering would be morally wrong, say having his newspaper wage a smear campaign against the GOP candidate's main opponent. Were Black's device set to insure the carrying out of this plan is it clear that it would be something for which Jones would be morally responsible should he 'do it on his own'? David Copp thinks not.8 That is because he believes that holding Jones accountable here would violate the Kantian maxim (the "Maxim") that "ought implies can." If spreading lies is immoral, then it is something from which Jones should refrain. But, according to the Maxim, if Jones should keep from defaming another person's character, then it is within his power to do so. But Jones would be incapable of avoiding such unscrupulous conduct, given the presence of Black's device. Thus, blame could not be attached to him, even if he were to engage in it at his own prompting. In that case, Jones would be, according to Copp, "morally lucky" in that but for Black's intervention he would be culpable for what he had done.9

A defender of Frankfurt should respond here that it would be odd if the Maxim could have the above consequence, since Kant explicitly denies the existence of moral luck.10 To eliminate the threatened inconsistency in Kant's system, then, an alternative reading of the Maxim must be found. I propose that it be taken as meaning that "ought implies can attempt," that is, that an action is obligatory for an agent only if she can take all of the antecedent "steps" that would lead to its performance under ordinary circumstances. Typically, a non-mental act is preceded by the determining and weighing of alternatives and the taking of a decision/forming of an intention. What constitutes a good faith attempt at performing some act is going to be settled by how far along its agent could get in this process. Where the possibility of carrying out one’s intention exists, an obligation can not be discharged merely by resolving to do the right thing. Performing this mental act would be sufficient in this regard, though, were it to mark, as in Copp’s case, the limit of what can be achieved by way of doing the right thing. (That such an act, being private, could not be viewed and assessed by others does not diminish its moral signifigance. It only means that we, as non-omniscient seekers of moral truth, must learn of it via our fallible understanding of its behavorial criteria.)

In light of Kant's remarks concerning the "good will" this interpretation seems independently motivated: he would say that the failure to effect a rescue is not to be held against a lifeguard, if it comes as the result of forces beyond her control. She has discharged her obligation and is thus to be praised as much as if she'd been successful, by doing all that she could under the circumstances.11 But then to say that she ought to rescue the child does not imply that she can do so, but only that it is within her power to take whatever lifesaving measures are then available to her, increasing the likelihood, but not guaranteeing, that the child will be saved. Given this reading of the Maxim, however, Jones would, contra Copp, be blameworthy for libeling his political opponent, since he could at least resolve to avoid the contemplated smear campaign, resisting his inclination to launch it, even though this effort would prove futile. His duty here consisted, not in refraining from such behavior, as Copp’s reading of the Maxim entails, but rather in doing all that he could to avoid it, which, given the presence of Black’s device, is limited to deciding to the right thing.

Expanding upon this Kantian theme, an agent is morally responsible for her action just in case its motive is a desire or belief upon which she has decided to act, thus making it effective. It is in virtue of having formed an act's motive that a person has rendered herself an appropriate candidate for either praise or blame: her moral responsibility for her action stems from what she has done to (her operation of) the will from which it issues, freely choosing as its object something good or evil. Here the will is being understood, ala Campbell, as the faculty or set of powers (most importantly, the ability to deliberate) that allows one to act as a free agent by determining which desires becomes one's motives and which ones do not (in virtue of the resistance supplied by competing urges).12 Thus, Jones makes himself responsible for performing the unavoidable act of spreading lies by not resolving to avoid playing dirty politics, a mental act (or series of acts) involving the exercise of his will that he could perform, assuming that his deliberative power is unimpaired, albeit with unintended consequences of the sort Kant deemed morally irrelevant.

In general, moral responsibility for a given act stems from what its agent has done or failed to do prior to its performance. (This point reemerges below in a discussion of a variation on the above case.) By employing this approach to moral responsibility, then, Frankfurt cases can be constructed involving impermissible actions: though they themselves would be unavoidable, holding their agents blameworthy for having performed them would not violate the Maxim, in that their duty was not to avoid the unavoidable but rather to resolve to do ‘the right thing’, an option that existed despite the counterfactual intervener’s presence.

David Widerker offers a different sort of criticism of Frankfurt's argument. Widerker concedes that Frankfurt cases do demonstrate the falsity of PAP in regards to "complex" actions like voting, "which involve the realization of a certain want (volition) or intention by events that are distinct from that want or intention and that occur after the latter is formed." He maintains that they are ineffective, however, against PAP as applied to "simple" mental acts such as intending or deciding. That is because a decision, unlike an act of voting, does not require an antecedent intention in order to be carried out. Thus, a counterfactual intervener attempting to forestall the making of a certain choice would be left without a sign within the act itself, which, eo ipso, would be something that is not causally sufficient for its performance, by which it can ascertain what is about to be done.13 This dichotomy leads to the following "dilemma" for a defender of Frankfurt's argument.

Either the factor the detection of which would allow a counterfactual intervener to forestall a decision is a causally sufficient condition for the taking of that decision or it is not. If it is, then the decision maker is neither committing an act for which she is morally responsible nor involved in a situation in which the "circumstances" that make her action unavoidable do not "bring it about" (henceforth a "U-situation"). (In the Frankfurt case discussed above, it is the salience of this dichotomy that belies PAP.) If it is not, then the making of the decision is not unavoidable. Thus, it is not possible that the making of a decision is unavoidable and something for which its agent is morally responsible.14

The defender of Frankfurt, however, should grasp this dilemma’s first horn. To begin with, Widerker begs the question in assuming that moral responsibility is incompatible with CD, since the issue is whether or not that doctrine sans PAP supports such a verdict. The libertarian can not use the doctrine that an action’s being unavoidable (for being causally determined) renders its agent "beyond good and evil" to establish that the agent in a Frankfurt case should not be held accountable for his conduct. A Frankfurt case is designed to show that CD's entailment of the unavoidability of a given action does not necessarily make it something for which its agent is not morally responsible. It establishes as much by showing that something else’s elimination of the ability to do otherwise, that would be the presence of a counterfactual intervener, does not leave an agent incapable of performing an action for which he should be held accountable. In evaluating the effectiveness of a Frankfurt case, then, we are to attend only to the question of whether or not the counterfacual intervener’s presence mitigates the agent’s responsibility merely by making his act unavoidable: the libertarian can not reject such a case as a counterexample to PAP on the grounds that there exists a causally sufficient condition for its agent’s action unless she has reasons for thinking so that are independent of the fact that CD eliminates the ability to do otherwise. Thus, since he does not detail why CD by itself is incompatible with moral responsibility, Widerker fails to refute the claim that the agent in such a case may justifiably be held accountable for making a decision even if it is effected by prior events.

Is CD alone incompatible with moral responsibility? I think not. Consider the following cases. In each, an agent should be held accountable for his action for having effected its cause, that is, for having determined 'on his own' that there be a specific causally sufficient condition for its performance. They point up that in attributing moral responsibility we are not concerned with the question of whether or not an action was caused but of how it was determined.

Suppose I decide that I ought to be motivated by the occurrence of a certain belief to perform a given action. Say I determine that I should count to ten if I come to believe that I am losing patience. I then work at carrying out this intention. As a result, it becomes true of me that I would count to ten if I were to find myself becoming eager, having a strong desire to do so: I develop a certain disposition. It seems, then, that its manifestations are actions for which I am responsible- even though there is a causally sufficient condition for each one's occurrence. After all, it is my intention that I become thus disposed.

It could be argued that beliefs are not effective in the way that this example presupposes. Perhaps it is only a desire or a desire along with a belief that is capable of producing behavior? No matter. The following case makes essentially the same point as the first. Suppose I decide not to oppose a desire that I know is controllable. That is, I determine to 'give in' to this urge whenever it occurs. It seems, then, that I ought to be held accountable for the conduct it causes, since but for my acquiescence it would not occur: I am partly responsible for the desire's effectiveness.

The moral of each case is that CD per se is compatible with moral responsibility: it is only if we think of unavoidability as entailing lack of responsibility that we see the agents as acting beyond "good and evil." Indeed, our judgments of responsibility here are based upon our understanding of the causal connection between the actions and their antecedent conditions. We are inclined to hold a person accountable for the effects of her character, that is, the dispositions she develops as the result of repeatedly engaging in various patterns of behavior, whether they be complex actions or simple mental acts such as the making of a decision. There appears to be nothing but PAP standing in the way of an acceptance of CD and common sense attributions of responsibility.

As for Widerker's charge that the agent of Frankfurt's case is not involved in a U-situation, it is overturned by a clarification of the example's point. Since it is designed to vindicate CD against the incompatibilist's reductio, a Frankfurt case must involve an action for which there is a causally sufficient condition, otherwise it would not show how CD and moral responsibility are consistent, that is, its unavoidability must be overdetermined. (Although, as noted above, in presenting the case it is only the counterfactual intervener’s role in excluding alternatives that is to be emphasized, its "punch line" being 'so you see, CD, which plays the same part, is not a threat to moral responsibility after all'). We see the action of Jones issuing from- being determined by- his character as it shows itself in the way he goes about making choices. That is why he acts freely (in the sense of being morally responsible for what he does): his deliberations and the decision issuing therefrom result in him voting Republican. Had Jones' action been a random occurrence, as Robert Kane and others have noted, it would not be something for which he is responsible.15 There is indeed a mechanism in place that is going to force his action to be performed whatever he has disposed himself to do. But it would not mitigate his responsibility were he to point out that his deed was unavoidable for this reason- that it would have been done, given the presence of Black’s device, even if he had developed a different character. (It is what makes Jones' action unavoidable in this sense that is germane to Frankfurt's purposes.) By drawing our attention to the hollowness of this excuse, Frankfurt demonstrates the falsity of PAP. This reading of Frankfurt’s argument along with the examples just given justify grasping the first horn of Widerker's dilemma. Thus, Frankfurt cases can be constructed involving mental acts such as deciding.

Let Black's device be ready to thwart Jones from even deciding to not vote Republican- effecting the opposite choice should the reasons against it become salient to Jones- if that decision is taken as the result of ordinary deliberation, it is a free mental act and, as such, something for he should be held accountable. Just as in the original case, Jones' will is obstructed- here it is the capacity to form a certain intention that is impaired- but not so as to have rendered him incapable of reasoning. Thus, this restriction does not preclude his being morally responsible for having decided to vote Republican, as this mental act is the outcome of the exercise of his intact deliberative capacity.

Moreover, where the decision in question would be to resolve to do evil, Copp's worry does not creep back into the picture because in saying that it is Jones' duty to not decide to do 'the wrong thing' we are not implying that he can avoid so deciding, only that he is capable of resisting the inclination to thus resolve, an ability that is not itself diminished by the futility of its exercise: one can think hard about not doing evil even if such thought could not issue the decision to do ‘the right thing’. Under ordinary circumstances, of course, merely considering alternatives to doing evil is not nearly enough to avoid becoming blameworthy. But, where one can not even decide to do the right thing, such thought would mark the limit of one’s capacity to act morally. Thus, as noted above, it would be Jones’ failure to exercise this option, which existed prior to his performance of the unavoidable act, that would make him morally responsible for forming an evil intention.

We can use this result to meet the "flicker of freedom" strategy favored by some of PAP’s defenders. Jones, in the original case, it is pointed out, was at least able to decide to do otherwise. His action was not externally determined, in that it was the outcome of a mental act from which he could have refrained, deciding instead to vote Democratic. It is this flicker of freedom that makes him responsible for carrying out his actual decision to vote Republican.16 It is assumed here that the following emendation of PAP must be true: ‘An agent must be able to decide against doing A in order to be morally responsible for its performance’ (or at least have available some alternative possibility, no matter how weak). But if, as we have just shown, Jones need not be able to decide against voting Republican in order to be morally responsible for taking that decision, he surely need not be able to decide to do otherwise in order to bear responsibility for carrying out his decision to vote Rebublican. Again, it is enough that this choice was the outcome of the employment of his normal deliberative methods.

But at this point the compatibilist runs into a problem. For, an incompatibilist, such as Robert Kane, would maintain that an agent whose character was shaped by a series of actions none of which she could have avoided committing should not be held morally responsible for the conduct issuing therefrom.17 An agent's will is unfree to the extent to which her possible motives (desires) are limited by circumstances not of her own making. To use Daniel Dennett's example,18 Luther acted freely at Wittenberg, "despite (then) being able to do no other," only because at an earlier date he had been able to avoid becoming the sort of person who would later have no option besides demanding ecclesiastical reform. Were it not true that he had had an alternative to developing the character that would thus manifest itself, his challenge of church authorities would not have been something for which he was "ultimately" responsible. PAP, Kane concludes, must be revised as follows: An agent acts freely iff her present conduct is undetermined or is the manifestation of a disposition whose origin is a series of unnecessitated actions.19 A free action is not necessarily avoidable at the time at which it is performed; but, if it is not, its agent must have had at some time or other options besides the ones whose exercising led to her having the character necessitating what she is doing.

Kane attempts to show how some actions are not unavoidable by reference to scientific findings concerning indeterminate processes in nature. It seems that the behavior of sub-atomic particles violates CD: replicating the circumstances in which a given quark did A and placing it therein would not guarantee a repeat performance: thus in the original trial it might not have done A. Some thinkers remain deeply skeptical of such claims, yet CD is no longer a working assumption of most scientists.20 Alternative possibilities become available to agents, according to Kane, as the result of indeterminate neural processes being involved in "efforts of the will": the non-equilibrium state that is a conflict between opposing desires responds to such processes generating a decision, which, for all that has physiologically occurred, might not have been taken.21

As Kane himself realizes, though, the existence of such processes is not enough to secure free agency on the incompatibilist's terms. From that perspective, an agent whose options are delimited by indeterministic forces beyond her control is no freer than one whose choices are narrowed by deterministic processes. Kane's principle of "ultimate responsibility" states that if an agent's character is the product of willings for which she is not solely responsible, its manifestations are unfree actions.22 His problem then becomes to explain how being responsible for an action is compatible with its being undetermined, that is, its not being a function of the willing of which it is the outcome. That such a process could yield a different result than the one it actually does seems at odds with its being something for whose consequences an agent is responsible. And it would not suffice, as Gary Watson points out, to simply avail oneself here of non-deterministic explanations of behavior in terms of "reasons, intentions, and purposes."23 Kane must also show how such things, without determining one's conduct, can be such as make one a responsible agent, which is something that, I believe, can not be done.

Consider a character forming action as some compatibilists would have it. It would be effected by the unhindered exercise of a faculty: the will, as described above. Its agent could thus be said to have guaranteed the realization of one possibility rather than others, thereby making herself responsible for her conduct. But if a character forming action must not determined by the will's exercise, as Kane insist, its agent's antecedent mental states are not its guarantor. But wherein, then, lies her responsibility? Nothing that she does or produces insures her action's performance; there are "possible worlds" in which she does otherwise even though she has not changed her mind nor had her plans interfered with. For all her reasoning, planning, and intending, her conduct might have been different. Lacking the power to guarantee that one thing be done rather than another, in which her libertarian "freedom" consists, the connection between an exercise of her will and the ensuing conduct would be too "loose" to entail the latter's being an action for which she is responsible. (A capacity that would take the same reasons as once appeared to warrant taking course of action C as reasons for doing something besides C on a different occasion of their becoming salient is merely unreliable and not to be regarded as making for free agency.) It thus appears that a compatibilism that emphasizes the dependence of free agency upon the possession of this power is in a better position to account for intuitions concerning moral responsibility than a view requiring undetermined actions.

Kane’s account of why the agent in a Frankfurt case acts freely may, then, be set aside. Neither Jones nor Luther is a free agent in virtue of once having performed undetermined actions. No, their freedom consists in the fitness of their wills- the fact that they can evaluate themselves, make reasoned decisions, and act on their choices. Lack of alternative possibilities is consistent with free agency because the necessity of an action does not preclude its issuing from an unhindered exercise of the will.

Granted, in a cosmic sense, no one ever acts freely, given that one’s temperament and the strength of the capacities making up one’s will are a function of hereditary factors over which one could have exercised no influence. Obviously, one can not be free in the sense of being able to provide for oneself alternative possibilities by altering the past in such a way as to change the genetic ‘blueprint’ that has delimited one’s personality and abilities. As Susan Wolf notes, "(w)e are not ... and never can be fully responsible for whether or how much we are responsible."24 Having conceded as much, we are still left with our "common sense" understanding of free agency (the one used in law courts where CD is never exculpatory and upon which "reactive attitudes," such as resentment and gratitude, are based25). It does not require the absence of causal connections between an action and preceding events, mandating only that the latter have left its agent's will intact. Moreover, no case can be made for the superiority of one of these concepts over the other: it is just as appropriate to view ourselves as being responsible for some of the things that we are and do as it is to see all of "our actions and characters merely as fortunate or unfortunate episodes," to use Nagel’s phrase. There is left, then, only the task of making clear just what it is that an agent has done in making herself an appropriate candidate for praise or blame as these are applied in our everday interactions with each other. It will not do to respond to this stalemate by positing a ‘fire wall’ between certain actions and the conditions antecedent to them that would make them inevitable but for its presence, as this move divorces such actions from the facts about their agents in virtue of which they are morally responsible for their commission.

In sum, Frankfurt cases sever all ties between alternative possibilities and the free agency we are capable of exercising. It is not the availability of the former per se that determines whether or not an agent acts freely. Rather, it is the manner in which they were eliminated that settles the issue. Being ruled out by CD alone does not imply unfreedom; to the contrary CD is a precondition of common sense assignments of moral responsibility. Only the exclusion of alternative possibilities by impairments of the will entails the absence of free agency. That is why, pace Watson, there is a difference "between ... natural and purposeful forms of determination that (is) relevant to freedom."26 Not rendering its exercise as it creates and evaluates one's purposes causally superflous, the former but not the latter leaves one's will intact. One is not an automaton for being subject to the (deterministic) laws governing, to use Camus' phrase, a benignly indifferent universe.

Notes

* Thanks to Randolph Clark, Michael Della Rocca, Gary Watson, Robert Kane, William Wainwright, and two referees from this journal for their helpful correspondence.

1. Frankfurt (1969), 150.

2. Campbell, 326.

3. Ibid., 324-6.

4. Ibid., 321. Cf. also Kane pp. 21-2.

5. Ibid., 322.

6. Ibid., 325-6. To a similar objection of David Lamb, (Lamb, 523-4) John Martin Fischer and Paul Hoffman respond that it commits him to the obviously false view that "a circuit can still overheat even though (a) fuse has been installed" (Fischer and Hoffman, 323). That is, despite the fact that its fuse will respond to a voltage overload so as to prevent its overheating, Lamb must maintain that a circuit can still do that which its fuse is supposed to forestall.

7. Fischer and Ravizza (1998), 31-4.

8. Copp, 422-4.

9. Ibid., 424.

10. Nagel, 24. Rosebury also presents a view of moral responsibility that manages to explain away the appearance of moral luck in the cases typically offered in support of its existence.

11. Ibid., 24.

12. I elaborate upon this view in Allen.

13. Widerker, 253-5. Though I shall adopt a different tack, a defender of Frankfurt might insist here that a decision is a complex mental action, our "everyday talk" notwithstanding. Cf. Stump (1998). Widerker overreaches in stating that it is not possible for a counterfactual intervener to take away one's ability to decide to do something other than what one actually decides to do (255). What he has established is that it can effect as much but only by taking as its cue something that is either causally determinative of a decision or is not. But, as shown below, this restriction exists in the original Frankfurt case as well without diminishing its effectiveness.

14. Ibid., 250-2.

15. Cf. Kane, 106-9.

16. This strategy was put forth recently in Della Rocca.

17. Kane, 38-40. William Wainwright (in correspondence) also suggests that we would not hold an agent morally responsible for an action if "we were to become convinced" that his having determined on his own that there be a causally sufficient condition for its occurrence "was itself causally determined." The detailing of an action's entire causal history, I would argue, is not a precondition of one's "common sense" judgment that its agent ought to be held accountable for its performance.

18. Dennett, 133. Dennett, though, unlike Kane, treats this case as just one more counterexample to PAP.

19. Kane, 42-3.

20. Ibid., (150) mentions Einstein, Planck, and De Broglie as notable holdouts.

21. Ibid., 129-30.

22. Ibid., 32-7.

23. Watson, 4. Watson goes on to point out (9-10, 12-3) that Kane, in responding to this concern, lays out conditions on free agency that could be met even if CD were true, making his view libertarian by conjoining them with the supposition that it is in fact false. But this move does not advance his position, for it puts him in the camp of the compatibilist if CD turns out to be true and leaves the initial concern unaddressed if it is false.

24. Wolf, 147.

25. Cf. Strawson for a discussion of how these attitudes bear on "common sense" ascriptions of moral responsibility. I elaborate upon this discussion in Allen.

26. Watson, 19.

References

Allen, Robert. (1997) "Responsibility and Motivation," Southern Journal of Philosophy, vol. XXXV: 289-99.

Campbell, Joseph. (1997) "A Compatibilist Theory of Alternative Possibilities," Philosophical Studies 88: 319-28.

Copp, David. (1997) "Defending the Principle of Alternate Possibilities: Blameworthiness and Moral Responsibility." Nous vol. XXXI: 441-456.

Della Rocca, Michael. 1998. "Frankfurt, Fischer, and Flickers."

Nous XXXII: 99-105.

Double, Richard. (1991) The Non-Reality of Free Will. Oxford: Oxford Univ. Press.

Fischer, John, ed. (1986) Moral Responsibility (Ithaca, NY: Cornell Univ. Press).

Fischer, John. (1995) The Metaphysics of Free Will (New York: Oxford Univ. Press).

Fischer, John and Mark Ravizza, eds.. (1993) Perspectives on Moral Responsibility , (Ithaca, NY: Cornell Univ. Press).

Fischer, John and Mark Ravizza. (1998) Responsibility and Control. (Cambridge: Cambridge University Press).

Fischer, John and Hoffman, Paul. (1994) "Alternative Possibilities: A Reply to Lamb," Journal of Philosophy 91: 321- 26.

Frankfurt, Harry. (1969) "Alternate Possibilities and Moral Responsibility," in Fischer (1986): 143-52.

. (1971a) "Freedom of Will and the Concept of a Person." in Fischer (1991): 650-668.

. (1971b) "Three Concepts of Free Action: II" in Fischer (1986): 113-23.

. (1976) "Identification and Externality," in The Identities of Persons, ed. Amelie Oksenberg Rorty (Berkeley: Univ. of California Press, 1976): 239-51.

. (1982) "What We Are Morally Responsible For," in Fischer and Ravizza (1993): 286-95.

. (1987) "Identification and Wholeheartedness, "in Fischer and Ravizza (1993): 170-87.

Kane, Robert. (1996) The Significance of Free Will. Oxford Univ. Press.

Lamb, James. (1993) "Evaluative Compatibilism and the Principle of Alternative Possibilities," Journal of Philosophy 90: 517-27.

Nagel, Thomas. (1976) "Moral Luck." Proceedings of the Aristotelian Society vol. L. Reprinted in Thomas Nagel. Mortal Questions. (Cambridge: Cambridge University Press, 1981) 24-38.

Pereboom, Derk. (1995) "Determinism Al Dente," Noûs 29: 21-45.

Reid, Thomas. (1983) Inquiry and Essays, Indianapolis: Hackett Publishing Company.

Rosebury, Brian. (1995). "Moral Responsibility and Moral Luck."

Philosophical Review 104: 499-524.

Strawson, Peter. (1963) "Freedom and Resentment," in Fischer and Ravizza: 45-66.

Stump, Eleanore. (1988) "Sanctification, Hardening of the Heart, and Frankfurt's Concept of Free Will." Journal of Philosophy 85: 395-420.

_______________. (1998) "Freedom and Alternative Possibilities." Presented at the 1998 meeting of the APA Pacific Division.

Watson, Gary. (1997) "Soft Libertarianism and Hard Compatibilism." Presented at the 1997 University of Arkansas Conference on Free Will.