Flickering the W-Defense Flickering the W-Defense

One way to defend the Principle of Alternative Possibil-ities (PAP) against Frankfurt-style cases is to challenge the claim that agents in these scenarios are genuinely morally responsible for what they do. Alternatively, one can grant that agents are morally responsible for what they do in these cases but resist the idea that they could not have done otherwise. This latter strategy is known as the flicker defense of PAP. In an argument he calls the W-Defense, David Widerker adopts the former approach. I argue that, while Widerker’s argument does a poor job showing that these agents are not morally responsible for what they do, it does a very good job highlighting the alternative possibilities that remain open to agents in these cases and illustrating their moral significance (or robustness ). In doing so, my aim is to co-opt Widerker’s argument to bolster the most promising versions of the flicker defense.


INTRODUCTION
The Principle of Alternative Possibilities (PAP) states that people are morally responsible for what they have done only if they could have done otherwise. The greatest challenge facing this seemingly intuitive principle comes from a class of putative counterexamples developed by Harry Frankfurt (1969). Here is Frankfurt's original case: his hand unnecessarily. So he waits until Jones is about to make up his mind what to do, and he does nothing unless it is clear to him (Black is an excellent judge of such things) that Jones is going to decide something other than what he wants him to do. If it does become clear that Jones is going to decide to do something else, Black takes effective steps to ensure that Jones decides to do, and that he does do, what he wants him to do. Whatever Jones's initial preferences and inclinations, then, Black will have his way . . . Now suppose that Black never has to show his hand because Jones, for reasons of his own, decides to perform and does perform the very action Black wants him to perform. (Frankfurt, 1969, pp. 835-836) Frankfurt argues that cases of this sort provide good grounds for rejecting PAP. Here, for instance, it is not possible for Jones to avoid performing the action he performs. If moral responsibility requires the ability to do otherwise, as PAP states that it does, we would have to conclude that Jones is not morally responsible for what he does. Since Jones acts on his own, however, and Black plays no causal role in Jones's decision, there appears to be no good reason to deny that Jones is genuinely morally responsible for his actions. It is just as though Black had not been there at all. Insofar as this is correct, PAP should be rejected.
The scenario just sketched provides a blueprint for a certain kind of case. A great many versions of this original case have appeared in the philosophical literature in the over fifty years since Frankfurt introduced it. 1 The agents, actions, and ensuring conditions vary, but here is what they all have in common: The person, process, or mechanism that exists to ensure that the agent will act in the desired way, by standing ready to force the agent to so act unless they do so on their own, ends up not playing any role in producing the agents' actual behavior. 2 The agents perform the relevant actions all on their own, without any interference, just as they would have in an ordinary situation with no ensuring conditions in place.
There are two main ways to try to defend PAP against the challenge posed by Frankfurt-style cases. The first way is to challenge the claim that the agents in these cases are genuinely morally responsible for what they do. After all, if the agents are not morally responsible for their actions, then it does not matter whether they could have avoided doing what they did; the cases will not pose any kind of threat to PAP. The second way to defend this principle is to grant that the agents are morally responsible for what they do in these cases but resist the idea that they could not have acted differently than they did. According to this line of argument, even if Frankfurt-style cases manage to significantly restrict agents' alternative possibilities, they do not eliminate them in the way that undermining PAP requires. This latter strategy has come to be known as the flicker defense of PAP. 3 In an argument he calls the W-Defense (short for "What-should-he-have-done defense") David Widerker (2000Widerker ( , 2003 adopts the former approach, arguing that agents in Frankfurt-style cases are not morally responsible for what they do because, given their circumstances, there is nothing that it is reasonable to expect them to have done instead of what they did. 4 Here I set out to show that the W-Defense turns out to have a significant role to play in defending PAP against Frankfurtstyle attack-just not in the way Widerker intended. Although Widerker's argument does a poor job of showing that agents in Frankfurt cases are not morally responsible for what they do, it does a very good job of highlighting the alternative possibilities that remain open to agents in these cases and underscoring their moral significance (or robustness). In demonstrating this, my goal is to co-opt Widerker's argument to bolster the most promising versions of the flicker defense (thereby "flickering" the W-Defense).  (2000, pp. 191-192) Since there is nothing that it is reasonable to expect Jones to have done instead of what he did, Widerker concludes, we should reject the claim that Jones is blameworthy for what he has done. Without the judgment that agents in Frankfurt-style scenarios are to blame for their behavior, these cases pose no threat to PAP. 6

FLICKERING THE W-DEFENSE
The question at the heart of the W-Defense-What should the agent have done instead?-is exactly the right question to ask. Where Widerker goes wrong is in answering the question. 7 Widerker is correct to maintain that it is unreasonable to expect Jones to have refrained from deciding to kill Smith, but he is mistaken in thinking that there is nothing else that it is reasonable to expect Jones to have done instead. The reason it seems so plausible to say that Jones is morally blameworthy for what he did in this case is that he did what he did on his own. Although Black was poised to force Jones to decide to kill Smith, forcing Jones to do so turned out not to be necessary. Jones made the decision to kill Smith (and followed through on it) without any actual involvement from Black (or whatever mechanisms Black put in place). It is because Jones did this all on his own, just as he would have in more ordinary circumstances, that it is fair to say that it was just as though Black had not been there at all. This is why we blame Jones. We blame him for the decision he made precisely because Black did not have to force him to make it. This also makes clear what Jones should have done instead: He should have refrained from making that decision without being forced to make it. That is, he should have omitted to decide on his own to kill Smith (and omitted to kill Smith on his own). Now, even if we suppose that is the correct answer to give to the question raised by the W-Defense, how is that useful to those looking to mount a flicker defense of PAP? To see this, start by recalling that the general strategy of the flicker defense is to try to show that agents in Frankfurt-style cases continue to possess alternative possibilities. But what sort of alternative is required to do the work flicker theorists need it to do? PAP states that a person is morally responsible for what he has done only if he could have done otherwise-that is, only if he could have done otherwise than that for which he is morally responsible. So, that which a person must have been able to do other than is determined by that for which the person is morally responsible. The referent of the latter fixes the referent of the former. It is also generally held that PAP is properly understood as a thesis about direct (or non-derivative) moral responsibility. This is why Frankfurt-style cases tend to focus on the agents' responsibility for (and inability to avoid) the decisions they make rather than for the overt actions they perform-since it is commonly held that responsibility for the latter derives from responsibility for the former. 8 Minimally, then, defending PAP against Frankfurtstyle cases will require showing that agents in these cases could have done otherwise than that for which they are directly morally responsible.
What is it for which agents in these cases are directly morally responsible? According to some of the most popular (and, in my view, most promising) versions of the flicker defense, it is not the decisions they make but, rather, the way that they make them for which these agents are directly morally responsible. Specifically, they argue, what these agents are directly morally responsible for is making these decisions on their own, without being forced to do so. Some proponents of this strategy insist that making the relevant decisions (and performing the relevant actions) on their own is the only thing for which agents in Frankfurt scenarios are morally responsible and that they are not at all morally responsible for making the relevant decisions (and performing the relevant actions) simpliciter (Naylor, 1984;O'Connor, 2000;Speak, 2002;Capes & Swenson, 2017). Other flicker theorists allow that these agents can also be morally responsible for making the decisions they make (and performing the actions they perform) simpliciter but only insofar as their responsibility for these things derives from their responsibility for making those decisions (and performing those actions) on their own, for which they are directly morally responsible (Robinson, 2012(Robinson, , 2019. The central point on which they all agree is that, in a case where Jones decides to kill Smith (and does so) without needing to be forced to do so by Black, it is deciding on his own to kill Smith for which Jones is directly morally responsible. To the extent that Jones is morally responsible for anything else-e.g., killing Smith on his own, deciding to kill Smith (simpliciter), or killing Smith (simpliciter)-his responsibility for it derives from his responsibility for making this decision on his own. 9 Assuming that PAP is properly understood as a claim about direct moral responsibility and that Jones is directly morally responsible for deciding on his own to kill Smith, PAP implies that Jones must have been able to do otherwise than decide on his own to kill Smith. Of course, this is something that was well within Jones's power to do. Even though the presence of Black and his special device meant that it was not possible for Jones to avoid making the decision to kill Smith, whether Jones would make this decision on his own or not, without needing to be forced to do so by Black, is something that remained entirely up to him. As Frankfurt himself observed, "it is in a way up to [Jones] whether he acts on his own or as a result of Black's intervention. That depends upon what action he himself is inclined to perform. But whether he finally acts on his own or as a result of Black's intervention, he performs the same action" (1969, p. 836). Indeed, it is precisely because it was possible for Jones not to decide on his own to kill Smith that Black and his special device are part of the case to begin with. Otherwise, these fanciful ensuring conditions would be superfluous. The same goes for all Frankfurt-style cases.
Contemporary advocates of Frankfurt's challenge to PAP are generally willing to grant that, so long as they are indeterministic, there are always going to be some alternatives that remain open in these cases-including the possibility of the agents being forced to make the decisions they make instead of making them on their own. 10 But doing the work set out for them in the flicker defense, they argue, requires more than showing merely that it was possible for things to have gone slightly differently than they actually did in these cases. In order to defend PAP against the threat posed by Frankfurt-style cases it needs to be shown that agents in these scenarios continue to have alternatives that are sufficiently robust. That is, these alternative possibilities need to be morally significant in a way that would help to explain why the agents in these cases are morally responsible for what they did. As John Martin Fischer puts it, it is not enough for the flicker theorist to analyze the relevant range of cases in such a way as to identify an alternative possibility. Although this is surely a first step, it is not enough to establish the flicker of freedom view, because what needs also to be shown is that these alternative possibilities play a certain role in the appropriate understanding of the cases. That is, it needs to be shown that these alternative possibilities ground our attributions of moral responsibility. (1994, p. 140; see also McKenna, 2003;Pereboom, 2012) The core sentiment underlying Fischer's remarks is that it is plausible to suppose that moral responsibility requires alternative possibilities only if this is because moral responsibility requires alternatives of a particular sort. It is implausible to maintain that moral responsibility requires that agents had alternative possibilities if those alternative possibilities are irrelevant to, and do not figure into an explanation of, why we should regard agents as being morally responsible for what they have done. It is for this reason that some of the most prominent Frankfurt defenders, rather than seeking to eliminate all alternative possibilities, have shifted their efforts to constructing scenarios that succeed in eliminating all robust alternatives, even if they leave open a range of non-robust alternative possibilities (see Pereboom, 2000Pereboom, , 2001Pereboom, , 2003Pereboom, , 2012McKenna, 2003;Hunt, 2005). If some Frankfurt-style cases are successful in showing that moral responsibility does not require alternative possibilities that are robust, that would be sufficient to significantly undermine the view that moral responsibility requires alternative possibilities of any sort. 11 Contemporary proponents of the flicker strategy are generally willing to grant that mounting a successful flicker defense of PAP against Frankfurt-style attack will have to involve showing that agents in these cases had alternative possibilities that are sufficiently robust. It is precisely at this point that the W-Defense is able to do work for versions of the flicker strategy that focus on the alternative possibility of agents omitting to make the relevant decisions on their own. As Widerker argues, it is essential to judging Jones to be morally blameworthy for what he did that there is something else we think he should have done instead. The fact that the intuitively correct thing to say about what Jones ought to have done instead is that he should have refrained from making the decision to kill Smith on his own, without needing to be forced to do this, indicates that this particular alternative possibility is one that really does help to ground our responsibility attributions in this case and is relevant to explaining Jones's blameworthiness for what he actually did in the way that robustness requires.
The same cannot be said of all the other alternative possibilities that various versions of the flicker defense have honed in on in the decades since these cases were first introduced. In some Frankfurt-style cases, for example, the way the ensuring conditions operated is that certain mechanisms were poised to force the agents to make the desired decisions at a certain time t unless the agents had already made that decision on their own by t (Mele & Robb, 1998. In other cases, the mechanisms were set to force the agents to make the desired decisions the moment the agents even started to consider (the reasons favoring) making a different decision, which was stipulated to be a necessary but not sufficient condition for making an alternative decision (Pereboom, 2000(Pereboom, , 2001(Pereboom, , 2003(Pereboom, , 2012Hunt, 2005). As some flicker theorists have pointed out, even though the ensuring conditions guaranteed that agents in these cases could not have avoided making the decisions they made, it was possible for them to have made those decisions at slightly different times than they did (Ginet, 1996(Ginet, , 2002Franklin, 2011;Palmer, 2011). Suppose, for instance, that Jones on his own made the decision at t-2 to kill Smith. There was nothing to prevent Jones from making that decision a moment earlier at t-3 or a moment later at t-1, and both of these options would have been well within his voluntary control. Returning to the central question from the W-Defense, though, it is not nearly as plausible to suppose that the intuitively correct thing to say about what Jones should have done instead of what he did is that he should have made the decision to kill Smith at a slightly different time than he did. Nor is it nearly as plausible to suppose that a significant part of the explanation for our willingness to blame Jones for what he did is the fact that it was possible for him to have decided to kill Smith a few seconds earlier or a couple seconds later (see Pereboom, 2012). Thus, while the W-Defense can be enlisted to help bolster versions of the flicker defense that focus on the possibility of agents refraining from making the bad decisions on their own, this strategy is not equally available to versions of the flicker defense that focus on other alternative possibilities.

AN OBJECTION AND A REPLY
Widerker has suggested that omitting to make the relevant decisions on their own is not something that it is morally reasonable to expect agents in Frankfurt-style cases to have done (even in indeterministic cases where this was well within their voluntary control). 12 To omit to make a given decision on one's own, he says, means either not to make it at all or to be forced to make it. Since agents in Frankfurt-style cases could not have avoided making the decisions they made, the expectation that they should have omitted to do so on their own "reduces to" the expectation that they should have made it the case that they were forced to make those decisions (2003, p. 65). According to Widerker, however, "A person can reasonably be expected to make it the case that p only if he knows (or should have known) or has good reasons to believe that he should make it the case that p" (p. 65). Thus, given that agents in Frankfurt-style cases did not know (nor should they have known) that they were in situations where they would be forced to make the relevant decisions unless they did so on their own, it would be morally unreasonable to expect them to have made it the case that they were forced to make the decisions they made. 13 The problem with this objection is that it rests on an equivocation between two different senses of (moral) expectation. Like 'know', 'expect' is an intensional verb which allows for ambiguity regarding its complements. Widerker's argument elides this fact and capitalizes on the resultant ambiguity. Once this ambiguity is resolved, the objection fades.
Consider, first, Widerker's claim that, since they could not avoid making the decisions they made, the expectation that agents in Frankfurt-style cases should have omitted to make those decisions on their own reduces to the expectation that they should have made it the case that they were forced to so decide. In what sense does (subjecting these agents to) the former expectation reduce to (subjecting them to) the latter? There are at least two ways to interpret this claim: 1a. To expect of agents in Frankfurt-style cases that they omit to decide on their own to A entails subjecting them to an expectation with the content "Make it the case that you are forced to decide to A." 1b. To expect of agents in Frankfurt-style cases that they omit to decide on their own to A is in effect to subject them to the expectation that they make it the case that they are forced to decide to A.
1a seems obviously false. Whereas an expectation with the content "Omit to decide on your own to A" would be met in any possible world where the agents omit to make the relevant decisions on their own-regardless of whether they are forced to make those decisions-an expectation with the content "Make it the case that you are forced to decide to A" would be fulfilled only in cases where the agents are forced to make the relevant decisions. Subjecting agents to the former expectation need not involve subjecting agents to the latter expectation-even in cases where, as a matter of contingent fact, acting in a way that meets the former expectation would also be sufficient for meeting the latter. Presumably, then, 1b is closer to what Widerker has in mind. In addition to being the only plausible alternative of the two, it is the only interpretation of this claim that can reasonably be thought to follow from the considerations Widerker offers in its favor. To better see this connection, it will be helpful to spell out more precisely what it is to in effect expect of a person that she make something the case (as opposed to in fact expecting of a person that she make something the case). Here is a proposal: To in effect expect of a person that she make it the case that p is to expect of her that she do (or omit to do) q and to subject her to that expectation under conditions in which her doing (or omitting to do) q will make it the case that p. With this understanding in place, we can further clarify 1b as follows: 1b*. To expect of agents in Frankfurt-style cases that they omit to decide on their own to A is to expect of them that they omit to decide on their own to A and to subject them to this expectation under conditions in which omitting to decide on their own to A will make it the case that they are forced to so decide.
Understood in this way (as 1b*), 1b appears to be uncontroversial. Now consider Widerker's claim that "a person can reasonably be expected to make it the case that p only if he knows (or should have known) or has good reasons to believe that he should make it the case that p" (p. 65). As before, there are at least two ways this can be interpreted: 2a. It is morally reasonable to subject a person to an expectation with the content "Make it the case that p" only if she knows (or should have known) or has good reasons to believe that she should make it the case that p.
2b*. It is morally reasonable to expect of a person that she do (or omit to do) q and to subject her to that expectation under conditions in which her doing (or omitting to do) q will make it the case that p only if she knows (or should have known) or has good reason to believe that she should make it the case that p. 14 Here, the precise nature of the problem with Widerker's argument begins to come into focus. Nothing logically follows from the conjunction of 1b* and 2a. The same goes for 1a and 2b*. Thus, even assuming that 2a is correct, this will show that omitting to decide on one's own to do that which one knows to be wrong is not a robust alternative in Frankfurtstyle cases only when taken together with 1a. As noted earlier, however, 1a is pretty plainly false.
The only other valid inference to the judgment that it is morally unreasonable to expect agents in Frankfurt-style cases to have refrained from making the relevant decisions on their own is from 1b* and 2b*. Like 1a, however, 2b* should be rejected. Consider, for instance, a case in which Lois has promised Clark that she will meet him for lunch. All else being equal, it is morally reasonable to expect of Lois that she keep her promise. Now suppose that Jimmy has wagered Clark twenty dollars that Lois will fail to show up. As a result, if Lois keeps her promise to meet Clark for lunch, then, unbeknownst to her, she will thereby make it the case that Clark wins twenty dollars. According to 2b*, this implies that it is morally unreasonable to expect of Lois that she keep her promise to meet Clark for lunch. But surely this is incorrect. Given that Lois is unaware (and had no good reason to be aware) of Clark's wager with Jimmy, it might well be morally unreasonable to (in fact) expect of her that she make it the case that Clark wins twenty dollars-that is, to subject to her to an expectation with the content "Make it the case that Clark wins twenty dollars." Furthermore, since, as a matter of contingent fact, keeping her promise to meet Clark for lunch will make it the case that Clark wins twenty dollars in this case, to expect of Lois that she keep her promise to Clark is in effect to expect of her that she make it the case that Clark wins twenty dollars (in the sense identified in 1b*). Even so, it is morally reasonable to expect of Lois that she keep her promise to meet Clark for lunch. After all, it is surely reasonable to expect this of her in a version of the case without the wager, and it is utterly mysterious how the mere addition of Jimmy's wager with Clark could do anything to change this. 15 The same is true of agents in (indeterministic) Frankfurt-style cases. Assuming they are morally responsible for what they did in these cases, they knew they were deciding to do something that was morally wrong for them to do, and it is reasonable to expect them not to have made the decision to do that on their own, without needing to be forced to do it-just as it would be reasonable to expect this of them in ordinary scenarios. Of course, since, as Widerker points out, it is a crucial feature of these cases that the agents neither knew nor ought to have known that they would be forced to make these decisions if they did not do so on their own, it might well be true that it is unreasonable to expect them to have intentionally or knowingly brought it about that they were forced to make these decisions. But this does nothing to diminish the reasonableness of expecting them to have omitted to decide on their own to do that which they knew to be wrong.
This last point bears emphasizing. In claiming that Jones ought to have made Black force him to decide to kill Smith by refraining from deciding on his own to kill Smith, I am not suggesting that Jones should have knowingly or intentionally made Black force him to make that decisionany more than I think Lois should have knowingly or intentionally made Clark win $20. Nor am I suggesting that Jones ought to have done something that he himself would have (explicitly, consciously) viewed as refraining from deciding on his own to kill Smith or as refraining from killing Smith on his own-any more than I think Lois ought to have done something that she herself would have viewed as making Clark win his wager with Jimmy. Instead, I am saying that Jones should have acted in a way that in fact would have involved refraining from deciding on his own to kill Smith and so would have made Black force him to decide to kill Smith.
At this point, one might wonder what exactly Jones would have viewed himself as doing if he were to have opted for the alternative possibility that on-your-own flicker theorists contend he should have. After all, it seems plausible to suppose that, for it to be reasonable to expect Jones not to have decided on his own to kill Smith and so made Black force him to make this decision, there must be some description under which it is reasonable to expect Jones to have considered that alternative. 16 In the above example, for instance, even though there is nothing it is reasonable to expect Lois to have done that she would have viewed as making it the case that Clark wins $20, it is reasonable to expect her to have done this because it is reasonable to expect her to have considered doing this under the description "meeting Clark for lunch" or "keeping her promise to Clark." That is, it is reasonable to expect this of her because it is reasonable to expect her to have done what she would have viewed as meeting Clark for lunch and keeping her promise to Clark (which, though she did not know this, would have sufficed for making Clark win $20).
From Jones's perspective, then, what is it he would be doing in the alternative sequence when he was in fact omitting to decide on his own to kill Smith and was in fact bringing it about that Black forces him to make that decision? And is that something that it is reasonable to expect Jones to have considered doing as an alternative to what he did? The best way to see the answer to this is to think about an ordinary, non-Frankfurt-style scenario. Since Jones does not know that he is in a Frankfurt scenario, his perspective is going to be the same as it would be in a normal situation. So, consider an ordinary case in which Jones deliberates about killing Smith but ultimately refrains from doing so. To make this even more analogous to Jones's situation in the Frankfurt-style cases, let us imagine this is a case in which, though he does not know it, Jones has a limited window of opportunity in which to (decide to) kill Smith. If he fails to (decide to) kill Smith by a certain time, the window will close and he will have omitted to (decide to) kill Smith. What would this have been like from Jones's point of view? From his perspective, Jones would be refraining from, or holding off on, killing Smith, at least temporarily. (In normal cases we typically think in terms of what we are going to do, not in terms of what we are going to decide to do, let alone what we are going to decide on our own to do.) He would be considering killing Smith, and perhaps trying to talk himself into doing it or against doing it, but he would have stopped short of committing (i.e., deciding) to kill Smith. That is not to say that he would have necessarily decided not to kill Smith. 17 Minimally, however, he would have held off on committing to kill Smith until that window of opportunity closed, at which point it would be true that he omitted to kill Smith.
From Jones's point of view, then, refraining from deciding on his own to kill Smith in the alternative sequence of a Frankfurt-style case would be phenomenologically indistinguishable from refraining from (or holding off on) deciding to kill Smith in a non-Frankfurt-style scenario-at least, right up to the point where Black's mechanism is activated, forcing Jones to decide to kill Smith, from which point, presumably, this would be phenomenologically indistinguishable from deciding on his own to kill Smith in a non-Frankfurt-style scenario. Is that something that it is reasonable to expect Jones to have considered as an alternative to what he did? Is it reasonable to expect Jones to have considered holding off on (committing to) killing Smith as an alternative to (committing to) killing him? Of course it is. Assuming that Jones is morally responsible for what he did, it is reasonable to expect him to have considered refraining from doing that which he knew to be morally wrong as a morally preferable alternative to doing it. And if Jones had done so-that is, if he had done what he would have viewed as holding off on (committing to) killing Smith (at least momentarily)-what he in fact would have been doing (even though he would not have known it or viewed it this way) is omitting to decide on his own to kill Smith and so making Black force him to make that decision. That is why it is reasonable to expect Jones to have made Black force him to decide to kill Smith by refraining from making that decision on his own.

CONCLUSION
The flicker strategy provides an important and intuitively forceful challenge to Frankfurt-style cases and the threat they pose to PAP. Versions of the flicker defense that focus on the alternative possibility of agents omitting to make the relevant decisions on their own are particularly promising and continue to receive attention in the philosophical literature (see, e.g., Sartorio, 2019;Chevarie-Cossette, 2021;Stockdale, 2022;Cyr, 2022;Haji, 2023). While contemporary critics of this strategy are generally willing to grant that this alternative possibility is one that cannot easily be eliminated from Frankfurt-style cases, they resist claims that this alternative is sufficiently robust. 18 It is precisely at this point, the crux of the current debate on this issue, that the W-Defense can be brought to bear in support of this version of the flicker defense by throwing into sharp relief the moral significance of this alternative possibility and the central role it plays in explaining our attributions of moral responsibility. For those of us who share the view that the agents are blameworthy for their behavior in (at least some of) these cases, it is reasonable to ask what we think these agents ought to have done differently than they did. The intuitively obvious answer here is that the agents should have refrained from making the bad decisions on their own. They should have had to be forced to make those decisions, and they should have behaved in a way that resulted in their being so forced. As we have seen, this alternative possibility is one that is embedded in the core structure of Frankfurt-style cases and continues to remain open to agents in these scenarios. Insofar, then, as the W-Defense is able to help demonstrate that this particular alternative possibility is robust, that will go a long way towards neutralizing the threat that these cases are widely thought to pose to PAP. 19  Widerker (1995) has also challenged Frankfurt-style cases on the grounds that they fail in their intended aim to constitute genuine "IRR-situations" in which the conditions that make it impossible for agents to avoid performing certain actions in no way bring it about that they perform those actions. For more on this, see n. 10. 5 The actions at issue in Frankfurt-style cases are nearly always mental acts of deciding. This is because PAP is widely understood to be a thesis about direct (or non-derivative) moral responsibility and responsibility for overt actions (e.g., stealing a car) is commonly taken to derive from responsibility for mental acts of deciding (e.g., deciding to steal the car). Also, if it were possible for agents in Frankfurt-style cases to have made different decisions, then, even if the circumstances prevented them from following through on those different decisions, that would pretty clearly suffice for the agents' continuing to possess significant alternative possibilities. 6 Frankfurt-style cases typically focus on cases of blameworthiness in which agents decide to act in ways that they know to be morally wrong. 7 But see Capes (2010) for a challenge to the idea that Widerker's question is the right question to ask. For a reply, see Palmer (2013). 8 To say that a person is directly (or non-derivatively) morally responsible for something, here, should be taken to mean that their responsibility for it is not inherited from (or in virtue of) their being morally responsible for anything else. Moral responsibility that is inherited in this way is indirect (or derivative) (Mele, 2006, p. 86). 9 Proponents of this version of the flicker strategy have offered various arguments in support of the claim that agents in Frankfurt-style cases are directly morally responsible not for making the decisions they make but for making those decisions on their own. (See Naylor, 1984;Robinson, 2012Robinson, , 2019Capes & Swenson, 2017.) I will not rehearse those arguments here, as my primary goal in this paper is not to offer a full defense of this version of the flicker strategy so much as to demonstrate how the W-Defense can be brought to bear on its behalf. 10 Early Frankfurt-style cases relied on employing signals about what actions the agents would perform if left to act on their own in order to indicate whether intervention was going to be necessary. These so-called "prior sign" cases faced a dilemma: Either these signs (or that which they indicated) causally determined the agents' actions or they were merely very reliable indicators of this. If the former, then these cases would be dialectically ineffective with incompatibilists, who could not be expected to share the intuition that agents in these cases are morally responsible for their actions. If the latter, then the occurrence of the signals would not strictly foreclose possibility that the agents might act differently than indicated. Following the publication of this problem (Kane, 1985(Kane, , 1996Widerker, 1995;Ginet, 1996), cases in the literature have tended to be explicitly indeterministic. 11 See Robinson (2014) for more on this distinction between direct Frankfurt-style attack (aimed at falsifying PAP) and indirect Frankfurt-style attack (aimed at undermining the plausibility of and the motivation for PAP). See also Speak (2002). 12 Widerker couches his discussion of this alternative possibility in terms of an agent's not making (or bringing it about that she does not make) the relevant decision freely, whereas I have been describing this as an agent's omitting (or seeing to it that she does not make) the relevant decision on her own, but this difference in terminology does not appear to amount to a substantive difference. We are both addressing the same alternative. 13 The kind of expectation at issue here is moral in nature and belongs to a class of what we might call normative expectation. To subject someone to an expectation of this sort is to hold her to a certain standard of behavior. It should not be confused with what we might call predictive or anticipatory expectation, which is a matter of believing that things will go a certain way in the future. 14 As with 1b*, 2b* is the claim (which we could label 2b) that "it is morally reasonable to in effect expect a person to make it the case that p only if she knows (or should have known) or has good reasons to believe that she should make it the case that p" interpreted in light of the proposed understanding of what it means to in effect expect something of someone. 15 Similarly, we could suppose that, by keeping her promise to meet Clark for lunch, Lois would make it the case that she is the second person to sit in a particular chair that day, or that the temperature in the restaurant is 0.2 degrees higher than it otherwise would have been, or that any number of other equally obscure states of affairs obtain. It seems passing strange to think that the reasonableness of expecting of Lois that she keep her promise to Clark would hinge on whether she is aware (or ought to be aware) of any of these things. 16 Thanks to Michael McKenna for encouraging me to say more about this concern, which is similar to the main point he makes in McKenna (2003) in his discussion of "deliberative significance." 17 Since the alternative of deciding not to kill Smith can be blocked in indeterministic Frankfurt-style cases, flicker theorists do well to avoid hanging their hat on that alternative (or any alternative decision or intention, for that matter). For more on this point, see Robinson (2014). 18 But see Stockdale (2022) for a recent attempt to construct an indeterministic Frankfurt-style case in which the agent is unable to avoid making the desired decision on his own. 19 Thanks to Al Mele, Randy Clarke, and Michael McKenna for helpful comments and discussion. An early version of this paper was presented at the 2013 meeting of the Florida Philosophical Association and the 2014 Pittsburgh Area Philosophy Colloquium. I am grateful to those audiences for their feedback.