<%BANNER%>

Dilemma for Frankfurt-Style Cases, and the Definition of 'Robustness'


PAGE 1

A DILEMMA FOR FRANKFURT-STYLE CASES, AND THE DEFINITION OF 'ROBUSTNESS' By JACLYN GENCHI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by JACLYN GENCHI

PAGE 3

iii TABLE OF CONTENTS page ABSTRACT....................................................................................................................... iv INTRODUCTION...............................................................................................................1 AN OBJECTION TO FRANKFURT-STYLE CASES, AND REPLIES.........................12 The First Horn of the Dilemma: Assumption that the Prior Sign Determines the Agents Choice (i.e., That the Prior Sign Deterministically Causes the Agents Choice)...................................................................................................................20 The Second Horn of the Dilemma: Assumption that the Prior Sign is Not Sufficient for the Agents Choice (i.e., That the Agents Choice Is Indeterministically Caused)...................................................................................29 No-Prior-Sign Cases............................................................................................30 Blockage Cases....................................................................................................34 Internal-Sign Cases..............................................................................................40 Necessary-Condition Cases.................................................................................43 THE DEFINITION OF ROBUSTNESS.........................................................................49 LIST OF REFERENCES...................................................................................................69 BIOGRAPHICAL SKETCH.............................................................................................71

PAGE 4

iv Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts A DILEMMA FOR FRANKFURT-STYLE CASES, AND THE DEFINITION OF 'ROBUSTNESS' By Jaclyn Genchi August 2006 Chair: Jon Tresan Major Department: Philosophy The purpose of this thesis is to argue that one particular objecti on to Frankfurt-style cases has thus far not been shown to be mistaken. I defend the objection called the dilemma defense of the Principle of Altern ate Possibilities objection from the replies given by proponents of Frankfurt-style cases to defend these cases from this objection.

PAGE 5

1 INTRODUCTION Frankfurt-style cases are traditionally taken to show that the Principle of Alternative Possibilities (PAP) is false1 – to show that alternative possibilities are not required for moral responsibility. In this essa y I will be discussing a particular objection to Frankfurt-style cases, the dilemma defense of the Principle of Alternate Possibilities. I will proceed by first explaining the role of Fr ankfurt-style cases in the debate concerning moral responsibility, and will follow with an explanation of the objection to Frankfurtstyle cases I will be discussing, the dilemma defense of PAP. Proponents of Frankfurtstyle cases have responded to both horns of the dilemma defense of PAP and, after reviewing these responses, I will argue that such responses are inadequate and, so, the dilemma defense of PAP stands as an object ion to Frankfurt-style cases. I will give substantial attention to the strategy used by Derk Pereboom as a response to the dilemma defense of PAP, arguing that it fails. Throughout this essay I will be discussing one of the many facets of the debate concerning moral responsibility, whether Fra nkfurt-style cases succeed in showing that the Principle of Alternate Possibilities is fals e. In considerations of moral responsibility, much of the debate has been focused on what is called the Compatibility Question: can we be morally responsible agen ts in a deterministic world? We are concerned to answer this question because there seems to be a c onflict between our picture of ourselves as 1 The Principle of Alternate Possibilities (PAP) is as follows: having an alternative possibility to choosing/doing x is necessary for an agent to be morally responsible for choosing/doing x

PAGE 6

2 agents that have control over what we c hoose and do, and our picture of ourselves as parts of a world governed by natu ral-law-abiding forces. We belie ve that what we do is in some way up to us, while at the same time we believe that what we do is a part of the causal order of the world, and these two pictures of ourselves seem to be in conflict. The thesis of determinism2 is, roughly, that the events of the past together with the laws of nature entail the events of the present and futu re; if all events of the present and future are entailed by the past, it seems that all of our choices and actions are entailed by events (together with the laws of nature) that happe ned before we were born. If the thesis of determinism is true, we are challenged to fi nd a way in which we can maintain our view that we have any control over our choices, our actions, the way in which we perform our actions, whether we perform our actions, etc. Two positions are taken in response to the Compatibility Question: Compatibilism and Incompatibilism; the Compa tibilist claims that moral re sponsibility and determinism are compatible (i.e., that the truth of the thes is of determinism would not entail that we cannot be morally responsible for what we do), while the Incompatibilist claims that moral responsibility and determinism are not co mpatible (i.e., that the truth of the thesis of determinism would entail that we cannot be morally responsible for what we do). There are two approaches a Compatibilist position may take. On the one hand, a Compatibilist may claim that whether or not de terminism obtains is ir relevant to whether or not we can be morally responsib le for what we do. P.F. Strawson3 and Harry G. 2 I like Peter van Inwagen’s statement of the thesis of determinism in An Essay on Free Will According to van Inwagen, the thesis of determinism is the conjunction of the following two theses: (1) For every instant of time, there is a proposition that expresses the state of the world at that instant, and (2) If p and q are any propositions that express the state of the world at some instants, then the conjunction of p with the laws of nature entails q [van Inwagen, 1983, p. 65]. 3 Strawson, 1982.

PAGE 7

3 Frankfurt4 both present Compatibilist views of this sort. On the other hand, a Compatibilist may claim that we can be mo rally responsible for what we do because whatever condition(s) is (are) necessary and sufficient for being morally responsible are compatible with determinism. Compatibilism of the latter sort has been the most prominent Compatibilist approach advocated in recent literature, notably by John Martin Fischer.5 There are three kinds of Inco mpatibilist: the Libertarian, the Hard Determinist, and the Hard Incompatibilist. The Libertarian claims that at least some of the things we do are not determined and, so, we can be morally re sponsible for such things, while both the Hard Determinist and the Hard Incompatib ilist claim that we cannot be morally responsible for any of the things we do. The Hard Determinist claims (disagreeing with the Libertarian) that all of the things we do are in fact determined and, so, we cannot be morally responsible for anything that we do. The Libertarian and the Hard Determinist agree, though, on the following point: that a necessary condition on being morally responsible for some action is that it be not determined. This view of moral responsibility is generally taken to be the Incompatibilist intuition, the core of the Incompatibilist position. Traditionally, I believe, most Incompatib ilists were led to ei ther the Libertarian or Hard Determinist position by first believing that a requirement of my being morally responsible for something I do is that in so doing I was not de termined to do so. There is, though, a variant of the Incompa tibilist position, the Hard In compatibilist, which adds that even an action that was not determined, but was undetermined, is not something that 4 Frankfurt, 1971. 5 Fischer, 1994.

PAGE 8

4 we can be held morally responsible for. The Hard Incompatibilist claims that we cannot be morally responsible for anything we do, rega rdless of whether or not all of what we do is determined. The Hard Incompatibilist clai ms (disagreeing with both the Libertarian and Hard Determinist) that regardless of whether the thesis of determinism is true or false we are not morally responsible for what we do b ecause we cannot be morally responsible for undetermined choices/actions as well as determined choices/actions. The Hard Incompatibilist claims that indeterminacy w ould not allow us the sort of control we believe is required for moral responsibility a nd which the Incompatibilist believes to be missing in a deterministic world. There is another question (one related to the Compatibility Question) that is central to the debate concerning moral responsibility: what condition(s) must be satisfied in order for an agent to be morally responsible for so mething he has done? In tuitively, there is at least one condition that must be satisfied in order for an agent to be responsible for what he does: the agent must have some kind of cont rol with respect to th is choice/action. Both Compatibilists and Incompatibilists (for the mo st part) agree that an agent’s having some kind of control over what he does is necessary for that agent to be morally responsible for so doing. There is major disagreement, t hough, over just what kind of control is necessary for being morally responsible. Ther e are two main positions that are taken on this issue. One position claims that having th e relevant sort of control amounts to an agent being the source or orig inator of his actions; one who takes this position is either a Source Compatibilist or a Source Incompatibilis t. The Source Compatibilist holds that the sort of control associated with being the source or originator of one’s actions is a sort of control that both is sufficient for moral responsibility and is compatible with

PAGE 9

5 determinism, whereas the Source Incompatibilist holds that the sort of control associated with being the source or orig ination of one’s actions is undermined by determinism. Another position taken with re spect to the question of what kind of control is required for moral responsibility is that th is kind of control amounts to having and/or exercising the ability to choose a course of action from alte rnatives; one who takes this position is either a Leeway Compatibilist or a Leeway Incompatibilist. Both the Leeway Compatibilist and the Leeway Incompatibilist ho ld that what is (at least partly) relevant to an agent’s being morally responsible fo r something is that he could have done otherwise than he actually did, that he had ope n to him alternative possibilities. Both the Leeway Compatibilist and the Leeway Inco mpatibilist adhere to the Principle of Alternate Possibilities (PAP): ha ving alternate possibilities to what one does is necessary for being morally responsible for so doing. Th e Leeway Compatibilist holds that having alternatives to what one does is necessary for being morally responsible, and that having them is compatible with determinism. The L eeway Compatibilist will typically claim that to claim that one has alternatives to what one does is to claim th at one would have done otherwise if one had so chosen, where this ability is taken as compatible with one’s actual choice being determined. The Leeway Incompa tibilist holds that the ability to choose/do otherwise is necessary for being morally res ponsible, and that this ability is undermined by determinism; if determinism is true, th en one never really has any alternative possibilities to what one actually chooses/does. If determinism is true, then the past together with the laws of nature entails wh at one chooses/does at any moment and, so, if determinism is true, one never can do anything other than what one does do.

PAGE 10

6 1. We can see that the truth of PAP is playing a crucial role in the Leeway Incompatibilist’s argument for the Incomp atibility of moral responsibility and determinism. The argument presented by the Leeway Incompatibilist for the incompatibility of moral responsibil ity and determinism is as follows: 2. Moral responsibility requires havi ng alternate possibilities (PAP). 3. If the thesis of determinism is true, then one never can do otherwise than what one actually does, one never has a ny alternative possibilities. 4. Therefore, the truth of the thesis of dete rminism entails that no one is ever morally responsible for what one does (i.e., mora l responsibility and determinism are incompatible). Thus, we can see that if PAP is false, the Leeway Incompatibilist’s argument for the incompatibility of moral res ponsibility and determinism fails. In his 1969 essay “Alternate Possibiliti es and Moral Responsibility,” Harry Frankfurt proposed that a case could be given to show that PAP is false. David Widerker gives a nice description of th e sort of case that Frankfur t is supposing would show PAP false: (FR) There may be circumstances in wh ich a person performs some action which, although [these circumstances] make it impo ssible for him to avoid performing that action, [these circumstances] in no way bring it about that he performs it.6 If a case could be given in which (i) a pe rson cannot do otherwise than what he does do (i.e., in which the agent has no alternate possibilities) and (ii) whatever explains why he cannot do otherwise is not what causes or brings about the action he performs, then if in such a case (iii) the agent is morally responsib le for what he does, this would show that having alternate possibilities is not required for moral responsibility. If a case could be given in which the agent has no alternate possi bilities to what he actually does (condition (i) of an FR scenario), and yet he is mora lly responsible for what he does (condition (iii) of an FR scenario), this would show that ha ving alternate possibilities to what one does is 6 Widerker, 2002, p. 323.

PAGE 11

7 not required for being morally responsible for so doing. In order to give a case that meets both conditions (i) and (iii ) of an FR scenario, it must be that the thing that makes it the case that the agent cannot do othe rwise is not the same thing th at brings about the agent’s choice or action (condition (ii) of an FR scen ario), because if what removes the agent’s alternate possibilities is the same thing that brings about the agent’ s choice or action, we would claim that agent was compelled or constr ained to choose or act as he did, in which case he would not be morally responsible for so doing. If a case could be given that meets all three conditions of an FR scenario, then such a case would be a counterexample to PAP. The following case, which is John Mart in Fischer’s standard example, is a Frankfurt-style case, a case meant to satisfy all three conditions of an FR scenario: Suppose Jones is in a voting booth deliberat ing about whether to vote for Gore or Bush. After reflection, he chooses to vote for Gore and does vote for Gore by marking his ballot in the normal way. Unbeknownst to him, Black, a liberal neurosurgeon working with the Democrat ic Party, has implanted a device in Jones’s brain which monitors Jones’s brain activities. If he is about to choose to vote Democratic, the device simply conti nues monitoring and does not intervene in the process in any way. If, however, J ones is about to choose to vote, say, Republican, the device triggers an interven tion that involves electronic stimulation of the brain sufficient to produce a ch oice to vote for the Democrat (and a subsequent Democratic vote). How can th e device tell whether Jones is about to choose to vote Republican or Democratic? Th is is where the “prior sign” comes in. If Jones is about to choose at T2 to vote for Gore at T3, he shows some involuntary sign – say a neurological pa ttern in his brain – at T1. Detecting this, Black’s device does not intervene. But if J ones is about to choose at T2 to vote for Bush at T3, he shows an involuntary sign – a different neur ological pattern in his brain – at T1. This brain pattern would trigger Black’s device to intervene and cause Jones to choose at T2 to vote for Gore, and to vote for Gore at T3.7 So, in this case, Jones does show the relevant prior sign and does vote for Gore, without any intervention from the device. It s eems, then, plausible to claim that Jones is morally responsible for his choice to vote for Gore. But, had Jones not shown the relevant 7 Fischer, 2002, p. 282.

PAGE 12

8 prior sign (i.e., had Jones shown a sign indica ting that he would do otherwise than choose to vote for Gore), the device would have inte rvened and caused Jones to choose to vote Gore. So, this shows that Jones could do not do otherwise than choose to vote for Gore; in this case, Jones has no alte rnate possibility to his actual choice. This case, then, seems to be an FR scenario, since this is a case in which (i) Jones has no alternate possibility to what he does (where “does” in this case refe rs to both his choice to vote for Gore and his subsequent action of voting), (i i) whatever makes it the case that Jones has no alternate possibility in this case the intervening device in no way brings about Jones’s actual choice, and (iii) Jones is morally responsible for what he actually does. Since all three conditions of an FR scenario seem to be sa tisfied by Fischer’s case, this Frankfurt-style case seems to be a counterexample to PAP. Before any evaluation of whether or not such Frankfurt-style cases, cases like Fischer’s given above, succeed in showing that PAP is false (i.e., succeed in meeting all three conditions of an FR s cenario), an important limitation on condition (i) of an FR scenario must be discussed. When the propone nt of PAP claims that having alternate possibilities to what one does is required fo r being morally responsible for so doing, the proponent of PAP is not claiming that having just any altern atives is what is necessary for moral responsibility. Consider a case in which Jones chooses at t0 to rob a bank and, consequently, robs a bank at t1, and does so under no compulsion from any internal or external sources. It could always be cl aimed that, rather than choosing at t0 to rob a bank, it was possible that a mete or fall on Jones at t0, or that Jones is teletransported by aliens to another universe at t0, or that Jones spontaneously combusts at t0, etc. Although it is possible that any of these things happen at t0 as an alternative to what Jones actually does

PAGE 13

9 at that time, none of these things would c ount as an alternate possibility that could plausibly be claimed relevant to explaining why Jones is morally responsible for what he actually does at t0, choose to rob a bank. More pl ausibly, the proponent of PAP is claiming that having only a certain kind of a lternate possibility to wh at one does is what is necessary for being responsible for so doing. We call this certain kind of alternate possibility a robust alte rnate possibility, which is the sort of possibility that it would be reasonable or intuitive to claim that having th is kind of alternate possibility is what (at least partially) explains why an agent is responsible for what he actually does; according to the proponent of PAP, an agent is mora lly responsible for doing something (at least partly) in virtue of having a robust alternate possib ility (i.e., having a robust alternate possibility is explanatorily relevant to an agent’s moral responsibility). A way to understand what the proponent of PAP means by ‘robust alternative possibility’ is to imagine a case in which all other conditions one believes necessary for moral responsibility are met (that is, all conditi ons other than having robust alternative possibilities; for example, one may believe th at it is necessary that the agent does not have certain forms of mental illness that resu lt in compulsion, that the agent is not under any form of constant direct br ain manipulation by another agen t, etc.), and then consider what kind of alternative(s) would be releva nt to assessing whether or not the agent is morally responsible for what he actually does; those alternatives that would be relevant are robust alternatives and thos e that would not be relevant are not robust alternatives. According to the proponent of PAP, then, havi ng a robust alternative possibility to what one does is required for bei ng responsible for so doing.8 8 Thus, PAP should be reformulated as the following: having a robust alternative possibility to choosing/doing x is necessary for an agent to be morally responsible for choosing/doing x

PAGE 14

10 If, then, an FR scenario is meant to be a counterexample to PAP, where PAP is the claim that having robust alternative possibili ties to what one does is required for being morally responsible for so doing, the first conditi on of an FR scenario must be changed to read: the agent has no robust alternate possibili ties. So, in order for a case to be a counterexample to PAP, the following three conditions of an FR scenario must be satisfied: (1) the agent has no r obust alternate possibil ities to what he does, (2) whatever circumstances make it the case that the agen t has no robust alternate possibilities are not what causes him to do what he does, or brings about what he does, and (3) the agent is morally responsible for what he does. Whethe r or not Frankfurt-style cases satisfy all three conditions of an FR scenario and, s o, whether or not Frankfurt-style cases are counterexamples to PAP, will be the topic of the rest of this essay. In the second section I will discuss the main line of objection to Frankfurt-type cases made by proponents of PAP, as well as the replies given by the proponents of Frankfurt-style cases. The proponent of PAP obj ects that Frankfurt-st yle cases are faced with a dilemma. I will discuss both horns of the dilemma, as well as the strategies used by the proponents of Frankfurtstyle cases to respond to the dilemma, focusing most of my discussion on the second horn of the dilemma (which, I believe, is more interesting). Concerning the second horn of the dilemma, the proponents of Frankf urt-style cases have devised four strategies for finding a way out of the dilemma. I will discuss the first three strategies, showing that they are all problematic strategies to take as responses to the dilemma defense of PAP. The fourth strategy s eems to me to be the strongest strategy and is taken by Derk Pereboom but, as I will s how, still problematic. I will show that the success of Pereboom’s case depends upon th e correctness of his definition of

PAGE 15

11 ‘robustness.’ In the third section I will argue that Pereboom’s definition of ‘robustness’ is not correct, and I will argue fo r an alternative definition of ‘robustness,’ which if correct, would show that Pereboom’s Frankfurtstyle case does include robust alternate possibilities and, so, does not satisfy conditi on (1) of an FR scenario. If, then, the definition of ‘robustness’ that I propose is co rrect, then Pereboom’s Frankfurt-style case will not satisfy all three conditions of an FR scenario and, so, will not be a counterexample to PAP.

PAGE 16

12 AN OBJECTION TO FRANKFURT-STYLE CASES, AND REPLIES There are many sorts of objections to Fra nkfurt-style cases given by defenders of PAP and in what follows I will be discussi ng the most prominent objection to Frankfurtstyle cases, the objection that has been given th e most attention in the recent literature, and which I take to be the most serious obj ection to Frankfurt-styl e cases. This objection is called the dilemma defense of PAP1, initially suggested by Robert Kane2 and developed by David Widerker3. The dilemma defense of PAP begins by firs t pointing out that a standard Frankfurtstyle case, such as Fischer’s case given in the previous section, does not specify whether or not the occurrence of the i nvoluntary prior sign is suffici ent for the agent to make the choice he actually makes, C Recall that in a standard Frankfurt-style case, an agent makes some choice, C and that some device is in place to prevent the agent from choosing otherwise. The device does its j ob of preventing the agent from choosing otherwise in the following way: if the agent will make choice C he will show some involuntary sign prior to making choice C (and with the occurrence of this prior sign, the device will not intervene, will remain idle, a nd will not affect what the agent does), such that the absence of this sign is the cue for the device to intervene and bring about choice C in the agent. In the actual scenario the i nvoluntary prior sign is displayed, the device 1 Widerker and McKenna, 2003, p. 8. 2 Kane, 1985, p. 51 and, Kane, 1996, pp. 142-145. 3 Widerker, 1995, pp. 247-261.

PAGE 17

13 does nothing and the agent makes choice C but if this sign had not been displayed, the device would have intervened and brought about choice C (i.e., the absence of the prior sign is sufficient for the device to intervene and bring about choice C ). Although the absence of the prior sign is sufficient for the device to intervene, we can see that it is not specified whether or not the occurrence of the involuntary prior sign in the actual scenario is sufficient for the agent to make choice C Suppose that it is stipulated that the occu rrence of the involuntary prior sign in a standard Frankfurt-style case is suffi cient for the agent to make choice C4, the choice the agent actually makes without the intervention of the device. If this is stipulated, then the occurrence of the involuntary prior sign dete rmines that the agent will make choice C If the agent’s choice is determined by the occurr ence of the involuntary prior sign, then it is not so obvious that the standard Frankfurt-st yle case meets the third condition of an FR scenario, that the agent is mora lly responsible for his actual ch oice. If an agent’s choice is determined by the occurrence of an involuntary prior sign, it is not clear that we should claim that the agent is morally responsible for making such a choice, or for performing the action that is chosen, because it is not cl ear that we should hold people responsible for choices and actions that are the product of something invol untary, something the person had absolutely no control over. Aside from it not being clear that the agent is morally responsible for his actual choice if it is stipulated that th e occurrence of the involuntary prior sign is sufficient for the actual choice made by the agent in a standard Frankfurtstyle case in which case a standard Frankf urt-style case would not succeed in showing 4 Hereafter, the suggestion that the occurrence of the involuntary prior si gn is sufficient for the agent to make choice C is to be taken to mean either (a) the occurren ce of the involuntary prior sign is sufficient for the agent to make choice C or (b) the condition(s) that produces the involuntary prior sign is (are) sufficient for the agent to make choice C

PAGE 18

14 PAP false because such cases would not meet the third condition of an FR scenario to claim that the agent is morally responsible und er such circumstances would be to beg the question against the Incompatibil ist who claims that we cannot be morally responsible for choices and actions that are determined. So, if we assume that the occurrence of the involuntary prior sign is sufficient for the agen t’s actual choice in a standard Frankfurtstyle case, then (a) it is not clear that a standard Frankfurt-style case shows PAP false because it is not clear that the third condition of an FR scenario, that the agent is morally responsible for his actual choice, is met by su ch cases, and (b) to claim that the third condition is met by such cases would beg th e question against th e Incompatibilist. That (a) and (b) result from stipulating that the occurrence of th e involuntary prior sign is sufficient for the agent’s actual c hoice in a standard Frankfurt-style case is problematic for the success of such case s as counterexamples to PAP under this stipulation. Concerning (a), part of the appeal of Frankfurt-styl e cases is that the intuition that the agent is morally res ponsible for the actual choice he makes is a very plausible intuition. Since in the actual scenario of a standard Frankf urt-style case the agent makes his choice as we believe c hoices are normally made by a dults – free from compulsion, constraint, intervention by any devi ce, etc. – the intuition that he is morally responsible is very reasonable. Part of the appeal of Fra nkfurt-style cases as c ounterexamples to PAP, where PAP is itself also an intuitively plausi ble principle, rests on the high degree of reasonability of claiming that the agent is mo rally responsible for his actual choice even though he could not have chosen otherwise. So, in so far as the success of Frankfurt-style cases as counterexamples to PAP relies on th e high degree of reasonability of claiming that the agent is morally res ponsible for his actual choice, gi ven that (a) results from the

PAGE 19

15 stipulation that the occurrenc e of the involuntary prior sign is sufficient for the agent’s actual choice in standard Frankfurt-style cas es, stipulating this is problematic for the success of such cases as counterexamples to PAP. Concerning (b), if the stipulation that th e occurrence of the in voluntary prior sign is sufficient for the agent’s actual choice in a standard Frankfurt-style case means that the agent’s actual choice was determined, then to claim that the agent is morally responsible for his choice would be to beg the question ag ainst the Incompatibilist. To see why this is a problem for the use of Frankf urt-style cases as counterexampl es to PAP, recall that PAP is one premise of the Leeway Incompatibilist ’s argument for the incompatibility of moral responsibility and determinism: Having robust alternate possibilities to something one does is necessary for being morally responsible for so doing (PAP), but determinism removes alternate possibilitie s, and, so, moral responsib ility and determinism are incompatible. If using a standard Frankfurtstyle case as a counterexample to PAP is to show that the Leeway Incompatibilist’s argument fails because Frankfurt-style cases show that PAP, one premise of this argument, is false it would be problematic if the success of such cases was based upon the fals ity of the Incompatib ilist’s conclusion. If the proponent of Frankfurt-style cases wishes to use such cases to show that the Leeway Incompatibilist’s argument for the conclusion that moral responsibility and determinism are incompatible fails, then it cannot be claime d both that the agent’s choice in such a case was determined and that the agent is mo rally responsible for this choice. To claim both that the agent’s choice was determined a nd that he is morally responsible for this choice as a part of an argument against an argument that concludes that moral responsibility and determinism are incompatible would be to beg th e question against the

PAGE 20

16 Incompatibilist. We can see, then, that in so far as the proponent of Frankfurt-style cases wishes to use such cases to disprove a prem ise of the Leeway Incompatibilist’s argument for the incompatibility of mora l responsibility and determinis m, the stipulation that the occurrence of the involuntary prior sign is su fficient for the agent’s choice in Frankfurtstyle cases is problematic for the success of such cases as counterexamples to PAP. In the above given explanation of the fi rst part, or first horn, of the dilemma defense of PAP, I have attempted to de fend the following claim made by the proponents of the dilemma defense of PAP: if the i nvoluntary prior sign determines the agent’s choice in a standard Frankfurt-style case, then to claim that the agent is morally responsible for that choice be gs the question against the In compatibilist. While I do find it problematic for the proponent of Frankfurt-styl e cases to claim that the agent is morally responsible for his choice if this choice wa s determined by an i nvoluntary prior sign, I do not believe that it is correct to claim that it begs the question against the Incompatibilist to do so. Considering that th e Incompatibilist has his own arguments to support the claim that moral responsibility and determinism are not compatible, the Incompatibilist may not share the intuition that the agent in a Frankfur t-style case is morally responsible for his choice if this choice is determined. But, ne vertheless, many will have this intuition and this intuition is plausible. Fr ankfurt-style cases are intended to be counterexamples to one premise (PAP) of the Leeway Incompatibilist’s argument for the incompatibility of moral responsibility and determinism, where PAP has its support in intuiti on, just as the claim that the agent is morally responsib le has its suppor t in intuition. Begging the question against one’s opponent occurs when one uses the falsity of the conclusion of one’s opponent to prove th e truth of one’s own conclusion, without

PAGE 21

17 having any reason/support for using the falsity of the opponent’s conclusion in one’s own argument. But, the proponent of Frankfurt-st yle cases does have a reason for denying that moral responsibility and determinism are incomp atible: there is a case, namely a standard Frankfurt-style case, in which it is plausible to claim that an agent is morally responsible for this choice, even if this choice was dete rmined. Unless, then, we want to claim that counterexamples to a premise of an argumen t (here PAP) beg the question against the conclusion of that argument (here the Leeway Incompatibilist’s conc lusion), it does not seem as if Frankfurt-style cases beg the question against the Incompatibilist if we suppose that his choice was determin ed by an involuntary prior sign. This, though, is not to say that the stipulat ion that the agent’s choice is determined is not in any other way problematic for th e proponent of Frankfurt-style cases. The stipulation that the agent is morally responsible for his choice if that choice is determined is problematic if the proponent of Frankfurtstyle cases intends for a Frankfurt-style case to be used in an argument for Compatibilism in something like the following way: 1. It is possible to give a case in which an ag ent, Jones, chooses to act and acts in the way we believe normal adults to act free from compulsion, constraint, mental illness, etc., while his choice is determ ined by some involuntary prior sign. 2. If Jones chooses to and acts in this norm al way, he is morally responsible for so doing. 3. Suppose that there is an intervening devi ce that would have made Jones choose as he in fact does had he shown some sign of choosing otherwise. Given the placement of this intervening device, Jones could not have done otherwise than he actually did (i.e., he has no alternate possibilities). 4. Therefore, since Jones is morally resp onsible for his choice and action despite having no alternate possibilities, alternate possibilities are not required for moral responsibility (i.e., PAP is false). 5. Determinism is merely one way of removi ng alternate possibilities, but (given 4) this removal of alternate possibilities is irre levant to an agent’s moral responsibility for what he did (and, there is no other featur e of determinism that is relevant to an agent’s moral responsibility). 6. Therefore, determinism and moral re sponsibility are compatible (i.e., Compatibilism is true).

PAGE 22

18 The problem with this argument for the propon ent of Frankfurt-style cases is that it makes the use of the special feature (the counterfactually inte rvening device that eliminates alternative possibilities) of Fra nkfurt-style cases unnece ssary. Notice that the conclusion, that determinism and moral responsibility are compatible, follows from the first and second premise of the argument given above because the first premise stipulates that the agent’s choice was determined, while the second premise states that the agent is morally responsible for his choice/action. Th is means that the special feature of a Frankfurt-style case, that feature being the c ounterfactual intervener that eliminates the agent’s ability to do otherwise, is playi ng no role in supporting the Compatibilist conclusion and, so, the use of Frankfurt-styl e cases become superfluous in proving the Compatibilist conclusion. If, though, the proponent of Frankfurt-style cases claims that Frankfurt-style cases in some way advances the debate concerni ng moral responsibility (or plays some role in this de bate), the special features of a Frankfurt-style case must be doing some work that an ordinary case of an agent choosing to perform and performing some action would not do. Thus, it is problema tic for the Compatibili st to use Frankfurtstyle cases to make the argument outlined a bove. We will see later that John Martin Fischer uses an argument similar to the one outlined above in order to defend Frankfurtstyle cases from the first horn of the dilemma defense of PAP. The dilemma defense of PAP continues by proposing that if it is stipulated in Frankfurt-style cases that the occurrence of the involuntary prior sign is not sufficient for the agent’s choice, it is not the case that th e agent lacks the ability to choose otherwise, and, so, the first condition of an FR scenario would not be satisfied. Under this stipulation, the occurrence of the involuntary pr ior sign is merely a reliable indicator that

PAGE 23

19 the agent will make choice C and, so, it is possible that the involuntary prior sign be displayed and yet the agent make some choice other than C or make no choice at all. If the occurrence of the involuntary prior sign doe s not determine that the agent will make choice C then, it seems that the agent will have the possibility of making a choice other than C and, so, if the occurrence of the involun tary prior sign is not sufficient for the agent to make choice C it is not the case that Frankfurt-style cases satisfy the first condition of an FR scenario, that an agent ha ve no (robust) alternat e possibilities to what he chooses. It seems, then, that the su ccess of Frankfurt-style cases as counterexamples to PAP is jeopardized by the st ipulation that the occurrence of the involuntary prior sign is not sufficient for the agent to make choice C But, as we saw in the previous discussion, the stipulation that the involuntar y prior sign is sufficient for the agent to make choice C also jeopardized the success of Frankfurt-style cases as c ounterexamples to PAP. The proponent of Frankfurt-style ca ses is, then, posed with the following dilemma: If, on the one hand, it is stipulated that the occurrence of the involuntar y prior sign is sufficient for the agent to make choice C then the agent cannot (unproblem atically) be claimed to be morally responsible for his choice, and, so, Fr ankfurt-style cases do not satisfy the third condition of an FR scenario. If, on the other ha nd, it is stipulated th at the occurrence of the involuntary prior sign is not suffi cient for the agent to make choice C then the agent can choose otherwise and, so, Fr ankfurt-style cases do not sati sfy the first condition of an FR scenario. Either way, it is argued by the proponent of the dilemma defense of PAP, standard Frankfurt-style cases cannot satisfy all three conditions of an FR scenario and, so, such cases are not counterexamples to PAP.

PAGE 24

20 The First Horn of the Dile mma: Assumption that the Pr ior Sign Determines the Agent’s Choice (i.e., That the Prior Sign Deterministically Causes the Agent’s Choice) One way that the proponent of Frankfurtstyle cases can respond to the dilemma defense of PAP is to argue that the questi on is not begged agains t the Incompatibilist5 if it is assumed that the occurrence of the prior sign determines the agent’s choice. The proponent of Frankfurt-style ca ses can argue his way out of the dilemma by arguing that to assume that the agent in a Frankfurt-st yle case is morally responsible for his choice given the stipulation that the occurrence of the prio r sign is sufficient for (or, determines) the agent’s choice is not to beg the question against the Incompatibilist. If the proponent of Frankfurt-style cases coul d successfully argue that doing so would not beg the question against the Incompatibilist, then th e proponent of Frankf urt-style cases could claim both that the occurrence of the prior sign is sufficient for the agent’s choice and that all three conditions of an FR scenario are satisfied, thus showing that he is not faced with a dilemma. In “Frankfurt-type Examples and Semi-C ompatibilism” John Martin Fischer argues that the use of standard Fra nkfurt-style cases in an argum ent against PAP need not be seen as begging the question against the In compatibilist if it is assumed that the occurrence of the prior sign de termines the agent’s choice: Begin with the first horn of the dilemma: the assumption that causal determinism obtains. I agree that one cannot now simply and precipitously conclude, from 5 While I have argued above that the problem with the assumption that the prior sign determines the agent’s choice is not that the question is begged against the Incompatibilist but, rather, that the use of Frankfurtstyle cases becomes unnecessary to establish the Comp atibilist’s conclusion, my main objective in these next few paragraphs is to explain the debate concerning this horn of the dilemma as it has played out in the recent literature as a debate between John Martin Fischer and Stuart Goetz. Both Fischer and Goetz have framed the discussion as a discussion of whether or not the question is begged against the Incompatibilist and, so, I will frame the debate as they do for the ti me being, and then make some critical remarks after explaining the discussion between Fischer and Goetz.

PAGE 25

21 consideration of the examples, that the ag ent is morally responsible for his choice and behavior. But in any case this is not the way I would have proceeded; I never have envisaged a simple “one-step” argumen t to the conclusion that (say) Jones is morally responsible for his choice and ac tion. Rather, I employ the Frankfurt-type examples as the first (but obviously impor tant) step of a slightly more complex argument to the conclusion that Jones is morally responsible for his choice and action (despite lacking alternative possibilities). The argument goes as follows. First, one carefully considers the Frankfurt-type cases. Upon reflection, I believe that one s hould conclude that in these cases the lack of alternative possibili ties does not in itself ground a claim that the agent is not morally responsible for his choice and ac tion. In other words, I think that the examples make highly plausible the preliminary conclusion that if Jones is not morally responsible for his choice and action, this is not simply because he lacks alternative possibilities. After all, everyt hing that has any causal (or any other kind of) influence on Jones would be exactly the same, if we “subtracted” [the intervening device] from th e scene. And Jones’s moral responsibility would seem supervenient on what has influen ce or impact on him in some way. So the relevant (preliminary) conclusion is if Jones is not morally responsible for his choice and action, the reason is not simply that he lack s alternative possibilities. And it does not appear to beg the question to co me to this conclusion, even if causal determinism obtains. The first step is to argue, based on the Frankfurt-type examples, that intuitively it is plausible that alternative possibilities are irrelevant to ascriptions of moral responsibility . The second step in the argument consists in asking whether causal determinism in itself and apart from ruling out alternative possibilities threatens moral responsibility. I have considered various possible reasons why someone might think that causal determinism does threaten moral responsibility in itself and ap art from ruling out alternativ e possibilities, and I have come to the conclusion that it is not pl ausible to accept any of these reasons. It seems to me that this two-stage argument is highly plausible and does not beg the question against the Incompatibilist, even on the assumption of causal determinism. Thus I believe that the use of the “prior sign” cases can be defended against the charge of begging the question.6 I have given this lengthy quote from Fisc her’s essay because this argument is the starting point of Stewart Goetz’s criticism of defending Frankfurt-style cases from the charge of begging the question. Be fore explaining Goetz’s criti cism it is important to get clear on what Fischer’s two-step argument is. I take Fischer’ s two-step argument against 6 Fischer, 2002, pp. 291-292.

PAGE 26

22 PAP to be the following, where 7-9 constitutes Fischer’s first step and 10-12 constitutes his second step: 7. An agent’s moral responsibility, or lack thereof, for some choice and/or action can only be based upon those things which ex erted influence over the agent’s making this choice or performing this action. 8. In the actual scenario of a Frankfurt-st yle case, that the agent has no alternative possibilities does not influen ce the agent’s actual choi ce and action, because this lack of alternative possibilities results fr om the intervening device, which does not exert any influence over the agen t’s actual choice and action. 9. Therefore, in a Frankfurt-style case, the agent’s lack of alternative possibilities cannot be the basis for the claim that the agent is not morally responsible for his choice and action. (i.e., lack of alternative po ssibilities is not sufficient for not being morally responsible, i.e., PAP is fa lse, i.e., alternat ive possibilities are irrelevant to ascriptions of moral responsibility). 10. The agent in a Frankfurt-st yle case is not morally res ponsible for his determined choice only if either (c) alternative possibi lities are relevant to ascriptions of moral responsibility, or (d) determinism dir ectly, and not by ruling out alternative possibilities, undermines moral responsibility. 11. I (Fischer) have considered various reasons for believing (d) and have come to the conclusion that (d) is false, and it has b een shown in 3 above that (c) is false. 12. Therefore, the agent in Frankfurt-style cas es is morally responsible for what he does, even if his choice is determined by the occurrence of the prior sign. In his “Frankfurt-style Counterexamples and Begging the Question,” Stewart Goetz replies to Frankfurt’s two-step argument by arguing that this twostep argument itself begs the question against the Incompatibilist by claiming that the agent is morally responsible for his choice while assuming de terminism in the actual sequence. The heart of Goetz’s criticism of Fisher’s two-step argument can be seen in the following three passages: What is one to think of Fischer’s two-step argument? In the end it is difficult to see how it advances the debate between Compa tibilists and Incompatibilists. This is because contrary to what Fischer claims, his two-step argument begs the question against the Incompatibilist in the same way that the one-step arguments do: it assumes, because it requires, the truth of causal determinism in the actual sequence of events. It requires the truth of causal determinism to create the illusion that it is the presence of something in the alte rnative sequence of events (e.g., Black’s device) that makes it the case that Jones is not free to choose otherwise. It is only through the creation of this illusion and the fact th at Black’s device is not explaining Jones’ actual c hoice that one is tempted or inclined to endorse the

PAGE 27

23 conclusion of the first step of Fischer’ s argument, which is that the lack of alternative possibilities is not sufficient for the lack of moral responsibility (PAP is false) and, thereby, that Jones might be morally responsible even though he is not free to choose otherwise. Once this illusi on is exposed, one’s in itial conviction that the lack of alternativ e choice is sufficient for the lack of moral responsibility (PAP is true) is vindicated. It is true that the advocate of PAP believ es that the lack of the freedom to choose otherwise is sufficient for the lack of moral responsibi lity, and in that sense the latter obtains simply because of the ob taining of the former. It is incorrect, however, to assume that a person who affi rms PAP believes that the lack of moral responsibility obtains simply because of the obtaining of the freedom to choose otherwise in the sense that the obtaining of the latter by itself that is, apart from what explain it, explains the obtaining of the former. The proponent of PAP thinks that the lack of the freedom to choos e otherwise does not by itself explain the absence of moral responsibility. This is be cause he believes that when this lack obtains, its obtaining is itself explai ned by, and only can be explained by, the occurrence of causal determinism in the actual sequence of events. When an agent is not free to choose otherwise, not only does causal determinism explain what occurs in the actual sequence of events, but also it explains the lack of the freedom to choose otherwise. Frankfurt himself recognized this prima facie dual explanatory role of causal determinism in his classic discussion of PAP. Thus, he states that “[i]n seeking illustrations of the principle of alternative possibilities, it is most natural to think of situations in which the same circumstance both bring it about that a person does something [choos es] and make it impossible for him to avoid doing it [choosing otherwise]”7 Implicit in Frankfurt’s remark is the idea that it is intuitively plausible to think not onl y that causal determinism in the actual sequence of events is capable of producing situations wherein the same circumstances produce the actual choice a nd make it impossible for the agent to choose otherwise, but also that causal de terminism in the actual sequence of events is the only way to bring about situations in which an agent does not have the possibility of choosing otherwise. It is b ecause it is intuitively plausible to think this that Frankfurt assumes the burden of constructing examples wherein the circumstances that make it impossible for Jones to choose otherwise are not the same as or identical with what is goi ng on in the actual sequ ence of events. The reason why Frankfurt’s paper is so controversial is that it claims to provide an example where an agent does not have an alternative choice and the lack of this alternative is not explained by causal determinism in the actual sequence. Until such an example is provided, the incompa tibilist is on firm ground in believing that the lack of the freedom to choose othe rwise is explained by and can only be 7 Frankfurt, 1969.

PAGE 28

24 explained by causal determinism in the actu al sequence of events and that PAP is true.8 Essentially, Goetz is arguing that premise 8 (the second premise) of Fischer’s twostep argument against PAP is false, and that th e initial plausibility of this premise is due to the following illusion: that what suffices fo r the agent’s lack of alternative possibilities in a Frankfurt-style case is only effective (or, active) in the counter factual scenario (i.e., does not exert any influence in the actual scenar io). Goetz argues that the truth of premise 8 above relies on the illusion that what expl ains, or suffices for, the agent’s lack of alternative possibilities is the intervening de vice, which does not exert any influence over the agent’s actual choice and action (this is the point of the first passage above), when what really and only can e xplain, or suffice for, the agent’s lack of alternative possibilities is the obtaining of determinism (this is the point of the second and third passages above), which does influence – by nece ssitating the agent’s actual choice and action. While Goetz is, mistakenly I believe (for the reasons laid out in the earlier discussion of this first horn of the dilemma), ultimately trying to argue that Fischer’s twostep argument begs the question against the Incompatibilist and, s o, does not help the proponent of Frankfurt-style ca ses respond to the first horn of the dilemma defense of PAP, I believe that Goetz is pointing to something that is problematic with Fischer’s defense of Frankfurt-style cases. So, even though Goetz mistakenly charges Fischer’s two-step argument with begging the question against the Incomp atibilist, Goetz is correct to point out that premise 8 of Fischer’s twostep argument relies on an illusion and, so, is 8 Goetz, 2005, pp. 87-89.

PAGE 29

25 unsupported. A look at Fischer’s reply to Goet z’s criticism will bring out this mistake more clearly. Fischer replies that: In the Frankfurt-type scenarios, two causes make it the case that Jones is unable to choose otherwise at T2: the prior condition of the world (together with the laws of nature) and Black’s counterfactual interven tion. What the examples show is that the mere fact that Jones is unable to choose otherwise does not in itself establish that Jones is not morally responsible for his choice. This is because Black’s counterfactual intervention is one of the factors that make it the case that Jones is unable to choose otherwise at T2, and yet it is irrelevant to the grounding of Jones’s moral responsibility. Considering this fact or (the counterfactua l intervention), and bracketing any other factor that might ma ke it the case that Jones is unable to choose otherwise at T2, it seems to me that Jones may well be morally responsible for his action. The mere fact that he lacks alternative po ssibilities, then, cannot in itself be the reason Jones is not morally re sponsible, if indeed he is not morally responsible. Now, of course, it is also true that the prior condition of the world, together with the natural laws, make it the case that J ones lacks alternative possibilities. But, given that the mere fact of lacking alterna tive possibilities does not in itself rule out moral responsibility, why should this way of lacking alternative possibilities rule out moral responsibility? Why exactly should the significance of causal determinism be that it rules out alternative possibilities? This is exactly what the Frankfurt-type exampl es call into question.9 I take it that Fischer’s reply to Goet z’s argument amounts to the following: the agent’s lack of alternative possibilities in a Frankfurt-st yle case is brought about by to two things, one of which is the interveni ng device that does not influence the agent’s actual choice and action and, so, is irrelevant to the agent’s moral responsibility. Given that one way of removing the agent’s alternativ e possibilities is irrelevant to the agent’s moral responsibility, there seems no reason to believe that any other particular way of removing alternative possibilities is rele vant to an agent’s moral responsibility. Therefore, the obtaining of determinism, which is merely one other way of removing alternative possibilities, is irrelevant to an agent’s moral responsibility and, so, it would 9 Fischer, 2006, pp. 340.

PAGE 30

26 not beg the question against th e Incompatibilist to claim th at the agent in a Frankfurtstyle case is morally responsible if his c hoice was determined. (Notice the similarity between this argument and the argument disc ussed earlier, the argument I claimed to be problematic for the Compatibilist). Fischer’s reply to Goetz is mistaken, a nd the mistake lies in his first claim: “In the Frankfurt-type scenarios, two causes make it the case that Jones is unable to choose otherwise at T2: the prior condition of the world (toge ther with the laws of nature) and Black’s counterfactual intervention.”10 When Fischer claims that “two causes make it the case that Jones is unable to choose otherwise,”11 he is claiming that there are two causes that are individually sufficient for Jones being unable to choose othe rwise: the obtaining of determinism and the counterfactual interv ener. Fischer must be claiming that the obtaining of determinism and the interveni ng device are each sufficient to make it the case that Jones cannot choose otherwise, becaus e if he is not claiming this, premise 9 of his two-step argument (that the lack of alterna tive possibilities is irrelevant to ascriptions of moral responsibility ) would not follow from premise 7 (only factors which exert influence over what an agent does are relevant to ascriptions of mo ral responsibility) and premise 8 (the presence of the intervening device does not exert any influence over the agent’s choice in a Frankfurt-style case). If he is not relying on the claim that the intervening device is sufficient for the agent’ s lack of alternative possibilities, then it would still be possible that there is somethi ng that is both sufficient for it and does exert influence over the agent’s choice and, so, the conclusion drawn in premise 9 would not 10 Fischer, 2006, pp. 340. 11 Fischer, 2006, pp. 340.

PAGE 31

27 follow from premises 7 and 8. In order for hi s two-step argument to be sound, then, when Fischer claims that “two causes make it the case that Jones is unable to choose otherwise,”12 we must take him to be claiming th at the presence of the counterfactual intervener is alone sufficient for Jone s’s inability to choose otherwise. It is not correct, though, to cl aim that the intervening device is sufficient for the agent’s lack of alternative po ssibilities in a Frankfurt-style case. The intervener cannot do its job of being a counterfactua l intervener unless it is the ca se that the occurrence of the involuntary prior sign determines that the agent will make choice C and, so, the presence of the device is not sufficient to remove the agent’s ability to do othe rwise. It must also be the case that a determinis tic relationship exists between the occurrence of the prior sign and the agent’s making choice C If this deterministic relationship is not present between the occurrence of the prior sign and the agent’s making choice C then it could be the case that the prior sign is displaye d (in which case the devi ce does not intervene) and yet the agent makes a choice other than C So, Fischer is not corr ect to claim that the presence of the intervening device is suffi cient for the agent’s lack of alternative possibilities, because it must also be the case that the occurrence of the prior sign determines that the agent will make choice C This is the illusion th at Fischer is relying on in making his two-step argument, namely, that the cause of the agent’s lack of alternative possibilities has no (causal) influe nce over what the agent chooses; but, surely that the occurrence of the prior sign dete rmines that the agent will make choice C does have (causal) influence over what the ag ent chooses. We can see, then, that although 12 Fischer, 2006, pp. 340.

PAGE 32

28 Goetz mistakenly labels Fischer’s two-step argument as begging the question against the Incompatibilist, Goetz has pointed to a seri ous mistake in Fischer’s two-step argument. According to Goetz, Fischer’s two-step argument does not give the proponent of Frankfurt-style cases a satisfactory response to the dilemma defense of PAP. Goetz gives the following summary of the debate concerni ng the success of Frankfurt-style cases as counterexamples to PAP, and presents th e following challenge to the proponent of Frankfurt-style cases: So, where are we? Right back where we began, because the dilemma for proponents of FSCs is still this: either cau sal determinism obtains in the actual sequence, in which case Black’s device is not causally efficacious in preventing Jones from choosing otherwise, or causal determinism fails to obtain in the actual sequence, in which case Black’s device is also not causally efficacious in preventing Jones from choosing otherwise. If my foregoing reasoning is correct, it is still the case that Fischer has not pr ovided any satisfactor y Compatibilist way through the horns of this dilemma and, as a result, PAP is not undermined. To find a way through the horns of this dilemma and undermine the intuitively powerful that PAP captures, it remains the case that Fischer (any Frankfurtian) must provide an example where what explains the ag ent’s lack of the freedom to choose otherwise is something other than causal determinism in the actual sequence of events.13 According to Goetz, then, in order for a Frankfurt-style case to be successful in showing that PAP is false, it must be the case that the prior sign is not sufficient for the agent’s actual choice (i.e., that the occurrence of the prior sign does not determine that the agent will make choice C ), because if the prior sign is sufficient for the agent’s choice, then this sufficiency is what explains the agent’s lack of a lternative possibilities, and this sufficiency is causally relevant to what the agent chooses (in which case the second condition of a FR scenario would not be satisfied). 13 Goetz, 2005, p. 94.

PAGE 33

29 Given the above discussion, it seems to me th at an attempt to rescue Frankfurt-style cases from the dilemma defense of PAP by re sponding to the first horn of the dilemma is not a promising way for the proponent of Frankf urt-style cases to re spond to the dilemma defense of PAP. As an alternative to res ponding in this way, the proponent of Frankfurtstyle cases can try to show that Frankfurtstyle cases can be rescued from the dilemma defense of PAP by giving a case in which the prior sign does not de termine the agent’s choice (or, in which there is no prior si gn) and in which the agent cannot choose/do otherwise and is morally res ponsible for so choosing/doing. The Second Horn of the Dilemma: Assumpti on that the Prior Sign is Not Sufficient for the Agent’s Choice (i.e., That the Agen t’s Choice Is Indeterministically Caused) Recall that the second horn of the dile mma defense of PAP brings the following problem to the table: if it is stipulated that the occurrence of the prior sign is not sufficient for the agent’s choice and, so, the ag ent’s choice is not determined by the prior sign (i.e., if the agent’s choice is brought about indeterministically), then it seems that the prior sign is merely a reliable indicator that the agent will make choice some choice, C If the occurrence of the prior sign is merely a reliable indicator that the agent will make choice C it would not follow that the agen t could not choose otherwise than C if the prior sign occurs. So, the problem for the proponent of Frankfurt-style cases is that if the prior sign is not sufficient for the agent’s actual c hoice, the first condition of an FR scenario (that the agent cannot choose otherwise, that the agent has no robust alternative possibilities) cannot be met, and, so, a Frankfur t-style case cannot show that PAP is false. Many proponents of Frankfurt-style cases though, have attempted to rescue Frankfurt-style cases from the dilemma defe nse of PAP by showing that a case can be given in which both the agent’s choice is br ought about indeterministically and in which

PAGE 34

30 the first condition of an FR scenario is met – the agent cannot choose otherwise, or, does not have any robust alternate possib ilities. In their introduction to Moral Responsibility and Alternative Possibilities editors David Widerker and Michael McKenna give brief summaries of the four strategi es that have developed to gi ve such cases, cases in which both the agent’s choice is brought about indete rministically and in which the agent cannot choose otherwise, or has no robust alternative possibilities. In what follows I will explain each of these four strategies, giving an example of a case that uses each strategy. I believe that the fourth strategy to be discussed, th e strategy taken by Pereboom, is the strongest strategy, and so I will devote most of th e discussion that follows to evaluating Pereboom’s Frankfurt-style case. I will, though, briefly discuss the problems I see with the first three strategies as well. No-Prior-Sign Cases Those proponents of Frankfur t-style cases that give No-Prior-Sign cases try to develop a situation in which an agent cannot choose otherwise, and this is ensured without the use of a prior sign as a basis for interven tion. In No-Prior-Sign cases, the agent cannot choose otherwise because th ere are two processes that proceed independently of one another and which bot h end in the agent’s making some choice, C One process, the process that actually re sults in the agent (freely) making choice C is an indeterministic process, and the other process is a deterministic process that will bring about choice C in all cases except the case in which the other process, the indeterministic process, brings about this choi ce. Since the agent makes choice C as a result of an indeterministic process, we can hold the agent morally responsible for his choice and, yet, the agent could not do othe rwise than make choice C A No-Prior-Sign case is given by Alfred Mele and David Robb:

PAGE 35

31 At t1, Black initiates a certain deterministic process P in Bob’s brain with the intention of thereby causi ng Bob to decide at t2 (an hour later, sa y) to steal Ann’s car. The process, which is screened off from Bob’s cons ciousness, will deterministically 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 at t2). (Black is unaware th at it is open to Bob to decide on his own at t2 to steal the car; he is confident that P will cause Bob to decide as he wants Bob to decide.) The pr ocess is in no way sensitive to any “sign” of what Bob will decide. As it happens, at t2 Bob decides on his own to steal the car, on the basis of his own indeterministi c deliberation about whether to steal it, and his decision has no deterministic cause. But if he had not just then decided on his own to steal it, P would have determinis tically issued, at t2, in his deciding to steal it. Rest assured that P in no way influences the indeterministic decisionmaking process that actually issues in Bob’s decision.14 If this case is plausible, it seems to meet all three conditions of an FR scenario: (1) the agent cannot choos e otherwise at t2 than to steal Ann’s car, (2) what brings about the circumstance that the agent cannot choos e otherwise – the deterministic process P – in no way influences the agent’s actual choice, and (3) the agent is morally responsible for his actual choice to steal Ann’s car, because he did so on his own through an indeterministic decision-making process. I do not, though, believe that this case can meet all three conditions of an FR scenario given that it is faced with the foll owing dilemma: in terms of processes x (the indeterministic process that resu lts in the agent’s choice to steal the car) and P (the deterministic process), if we unde rstand each process as being a temporal process with a beginning and an end, where at the instant the pro cess ends, the choice made by the agent, the processes can either be such that: (e) process P is faster than process x (i.e., the deterministic process will end before the indeterministic process and, so, w ill result in the agent’s choice at t2, before the indeterministic process has reached its end), or (f) both processes can be simultaneous (i.e ., both processes woul d simultaneously end and, so, simultaneously produce the choice at t2, in which case the choice is overdetermined), or 14 Mele and Robb, 1998, pp. 101-102.

PAGE 36

32 (g) process x can be faster than process P (i.e., the indeterministic process will end before the deterministic process and, so, result in the agent’s choice at t2, before the deterministic process has reached its end). In neither situation (e), (f), or (g) will all three conditions of an FR scenario be satisfied. Suppose that (e) is the case and, so, the determ inistic process results in the agent’s choice at t2 to steal the car. If this is the case, then (2) ab ove is not satisfied (and it would beg the question to claim that (3) is satisfied) and, so, Mele and R obb’s case would not be an FR scenario. Suppose, next, that (f) is the case, and the two processes are simultaneous so that the agent’s choice at t2 is overdetermined. It then s eems unclear whether (2) above is satisfied, and this lack of clarity, I believe results from it being not clear whether the existence of a deterministic process usurps a ny other process as the (and the only) causal explanation of the end resu lt of such a process. If both a deterministic and an indeterministic process simultaneously result in the very same thing, it is hard to see what role the non-deterministic process is playi ng, since the deterministic process alone is sufficient for, while the non-deterministic pr ocess is not sufficient for, the end in question. Even though (f) stipulates that bot h the deterministic and indeterministic processes end at the same time and, so, both cau se the end in question, it is not the case that the indeterministic process is sufficient for the end in question, while it is the case that the deterministic process is sufficient for the end in question. A deterministic process that has some result x at time tn is one which, given the past and the laws of nature, could not have had any result other than x at tn, whereas an indeterministic process that has result x at time tn is one which, given the past and th e laws of nature, could have had a result other than x at tn; indeterministic processes, then, ar e not sufficient for their ends or results. It is not clear, th en, that if the two processes occur simultaneously, that the deterministic process is not infl uencing the agent’s choice at t2, because it seems

PAGE 37

33 plausible to claim that if th e two processes simultaneously result in the choice, then the deterministic process has (at least, and perhaps all there is to it) so mething to do with the agent’s making that choice.15 Suppose, finally, that (g) is th e case and the indeterministic 15 Mele and Robb address this in their paper, and I take it that they intend that (f) is the case, that the two processes occur simultaneously and yet the indeterministic process is THE process that results in the agent’s choice at t2 to steal the car: “The coherence of our scenario may, however, be called into question. How, one might wonder, can it happen that Bob decides on his own at t2 to steal Ann’s car, given the presence of the deterministic process we mentioned? One can understand how, prior to t2, Bob might decide on his own to steal the car. (Notice that in that case, other things being equal, Bob could have decided otherwise at this earlier time . ). But how ca n it happen that Bob decides on his own at t2 to steal the car, and that P does not produce the d ecision, given what we said about P? Consideration of the following fanciful machine w ill prove useful in answering this question. The machine, designed by a specialist in machine art, produces artistic widg ets of different shapes and colors. The colors of the widgets produced are determined by the color of a ball bearing (bb) that hits the machine’s receptor at a relevant ti me. The machine, M, is surrounded by several automatic bb guns, each containing bbs of various colors. The relevant aspect of M’s mechanical design, for our purposes, is relatively simple. First, with one qualification, if a bb of color x h its M’s receptor, and M is not already in the process of making a widget, M at once starts a process designed to result in the production of an xcolored widget. Second, because two or more bbs some times hit the receptor simultaneously, the artist has designed his machine in such a way that whenever this happens (while M is not busy making a widget) M at once starts a process designed to result in the production of a widget the color of the right-most bb. Bob is analogous to M in an important respect. He is physically and psychologically so constituted that if an unconscious deterministic process in his br ain and an indeterministic decision-making process of his were to “coincide” at the moment of decision, he would indeterministically decide on his own and the deterministic process would have no effect on his decision. This situation is an analogue of a case in which two bbs of the same color simultane ously hit M’s receptor (w hile M is not busy making a widget)” [Mele and Robb, 1998, pp. 103-104]. It seems to me that the consideration of this fanciful widget-making machine does not help Mele and Robb’s case to be an FR scenario. Recall that Me le and Robb’s case is intended to be a case in which the agent could not have chosen otherwise than to steal Ann’s car, whatever makes it the case that the agent could not choose otherwise does not exert influence ove r what the agent chooses, and the agent is morally responsible for his choice because he made it indeterministically. Stipulating that the agent makes the choice to steal Ann’s car indeterministi cally is what allows Mele and Robb’s case to be a case that satisfies the second and third of these conditions despite the fact that a deterministic process makes it the case that the agent could not have chosen otherwise. One should note, though, that merely stipulating that the agent’s choice was made indeterministically does not amount to the stipulation that the agent’s choice was made indeterministically in the right sort of way, where the right sort of way is the way such that the agent can be claimed morally responsible (at least partly) in virtue of the choice arising in this indeterministic fashion. In order that the agent’s choice being brought about in deterministically is the reason, or one of the reasons, for which we claim that the agent is morally responsi ble for this choice, it must be the case that the indeterminacy is located in a proper place in the ca usal history of the choice. For example, if the indeterminacy was located at some point in the causal history of the choice that occurred billions of years ago, we would not claim that this is the right sort of indeterminacy to allow for the agent’s moral responsibility for his choice billions of years later. Althou gh it is very difficult to determine at what point in the causal history of a choice the indeterminacy must be in order that the agent is morally responsible for this choice, it seems most reasonable to claim that th e indeterminacy must occur at the very moment the choice is made. It seems to me that when we claim that some choice was made indeterministically, this does not mean that something else happened indeterm inistically from which follo wed a choice; it seems to me that when we claim that some agent is morally responsible for some choice (at least partly) in virtue of the fact that the choice was made in deterministically, we do not mean to claim that some agent is morally

PAGE 38

34 process results in the agent’s choice at t2 because it reaches its end before the deterministic process reaches its end. In this case, the indeterministic process will result in the agent’s choice and, so, it is possible that the agent wi ll make some choice at t2 other than to steal the car. If (g) is the case, then, (1) above is not true and, so, Mele and Robb’s case would not be an FR scenario. Either way, then, whether (e), (f), or (g) is the case, Mele and Robb’s Frankfurt-style case doe s not meet all three conditions of an FR scenario.16 Blockage Cases Blockage Examples also attempt to meet all the conditions of an FR scenario without the use of a pr ior sign. In a blockage case, all neural pathways other than the pathway that is (coincidentally) actually take n are blocked, so that (1) the agent could not responsible for his choice (at least partly) in virtue of the fact that something happened indeterministically from which followed the agent’s choice. The claim that the agent is mo rally responsible for some choice made only makes sense to me if it is the occurrence of the choice itself that is indeterministic; if it is not the case that up until the choice was made it could have been different than it was (i.e., if it is not the case that the occurrence of the choice is itself indeterministic), it is hard to see the relevance of the claim that an agent made some choice indeterministically to the claim that an agent is morally responsible for that choice. In order, then, for the agent to be morally responsi ble for his choice at tn, it must be the case that everything in the world could have been just as it was up to tn and yet it was possible that the agent makes some other choice at that time. I would like to suggest that, given the setup of their case, the indeterministic process x that results in the agent’ s choice to steal Ann’s car does not meet this condition. Taking into account the analogy with the widget-m aking machine, and given that (f) is the case and the deterministic process and the indeterministic process both hit the decision-making node(s) simultaneously, it must be the case that the processes hitting the decision-making node is a distinct event from the event of the choice being made in the same way that two bbs of the same color hitting the right and left receptors is a distinct event from the initiation of or the making of the widget. Since the indeterministic process hitting the decision-making node is a separate event from the event of the choice be ing made, in order for the choice to be made indeterministically in the right sort of way to allow for the agent to be morally responsible for this choice, it must be the case that everything pr ior (including the indeterministic process hitting the decision-making node) could have been the same as it was, and yet the agent could have chosen otherwise. In Mele and Robb’s case, though, this is not so; once the indeterministic process hits the decision-making node, the agent will choose to steal Ann’s car. And, so, I believe that we should not claim that the agent’s decision in Mele and Robb’s case is brought about inde terministically in the right sort of way to allow for the agent to be morally responsible for his choice and, so, Mele and Robb’s case does not satisfy the conditions of an FR scenario and, so, is not a counterexample to PAP. 16 It should be noted that others have criticized Mele and Robb’s case as well: See Pereboom, 2001, pp. 1415; See also Kane, 2003; See also Widerker, 2000, pp. 183-185.

PAGE 39

35 have done anything other than he did, but (2 ) the blockage is such that it in no way affects what the agent actually does and, so, (3) the agent is morally responsible for what he does. In “Moral Responsibility and Una voidable Action,” David Hunt suggests such a blockage case that is based upon Locke’s locked door example: The original “Frankfurt scenario” is of course John Locke’s famous example in which “a man be carried whilst fast asleep into a room where is a person he longs to see and speak with, and be there locked fast in, beyond his power to get out; he awakes and is glad to find himself in so desirable company, which he stays willingly in, i.e. prefers his stay to going away.”17 In this case the man is morally responsible for what he does even th ough he cannot do otherwise . But the condition that makes for unavoidability in this example is not counterfactual in nature: the door is actually locked; it doesn’t lock only when someone approaches the door and tries to leave. What makes th e locked door compatible with the man’s moral responsibility is simply that it is not among the conditions actually leading the man to stay in the room . Locke’s example does provide some initial encouragement that the unavoidability esse ntial to a Frankfurt scenario does not have to rest on a counterfactual device. Thus encouraged . Imagine then a device that blocks neural pathways rather than doorways . The mechanism is not interven ing directly in the series itself; it is allowing the series to unfold on its own, but simply blocking all alternatives to the series. Of course it can ’t block alternatives in response to the way the series is unfolding, because then the blockage would be coming too late to have any effect on the avoidability or unavoidability of Jones’s actions. Instead the mechanism blocks alternatives in advance, but owni ng to a fantastic coin cidence the pathways it blocks just happen to be all the ones th at will be unactualized in any case, while the single pathway that remains unblocked is precisely the route the man’s thoughts would be following anyways (if all neural pathways were unblocked). Under these conditions, the man appears to remain res ponsible for his thoughts and actions . 18 It is clear that Hunt takes his case to satisfy all three conditions of an FR scenario: (1) the agent cannot do otherwise than he does (t his is ensured by the mechanism, which, coincidentally, blocks all unact ualized pathways before the actual pathway is actualized), 17 Hunt’s footnote reads: John Locke, An Essay Concerning Human Understanding Bk.II, ch. XXI, § 10. Of course Locke has his own agenda here, which is not entirely the same as Frankfurt’s. 18 Hunt, 2000, pp. 217-218.

PAGE 40

36 (2) what brings about the agent’s inability to do otherwise in no way influences what he does (because the operation of the mechanism does not affect the actual pathway), and (3) the agent is morally responsible for what he does. Robert Kane has challenged the intuition elicited by Hunt’s blockage cas e, and Pereboom elaborates upon Kane’s challenge, arguing that our intuition that (3) th e agent is morally responsible in this case is not clearly correct, because it is far from clear whether it is reasonable to claim, given the circumstances, that the agent’s choice wa s brought about indeterministically. Kane’s challenge to such blockage cases is the following: In [a case in which every other alterna tive is blocked except the agent’s making A at t], of course there are no alternative possibilities left to the agent; every one is blocked except the agent’s choosing A at t. But now we seem to have determinism pure and simple. By implanting the mechanis m in this fashion, a controller would have predetermined exactly what the agent would do (and when); and, as a consequence, the controller, not the agent, would be ultimately responsible for the outcome. Blockage by a contro ller that rules out alternat ive possibilities is simply predestination; and on my view at least, predestination runs afoul of ultimate responsibility.19 Kane’s claim here is that given the me thod of the mechanism, even though it only coincidentally blocks the pathways that ar e unactualized, since the mechanism does this before the actualized pathway is actualized, the blockage amounts to predestination of the agent’s choice. If an agent’s choice being predestined means, as it means according to Kane, that the agent is not ultimately responsib le for such a choice, then, the presence of the blockage removes the agent’s responsib ility. Kane’s objection to blockage cases, then, is that blocking alternat ive pathways prior to the actu alization of the pathway that leads to the agent’s choice amounts to determin ing the agent’s choice, in which case, the 19 Kane, 2000, p. 162.

PAGE 41

37 third condition of an FR scenario would not be met by such blockage cases (without begging the question). In Living Without Free Will Pereboom strengthens this criticism by first suggesting that the plausibility of claiming th at the agent in a bloc kage case is morally responsible is due to the fact that the presence of the blocka ge does not have any part in the actual causal history of the agent’s choice – it is purely coincident al that just those pathways were blocked that were unactualized (and would have been unactualized even if the blockages were not present). Consider though, the following two situations: In situation A the agent makes choice x with no blockages or anyt hing else unusual present (and so could have at that time chosen differe ntly), while in situation B the agent makes choice x with the same causal history as in situ ation A, but in situation B the blockages are in place in the way that Hunt suggests – ahead of time such that coincidentally those pathways other than the act ualized pathway are blocked.20 Concerning situations A and B, Pereboom says that “cases of the above sort mi ght be misleading just because it is natural to assume that the actual causal history of an event is essentially the same in each, given that the only difference between them is a rest riction that would seem to have no effect 20 I have paraphrased the relevant aspects of Situation A (the first case) and Situation B (the second case), the cases Pereboom uses at this point of the discussi on. Situation A and Situation B are given on p. 16 of [Pereboom, 2001] as follows: “Situation A: Ms. Scarlet deliberately decides to kill Colonel Mustard at t1, and there are no factors beyond her control that determ inistically produce her choice. When she chooses to kill the Colonel, she could have chosen not to kill him. There are no causal factors that would prevent her from not making the choice to kill Colonel Mustard. Situation B: Ms. Scarlet’s choice to kill Colonel Mustard has precisely the same actual causal history as in A. But before she even started to think about killing Colonel Mustard, a neurophysiologist has blocked all the neural pathways not used in Situation A, so that no neural pathway other than the one employed in that situation could be used. Let us supposed that it is causally determined that she remain a living agent, and if she remains a living agent, some neural pathway has to be used. Thus every alternative for Ms Scarlet is blocked except the one that realizes her choice to kill the Colonel. But the blockage does not aff ect the actual causal history of Ms. Scarlet’s choice, because the blocked pathways w ould have remained dormant.”

PAGE 42

38 on the event.”21 Consider, though, the following tw o-situation case discussed by Pereboom22, which seems to show that two situat ions with the (seemingly) same actual causal history may differ in that the one with (something relevantly similar to) a blockage may be essentially determined to proceed as it does: Imagine a universe correctly described by Epicurean physics: At the most fundamental level all that ex ists is atoms and the fric tionless void, and there is a determinate downward direction in which al l atoms naturally fall – except if they undergo uncaused swerves. Situation C: A spherical atom is falli ng downward through space, with a certain velocity and acceleration. Its actual causal history is indeterministic because at any time the atom can be subject to an un caused swerve. Suppose that the atom can swerve in any direction other than upwards. In actual fact, from t1 to t2 it does not swerve. A counterfactual situation diverges from C only by vi rtue of a device that eliminates alternative possibilities and all differences thereby entailed: Situation D: The case is identical to C, except that the atom is falling downward through a straight and vertically oriented tube whose interior surface is made of frictionless material, and whose interior is precisely wide enough to accommodate the atom. The atom would not have swerved during this time interval, and the trajectory, velocity, and acceler ation of the atom from t1 to t2 are precisely what they are in C.23 So, what is the lesson from Situations C and D? One might initially want to claim that the causal histories in these two situ ations are identical between t1 and t2, because the same thing happens to the particle during this time interval in both situations. This intuition, though, is questionable because it might be that the presence of the tube in Situation D makes the causal history of the particle in Si tuation D essentially deterministic. If the tube prevents the particle from swerving, and the possibility of swerving is what 21 Pereboom, 2001, p. 17. 22 Pereboom notes on p. 17 of Pereboom, 2001 that this two-situation case is “modeled on a reflection of Hunt’s” from Hunt’s personal correspondence with Fischer, cited on pp. 119-120 of Fischer, 1999. 23 Pereboom, 2001, p. 17.

PAGE 43

39 constitutes the indeterminate nature of the particle’s movement, the tube precludes the movement of the particle from being indeterm inate in nature; if the tube precludes the movement of the particle from being indeterm inistic, then the tube makes it the case that the particle’s movement is essentially dete rmined and, so, makes the particle’s causal history deterministic in Situation D. In order fo r it to be the case that the movement of the particle through the tube between times t1 and t2 is indeterministic, it must be the case that at t1, t2, and any moments between (if there are any) the world could have been just as it was up until that moment, and yet the part icle could have swerved (rather than not swerved, which is what it actually did). Given the placement of the tube, though, the particle could not have swerved and, so, it movement between t1 and t2 is not indeterministic. This would show that an obj ect in two situations, here C and D, can (seemingly) have the same thing happen to it in both situations, and yet in one situation, C, what happens to the object is indeterminis tic, whereas in the other, D, the presence of a blocking mechanism (here, a tube) makes it th e case that what happens to the object is determined. If this is the correct thing to sa y about situations C and D, this lesson applies to Hunt’s blockage case in the following way: the intuition that the agent is morally responsible in Hunt’s blockage case is taken to be reasonable, given that the actual causal history of the agent’s choice is not affected by the presence of the blockages. The lesson from situations C and D is that the presence of such a blockage may affect the causal history (making it deterministic), even though the agent’s choice would have been the same had the blockage not been present. Concerning what the lesson from situations C and D shows us about blockage cases, Pereboom says the following: Sympathy for Frankfurt-style arguments is generated by the sense that moral responsibility is very much a function of the features of th e actual causal history of

PAGE 44

40 an action, to which restricti ons that exist but would seem to play no actual causal role are irrelevant. However, in a scenario in which such restrictions, despite initial appearances, could be relevant to the nature of the actual causal history of an action after all, one’s intuitions about whether the agent is morally responsible might become unstable. My own view is not that actual causal histories in blockage cases are clearly deterministic, but only that thes e considerations show that they may be. This type of problem should make on e less confident when evaluating these difficult kinds of Frankfurt-style cases.24 Internal-Sign Cases Unlike No-Prior-Sign examples and Blocka ge examples, Internal-Sign examples do not attempt to eliminate a sign whose presen ce indicates that the agent will make the desired choice, C and whose absence is sufficient fo r the intervention of the device to produce choice C Rather, Internal-Sign examples are se t up so that there is a sign of this sort, but, rather than occurring prior to the ag ent’s action (i.e., rather than being a prior sign), the sign is internal to the event of the agent’s choice /action. Such a case is given by Eleonore Stump in “Alternative Possibilitie s and Moral Responsibil ity: The Flicker of Freedom”: Suppose that a neurosurgeon Grey wants hi s patient Jones to vote for Republicans in the upcoming election. Grey has a neur oscope which lets him both observe and bring about neural firings which correlate with acts of will on Jones’s part. Through his neuroscope, Grey ascertains that ever y time Jones wills to vote for Republican candidates, that act of will correlates with the completion of a sequence of neural firings in Jones’s brain that always incl udes, near its beginning, the firing of neurons a b c (call this neural sequence “ R ”). On the other hand, Jones’s willing to vote for Democratic candidates is correla ted with the completion of a different neural sequence that always includes, n ear its beginning, the firings of neurons x y z none of which is the same as those in neural sequence R (call this neural sequence “ D ”). For simplicity’s sake, suppose that neither neural sequence R nor neural sequence D is also correlated with any furt her set of mental acts. Again, for simplicity’s sake, suppose that Jones’s only relevant options are an act of will to vote for Republicans or an act of will to vote for Democrats. Then Grey can tune his neuroscope acco rdingly. Whenever the neuroscope detects the firing of x y and z the initial neurons of neural sequence D the neuroscope 24 Pereboom, 2001, p.18.

PAGE 45

41 immediately disrupts the neural sequence, so that it isn’t brought to completion. The neuroscope then activates the coerci ve neurological mechanism which fires the neurons of neural sequence R thereby bringing it about that Jones wills to vote for Republicans. But if the neuroscope detects the firing of a b and c the initial neurons in neural sequence R which is correlated with the act of will to vote for Republicans, then the neuroscope does not in terrupt that neural sequence. It doesn’t activate the coercive neurological mechanism, and neural sequence R continues, culminating in Jones’s will ing to vote for Republican s, without Jones’s being caused to will this way by Grey. And suppose that [in this case] Grey does not act to bring about neural sequence R but that Jones wills to vote for Republican s, without Grey’s coercing him to do so.25 In setting up this case, Stump makes two presuppositions. The first presupposition is that there is a correlation between the mind and the brain, that mental events are somehow correlated with neural ev ents. The nature of this corre lation is left vague so that various different theories of mind/brain interaction can be accommodated by this Frankfurt-style case. This presupposition leav es it open whether each mental event is a temporally extended event that occurs over th e time period from th e beginning to the end of the corresponding neural even t (in which case, the mental event does not occur unless the entire neural sequence occurs), or whether each mental event occurs after the completion of the corresponding neural event (in which case the mental event does not occur until the entire neural se quence occurs, when the last neuron firing of the neural event is completed). The second presupposition is that the correlation between a mental event and the neural event it is correlated w ith is a one-many relation; each mental event (say, a choice) is correlated to a neural event that consists of many neural firings.26 25 Stump, 1999a, pp. 303-305. 26 “When I suddenly recognize my daughter’s face across a crowded room, th at one mental act of recognition, which feels sudden, or even instantaneous, is correlated with many neural firings as information from the retina is sent through the optic ne rve, relayed through the lateral geniculate nucleus of the thalamus, processed in various parts of the occi pital cortex, which take account of figure, motion, orientation in space, and color, and then processed further in cortical association areas. Only when the whole sequence of neural firings is complete, do I have the mental act of recognizing my daughter.

PAGE 46

42 It seems to me that Stump’s Frankfurt-styl e case is in no better shape than Mele and Robb’s, and is faced with a similar problem (refer back to footnote 20). In Stump’s case, it must be the case that there is indetermin acy in the event of the agent’s choice being made such that this indeterminacy is the right kind of indeterminacy, meaning that (at least partly) in virtue of his choice being undetermined in this way the agent can be morally responsible for his choice. We can see that it must be the case that the indeterminacy in Stump’s case lies in the initia tion of the neural sequ ence that correlates with the mental event ‘choosing to vote Republ ican’ (this neural sequence begins with firings a b c ), or in the initiation of the neural sequence that correlates with the mental event ‘choosing to vote Democrat’ (this neural sequence begi ns with firings x y z ).27 Stump wants to claim (see footnote 30) that unless a neural event is brought to completion, nothing mentally signif icant has occurred (i .e., if a neural event is terminated prior to its completion, no mental event occurs at all); so, for exam ple, the firings of a b c alone are not mentally significant, because these firings are only the beginning of a neural event. If the indeterminacy occurs in the initiation of the neural sequence, and nothing mentally significant occurs until after the completion of the neural sequence, then Whenever neural firings are associated with an act of w ill, I take it that in this case, as in all others, the correlation between the mental act and the firing of the relevant neurons is a one-many relation” [Stump, 1999a, p. 306]. 27 Stump intends for the indeterminacy to be located in the initiation of the neural sequence, as is noted in a later essay of hers: “So, as I origin ally presented the case, the initiating cause of the neural sequence in [the case] – whatever exactly that is – is itself indeterm inistic and to be understood in a way that doesn’t preclude the indeterministic firing of the initial neurons of the sequence” [Stump, 2003, p. 142]. It must be that the indeterminacy in Stump’s case is at no point later than the initiation of the neural sequence. It cannot be the case, for example, that the indeterminacy occurs after neurons a b c have fired, because then it would be the case that neurons a b c fire (in which case the device does not intervene) and yet a different neural event occurs, one other than that which correlates with the mental event ‘choosing to vote Republican’ (in which case the agent could choose otherwise, even though the intervening device is in place).

PAGE 47

43 the indeterminacy is located prior to the occurrence of anything mentally significant.28 For an agent’s choice to be undetermined in the right kind of way, to allow for the agent to be morally responsible for that choice, it must be the case that up until the moment that choice is made, the past (which in Stump’ s case would include every neuron firing up until the last one of the neural sequen ce, everything up until something mentally significant occurs) could have been just as it was and yet it be possible for the agent to choose otherwise. This is not so in Stump’s case, because once firings a b c occurs, the rest of the sequence will be complete d and the mental event ‘choosing to vote Republican’ will occur. So, in Stump’s Frankf urt-style case the agent’s choice is not undetermined in the right way to allow for the agent to be morally responsible for this choice and, so, the conditions of an FR scenar io are not met and, so, Stump’s case is not a counterexample to PAP. Necessary-Condition Cases Necessary-Condition Cases make use of a devi ce that relies on a prior sign in order to guarantee that the agent cannot choose othe rwise than he does, but this prior sign is merely a necessary condition for the agent doing otherwise, rather than a sufficient condition of the agent doing otherwise. In a Necessary-C ondition example, the agent (indeterministically) makes choice C but in order for him to fail to make choice C he 28 Suppose that Stump were to claim that the beginning of a neural sequence, for example firings a b c is mentally significant and, so, the indeterminacy lies in the initiation of something mentally significant. This claim, though, would be problematic as well. If the beginning of a neural sequence is mentally significant, then the agent in Stump’s case would have an alternative possibility: firings x y z could have happened, which would be the mentally significant beginning of the mental event ‘choosing to vote Democrat’ (perhaps the mental event in this case would be ‘trying to vote Democrat’), in which case the device would intervene and bring about firings a b c and the mental event ‘choosing to vote Republican’ would also occur. Since there is this alternative possibility under the stipulation that the neural firings at the beginning of a neural sequence are mentally significant, under this stipulation Stump’s case would not satisfy the conditions of an FR scenario and, so, would not be a counterexample to PAP.

PAGE 48

44 would have has to (indeterministically) choose/do A If the agent chooses/does A though, the device would have interven ed and brought about choice C A Necessary-Condition example is given by Derk Pereboom, in its most revised version in “Defending Hard Incompatibilism.” Tax Evasion (2): Joe is considering wh ether to claim a tax deduction for the substantial local registration fee that he paid when he bought a house. He knows that claiming the deduction is illegal, that he probably won’t be caught, and that if he is, he can convincingly plead ignorance. Suppose he has a very powerful but not always overriding desire to a dvance his self-interest regardle ss of the cost to others, and no matter whether advancing his se lf-interest involve s illegal activity. Crucially, his psychology is such that the on ly way in this situation he could fail to choose to evade taxes is for moral reas ons. His psychology is not, for example, such that he could fail to choose to evad e taxes for no reason or simply on a whim. In addition, it is causally nece ssary for his failing to choose to evade taxes in this situation that he attain a certain level of attentiveness to these moral reasons. He can secure this level of attentiveness voluntarily. However, his attaining this level of attentiveness is not causally sufficient for his failing to choose to evade taxes. If he were to attain this level of attentiven ess, Joe could, with his libertarian free will, either choose to evade taxes or refrain fr om so choosing (without the intervener’s device in place). More generally, Joe is a li bertarian free agent. But to ensure that he chooses to evade taxes, a neuroscientis t now implants a device, which, were it to sense the requisite level of attentiveness, would electroni cally stimulate his brain so that he would choose to evade taxes. In act ual fact, he does not attain this level of attentiveness, and he chooses to evade taxes while the device remains idle.29 In Tax Evasion 2, Joe (1) cannot choose otherwise than he does, choose to evade taxes, because if he met a necessary conditi on of choosing otherwise – attained a certain level of attentiveness to his moral reasons – the device would inte rvene and bring about the choice to evade taxes, (2) what brings about the situation described in (1), the device, in no way affects what Joe actually does, and (3) Joe is morally responsible for what he does because he did so on his own, was not determined to do so, etc. Some proponents of PAP, though, may challenge (1) and claim that Joe does have a relevant alternative open to him: he can volunt arily choose to attain the requisite level of 29 Pereboom, 2005, p. 231-232.

PAGE 49

45 attentiveness to moral reasons for not evading ta xes. His ability to so choose is a morally relevant alternative because had Joe chosen to reach the requisite level of attentiveness to moral reasons, the device would have interven ed and forced Joe to choose to evade taxes. Under these circumstances, though, Joe would not be morally responsible for choosing to evade taxes, because this choice would be the product of the device and not made freely by Joe. Joe’s ability to choose to attain this requisite leve l of attentiven ess to moral reasons, then, is a relevant al ternative – one that explains why Joe is morally responsible for his actual choice to evade taxes – because in this counterfactual scenario Joe is not morally responsible for choosing to evade ta xes, while in the actual scenario he is morally responsible for choosing to evade taxe s; thus, Joe has an alternative open to him in which he is not responsible for what in the actual scenario he chooses to do (i.e., he has an alternative open to him in which he is not blameworthy for what he does).30 Pereboom, though, argues that having an al ternative in which the agent is not blameworthy for what he does is not the right kind of alternative possibility that could explain why the agent is morally responsible for what he does; Pereboom argues that the alternative possibility that the objectors are referring to is not a robust alternate possibility and, so, his Tax Evasion 2 does sati sfy the first condition of an FR scenario. According to Pereboom, in order for an alterna tive possibility to be robust it must satisfy the following conditions: Robustness: For an alternative possibility to be relevant per se to explaining an agent’s moral responsibility for an action it must satisfy the following characterization: she could have willed something other than what she actually 30 Otsuka, 1998 and, Wyma,1997.

PAGE 50

46 willed such that she understood that by willing it she would thereby have been precluded from the moral responsibil ity she actually has for the action.31 So, according to Pereboom, in order for an alternative possibility to be the kind of possibility that is relevant to explaini ng an agent’s moral responsibility for doing R the alternative possibility must satisfy the two following conditio ns: (i) the agent could have chosen something, Q such that Q is different from R AND the choosing of Q would thereby preclude the agent of the moral re sponsibility she has for what she actually chooses, R and (ii) the agent unders tood that had she chosen Q she would thereby be precluded from the moral responsibility she has for doing R In Tax Evasion 2, the alternative possibility that is claimed by the objector to be morally relevant (the possibility of choosing to reach the requisite level of attentiveness, which would cause the device to intervene and produ ce the choice to evade taxes), while it satisfies condition (i) of Pereboom’s definition of ‘robustne ss,’ does not satisfy condition (ii). Had Joe chosen to reach the requisite level of atten tiveness to moral reasons, this choice would be different from his actual choi ces (Joe actually chooses to not pay attention to moral reasons and, so, chooses to evade taxes) a nd choosing to reach the requisite level of attentiveness would pr eclude him from the moral respons ibility he has for his actual choices (as explained above, had Joe chosen to reach the requisite level of attentiveness, the device would have caused him to choose to evade taxes and Joe would not be morally responsible for this choice). This possibility of choosing to reach the requisite level of attentiveness, though, accord ing to Pereboom, does not meet condition (ii) above: But, Joe does not have an alternative possi bility available to him that is robust. First, he does not even believe that if he had achieved the requisite level of attentiveness he would th ereby have been precluded from responsibility for 31 Pereboom, 2005, p. 230.

PAGE 51

47 deciding to evade taxes. For he believes th at achieving this level of attentiveness is compatible with his never refraining from making this decision, or being seriously inclined so to refrain, and deciding to ev ade taxes instead. In addition, Joe does not know enough to understand that voluntarily achieving the requisi te attentiveness would preclude him from responsibility for choosing to evade taxes. True, were he voluntarily to achieve this attentiveness, the intervention would take place, and he would not then have been responsible for this choice. Nevertheless, Joe does not understand that the intervention would then ta ke place, or that as a consequence of this intervention he would be precluded from responsibility for choosing to evade taxes. Hence, no robust alternative possibili ty is available to him. Still, Joe is morally responsible for deciding to evade taxes.32 Pereboom, then, maintains that Tax Evasi on 2 meets the first condition of an FR scenario and, so, (since it meets the second and third conditions as well) is a counterexample to PAP. But, we can see that Pereboom’s argument for the claim that Joe has no robust possibilities is dependent upon the correctness of his definition of ‘robustness.’ Pereboom’s argument for the cl aim that Joe has no robust alternative possibilities in Tax Evasion 2 is just the fo llowing: the alternative possibility open to Joe does not meet my (Pereboom’s) conditions on robustness, and, so, Joe has no robust alternative possibilities. Speci fically, the alternative possibility open to Joe in Tax Evasion 2 does not meet what I will call Pereboom’s epistemic c ondition on robustness, condition (ii) of the definition of ‘robustness’: in order for an alternat ive possibility to be robust the agent must have understood that by making a choice other than the one he actually makes he would thereby be precluded from the moral responsibility he has for what he actually does. In the following section, I will argue that Pereboom’s epistemic condition on robustness is not correct. While I will argue that there is some kind of an epistemic condition that must be included in a defi nition of ‘robustness,’ I will argue that 32 Pereboom, 2005, pp. 232-233.

PAGE 52

48 Pereboom’s epistemic condition demands too much of the agent in order for the alternative possibility open to him to be rele vant to explaining the moral responsibility he has for what he actually does. If the defin ition of ‘robustness’ that I will propose is correct, the alternative open to Joe in Tax Evasion 2 is a robu st alternative possibility. If, then, the definition of ‘robustness’ that I wi ll propose is correct, Pereboom’s Tax Evasion 2 will not meet the first condition of an FR scenario and, so, will not be a counterexample to PAP.

PAGE 53

49 THE DEFINITION OF ‘ROBUSTNESS’ In this section, I will argue that Pereboom’s definition of ‘robustness’ is not correct, and given what I take is a more accurate account of ‘robustness,’ Pereboom’s Frankfurtstyle case, Tax Evasion 2, will leave the agent with robust alternative possibilities and, so, will not be a counterexample to PAP. If Pereboom’s Tax Evasion 2 is not a counterexample to PAP, then the proponent of Frankfurt-style cases will either need to formulate a Necessary-Condition Fr ankfurt-style case that does meet all three conditions of an FR scenario, or will need to formulat e a No-Prior-Sign, a Blockage, or an InternalSign example that meets the challenges previously made to such cases. To show that Pereboom’s definition of ‘robustness,’ specifically his epistemic condition, is not correct, I will first look at his argument to support his epistemic condition, this argument being based upon our intuitions about a particular case I will argue that our intuitions about this case do not need to be explained by positing Pereboom’s epistemic condition on ‘robustness,’ but, rather, are better explained by the episte mic condition(s) that I will suggest. I will spend the remainder of this essay discussing the condition(s) that I propose, and how given the definition of ‘r obustness’ that I propose, Pereboom’s Tax Evasion 2 is not a counterexample to PAP. Be fore taking a look at Pereboom’s argument for his epistemic condition on ‘robustness,’ I will briefly discuss the criteria of adequacy for (or, what we are looking fo r in) a definition of ‘robustness.’ The term ‘robustness’ is a technical term that picks out a certain feature (or, a set of features) of some, but not al l, of the counterfactual/altern ative possibilities that we

PAGE 54

50 believe are open to agents. We will say that an agent has a robust alternative possibility if and only if there is a counterfactual sequence ope n to him that has this certain feature (or set of features). In order for a definition of ‘robustness’ to be an adequate definition, it must spell out the necessary and sufficient c onditions for when some agent has a robust alternative possibility, and th e definition must do this in a way that accords with our intuitions about particular cas es. Intuitively, it seems to me that an agent has a robust alternative possibility in those cases in whic h the addition or remova l of this possibility affects our judgment of the agent’s moral respons ibility. It seems to me that our intuitions concerning when some alternative possibility is robust are guided, then, in the following way: An agent has a robust alternative possibility to choosing/doing x in those cases in which the removal of that alternative possibili ty would change our judgment of him from morally responsible for choosing/doing x to not morally respons ible for choosing/doing x ; and, an alternative possibility to x is robust in those cases in which the addition of that alternative possibility would change our judgment of an agent from not morally responsible for choosing/doing x to morally responsib le for choosing/doing x It is the job of one who pr oposes a definition of ‘robustness’ to give the necessary and sufficient conditions that must be met in order for an agent to have a robust alternative possibility, and th e satisfaction of these condi tions should occur when and only when our intuition, guided in the way outlin ed above, tells us that the agent has an alternative open to him that affects our j udgment of his moral re sponsibility. In what follows, I will argue that Pereboom’s argumen t for his definition of ‘robustness’ is flawed, and I will propose a definition of ‘robustness’ that will do a better job of explaining our intuitions about an agent’s moral responsibility in particular cases.

PAGE 55

51 In arguing for his epistemic condition on robustness, Pereboom gives a case in which (a) (the agent presumably knows that) th ere was an alternative open to the agent, that he could have chosen otherwise, and (b) it is clear that the agent is morally responsible, and (c) the agent does not understand th at if he had chosen otherwise he would thereby be precluded fr om the moral responsibility he has for what he does. Roughly, Pereboom’s argument for his ep istemic condition on robustness is the following: in this case, (a) the agent’s ability to choose otherwise is not relevant to (b) the agent’s moral responsibility for what he does because (c) the agent does not understand that by choosing otherwise he would thereby be precluded from the responsibility he has for what he does: Robustness also has an epistemic dimensi on, and it is important that it be made explicit in the characteriza tion of this notion. Imagine that the only way in which Jones could have voluntarily avoided decidi ng to kill Smith is by taking a sip from his coffee cup prior to making the deci sion, and this is only because it was poisoned so that taking a si p would have killed him in stantly. Suppose that Jones does not understand that this action would preclude his deci ding to kill, because he has no idea that the coffee is poisoned. In this situation, Jones could have voluntarily behaved in such a manner that would have preclu ded the action for which he was in fact blameworthy, as a result of which he would have avoided the moral responsibility he actually has. Bu t whether he could have voluntarily taken the sip from the coffee cup, not understandi ng that it would render him blameless in this way, is irrelevant qua alternative possibility to explaining why he is morally responsible for deciding to kill. Despite th e fact that Jones could have voluntarily taken a sip from his coffee cup, and doing so would have rendered him not morally responsible for deciding to kill, this a lternative possibility is nevertheless insufficiently robust to have an importa nt role in grounding the agent’s moral responsibility.1 So, Jones (b) is morally responsible for choos ing to kill Smith, despite the fact that (a) Jones could have voluntarily taken a si p of poisoned coffee, which would thereby preclude him from the moral responsibility he has for killing Smith (had Jones 1 Pereboom, 2005, p. 230.

PAGE 56

52 voluntarily chosen to take a sip of poisoned coffee, he would die and, so, he would not kill Smith and, so, he would not be morally responsible for killing Smith). The reason that (a) Jones’s having the ability to choose otherwise – to choos e to take a sip of coffee is irrelevant to explaining (b ) Jones’s moral responsibility fo r what he does is because (c) Jones did not understand that had he chosen otherwise he would be precluded from the moral responsibility he has for killing Sm ith. Jones does not know that his coffee is poisoned and, so, does not understand that by ta king a sip he would thereby be precluded from the moral responsibility he has for hi s choice to kill Smith. What this suggests, according to Pereboom, is that in order for an alternative possibility to be relevant to explaining an agent’s moral responsibility, it must be the case that the agent understands that by choosing otherwise he would thereby be precluded from the responsibility he has for what he does; in order for an alternative possibility to be robust, it must be the case, according to Pereboom, that the agent understa nds that he could choose otherwise and understands that by choosing ot herwise he would thereby be precluded from the moral responsibility he has for what he actually does. It seems to me that Pereboom’s reasoning desc ribed above is flawed in that (c), that the agent does not understand th at by choosing otherwise he would thereby be precluded from the moral responsibility he has for what he does, is not the only feature of the case that could be appealed to in order to e xplain why (a) the agent’s ability to choose otherwise is irrelevant to (b) the agent’s mo ral responsibility. There is another feature of this case given above, a feature that better e xplains why (a) the agent’s ability to choose otherwise is irrelevant to (b) the agent’s mo ral responsibility: (d) th e agent does not have an alternative (to what we can assume he is being claimed morally responsible for,

PAGE 57

53 choosing to kill and killing Smith) open to him such that he believes that as a result of choosing otherwise, the likelihood will be differe nt (than if he did not choose otherwise) that he choose to kill and kill Smith. I believ e that feature (d) of the case more plausibly explains why (a) Jones’s ability to choose otherwise is irrelevant to (b) his moral responsibility for choosing to ki ll and killing Smith. It is tr ue that by choosing to take a sip of coffee, Jones would not choose to kill Smith (and would not kill Smith), but Jones does not believe that this is the case, nor should we say he has any good reason to believe so. This explains why Jones’ s ability to choose otherwise is irrelevant to his moral responsibility for what he chooses and does. Given this analysis of Pereboom’s poisoned coffee case, I propose the following definition of ‘robustness’: Robustness: an agent has a robust possibility, y to x at tn (where tn is not necessarily the time x was chosen/done) if and only if (i) had he done y at tn he would not have done x and, (j) he believes that he coul d have chosen and done otherwise, y at tn and, (g) the agent believes that he can choose/do y at tn such that the likelihood that he choose/do x is different than it would be if he does not choose/do y at tn.2 So, the reason that Jones’s ability to choose to take a sip of coffee is irrelevant to his moral responsibility for choosing to kill and killing Smith is because Jones does not believe that taking a sip will change the likeli hood that he chooses to kill and kills Smith. According to the proponent of PAP, then, I am suggesting that unless Jones has an alternative to killing Smith that meets the criteria give n above, Jones is not morally responsible for killing Smith in the poisoned coffee case. 2 Recall that the reformulated version of PAP is as follows: having a robust alternative possibility to choosing/doing x is necessary for an agent to be morally responsible for choosing/doing x With this definition of ‘robustness,’ then, what I am suggesting is that the intuition behind PAP really amounts to the following: an agent is morally responsible for choosing/doing x at t1 only if he could have either (a) chosen/done otherwise at t1 or, (b) chosen/done otherwise, y at some time prior to t1, where y satisfies my definition of ‘robustness.’

PAGE 58

54 Let’s test this definition against our intuition, guided in the way previously outlined, about Jones’s moral responsibility in a slightly modified version of the poisoned coffee case. In Poisoned Coffee Case 2, Jones chooses to not and does not take a sip of coffee and then chooses to kill and does kill Smith. Unbeknownst to Jones, though, his coffee is poisoned and, so, had he taken a si p he would have died instantly. Suppose, though, that Jones could not have done anything other than c hoose to not and not take a sip of coffee and then choose to kill and ki ll Smith. Suppose that, unbeknownst to Jones, he has a mental illness such that he always acts on the first course of action that comes to his mind (perhaps he has some kind of attenti on deficit disorder that prevents him from going through an adequate deliberative proc ess of weighing reasons for and against various courses of action), and in this case th e first and only course of action he considers is to not take a sip of coffee and then go kill Smith. Suppose also that, unbeknownst to Jones, there is a device in his brain that is indeterministically stimulating his brain in such a way that the device is indeterministically s upplying his brain with the first course of action that he considers, and, s o, as a result of the combina tion of this device and Jones’s mental illness, Jones cannot do anything othe r than carry out the course of action supplied to him by the device. In Poisoned Co ffee Case 2, the device supplied Jones with the choice to not take a sip of coffee and the choice to kill Smith and, so, Jones follows through with this course of action and does not take a sip of coffee and does kill Smith. It seems plausible that in Poisoned Coffee Case 2 Jones is not responsible for choosing to kill Smith. Now consider Poisoned Coffee Case 3, whic h is just like Poisoned Coffee Case 2, except that the device is not present in Pois oned Coffee Case 3. In Poisoned Coffee Case

PAGE 59

55 3 Jones could have taken a sip of poisoned co ffee had he chosen to do so, i.e., had this been the first course of action he considered (in which case, he would not choose to kill and would not kill Smith, because he would have died instantly). In Poisoned Coffee Case 3, it is also the case that Jones believes th at taking a sip of coff ee will make it less likely that he chooses to kill and kills Smith (i.e., suppose Jones has an alternative possibility that satisfies my proposed defin ition of ‘robustness’); perhaps Jones believes this based on past experiences with drinki ng coffee and making important decisions (such as the decision to kill someone) and past experience with drinking coffee and shooting – Jones has previously taken a sip of coffee be fore practicing his shooting and this has in the past reduced his accuracy. (It is still th e case, though, that Jones, due to his mental illness, can only act on the first course of acti on he considers, where in this case the first course of action he considers is not taking a sip of coffee and then killing Smith). In Poisoned Coffee Case 3, the case in which J ones has an alternative possibility that satisfies my proposed definition of ‘robustness, ’ I believe that the intuition will be that Jones is responsible for choosing to kill Smith whereas he is not responsible for doing this in Poisoned Coffee Case 2, the case in which he had no such alternative. That Jones has an alternative possibility such that he believes that by choosing otherwise he will be less likely to do something – something that he knows is wrong, choosing to kill Smith – affects our judgment about whet her or not Jones is morally re sponsible for what he does. If I am right that one will have the intuition that Jones is morally responsible for choosing to kill and killing Smith in Poisoned Coffee Ca se 3, and I am right that one will have the intuition that he is not morally responsibl e for this in Poisoned Coffee Case 2, and the difference between these two cases is that in Poisoned Coffee Case 2 Jones does not and

PAGE 60

56 in Poisoned Coffee Case 3 Jones does have an alternative open to him that satisfies my proposed definition of ‘robustness,’ then the ad dition of such an alternative to a case affects our judgment of an agent’s moral re sponsibility in the outlined above (i.e., the addition of an alternative that satisfies my definition of ‘robustness’ changes our judgment from not morally responsible for choosing and doing x to morally responsible for choosing and doing x ). One might object that in Poisoned Coffee Case 3, it is still the case that Jones has a mental illness that makes it the case that he can only consider one course of action at a time. One might claim that, despite having a r obust alternative possibi lity, since Jones has this mental illness, it is unjustifiable to hold him morally responsible for what he does. Since Jones is not morally responsible in th e case in which he is given an alternative possibility that satisfies my definition of ‘r obustness,’ my definition of ‘robustness’ is not correct. I believe that this objection is mistaken, and that there is no reason that an agent with a mental illness cannot be held morally responsible for choices and actions when that agent has an alternate po ssibility to what he does and the agent meets a certain epistemic condition concerning the alternative( s) open to him. If an agent could have chosen otherwise and believes that he could do (could have done) so such that choosing otherwise will (would have) make (made) it le ss likely that he kill another person, it is reasonable to hold him responsible for kil ling that person. In Poisoned Coffee Case 3, even though Jones has a mental illness, this mental illness does not prevent Jones from choosing otherwise; in Poisoned Coffee Case 3, Jones can choose otherwise than to not take a sip of coffee and to kill Smith. It could have been the case that the first course of

PAGE 61

57 action Jones considered was to take a sip of coffee and then not kill Smith, in which case Jones would have taken a sip of poisoned coffe e and then died (and, so, not killed Smith). It seems, then, that an agent that has a mental illness will only be exempt from moral responsibility for doing something because of that mental illness if it is the case that his mental illness either prevents him from doing anything other than what he does, or if it prevents the agent from having the belief asso ciated with condition (g) of my definition of ‘robustness.’ The mental illness Jones has in Poisoned Coffee Case 3 neither prevents him from doing anything othe r than what he does, nor doe s it prevent him from having the belief associated with condition (g) of my definition of ‘robustness’ and, so, Jones’s mental illness will not exempt him from mora l responsibility for choosing to kill and killing Smith. So, in Poisoned Coffee Case 3, Jones is morally responsible for choosing to not take a sip of coffee, choosing to kill Smith, and killing Smith, because he has a robust alternative possibility to choos ing and doing these things: he c ould have chosen to take a sip of coffee such that (j) if he chose to take a sip of coffee he would not have chosen to not take a sip of coffee, chosen to kill Sm ith, or killed Smith, and (g) he believed that choosing to take a sip of coffee would make it less likely that he choose to kill Smith and kill Smith. Notice, though, that according to Pereboom’s definition of ‘robustness’ Jones would not have a robust alternate possibility to choosing to kill and killing Smith in Poisoned Coffee Case 3. Recall that accordi ng to Pereboom, in order to have a robust alternative possibility to what an agent does, that agent must be able to choose otherwise such that he understands that by choosing ot herwise he would be precluded from the responsibility he has for what he actually doe s. So, in Poisoned Coffee Case 3, in order

PAGE 62

58 for Jones’s ability to take a sip of coffee to be a robust altern ate possibility to choosing to kill and killing Smith, according to Pereboom Jones would need to understand that by taking a sip of coffee he woul d be precluded from the moral responsibility he has for choosing to kill and killing Smith. Jones does not, though, understand this in Poisoned Coffee Case 3, because Jones believes only th at the likelihood of choosing to kill and killing Smith is reduced if he takes a sip of coffee. In Poisoned Coffee Case 3 Jones believes that if he takes a sip of coffee, it is still possible he will choose to kill and kill Smith. So, since Jones does not understand th at by taking a sip of coffee he will be precluded from the moral responsibility he has for choosing to kill and killing Smith, Jones does not, according to Pereboom’s definition of ‘robustness,’ have a robust alternative possibility. If, then, one agrees that Jones has an alternate possibility that is relevant to his moral responsibility for choosi ng to kill and for killing Smith in Poisoned Coffee Case 3, one cannot accept Pereboom’s defi nition of ‘robustness.’ I believe that one will have the intuition that Jones has an alternative open to him in Poisoned Coffee Case 3 that is relevant to his moral responsibility for choos ing to kill and for killing Smith and, so, one will see that my proposed de finition of ‘robustness’ is more accurate than Pereboom’s defi nition of ‘robustness.’ Given the arguments presented above fo r the definition of ‘robustness’ I am proposing, we can see how Pereboom’s Tax Evas ion 2 Frankfurt-style case leaves the agent with a robust alterna tive possibility and, so, is not a counterexample to PAP because it does not satisfy the first c ondition on an FR scenario. As given, though, Pereboom’s Tax Evasion 2 is underdescribed and, so, I will make a stipulation about the agent’s psychology and show that, given th is stipulation, the agent has a robust

PAGE 63

59 alternative possibility. Recall that in Tax Evasion 2, Joe chooses to not pay attention to his moral reasons for not evading taxes and, s o, chooses to and does evade his taxes. The only way that Joe could have not evaded his taxes was by choosing to pay sufficient attention to his moral reasons for not ev ading. Achieving this requisite level of attentiveness to his mo ral reasons is necessary but not sufficient for his choosing to not evade taxes; had he reached this requisite le vel of attentiveness (w ithout the device being in place) he could either choos e to evade or fail to choose to evade. It is the case, though, that a device is in place such that if Joe reach es the requisite level of attentiveness to his moral reasons for not evading, the device wi ll intervene and bring about the choice to evade and Joe would evade his taxes. Joe act ually chooses to not pay attention to his moral reasons and, so, chooses to evade and does evade, even though Joe could have chosen to pay attention to his moral reasons fo r not evading taxes. Recall that at this point Pereboom argues that this alternative open to Joe – the ability to choose to pay sufficient attention to moral reasons is not a robust alte rnative to what he ac tually does. Given the following reasonable supposition I will make a bout Joe’s psychology, if the definition of ‘robustness’ I have proposed is correct, then Joe’s availa ble option of choosing to pay sufficient attention to his moral reasons fo r not evading taxes is a robust alternative possibility to his choosing to not pay atten tion to moral reasons: Joe believes that he could have chosen otherwise chosen to pa y sufficient attention to his moral reasons – such that if he chose to pay sufficient attent ion to his moral reasons, he would have been less likely to evade his taxes. I will call this version of Tax Evasion 2, in which Jones has a robust alternate possibility to choosing to not pay attention to moral reasons, Tax Evasion 2. In Tax Evasion 2 Joe’s ability to choose to pa y sufficient attention to his

PAGE 64

60 moral reasons is an alternative to what he ac tually chooses, the availability of which is relevant to explaining his moral responsibility for choosing to not pay attention to his moral reasons. So, in Tax Evasion 2, the defender of PAP can argue that Joe is morally responsible for choosing to not pay attention to his moral reasons (on the supposition that Tax Evasion 2 is just like Tax Evasion 2), because he had a robust alternate possibility to doing so. Perhaps, though, this is all Joe is morally responsible for in Tax Evasion 2; perhaps Joe is not morally responsible for c hoosing to evade and for evading his taxes. After reading Tax Evasion 2, I believe one ha s the intuition that Jo e is responsible for something, and the most obvious candidate for this is the action he performs – evading his taxes. It seems to me, though, that it is ope n to the proponent of PAP to claim that this is not what Joe is morally responsible for in Tax Evasion 2, and that he is only morally responsible for his choice to not pay atten tion to moral reasons, in which case Tax Evasion 2 is not a counterexample to PAP because Joe does have a robust alternative possibility to what he is morally responsible for, choosing to not pay attention to his moral reasons. What, though, would be the case if Joe does not have such a belief, if Joe does not believe that choosing to pay sufficient attention to his moral reasons will make it less likely that he choose to evade taxes? It seem s very hard to imagine that a person could not have any beliefs concerning how the consid eration of moral reasons will affect one’s choices and behavior. Cons ider case Tax Evasion 2~ in which we suppose that Joe does not have any such beliefs that he does not believe that paying sufficient attention to his moral reason will in any way change the likelihood that he will evade taxes – would we

PAGE 65

61 still want to claim that Joe is morally respons ible for anything he doe s? If we would want to claim that Joe is morally responsible fo r choosing to not pay attention to his moral reasons and for choosing to evade and evading hi s taxes, despite the f act that he does not, according the definition of ‘robustness’ I have proposed, have a robust alternate possibility to choosing/doing as such, then eith er my definition of ‘robustness’ or PAP (or both) would be false. I think, though, that both my definition of ‘robustness’ and PAP can be upheld in the case at hand, the case in whic h Joe’s alternative possi bility – the ability to choose to pay sufficient attention to his moral reasons – does not satisfy my definition of ‘robustness.’ Tax Evasion 2~ is just like the original poi soned coffee case given by Pereboom, even though it is very difficult for us to imag ine that any person coul d hold the choice of whether or not to consider moral reasons as on par (with respect to the effect that such consideration will have on one’s future choi ces and actions) with the consideration of whether or not to take a sip of coffee. If we can picture that, with respect to the effect that the agent believes that the cons ideration of whether or not to x will have on his future choices and actions, Joe believes that the c hoice of whether or not to pay sufficient attention to his moral reasons is on par with th e choice of whether or not to take a sip of coffee (this is what Joe believes in Tax Evasion 2~ ), I believe the intuition that Joe is morally responsible for choosing to not pay sufficient attention to his moral reasons, choosing to evade, and evading taxes will fade away. In what follows, I will argue that in Tax Evasion 2~ Joe is neither responsible for choosi ng to not pay sufficient attention to his moral reasons, nor for choosing to evade and evading his taxes.

PAGE 66

62 Consider Tax Evasion 2* which is set up ju st like Tax Evasion 2, except that Joe* will choose to evade and will evade taxes only unless he chooses to and takes a sip of coffee; if Joe* does choose to and take a sip of coffee, under normal circumstance, he could then choose whether or not to evade taxes. Suppose, though, th at there is a device such that if Joe* will choose to take a sip of coffee, the device will prevent him from making this choice and will bring about the choice to not take a sip. Joe*, then, cannot choose to do anything but not take a sip of coffee and evade his taxes. Suppose further that Joe* does not believe that the decision of whether or not to take a sip of coffee will in any way affect the likelihood of making a ny future choice or performing any future action. Suppose that for his own reasons, Joe* does not take a sip of coffee and chooses to evade and does evade his taxes. In Tax Ev asion 2*, I cannot see what could justify the claim that he is morally responsible for not ta king a sip of coffee. Taking or not taking a sip of coffee is not normally the kind of activity that one is morally responsible for, unless of course, the agent has some kind of a belief associated with the effects of taking or not taking a sip on his beha vior (and, Joe* does not have any beliefs like this in Tax Evasion 2*). If Joe* is not morally responsible for not taking a sip of coffee in Tax Evasion 2*, and if the decision of whether or not to take a sip of co ffee is considered by the agent as on par (with respec t to the affect on one’s future choices and behavior) with the decision of whether or not to consider moral reasons, then Joe is not morally responsible for choosing to not pay attenti on to his moral reasons in Tax Evasion 2~ the case in which he does not believe that paying attention to his moral reasons will affect the likelihood that he evade taxes. So, if Joe did not have a robust, according to my definition of ‘robust,’ alternative to choosing to not pay attention to his moral reasons (as he would

PAGE 67

63 not in Tax Evasion 2~ ), he would not be morally res ponsible for choosing to not pay attention to his moral reasons. A defender of Frankfurt-style cases might object at this point by claiming that the previous argument ha suggested a Frankfurt-style case that sa tisfies all three conditions of an FR scenario: Tax Evasion 2~ In Tax Evasion 2~ Joe does not have a robust alternate possibility to choosing to evade a nd evading his taxes (given the placement of the device Joe cannot do anything but choose to evade and evade taxes), what makes it the case that he cannot do otherwise does not bring about his actual choice/action (in Tax Evasion 2~ it is still the case that Joe chooses to not pay attention to his moral reasons and, so, chooses to evade and evades his ta xes, and he chooses/does so without the intervention of the device), and it is plausible to claim that Joe is morally responsible for choosing to evade and for evading taxes since he did so indeterminis tically (jus t like Joe did so indeterministically in Tax Evasion 2). I do not, though, believe that Tax Evasion 2~ is a case in which the indeterminacy in Joe’s choice to evade taxes is the right kind of indeterminacy to allow for moral responsibility for so choosing and doing. We must remember what role Tax Evasion 2 (or any Necessary-Condition Frankfur t-style case) is playing in the debate. Recall that Pereboom’s Tax Evasion 2, and any Necessary-Condition example for that matter (for example, Tax Evasion 2~ ), is a response on behalf of the proponent of Frankfurt-style cases to the second horn of the dilemma defens e of PAP. The challenge set forth in the second horn of the dilemma is to give a case in which the agent’s choice is in some relevant way undetermined and yet the agent has no robust alternative possi bilities. The whole motivation for challenging the proponent of Frankfurt-style cases in this way

PAGE 68

64 comes from the Incompatibilis t who claims that in order for an agent to be morally responsible for what he chooses or does, his choice must be in some relevant way undetermined. The Incompatibilist who makes such a challenge, though, will not be satisfied with the presentation of an agent w hose choice is undetermined in just any way. For example, consider a case in which an ag ent’s choice is undete rmined in just the following way: a particular agent who is 21 years old will at age 24 choose to go to medical school only unless at age 22 he choos es that red is his favorite color. Suppose that the agent does choose at age 24 to go to medical school. The Incompatibilist, though, will not accept that the kind of indeterminacy present in this case is the right kind of indeterminacy to allow for moral responsibilit y, and I believe that the Incompatibilist will not accept this kind of indeterminacy for the following reason: the agent has no reason to believe that there is any connection between his choice at age 22 of a favorite color and his choice at age 24 of what ca reer to pursue. In order for the indeterminacy present in a case to be the kind of indeterminacy required for an agent to be morally responsible for some choice C it must be the case that whatever explains the indeterminacy (here, the ability to choose or not choose red as his favorite color) is in some way relevant to the undetermined choice for which the agent is clai med to be morally responsible (here, the choice to go to medical school). Contrast th e kind of indeterminacy present in the case just given with the indeterminacy presen t in Pereboom’s Tax Evasion 2 case. In Pereboom’s case, what explains the fact th at Joe’s choice concer ning whether to evade taxes is undetermined is that he has the opti on of (voluntarily) payi ng attention to moral reasons for and against doing so. The reason, I believe, that this kind of indeterminacy will satisfy the Incompatibilist proponent of the dilemma defense of PAP is because the

PAGE 69

65 weighing of moral reasons for and against doin g something is taken to be (assumed to be) relevant to the undetermined choice Joe is claimed to be morally responsible for, choosing to evade taxes. Alt hough the kind of indeterminacy pr esent in Tax Evasion 2 is the right kind to allow for the agent to be morally responsib le for what he does, it is not necessarily the case that this kind of indeterminacy suffices for moral responsibility and, so, one can still argue that Jo e is not morally responsible for choosing to evade and for evading taxes in Pereboom’s Tax Evasion 2. (It is still the case, then, that it is open to the proponent of PAP to argue that Joe is not mo rally responsible for choosing to evade and for evading taxes in Tax Evasion 2, and that what he is morally responsible for is choosing to not pay sufficient attention to moral reasons, something he has a robust alternative to choosing/doing.). The indeterminacy in Tax Evasion 2~ the case in which Joe has no belief concerning the effect, on his future choices a nd actions, of paying sufficient attention to his moral reasons is like the indeterminacy in the case of the agent who will choose at age 24 to go to medical school only unless he chooses at age 22 that red is has favorite color. In the case in which Joe has no belief con cerning the effect, on hi s future choices and actions, of paying sufficient attention to his mo ral reasons, what explai ns that his choice to evade taxes is undetermined is that he c ould fail to so choose if he chooses to pay sufficient attention (if the devi ce were not in place). But, if his belief concerning the effect, on his future choices and actions, of considering moral reasons is on par with his belief concerning the effect, on his future choi ces and actions, of considering whether or not to take a sip of coffee, then the indeterminacy present in Tax Evasion 2~ is just like the indeterminacy present in the case of the student who w ill choose at age 24 to go to

PAGE 70

66 medical school only unless he chooses at age 22 th at red is his favorite color. If it is the case, then, that Joe does not believe that payi ng attention to moral reasons will affect the likelihood of choosing to evade and evading ta xes, there will not be the right kind of indeterminacy present in th e case to allow for moral re sponsibility. Tax Evasion 2~ (i.e., the case in which Joe does not have an altern ative open to him that satisfies my proposed definition of ‘robustness’), then, is not a case that is a respon se to the dilemma defense of PAP and, so, will not help the proponent of Fra nkfurt-style cases argue his way out of the dilemma. In Tax Evasion 2~ then, Joe is neither morally responsible for choosing to not pay sufficient attention to moral reasons, nor is he morally responsible for choosing to evade and for evading taxes. That Joe has no robust alternative in Tax Evasion 2~ to so choosing/doing according to my proposed definition of ‘robustness,’ then, is neither problematic for my proposed definiti on of ‘robustness,’ nor for PAP. On the one hand, then, we can see that if Joe has an alternative possibility open to him that satisfies my proposed definition of ‘robustness,’ (the versi on called Tax Evasion 2) Tax Evasion 2 will not satisfy the first c ondition of an FR scenar io (the agent must have no robust alternative possibilities to what he actually chooses to do) and, so, will not be a counterexample to PAP. On the other ha nd, if the alternative pos sibility open to Joe does not meet my proposed definition of ‘r obustness,’ (the version called Tax Evasion 2~ ) Tax Evasion 2 will not satisfy the third co ndition of an FR scenario (the agent must be morally responsible for his actual choice) and, so, will not be a counterexample to PAP. I conclude, then, that Pereboom’s Fra nkfurt-style case does not satisfy all three conditions of an FR scenario and, so is not a counterexample to PAP.

PAGE 71

67 The discussion of section three leads me to conclude that the proponents of Frankfurt-style cases have yet to give a cas e that is a counterexample to PAP because they have yet to give a case that satisfies all three conditions of an FR scenario. It is open to the proponent of Frankfurt-style cases to support Pereboom’s definition of ‘robustness’ or some other definition that shows that Tax Evasion 2 does not include such alternatives open to the agent. The proponent of Frankfurtstyle cases could also support one of the three other styles of cases – the No-Prior-S ign example, the Blockage example, or the Internal-Sign example – by showing that someth ing is mistaken with the criticisms of such cases presented in the second section of this essay. It is also open to the proponent of Frankfurt-style cases that finds the defi nition of ‘robustness’ I have here proposed attractive to give a Neces sary-Condition example in which the agent’s choice is undetermined in the right kind of way a nd yet the agent has no robust alternative possibilities given the definition of ‘robustness’ that I have proposed. In this essay I have argued that the pr oponent of Frankfurt-style cases has yet to give a case that satisfies all three conditions of an FR scenario and, so, has yet to give a case that is a counterexample to PAP. I have argued this by discussion a dilemma that is posed to the proponent of Fra nkfurt-style cases, the dilemm a defense of PAP. I have argued that the proponent of Frankfurt-style cases has not adequately responded to the first horn of the dilemma; if it is supposed that the prior sign is sufficient for the agent’s choice in a Frankfurt-style case, the use of a Frankfurt-style case to support the Compatibilist’s conclusion is unnecessary and, s o, this first horn is problematic if the Compatibilist claims that Frankfurt-style ca ses go some way to advancing the debate, between Compatibilists and Incompatibilist s, concerning moral responsibility. Although

PAGE 72

68 the problem with the supposition that the prio r sign suffices for the agent’s choice has, up until now, been posed as a problem of begging the question against the Incompatibilist, I have argued that this is not the real problem with the suppos ition but that, regardless, the proponent of Frankfurt-style ca ses (Fischer) has not adequa tely responded to the real problem posed by the supposition of the first horn of the dilemma defense of PAP. I have also argued that the proponent of Frankfurt-style cases has not adequately responded to the second horn of the dilemm a defense of PAP. I have done this by discussing the four types of cases that propone nts of Frankfurt-style cases have used to respond to this horn of the dilemma, showing that neither strategy succeeds in satisfying all three conditions of an FR scenario and, so, neither strategy succeeds in giving a counterexample to PAP. I have paid sp ecial attention to the Necessary-Condition strategy, used by Derk Pereboom, showing that the proponent of PAP can use my proposed definition of ‘robustness’ to argue that Pereboom’s Tax Evasion 2 is not a counterexample to PAP. I have gone some wa y to showing that my proposed definition of ‘robustness’ is more accu rate than Pereboom’s definition of ‘robustness,’ and, since little attention has been paid in the literature to attempting to define this term, I have (hopefully) helped us to understand more about the Principle of A lternate Possibilities and when it is satisfied.

PAGE 73

69 LIST OF REFERENCES John Martin Fischer, The Metaphysics of Free Wi ll: An Essay on Control (Cambridge, MA.: Blackwell, 1994). John Martin Fischer, “Recent Work on Moral Responsibility,” Ethics 110 (1999): 93-139. John Martin Fischer, “Frankfurt-type Examples and Semi-Compatibilism,” in The Oxford Handbook of Free Will ed. Robert Kane, ed. (New York: Oxford University Press, 2002), 281-308. John Martin Fischer, “Free Will and Moral Responsibility,” in The Oxford Handbook of Ethical Theory ed. David Copp (New York: Oxford University Press, 2006), 321354. Harry G. Frankfurt, “Alternate Po ssibilities and Moral Responsibility,” Journal of Philosophy 66 (1969): 829-839. Harry G. Frankfurt, “Freedom of th e Will and the Concept of a Person,” Journal of Philosophy 68 (1971): 5-20. Stewart Goetz, “Stumping for Widerker,” Faith and Philosophy 16 (1999): 83-89. Stewart Goetz, “Frankfurt-style Counter examples and Begging the Question,” in Midwest Studies in Philosophy Volume XXIX: Free Will and Moral Responsibility, eds. Peter A. French and Howard K. Wettstein, gue st ed. John Martin Fischer (Boston: Blackwell Publishing, 2005), 83-105. David Hunt, “Moral Responsibil ity and Unavoidable Action,” Philosophical Studies 97 (2000): 195-227. Robert Kane, Free Will and Values (Albany: State University of New York Press, 1985). Robert Kane, The Significance of Free Will (New York: Oxford Un iversity Press, 1996). Robert Kane, “Responses to Bernard Be rofsky, John Martin Fischer and Galen Strawson,” Philosophy and Phenomenological Research 60 (2000): 157-167. Robert Kane, “Responsibility, I ndeterminism and Frankfurt-styl e Cases: A Reply to Mele and Robb,” in Moral Responsibility and Alternate Possibilities eds. David Widerker and Michael McKenna (Burlington, VT: Ashgate, 2003), 91-105.

PAGE 74

70 Michael McKenna, “Alternative Possibilitie s and the Failure of the Counterexample Strategy,” Journal of Social Philosophy 28 (1997): 71-85. Alfred R. Mele and David Robb, “R escuing Frankfurt-Style Cases,” The Philosophical Review 107 (1998): 97-112. Michael Otsuka, “Incompatibilism a nd the Avoidability of Blame,” Ethics 108 (1998): 685-701. Derk Pereboom, Living Without Free Will (New York: Cambridge University Press, 2001). Derk Pereboom, “Defending Hard Incompatibilism” in Midwest Studies in Philosophy Volume XXIX: Free Will and Moral Responsibility, eds. Peter A. French and Howard K. Wettstein, guest ed. John Martin Fischer (Boston: Blackwell Publishing, 2005), 228-247. P.F. Strawson, “Freedom and Resentment,” in Free Will ed. Gary Watson (New York: Oxford University Press, 1982), 59-80. Eleonore Stump, “Alternative Possibilities and Moral Resp onsibility: The Flicker of Freedom,” The Journal of Ethics 3 (1999a): 299-324. Eleonore Stump, “Dust, Determinism, and Frankfurt: A Reply to Goetz,” Faith and Philosophy 16 (1999b): 413-422. Eleonore Stump, “Moral Responsibility wi thout Alternative Possibilities,” in Moral Responsibility and Alternative Possibilities eds. David Widerker and Michael McKenna (Burlington, VT: Ashgate, 2003), 139-158. Peter van Inwagen, An Essay on Free Will (New York: Oxford University Press, 1983). David Widerker, “Libertarianism and Frankfur t’s Attack on the Prin ciple of Alternate Possibilities,” Philosophical Review 104 (1995): 247-261. David Widerker, “Frankfurt’s Attack on the Principle of Alternative Possibilities: A Further Look,” Philosophical Perspectives 14 (2000): 181-201. David Widerker, “Responsibility an d Frankfurt-type Examples,” in The Oxford Handbook of Free Will ed. Robert Kane (New York : Oxford University Press, 2002) 323-336. David Widerker and Michael McKenna, eds., Moral Responsibility and Alternative Possibilities (Burlington, VT: Ashgate, 2003). Keith Wyma, “Moral Responsibility and Leeway for Action,” American Philosophical Quarterly 34 (1997): 57-70.

PAGE 75

71 BIOGRAPHICAL SKETCH Jaclyn Genchi was born in Palm Beach Gard ens, Florida, and went to elementary, junior high, and high school in the surrounding areas. She at tended Suncoast Community High School, where she received an Internati onal Baccalaureate degree. Jaclyn went to Elon College, in Elon, North Carolina, for her first year of underg raduate education, and then transferred to the Univers ity of Florida to complete a Bachelor of Arts in philosophy in May 2004. After graduating with her Mast er of Arts in philosophy in August 2006, she will start law school at Duke Law as a 1L in the graduating class of 2009.


Permanent Link: http://ufdc.ufl.edu/UFE0015771/00001

Material Information

Title: Dilemma for Frankfurt-Style Cases, and the Definition of 'Robustness'
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0015771:00001

Permanent Link: http://ufdc.ufl.edu/UFE0015771/00001

Material Information

Title: Dilemma for Frankfurt-Style Cases, and the Definition of 'Robustness'
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0015771:00001


This item has the following downloads:


Full Text












A DILEMMA FOR FRANKFURT-STYLE CASES, AND THE DEFINITION OF
'ROBUSTNESS'
















By

JACLYN GENCHI


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

JACLYN GENCHI
















TABLE OF CONTENTS



A B S T R A C T ....................................................................................................................... iv

INTRODUCTION .................................... ........... .............................

AN OBJECTION TO FRANKFURT-STYLE CASES, AND REPLIES ......................12

The First Horn of the Dilemma: Assumption that the Prior Sign Determines the
Agent's Choice (i.e., That the Prior Sign Deterministically Causes the Agent's
Choice) ........................... ....... ........ .. ... ............ ............... 20
The Second Horn of the Dilemma: Assumption that the Prior Sign is Not
Sufficient for the Agent's Choice (i.e., That the Agent's Choice Is
Indeterm inistically C caused) ..................................... ...................... ................ 29
N o-P rior-Sign C cases .............................................. ....................... .... ......... 30
B lo ck ag e C ase s.................................................................................................... 3 4
Internal-Sign Cases .... ................................................ ............. 40
N ecessary-C condition C ases ............................................................ ................ 43

THE DEFINITION OF 'ROBU STNESS' ................................................. ................ 49

L IST O F R E FE R E N C E S .. ......................................................................... ................ 69

B IO G R APH ICAL SK ETCH .................................................................... ................ 71















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Arts

A DILEMMA FOR FRANKFURT-STYLE CASES, AND THE DEFINITION OF
'ROBUSTNESS'

By

Jaclyn Genchi

August 2006

Chair: Jon Tresan
Major Department: Philosophy

The purpose of this thesis is to argue that one particular objection to Frankfurt-style

cases has thus far not been shown to be mistaken. I defend the objection called the

dilemma defense of the Principle of Alternate Possibilities objection from the replies

given by proponents of Frankfurt-style cases to defend these cases from this objection.















INTRODUCTION

Frankfurt-style cases are traditionally taken to show that the Principle of

Alternative Possibilities (PAP) is false1 to show that alternative possibilities are not

required for moral responsibility. In this essay I will be discussing a particular objection

to Frankfurt-style cases, the dilemma defense of the Principle of Alternate Possibilities. I

will proceed by first explaining the role of Frankfurt-style cases in the debate concerning

moral responsibility, and will follow with an explanation of the objection to Frankfurt-

style cases I will be discussing, the dilemma defense of PAP. Proponents of Frankfurt-

style cases have responded to both horns of the dilemma defense of PAP and, after

reviewing these responses, I will argue that such responses are inadequate and, so, the

dilemma defense of PAP stands as an objection to Frankfurt-style cases. I will give

substantial attention to the strategy used by Derk Pereboom as a response to the dilemma

defense of PAP, arguing that it fails.

Throughout this essay I will be discussing one of the many facets of the debate

concerning moral responsibility, whether Frankfurt-style cases succeed in showing that

the Principle of Alternate Possibilities is false. In considerations of moral responsibility,

much of the debate has been focused on what is called the Compatibility Question: can

we be morally responsible agents in a deterministic world? We are concerned to answer

this question because there seems to be a conflict between our picture of ourselves as


1 The Principle of Alternate Possibilities (PAP) is as follows: having an alternative possibility to
choosing/doing x is necessary for an agent to be morally responsible for choosing/doing x.









agents that have control over what we choose and do, and our picture of ourselves as

parts of a world governed by natural-law-abiding forces. We believe that what we do is in

some way up to us, while at the same time we believe that what we do is a part of the

causal order of the world, and these two pictures of ourselves seem to be in conflict. The

thesis of determinism2 is, roughly, that the events of the past together with the laws of

nature entail the events of the present and future; if all events of the present and future are

entailed by the past, it seems that all of our choices and actions are entailed by events

(together with the laws of nature) that happened before we were born. If the thesis of

determinism is true, we are challenged to find a way in which we can maintain our view

that we have any control over our choices, our actions, the way in which we perform our

actions, whether we perform our actions, etc.

Two positions are taken in response to the Compatibility Question: Compatibilism

and Incompatibilism; the Compatibilist claims that moral responsibility and determinism

are compatible (i.e., that the truth of the thesis of determinism would not entail that we

cannot be morally responsible for what we do), while the Incompatibilist claims that

moral responsibility and determinism are not compatible (i.e., that the truth of the thesis

of determinism would entail that we cannot be morally responsible for what we do).

There are two approaches a Compatibilist position may take. On the one hand, a

Compatibilist may claim that whether or not determinism obtains is irrelevant to whether

or not we can be morally responsible for what we do. P.F. Strawson3 and Harry G.


2 like Peter van Inwagen's statement of the thesis of determinism inAn Essay on Free Will. According to
van Inwagen, the thesis of determinism is the conjunction of the following two theses: (1) For every instant
of time, there is a proposition that expresses the state of the world at that instant, and (2) If p and q are any
propositions that express the state of the world at some instants, then the conjunction of p with the laws of
nature entails q [van Inwagen, 1983, p. 65].

3 Strawson, 1982.









Frankfurt4 both present Compatibilist views of this sort. On the other hand, a

Compatibilist may claim that we can be morally responsible for what we do because

whatever conditions) is (are) necessary and sufficient for being morally responsible are

compatible with determinism. Compatibilism of the latter sort has been the most

prominent Compatibilist approach advocated in recent literature, notably by John Martin

Fischer.5

There are three kinds of Incompatibilist: the Libertarian, the Hard Determinist, and

the Hard Incompatibilist. The Libertarian claims that at least some of the things we do are

not determined and, so, we can be morally responsible for such things, while both the

Hard Determinist and the Hard Incompatibilist claim that we cannot be morally

responsible for any of the things we do. The Hard Determinist claims (disagreeing with

the Libertarian) that all of the things we do are in fact determined and, so, we cannot be

morally responsible for anything that we do. The Libertarian and the Hard Determinist

agree, though, on the following point: that a necessary condition on being morally

responsible for some action is that it be not determined. This view of moral responsibility

is generally taken to be the Incompatibilist intuition, the core of the Incompatibilist

position. Traditionally, I believe, most Incompatibilists were led to either the Libertarian

or Hard Determinist position by first believing that a requirement of my being morally

responsible for something I do is that in so doing I was not determined to do so. There is,

though, a variant of the Incompatibilist position, the Hard Incompatibilist, which adds

that even an action that was not determined, but was undetermined, is not something that


4 Frankfurt, 1971.

5 Fischer, 1994.









we can be held morally responsible for. The Hard Incompatibilist claims that we cannot

be morally responsible for anything we do, regardless of whether or not all of what we do

is determined. The Hard Incompatibilist claims (disagreeing with both the Libertarian and

Hard Determinist) that regardless of whether the thesis of determinism is true or false we

are not morally responsible for what we do because we cannot be morally responsible for

undetermined choices/actions as well as determined choices/actions. The Hard

Incompatibilist claims that indeterminacy would not allow us the sort of control we

believe is required for moral responsibility and which the Incompatibilist believes to be

missing in a deterministic world.

There is another question (one related to the Compatibility Question) that is central

to the debate concerning moral responsibility: what conditions) must be satisfied in order

for an agent to be morally responsible for something he has done? Intuitively, there is at

least one condition that must be satisfied in order for an agent to be responsible for what

he does: the agent must have some kind of control with respect to this choice/action. Both

Compatibilists and Incompatibilists (for the most part) agree that an agent's having some

kind of control over what he does is necessary for that agent to be morally responsible for

so doing. There is major disagreement, though, over just what kind of control is

necessary for being morally responsible. There are two main positions that are taken on

this issue. One position claims that having the relevant sort of control amounts to an

agent being the source or originator of his actions; one who takes this position is either a

Source Compatibilist or a Source Incompatibilist. The Source Compatibilist holds that the

sort of control associated with being the source or originator of one's actions is a sort of

control that both is sufficient for moral responsibility and is compatible with









determinism, whereas the Source Incompatibilist holds that the sort of control associated

with being the source or origination of one's actions is undermined by determinism.

Another position taken with respect to the question of what kind of control is

required for moral responsibility is that this kind of control amounts to having and/or

exercising the ability to choose a course of action from alternatives; one who takes this

position is either a Leeway Compatibilist or a Leeway Incompatibilist. Both the Leeway

Compatibilist and the Leeway Incompatibilist hold that what is (at least partly) relevant

to an agent's being morally responsible for something is that he could have done

otherwise than he actually did, that he had open to him alternative possibilities. Both the

Leeway Compatibilist and the Leeway Incompatibilist adhere to the Principle of

Alternate Possibilities (PAP): having alternate possibilities to what one does is necessary

for being morally responsible for so doing. The Leeway Compatibilist holds that having

alternatives to what one does is necessary for being morally responsible, and that having

them is compatible with determinism. The Leeway Compatibilist will typically claim that

to claim that one has alternatives to what one does is to claim that one would have done

otherwise if one had so chosen, where this ability is taken as compatible with one's actual

choice being determined. The Leeway Incompatibilist holds that the ability to choose/do

otherwise is necessary for being morally responsible, and that this ability is undermined

by determinism; if determinism is true, then one never really has any alternative

possibilities to what one actually chooses/does. If determinism is true, then the past

together with the laws of nature entails what one chooses/does at any moment and, so, if

determinism is true, one never can do anything other than what one does do.











1. We can see that the truth of PAP is playing a crucial role in the Leeway
Incompatibilist's argument for the Incompatibility of moral responsibility and
determinism. The argument presented by the Leeway Incompatibilist for the
incompatibility of moral responsibility and determinism is as follows:
2. Moral responsibility requires having alternate possibilities (PAP).
3. If the thesis of determinism is true, then one never can do otherwise than what one
actually does, one never has any alternative possibilities.
4. Therefore, the truth of the thesis of determinism entails that no one is ever morally
responsible for what one does (i.e., moral responsibility and determinism are
incompatible).
Thus, we can see that if PAP is false, the Leeway Incompatibilist's argument for

the incompatibility of moral responsibility and determinism fails.

In his 1969 essay "Alternate Possibilities and Moral Responsibility," Harry

Frankfurt proposed that a case could be given to show that PAP is false. David Widerker

gives a nice description of the sort of case that Frankfurt is supposing would show PAP

false:

(FR) There may be circumstances in which a person performs some action which,
although [these circumstances] make it impossible for him to avoid performing that
action, [these circumstances] in no way bring it about that he performs it.6

If a case could be given in which (i) a person cannot do otherwise than what he does do

(i.e., in which the agent has no alternate possibilities) and (ii) whatever explains why he

cannot do otherwise is not what causes or brings about the action he performs, then if in

such a case (iii) the agent is morally responsible for what he does, this would show that

having alternate possibilities is not required for moral responsibility. If a case could be

given in which the agent has no alternate possibilities to what he actually does (condition

(i) of an FR scenario), and yet he is morally responsible for what he does (condition (iii)

of an FR scenario), this would show that having alternate possibilities to what one does is


6 Widerker, 2002, p. 323.









not required for being morally responsible for so doing. In order to give a case that meets

both conditions (i) and (iii) of an FR scenario, it must be that the thing that makes it the

case that the agent cannot do otherwise is not the same thing that brings about the agent's

choice or action (condition (ii) of an FR scenario), because if what removes the agent's

alternate possibilities is the same thing that brings about the agent's choice or action, we

would claim that agent was compelled or constrained to choose or act as he did, in which

case he would not be morally responsible for so doing. If a case could be given that meets

all three conditions of an FR scenario, then such a case would be a counterexample to

PAP. The following case, which is John Martin Fischer's standard example, is a

Frankfurt-style case, a case meant to satisfy all three conditions of an FR scenario:

Suppose Jones is in a voting booth deliberating about whether to vote for Gore or
Bush. After reflection, he chooses to vote for Gore and does vote for Gore by
marking his ballot in the normal way. Unbeknownst to him, Black, a liberal
neurosurgeon working with the Democratic Party, has implanted a device in
Jones's brain which monitors Jones's brain activities. If he is about to choose to
vote Democratic, the device simply continues monitoring and does not intervene in
the process in any way. If, however, Jones is about to choose to vote, say,
Republican, the device triggers an intervention that involves electronic stimulation
of the brain sufficient to produce a choice to vote for the Democrat (and a
subsequent Democratic vote). How can the device tell whether Jones is about to
choose to vote Republican or Democratic? This is where the "prior sign" comes in.
If Jones is about to choose at T2 to vote for Gore at T3, he shows some involuntary
sign say a neurological pattern in his brain at T1. Detecting this, Black's device
does not intervene. But if Jones is about to choose at T2 to vote for Bush at T3, he
shows an involuntary sign a different neurological pattern in his brain at T1.
This brain pattern would trigger Black's device to intervene and cause Jones to
choose at T2 to vote for Gore, and to vote for Gore at T3.

So, in this case, Jones does show the relevant prior sign and does vote for Gore,

without any intervention from the device. It seems, then, plausible to claim that Jones is

morally responsible for his choice to vote for Gore. But, had Jones not shown the relevant


Fischer, 2002, p. 282.









prior sign (i.e., had Jones shown a sign indicating that he would do otherwise than choose

to vote for Gore), the device would have intervened and caused Jones to choose to vote

Gore. So, this shows that Jones could do not do otherwise than choose to vote for Gore;

in this case, Jones has no alternate possibility to his actual choice. This case, then, seems

to be an FR scenario, since this is a case in which (i) Jones has no alternate possibility to

what he does (where "does" in this case refers to both his choice to vote for Gore and his

subsequent action of voting), (ii) whatever makes it the case that Jones has no alternate

possibility in this case the intervening device in no way brings about Jones's actual

choice, and (iii) Jones is morally responsible for what he actually does. Since all three

conditions of an FR scenario seem to be satisfied by Fischer's case, this Frankfurt-style

case seems to be a counterexample to PAP.

Before any evaluation of whether or not such Frankfurt-style cases, cases like

Fischer's given above, succeed in showing that PAP is false (i.e., succeed in meeting all

three conditions of an FR scenario), an important limitation on condition (i) of an FR

scenario must be discussed. When the proponent of PAP claims that having alternate

possibilities to what one does is required for being morally responsible for so doing, the

proponent of PAP is not claiming that having just any alternatives is what is necessary for

moral responsibility. Consider a case in which Jones chooses at to to rob a bank and,

consequently, robs a bank at ti, and does so under no compulsion from any internal or

external sources. It could always be claimed that, rather than choosing at to to rob a bank,

it was possible that a meteor fall on Jones at to, or that Jones is teletransported by aliens

to another universe at to, or that Jones spontaneously combusts at to, etc. Although it is

possible that any of these things happen at to as an alternative to what Jones actually does









at that time, none of these things would count as an alternate possibility that could

plausibly be claimed relevant to explaining why Jones is morally responsible for what he

actually does at to, choose to rob a bank. More plausibly, the proponent of PAP is

claiming that having only a certain kind of alternate possibility to what one does is what

is necessary for being responsible for so doing. We call this certain kind of alternate

possibility a robust alternate possibility, which is the sort of possibility that it would be

reasonable or intuitive to claim that having this kind of alternate possibility is what (at

least partially) explains why an agent is responsible for what he actually does; according

to the proponent of PAP, an agent is morally responsible for doing something (at least

partly) in virtue of having a robust alternate possibility (i.e., having a robust alternate

possibility is explanatorily relevant to an agent's moral responsibility). A way to

understand what the proponent of PAP means by 'robust alternative possibility' is to

imagine a case in which all other conditions one believes necessary for moral

responsibility are met (that is, all conditions other than having robust alternative

possibilities; for example, one may believe that it is necessary that the agent does not

have certain forms of mental illness that result in compulsion, that the agent is not under

any form of constant direct brain manipulation by another agent, etc.), and then consider

what kind of alternatives) would be relevant to assessing whether or not the agent is

morally responsible for what he actually does; those alternatives that would be relevant

are robust alternatives and those that would not be relevant are not robust alternatives.

According to the proponent of PAP, then, having a robust alternative possibility to what

one does is required for being responsible for so doing.8


8 Thus, PAP should be reformulated as the following: having a robust alternative possibility to
choosing/doing x is necessary for an agent to be morally responsible for choosing/doing x.









If, then, an FR scenario is meant to be a counterexample to PAP, where PAP is the

claim that having robust alternative possibilities to what one does is required for being

morally responsible for so doing, the first condition of an FR scenario must be changed to

read: the agent has no robust alternate possibilities. So, in order for a case to be a

counterexample to PAP, the following three conditions of an FR scenario must be

satisfied: (1) the agent has no robust alternate possibilities to what he does, (2) whatever

circumstances make it the case that the agent has no robust alternate possibilities are not

what causes him to do what he does, or brings about what he does, and (3) the agent is

morally responsible for what he does. Whether or not Frankfurt-style cases satisfy all

three conditions of an FR scenario and, so, whether or not Frankfurt-style cases are

counterexamples to PAP, will be the topic of the rest of this essay.

In the second section I will discuss the main line of objection to Frankfurt-type

cases made by proponents of PAP, as well as the replies given by the proponents of

Frankfurt-style cases. The proponent of PAP objects that Frankfurt-style cases are faced

with a dilemma. I will discuss both horns of the dilemma, as well as the strategies used

by the proponents of Frankfurt-style cases to respond to the dilemma, focusing most of

my discussion on the second horn of the dilemma (which, I believe, is more interesting).

Concerning the second horn of the dilemma, the proponents of Frankfurt-style cases have

devised four strategies for finding a way out of the dilemma. I will discuss the first three

strategies, showing that they are all problematic strategies to take as responses to the

dilemma defense of PAP. The fourth strategy seems to me to be the strongest strategy and

is taken by Derk Pereboom but, as I will show, still problematic. I will show that the

success of Pereboom's case depends upon the correctness of his definition of






11


'robustness.' In the third section I will argue that Pereboom's definition of 'robustness' is

not correct, and I will argue for an alternative definition of 'robustness,' which if correct,

would show that Pereboom's Frankfurt-style case does include robust alternate

possibilities and, so, does not satisfy condition (1) of an FR scenario. If, then, the

definition of 'robustness' that I propose is correct, then Pereboom's Frankfurt-style case

will not satisfy all three conditions of an FR scenario and, so, will not be a

counterexample to PAP.
















AN OBJECTION TO FRANKFURT-STYLE CASES, AND REPLIES

There are many sorts of objections to Frankfurt-style cases given by defenders of

PAP and in what follows I will be discussing the most prominent objection to Frankfurt-

style cases, the objection that has been given the most attention in the recent literature,

and which I take to be the most serious objection to Frankfurt-style cases. This objection

is called the dilemma defense of PAP1, initially suggested by Robert Kane2 and

developed by David Widerker3.

The dilemma defense of PAP begins by first pointing out that a standard Frankfurt-

style case, such as Fischer's case given in the previous section, does not specify whether

or not the occurrence of the involuntary prior sign is sufficient for the agent to make the

choice he actually makes, C. Recall that in a standard Frankfurt-style case, an agent

makes some choice, C, and that some device is in place to prevent the agent from

choosing otherwise. The device does its job of preventing the agent from choosing

otherwise in the following way: if the agent will make choice C, he will show some

involuntary sign prior to making choice C (and with the occurrence of this prior sign, the

device will not intervene, will remain idle, and will not affect what the agent does), such

that the absence of this sign is the cue for the device to intervene and bring about choice

C in the agent. In the actual scenario the involuntary prior sign is displayed, the device

1 Widerker and McKenna, 2003, p. 8.
2 Kane, 1985, p. 51 and, Kane, 1996, pp. 142-145.

3 Widerker, 1995, pp. 247-261.









does nothing and the agent makes choice C, but if this sign had not been displayed, the

device would have intervened and brought about choice C (i.e., the absence of the prior

sign is sufficient for the device to intervene and bring about choice C). Although the

absence of the prior sign is sufficient for the device to intervene, we can see that it is not

specified whether or not the occurrence of the involuntary prior sign in the actual

scenario is sufficient for the agent to make choice C.

Suppose that it is stipulated that the occurrence of the involuntary prior sign in a

standard Frankfurt-style case is sufficient for the agent to make choice C4, the choice the

agent actually makes without the intervention of the device. If this is stipulated, then the

occurrence of the involuntary prior sign determines that the agent will make choice C. If

the agent's choice is determined by the occurrence of the involuntary prior sign, then it is

not so obvious that the standard Frankfurt-style case meets the third condition of an FR

scenario, that the agent is morally responsible for his actual choice. If an agent's choice is

determined by the occurrence of an involuntary prior sign, it is not clear that we should

claim that the agent is morally responsible for making such a choice, or for performing

the action that is chosen, because it is not clear that we should hold people responsible for

choices and actions that are the product of something involuntary, something the person

had absolutely no control over. Aside from it not being clear that the agent is morally

responsible for his actual choice if it is stipulated that the occurrence of the involuntary

prior sign is sufficient for the actual choice made by the agent in a standard Frankfurt-

style case in which case a standard Frankfurt-style case would not succeed in showing


4 Hereafter, the suggestion that the occurrence of the involuntary prior sign is sufficient for the agent to
make choice C is to be taken to mean either (a) the occurrence of the involuntary prior sign is sufficient for
the agent to make choice C, or (b) the conditions) that produces the involuntary prior sign is (are)
sufficient for the agent to make choice C.









PAP false because such cases would not meet the third condition of an FR scenario to

claim that the agent is morally responsible under such circumstances would be to beg the

question against the Incompatibilist who claims that we cannot be morally responsible for

choices and actions that are determined. So, if we assume that the occurrence of the

involuntary prior sign is sufficient for the agent's actual choice in a standard Frankfurt-

style case, then (a) it is not clear that a standard Frankfurt-style case shows PAP false

because it is not clear that the third condition of an FR scenario, that the agent is morally

responsible for his actual choice, is met by such cases, and (b) to claim that the third

condition is met by such cases would beg the question against the Incompatibilist.

That (a) and (b) result from stipulating that the occurrence of the involuntary prior

sign is sufficient for the agent's actual choice in a standard Frankfurt-style case is

problematic for the success of such cases as counterexamples to PAP under this

stipulation. Concerning (a), part of the appeal of Frankfurt-style cases is that the intuition

that the agent is morally responsible for the actual choice he makes is a very plausible

intuition. Since in the actual scenario of a standard Frankfurt-style case the agent makes

his choice as we believe choices are normally made by adults free from compulsion,

constraint, intervention by any device, etc. the intuition that he is morally responsible is

very reasonable. Part of the appeal of Frankfurt-style cases as counterexamples to PAP,

where PAP is itself also an intuitively plausible principle, rests on the high degree of

reasonability of claiming that the agent is morally responsible for his actual choice even

though he could not have chosen otherwise. So, in so far as the success of Frankfurt-style

cases as counterexamples to PAP relies on the high degree of reasonability of claiming

that the agent is morally responsible for his actual choice, given that (a) results from the









stipulation that the occurrence of the involuntary prior sign is sufficient for the agent's

actual choice in standard Frankfurt-style cases, stipulating this is problematic for the

success of such cases as counterexamples to PAP.

Concerning (b), if the stipulation that the occurrence of the involuntary prior sign is

sufficient for the agent's actual choice in a standard Frankfurt-style case means that the

agent's actual choice was determined, then to claim that the agent is morally responsible

for his choice would be to beg the question against the Incompatibilist. To see why this is

a problem for the use of Frankfurt-style cases as counterexamples to PAP, recall that PAP

is one premise of the Leeway Incompatibilist's argument for the incompatibility of moral

responsibility and determinism: Having robust alternate possibilities to something one

does is necessary for being morally responsible for so doing (PAP), but determinism

removes alternate possibilities, and, so, moral responsibility and determinism are

incompatible. If using a standard Frankfurt-style case as a counterexample to PAP is to

show that the Leeway Incompatibilist's argument fails because Frankfurt-style cases

show that PAP, one premise of this argument, is false, it would be problematic if the

success of such cases was based upon the falsity of the Incompatibilist's conclusion. If

the proponent of Frankfurt-style cases wishes to use such cases to show that the Leeway

Incompatibilist's argument for the conclusion that moral responsibility and determinism

are incompatible fails, then it cannot be claimed both that the agent's choice in such a

case was determined and that the agent is morally responsible for this choice. To claim

both that the agent's choice was determined and that he is morally responsible for this

choice as a part of an argument against an argument that concludes that moral

responsibility and determinism are incompatible would be to beg the question against the









Incompatibilist. We can see, then, that in so far as the proponent of Frankfurt-style cases

wishes to use such cases to disprove a premise of the Leeway Incompatibilist's argument

for the incompatibility of moral responsibility and determinism, the stipulation that the

occurrence of the involuntary prior sign is sufficient for the agent's choice in Frankfurt-

style cases is problematic for the success of such cases as counterexamples to PAP.

In the above given explanation of the first part, or first horn, of the dilemma

defense of PAP, I have attempted to defend the following claim made by the proponents

of the dilemma defense of PAP: if the involuntary prior sign determines the agent's

choice in a standard Frankfurt-style case, then to claim that the agent is morally

responsible for that choice begs the question against the Incompatibilist. While I do find

it problematic for the proponent of Frankfurt-style cases to claim that the agent is morally

responsible for his choice if this choice was determined by an involuntary prior sign, I do

not believe that it is correct to claim that it begs the question against the Incompatibilist

to do so. Considering that the Incompatibilist has his own arguments to support the claim

that moral responsibility and determinism are not compatible, the Incompatibilist may not

share the intuition that the agent in a Frankfurt-style case is morally responsible for his

choice if this choice is determined. But, nevertheless, many will have this intuition and

this intuition is plausible. Frankfurt-style cases are intended to be counterexamples to one

premise (PAP) of the Leeway Incompatibilist's argument for the incompatibility of moral

responsibility and determinism, where PAP has its support in intuition, just as the claim

that the agent is morally responsible has its support in intuition.

Begging the question against one's opponent occurs when one uses the falsity of

the conclusion of one's opponent to prove the truth of one's own conclusion, without









having any reason/support for using the falsity of the opponent's conclusion in one's own

argument. But, the proponent of Frankfurt-style cases does have a reason for denying that

moral responsibility and determinism are incompatible: there is a case, namely a standard

Frankfurt-style case, in which it is plausible to claim that an agent is morally responsible

for this choice, even if this choice was determined. Unless, then, we want to claim that

counterexamples to a premise of an argument (here PAP) beg the question against the

conclusion of that argument (here the Leeway Incompatibilist's conclusion), it does not

seem as if Frankfurt-style cases beg the question against the Incompatibilist if we

suppose that his choice was determined by an involuntary prior sign.

This, though, is not to say that the stipulation that the agent's choice is determined

is not in any other way problematic for the proponent of Frankfurt-style cases. The

stipulation that the agent is morally responsible for his choice if that choice is determined

is problematic if the proponent of Frankfurt-style cases intends for a Frankfurt-style case

to be used in an argument for Compatibilism in something like the following way:

1. It is possible to give a case in which an agent, Jones, chooses to act and acts in the
way we believe normal adults to act free from compulsion, constraint, mental
illness, etc., while his choice is determined by some involuntary prior sign.
2. If Jones chooses to and acts in this normal way, he is morally responsible for so
doing.
3. Suppose that there is an intervening device that would have made Jones choose as
he in fact does had he shown some sign of choosing otherwise. Given the
placement of this intervening device, Jones could not have done otherwise than he
actually did (i.e., he has no alternate possibilities).
4. Therefore, since Jones is morally responsible for his choice and action despite
having no alternate possibilities, alternate possibilities are not required for moral
responsibility (i.e., PAP is false).
5. Determinism is merely one way of removing alternate possibilities, but (given 4)
this removal of alternate possibilities is irrelevant to an agent's moral responsibility
for what he did (and, there is no other feature of determinism that is relevant to an
agent's moral responsibility).
6. Therefore, determinism and moral responsibility are compatible (i.e.,
Compatibilism is true).









The problem with this argument for the proponent of Frankfurt-style cases is that it

makes the use of the special feature (the counterfactually intervening device that

eliminates alternative possibilities) of Frankfurt-style cases unnecessary. Notice that the

conclusion, that determinism and moral responsibility are compatible, follows from the

first and second premise of the argument given above because the first premise stipulates

that the agent's choice was determined, while the second premise states that the agent is

morally responsible for his choice/action. This means that the special feature of a

Frankfurt-style case, that feature being the counterfactual intervener that eliminates the

agent's ability to do otherwise, is playing no role in supporting the Compatibilist

conclusion and, so, the use of Frankfurt-style cases become superfluous in proving the

Compatibilist conclusion. If, though, the proponent of Frankfurt-style cases claims that

Frankfurt-style cases in some way advances the debate concerning moral responsibility

(or plays some role in this debate), the special features of a Frankfurt-style case must be

doing some work that an ordinary case of an agent choosing to perform and performing

some action would not do. Thus, it is problematic for the Compatibilist to use Frankfurt-

style cases to make the argument outlined above. We will see later that John Martin

Fischer uses an argument similar to the one outlined above in order to defend Frankfurt-

style cases from the first horn of the dilemma defense of PAP.

The dilemma defense of PAP continues by proposing that if it is stipulated in

Frankfurt-style cases that the occurrence of the involuntary prior sign is not sufficient for

the agent's choice, it is not the case that the agent lacks the ability to choose otherwise,

and, so, the first condition of an FR scenario would not be satisfied. Under this

stipulation, the occurrence of the involuntary prior sign is merely a reliable indicator that









the agent will make choice C and, so, it is possible that the involuntary prior sign be

displayed and yet the agent make some choice other than C or make no choice at all. If

the occurrence of the involuntary prior sign does not determine that the agent will make

choice C, then, it seems that the agent will have the possibility of making a choice other

than C, and, so, if the occurrence of the involuntary prior sign is not sufficient for the

agent to make choice C, it is not the case that Frankfurt-style cases satisfy the first

condition of an FR scenario, that an agent have no (robust) alternate possibilities to what

he chooses.

It seems, then, that the success of Frankfurt-style cases as counterexamples to PAP

is jeopardized by the stipulation that the occurrence of the involuntary prior sign is not

sufficient for the agent to make choice C. But, as we saw in the previous discussion, the

stipulation that the involuntary prior sign is sufficient for the agent to make choice C also

jeopardized the success of Frankfurt-style cases as counterexamples to PAP. The

proponent of Frankfurt-style cases is, then, posed with the following dilemma: If, on the

one hand, it is stipulated that the occurrence of the involuntary prior sign is sufficient for

the agent to make choice C, then the agent cannot (unproblematically) be claimed to be

morally responsible for his choice, and, so, Frankfurt-style cases do not satisfy the third

condition of an FR scenario. If, on the other hand, it is stipulated that the occurrence of

the involuntary prior sign is not sufficient for the agent to make choice C, then the agent

can choose otherwise and, so, Frankfurt-style cases do not satisfy the first condition of an

FR scenario. Either way, it is argued by the proponent of the dilemma defense of PAP,

standard Frankfurt-style cases cannot satisfy all three conditions of an FR scenario and,

so, such cases are not counterexamples to PAP.









The First Horn of the Dilemma: Assumption that the Prior Sign Determines the
Agent's Choice (i.e., That the Prior Sign Deterministically Causes the Agent's
Choice)

One way that the proponent of Frankfurt-style cases can respond to the dilemma

defense of PAP is to argue that the question is not begged against the Incompatibilist5 if it

is assumed that the occurrence of the prior sign determines the agent's choice. The

proponent of Frankfurt-style cases can argue his way out of the dilemma by arguing that

to assume that the agent in a Frankfurt-style case is morally responsible for his choice

given the stipulation that the occurrence of the prior sign is sufficient for (or, determines)

the agent's choice is not to beg the question against the Incompatibilist. If the proponent

of Frankfurt-style cases could successfully argue that doing so would not beg the

question against the Incompatibilist, then the proponent of Frankfurt-style cases could

claim both that the occurrence of the prior sign is sufficient for the agent's choice and

that all three conditions of an FR scenario are satisfied, thus showing that he is not faced

with a dilemma.

In "Frankfurt-type Examples and Semi-Compatibilism" John Martin Fischer argues

that the use of standard Frankfurt-style cases in an argument against PAP need not be

seen as begging the question against the Incompatibilist if it is assumed that the

occurrence of the prior sign determines the agent's choice:

Begin with the first horn of the dilemma: the assumption that causal determinism
obtains. I agree that one cannot now simply and precipitously conclude, from

5 While I have argued above that the problem with the assumption that the prior sign determines the agent's
choice is not that the question is begged against the Incompatibilist but, rather, that the use of Frankfurt-
style cases becomes unnecessary to establish the Compatibilist's conclusion, my main objective in these
next few paragraphs is to explain the debate concerning this horn of the dilemma as it has played out in the
recent literature as a debate between John Martin Fischer and Stuart Goetz. Both Fischer and Goetz have
framed the discussion as a discussion of whether or not the question is begged against the Incompatibilist
and, so, I will frame the debate as they do for the time being, and then make some critical remarks after
explaining the discussion between Fischer and Goetz.









consideration of the examples, that the agent is morally responsible for his choice
and behavior. But in any case this is not the way I would have proceeded; I never
have envisaged a simple "one-step" argument to the conclusion that (say) Jones is
morally responsible for his choice and action. Rather, I employ the Frankfurt-type
examples as the first (but obviously important) step of a slightly more complex
argument to the conclusion that Jones is morally responsible for his choice and
action (despite lacking alternative possibilities).

The argument goes as follows. First, one carefully considers the Frankfurt-type
cases. Upon reflection, I believe that one should conclude that in these cases the
lack of alternative possibilities does not in itself ground a claim that the agent is not
morally responsible for his choice and action. In other words, I think that the
examples make highly plausible the preliminary conclusion that if Jones is not
morally responsible for his choice and action, this is not simply because he lacks
alternative possibilities. After all, everything that has any causal (or any other kind
of) influence on Jones would be exactly the same, if we "subtracted" [the
intervening device] from the scene. And Jones's moral responsibility would seem
supervenient on what has influence or impact on him in some way.

So the relevant (preliminary) conclusion is, if Jones is not morally responsible for
his choice and action, the reason is not simply that he lacks alternative possibilities.
And it does not appear to beg the question to come to this conclusion, even if
causal determinism obtains. The first step is to argue, based on the Frankfurt-type
examples, that intuitively it is plausible that alternative possibilities are irrelevant to
ascriptions of moral responsibility ... The second step in the argument consists in
asking whether causal determinism in itself and apart from ruling out alternative
possibilities threatens moral responsibility. I have considered various possible
reasons why someone might think that causal determinism does threaten moral
responsibility in itself and apart from ruling out alternative possibilities, and I have
come to the conclusion that it is not plausible to accept any of these reasons. It
seems to me that this two-stage argument is highly plausible and does not beg the
question against the Incompatibilist, even on the assumption of causal determinism.
Thus I believe that the use of the "prior sign" cases can be defended against the
charge of begging the question.6

I have given this lengthy quote from Fischer's essay because this argument is the

starting point of Stewart Goetz's criticism of defending Frankfurt-style cases from the

charge of begging the question. Before explaining Goetz's criticism it is important to get

clear on what Fischer's two-step argument is. I take Fischer's two-step argument against


6 Fischer, 2002, pp. 291-292.









PAP to be the following, where 7-9 constitutes Fischer's first step and 10-12 constitutes

his second step:

7. An agent's moral responsibility, or lack thereof, for some choice and/or action can
only be based upon those things which exerted influence over the agent's making
this choice or performing this action.
8. In the actual scenario of a Frankfurt-style case, that the agent has no alternative
possibilities does not influence the agent's actual choice and action, because this
lack of alternative possibilities results from the intervening device, which does not
exert any influence over the agent's actual choice and action.
9. Therefore, in a Frankfurt-style case, the agent's lack of alternative possibilities
cannot be the basis for the claim that the agent is not morally responsible for his
choice and action. (i.e., lack of alternative possibilities is not sufficient for not
being morally responsible, i.e., PAP is false, i.e., alternative possibilities are
irrelevant to ascriptions of moral responsibility).
10. The agent in a Frankfurt-style case is not morally responsible for his determined
choice only if either (c) alternative possibilities are relevant to ascriptions of moral
responsibility, or (d) determinism directly, and not by ruling out alternative
possibilities, undermines moral responsibility.
11. I (Fischer) have considered various reasons for believing (d) and have come to the
conclusion that (d) is false, and it has been shown in 3 above that (c) is false.
12. Therefore, the agent in Frankfurt-style cases is morally responsible for what he
does, even if his choice is determined by the occurrence of the prior sign.
In his "Frankfurt-style Counterexamples and Begging the Question," Stewart Goetz

replies to Frankfurt's two-step argument by arguing that this two-step argument itself

begs the question against the Incompatibilist by claiming that the agent is morally

responsible for his choice while assuming determinism in the actual sequence. The heart

of Goetz's criticism of Fisher's two-step argument can be seen in the following three

passages:

What is one to think of Fischer's two-step argument? In the end it is difficult to see
how it advances the debate between Compatibilists and Incompatibilists. This is
because contrary to what Fischer claims, his two-step argument begs the question
against the Incompatibilist in the same way that the one-step arguments do: it
assumes, because it requires, the truth of causal determinism in the actual sequence
of events. It requires the truth of causal determinism to create the illusion that it is
the presence of something in the alternative sequence of events (e.g., Black's
device) that makes it the case that Jones is not free to choose otherwise. It is only
through the creation of this illusion and the fact that Black's device is not
explaining Jones' actual choice that one is tempted or inclined to endorse the









conclusion of the first step of Fischer's argument, which is that the lack of
alternative possibilities is not sufficient for the lack of moral responsibility (PAP is
false) and, thereby, that Jones might be morally responsible even though he is not
free to choose otherwise. Once this illusion is exposed, one's initial conviction that
the lack of alternative choice is sufficient for the lack of moral responsibility (PAP
is true) is vindicated.

It is true that the advocate of PAP believes that the lack of the freedom to choose
otherwise is sufficient for the lack of moral responsibility, and in that sense the
latter obtains simply because of the obtaining of the former. It is incorrect,
however, to assume that a person who affirms PAP believes that the lack of moral
responsibility obtains simply because of the obtaining of the freedom to choose
otherwise in the sense that the obtaining of the latter by itself, that is, apart from
what explain it, explains the obtaining of the former. The proponent of PAP thinks
that the lack of the freedom to choose otherwise does not by itself explain the
absence of moral responsibility. This is because he believes that when this lack
obtains, its obtaining is itself explained by, and only can be explained by, the
occurrence of causal determinism in the actual sequence of events.

When an agent is not free to choose otherwise, not only does causal determinism
explain what occurs in the actual sequence of events, but also it explains the lack of
the freedom to choose otherwise. Frankfurt himself recognized this prima facie
dual explanatory role of causal determinism in his classic discussion of PAP. Thus,
he states that "[i]n seeking illustrations of the principle of alternative possibilities, it
is most natural to think of situations in which the same circumstance both bring it
about that a person does something [chooses] and make it impossible for him to
avoid doing it [choosing otherwise]"7 Implicit in Frankfurt's remark is the idea that
it is intuitively plausible to think not only that causal determinism in the actual
sequence of events is capable of producing situations wherein the same
circumstances produce the actual choice and make it impossible for the agent to
choose otherwise, but also that causal determinism in the actual sequence of events
is the only way to bring about situations in which an agent does not have the
possibility of choosing otherwise. It is because it is intuitively plausible to think
this that Frankfurt assumes the burden of constructing examples wherein the
circumstances that make it impossible for Jones to choose otherwise are not the
same as or identical with what is going on in the actual sequence of events. The
reason why Frankfurt's paper is so controversial is that it claims to provide an
example where an agent does not have an alternative choice and the lack of this
alternative is not explained by causal determinism in the actual sequence. Until
such an example is provided, the incompatibilist is on firm ground in believing that
the lack of the freedom to choose otherwise is explained by and can only be


Frankfurt, 1969.









explained by causal determinism in the actual sequence of events and that PAP is
true.8

Essentially, Goetz is arguing that premise 8 (the second premise) of Fischer's two-

step argument against PAP is false, and that the initial plausibility of this premise is due

to the following illusion: that what suffices for the agent's lack of alternative possibilities

in a Frankfurt-style case is only effective (or, active) in the counterfactual scenario (i.e.,

does not exert any influence in the actual scenario). Goetz argues that the truth of premise

8 above relies on the illusion that what explains, or suffices for, the agent's lack of

alternative possibilities is the intervening device, which does not exert any influence over

the agent's actual choice and action (this is the point of the first passage above), when

what really and only can explain, or suffice for, the agent's lack of alternative

possibilities is the obtaining of determinism (this is the point of the second and third

passages above), which does influence by necessitating the agent's actual choice and

action.

While Goetz is, mistakenly I believe (for the reasons laid out in the earlier

discussion of this first horn of the dilemma), ultimately trying to argue that Fischer's two-

step argument begs the question against the Incompatibilist and, so, does not help the

proponent of Frankfurt-style cases respond to the first horn of the dilemma defense of

PAP, I believe that Goetz is pointing to something that is problematic with Fischer's

defense of Frankfurt-style cases. So, even though Goetz mistakenly charges Fischer's

two-step argument with begging the question against the Incompatibilist, Goetz is correct

to point out that premise 8 of Fischer's two-step argument relies on an illusion and, so, is


8 Goetz, 2005, pp. 87-89.









unsupported. A look at Fischer's reply to Goetz's criticism will bring out this mistake

more clearly. Fischer replies that:

In the Frankfurt-type scenarios, two causes make it the case that Jones is unable to
choose otherwise at T2: the prior condition of the world (together with the laws of
nature) and Black's counterfactual intervention. What the examples show is that the
mere fact that Jones is unable to choose otherwise does not in itself establish that
Jones is not morally responsible for his choice. This is because Black's
counterfactual intervention is one of the factors that make it the case that Jones is
unable to choose otherwise at T2, and yet it is irrelevant to the grounding of Jones's
moral responsibility. Considering this factor (the counterfactual intervention), and
bracketing any other factor that might make it the case that Jones is unable to
choose otherwise at T2, it seems to me that Jones may well be morally responsible
for his action. The mere fact that he lacks alternative possibilities, then, cannot in
itself be the reason Jones is not morally responsible, if indeed he is not morally
responsible.

Now, of course, it is also true that the prior condition of the world, together with
the natural laws, make it the case that Jones lacks alternative possibilities. But,
given that the mere fact of lacking alternative possibilities does not in itself rule out
moral responsibility, why should this way of lacking alternative possibilities rule
out moral responsibility? Why exactly should the significance of causal
determinism be that it rules out alternative possibilities? This is exactly what the
Frankfurt-type examples call into question.9

I take it that Fischer's reply to Goetz's argument amounts to the following: the

agent's lack of alternative possibilities in a Frankfurt-style case is brought about by to

two things, one of which is the intervening device that does not influence the agent's

actual choice and action and, so, is irrelevant to the agent's moral responsibility. Given

that one way of removing the agent's alternative possibilities is irrelevant to the agent's

moral responsibility, there seems no reason to believe that any other particular way of

removing alternative possibilities is relevant to an agent's moral responsibility.

Therefore, the obtaining of determinism, which is merely one other way of removing

alternative possibilities, is irrelevant to an agent's moral responsibility and, so, it would


9 Fischer, 2006, pp. 340.









not beg the question against the Incompatibilist to claim that the agent in a Frankfurt-

style case is morally responsible if his choice was determined. (Notice the similarity

between this argument and the argument discussed earlier, the argument I claimed to be

problematic for the Compatibilist).

Fischer's reply to Goetz is mistaken, and the mistake lies in his first claim: "In the

Frankfurt-type scenarios, two causes make it the case that Jones is unable to choose

otherwise at T2: the prior condition of the world (together with the laws of nature) and

Black's counterfactual intervention."10 When Fischer claims that "two causes make it the

case that Jones is unable to choose otherwise,"" he is claiming that there are two causes

that are individually sufficient for Jones being unable to choose otherwise: the obtaining

of determinism and the counterfactual intervener. Fischer must be claiming that the

obtaining of determinism and the intervening device are each sufficient to make it the

case that Jones cannot choose otherwise, because if he is not claiming this, premise 9 of

his two-step argument (that the lack of alternative possibilities is irrelevant to ascriptions

of moral responsibility) would not follow from premise 7 (only factors which exert

influence over what an agent does are relevant to ascriptions of moral responsibility) and

premise 8 (the presence of the intervening device does not exert any influence over the

agent's choice in a Frankfurt-style case). If he is not relying on the claim that the

intervening device is sufficient for the agent's lack of alternative possibilities, then it

would still be possible that there is something that is both sufficient for it and does exert

influence over the agent's choice and, so, the conclusion drawn in premise 9 would not


10 Fischer, 2006, pp. 340.

1 Fischer, 2006, pp. 340.









follow from premises 7 and 8. In order for his two-step argument to be sound, then, when

Fischer claims that "two causes make it the case that Jones is unable to choose

otherwise,"12 we must take him to be claiming that the presence of the counterfactual

intervener is alone sufficient for Jones's inability to choose otherwise.

It is not correct, though, to claim that the intervening device is sufficient for the

agent's lack of alternative possibilities in a Frankfurt-style case. The intervener cannot do

its job of being a counterfactual intervener unless it is the case that the occurrence of the

involuntary prior sign determines that the agent will make choice C and, so, the presence

of the device is not sufficient to remove the agent's ability to do otherwise. It must also

be the case that a deterministic relationship exists between the occurrence of the prior

sign and the agent's making choice C. If this deterministic relationship is not present

between the occurrence of the prior sign and the agent's making choice C, then it could

be the case that the prior sign is displayed (in which case the device does not intervene)

and yet the agent makes a choice other than C. So, Fischer is not correct to claim that the

presence of the intervening device is sufficient for the agent's lack of alternative

possibilities, because it must also be the case that the occurrence of the prior sign

determines that the agent will make choice C. This is the illusion that Fischer is relying

on in making his two-step argument, namely, that the cause of the agent's lack of

alternative possibilities has no (causal) influence over what the agent chooses; but, surely

that the occurrence of the prior sign determines that the agent will make choice C does

have (causal) influence over what the agent chooses. We can see, then, that although


12 Fischer, 2006, pp. 340.









Goetz mistakenly labels Fischer's two-step argument as begging the question against the

Incompatibilist, Goetz has pointed to a serious mistake in Fischer's two-step argument.

According to Goetz, Fischer's two-step argument does not give the proponent of

Frankfurt-style cases a satisfactory response to the dilemma defense of PAP. Goetz gives

the following summary of the debate concerning the success of Frankfurt-style cases as

counterexamples to PAP, and presents the following challenge to the proponent of

Frankfurt-style cases:

So, where are we? Right back where we began, because the dilemma for
proponents of FSCs is still this: either causal determinism obtains in the actual
sequence, in which case Black's device is not causally efficacious in preventing
Jones from choosing otherwise, or causal determinism fails to obtain in the actual
sequence, in which case Black's device is also not causally efficacious in
preventing Jones from choosing otherwise. If my foregoing reasoning is correct, it
is still the case that Fischer has not provided any satisfactory Compatibilist way
through the horns of this dilemma and, as a result, PAP is not undermined. To find
a way through the horns of this dilemma and undermine the intuitively powerful
that PAP captures, it remains the case that Fischer (any Frankfurtian) must provide
an example where what explains the agent's lack of the freedom to choose
otherwise is something other than causal determinism in the actual sequence of
events.13

According to Goetz, then, in order for a Frankfurt-style case to be successful in

showing that PAP is false, it must be the case that the prior sign is not sufficient for the

agent's actual choice (i.e., that the occurrence of the prior sign does not determine that

the agent will make choice C), because if the prior sign is sufficient for the agent's

choice, then this sufficiency is what explains the agent's lack of alternative possibilities,

and this sufficiency is causally relevant to what the agent chooses (in which case the

second condition of a FR scenario would not be satisfied).


13 Goetz, 2005, p. 94.









Given the above discussion, it seems to me that an attempt to rescue Frankfurt-style

cases from the dilemma defense of PAP by responding to the first horn of the dilemma is

not a promising way for the proponent of Frankfurt-style cases to respond to the dilemma

defense of PAP. As an alternative to responding in this way, the proponent of Frankfurt-

style cases can try to show that Frankfurt-style cases can be rescued from the dilemma

defense of PAP by giving a case in which the prior sign does not determine the agent's

choice (or, in which there is no prior sign) and in which the agent cannot choose/do

otherwise and is morally responsible for so choosing/doing.

The Second Horn of the Dilemma: Assumption that the Prior Sign is Not Sufficient
for the Agent's Choice (i.e., That the Agent's Choice Is Indeterministically Caused)

Recall that the second horn of the dilemma defense of PAP brings the following

problem to the table: if it is stipulated that the occurrence of the prior sign is not

sufficient for the agent's choice and, so, the agent's choice is not determined by the prior

sign (i.e., if the agent's choice is brought about indeterministically), then it seems that the

prior sign is merely a reliable indicator that the agent will make choice some choice, C. If

the occurrence of the prior sign is merely a reliable indicator that the agent will make

choice C, it would not follow that the agent could not choose otherwise than C if the prior

sign occurs. So, the problem for the proponent of Frankfurt-style cases is that if the prior

sign is not sufficient for the agent's actual choice, the first condition of an FR scenario

(that the agent cannot choose otherwise, that the agent has no robust alternative

possibilities) cannot be met, and, so, a Frankfurt-style case cannot show that PAP is false.

Many proponents of Frankfurt-style cases, though, have attempted to rescue

Frankfurt-style cases from the dilemma defense of PAP by showing that a case can be

given in which both the agent's choice is brought about indeterministically and in which









the first condition of an FR scenario is met the agent cannot choose otherwise, or, does

not have any robust alternate possibilities. In their introduction to Moral Responsibility

and Alternative Possibilities, editors David Widerker and Michael McKenna give brief

summaries of the four strategies that have developed to give such cases, cases in which

both the agent's choice is brought about indeterministically and in which the agent cannot

choose otherwise, or has no robust alternative possibilities. In what follows I will explain

each of these four strategies, giving an example of a case that uses each strategy. I believe

that the fourth strategy to be discussed, the strategy taken by Pereboom, is the strongest

strategy, and so I will devote most of the discussion that follows to evaluating

Pereboom's Frankfurt-style case. I will, though, briefly discuss the problems I see with

the first three strategies as well.

No-Prior-Sign Cases

Those proponents of Frankfurt-style cases that give No-Prior-Sign cases try to

develop a situation in which an agent cannot choose otherwise, and this is ensured

without the use of a prior sign as a basis for intervention. In No-Prior-Sign cases, the

agent cannot choose otherwise because there are two processes that proceed

independently of one another and which both end in the agent's making some choice, C.

One process, the process that actually results in the agent (freely) making choice C, is an

indeterministic process, and the other process is a deterministic process that will bring

about choice C in all cases except the case in which the other process, the indeterministic

process, brings about this choice. Since the agent makes choice C as a result of an

indeterministic process, we can hold the agent morally responsible for his choice and, yet,

the agent could not do otherwise than make choice C. A No-Prior-Sign case is given by

Alfred Mele and David Robb:









At t1, Black initiates a certain deterministic process P in Bob's brain with the
intention of thereby causing Bob to decide at t2 (an hour later, say) to steal Ann's
car. The process, which is screened off from Bob's consciousness, will
deterministically 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 at t2). (Black is unaware that it is open to Bob to
decide on his own at t2 to steal the car; he is confident that P will cause Bob to
decide as he wants Bob to decide.) The process is in no way sensitive to any "sign"
of what Bob will decide. As it happens, at t2 Bob decides on his own to steal the
car, on the basis of his own indeterministic deliberation about whether to steal it,
and his decision has no deterministic cause. But if he had not just then decided on
his own to steal it, P would have deterministically 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.14

If this case is plausible, it seems to meet all three conditions of an FR scenario: (1)

the agent cannot choose otherwise at t2 than to steal Ann's car, (2) what brings about the

circumstance that the agent cannot choose otherwise the deterministic process P in no

way influences the agent's actual choice, and (3) the agent is morally responsible for his

actual choice to steal Ann's car, because he did so on his own through an indeterministic

decision-making process. I do not, though, believe that this case can meet all three

conditions of an FR scenario given that it is faced with the following dilemma: in terms

of processes x (the indeterministic process that results in the agent's choice to steal the

car) and P (the deterministic process), if we understand each process as being a temporal

process with a beginning and an end, where at the instant the process ends, the choice

made by the agent, the processes can either be such that:

(e) process P is faster than process x (i.e., the deterministic process will end before
the indeterministic process and, so, will result in the agent's choice at t2, before the
indeterministic process has reached its end), or
(f) both processes can be simultaneous (i.e., both processes would simultaneously end
and, so, simultaneously produce the choice at t2, in which case the choice is
overdetermined), or


14 Mele and Robb, 1998, pp. 101-102.









(g) process x can be faster than process P (i.e., the indeterministic process will end
before the deterministic process and, so, result in the agent's choice at t2, before the
deterministic process has reached its end).
In neither situation (e), (f), or (g) will all three conditions of an FR scenario be satisfied.

Suppose that (e) is the case and, so, the deterministic process results in the agent's choice

at t2 to steal the car. If this is the case, then (2) above is not satisfied (and it would beg the

question to claim that (3) is satisfied) and, so, Mele and Robb's case would not be an FR

scenario. Suppose, next, that (f) is the case, and the two processes are simultaneous so

that the agent's choice at t2 is overdetermined. It then seems unclear whether (2) above is

satisfied, and this lack of clarity, I believe, results from it being not clear whether the

existence of a deterministic process usurps any other process as the (and the only) causal

explanation of the end result of such a process. If both a deterministic and an

indeterministic process simultaneously result in the very same thing, it is hard to see what

role the non-deterministic process is playing, since the deterministic process alone is

sufficient for, while the non-deterministic process is not sufficient for, the end in

question. Even though (f) stipulates that both the deterministic and indeterministic

processes end at the same time and, so, both cause the end in question, it is not the case

that the indeterministic process is sufficient for the end in question, while it is the case

that the deterministic process is sufficient for the end in question. A deterministic process

that has some result x at time tn is one which, given the past and the laws of nature, could

not have had any result other than x at tn, whereas an indeterministic process that has

result x at time tn is one which, given the past and the laws of nature, could have had a

result other than x at tn; indeterministic processes, then, are not sufficient for their ends or

results. It is not clear, then, that if the two processes occur simultaneously, that the

deterministic process is not influencing the agent's choice at t2, because it seems










plausible to claim that if the two processes simultaneously result in the choice, then the

deterministic process has (at least, and perhaps all there is to it) something to do with the

agent's making that choice.15 Suppose, finally, that (g) is the case and the indeterministic



15 Mele and Robb address this in their paper, and I take it that they intend that (f) is the case, that the two
processes occur simultaneously and yet the indeterministic process is THE process that results in the
agent's choice at t2 to steal the car: "The coherence of our scenario may, however, be called into question.
How, one might wonder, can it happen that Bob decides on his own at t2 to steal Ann's car, given the
presence of the deterministic process we mentioned? One can understand how, prior to t2, Bob might
decide on his own to steal the car. (Notice that in that case, other things being equal, Bob could have
decided otherwise at this earlier time .. ). But how can it happen that Bob decides on his own at t2 to steal
the car, and that P does not produce the decision, given what we said about P?
Consideration of the following fanciful machine will prove useful in answering this question. The
machine, designed by a specialist in machine art, produces artistic widgets of different shapes and colors.
The colors of the widgets produced are determined by the color of a ball bearing (bb) that hits the
machine's receptor at a relevant time. The machine, M, is surrounded by several automatic bb guns, each
containing bbs of various colors. The relevant aspect of M's mechanical design, for our purposes, is
relatively simple. First, with one qualification, if a bb of color x hits M's receptor, and M is not already in
the process of making a widget, M at once starts a process designed to result in the production of an x-
colored widget. Second, because two or more bbs sometimes hit the receptor simultaneously, the artist has
designed his machine in such a way that whenever this happens (while M is not busy making a widget) M
at once starts a process designed to result in the production of a widget the color of the right-most bb.
Bob is analogous to M in an important respect. He is physically and psychologically so constituted
that if an unconscious deterministic process in his brain and an indeterministic decision-making process of
his were to "coincide" at the moment of decision, he would indeterministically decide on his own and the
deterministic process would have no effect on his decision. This situation is an analogue of a case in which
two bbs of the same color simultaneously hit M's receptor (while M is not busy making a widget)" [Mele
and Robb, 1998, pp. 103-104].
It seems to me that the consideration of this fanciful widget-making machine does not help Mele
and Robb's case to be an FR scenario. Recall that Mele and Robb's case is intended to be a case in which
the agent could not have chosen otherwise than to steal Ann's car, whatever makes it the case that the agent
could not choose otherwise does not exert influence over what the agent chooses, and the agent is morally
responsible for his choice because he made it indeterministically. Stipulating that the agent makes the
choice to steal Ann's car indeterministically is what allows Mele and Robb's case to be a case that satisfies
the second and third of these conditions despite the fact that a deterministic process makes it the case that
the agent could not have chosen otherwise. One should note, though, that merely stipulating that the agent's
choice was made indeterministically does not amount to the stipulation that the agent's choice was made
indeterministically in the right sort of way, where the right sort of way is the way such that the agent can be
claimed morally responsible (at least partly) in virtue of the choice arising in this indeterministic fashion.
In order that the agent's choice being brought about indeterministically is the reason, or one of the reasons,
for which we claim that the agent is morally responsible for this choice, it must be the case that the
indeterminacy is located in a proper place in the causal history of the choice. For example, if the
indeterminacy was located at some point in the causal history of the choice that occurred billions of years
ago, we would not claim that this is the right sort of indeterminacy to allow for the agent's moral
responsibility for his choice billions of years later. Although it is very difficult to determine at what point in
the causal history of a choice the indeterminacy must be in order that the agent is morally responsible for
this choice, it seems most reasonable to claim that the indeterminacy must occur at the very moment the
choice is made. It seems to me that when we claim that some choice was made indeterministically, this
does not mean that something else happened indeterministically from which followed a choice; it seems to
me that when we claim that some agent is morally responsible for some choice (at least partly) in virtue of
the fact that the choice was made indeterministically, we do not mean to claim that some agent is morally










process results in the agent's choice at t2 because it reaches its end before the

deterministic process reaches its end. In this case, the indeterministic process will result

in the agent's choice and, so, it is possible that the agent will make some choice at t2

other than to steal the car. If (g) is the case, then, (1) above is not true and, so, Mele and

Robb's case would not be an FR scenario. Either way, then, whether (e), (f), or (g) is the

case, Mele and Robb's Frankfurt-style case does not meet all three conditions of an FR

scenario.16

Blockage Cases

Blockage Examples also attempt to meet all the conditions of an FR scenario

without the use of a prior sign. In a blockage case, all neural pathways other than the

pathway that is (coincidentally) actually taken are blocked, so that (1) the agent could not

responsible for his choice (at least partly) in virtue of the fact that something happened indeterministically
from which followed the agent's choice. The claim that the agent is morally responsible for some choice
made only makes sense to me if it is the occurrence of the choice itself that is indeterministic; if it is not the
case that up until the choice was made it could have been different than it was (i.e., if it is not the case that
the occurrence of the choice is itself indeterministic), it is hard to see the relevance of the claim that an
agent made some choice indeterministically to the claim that an agent is morally responsible for that
choice.
In order, then, for the agent to be morally responsible for his choice at tn, it must be the case that
everything in the world could have been just as it was up to tn and yet it was possible that the agent makes
some other choice at that time. I would like to suggest that, given the setup of their case, the indeterministic
process x that results in the agent's choice to steal Ann's car does not meet this condition. Taking into
account the analogy with the widget-making machine, and given that (f) is the case and the deterministic
process and the indeterministic process both hit the decision-making node(s) simultaneously, it must be the
case that the processes hitting the decision-making node is a distinct event from the event of the choice
being made in the same way that two bbs of the same color hitting the right and left receptors is a distinct
event from the initiation of or the making of the widget. Since the indeterministic process hitting the
decision-making node is a separate event from the event of the choice being made, in order for the choice
to be made indeterministically in the right sort of way to allow for the agent to be morally responsible for
this choice, it must be the case that everything prior (including the indeterministic process hitting the
decision-making node) could have been the same as it was, and yet the agent could have chosen otherwise.
In Mele and Robb's case, though, this is not so; once the indeterministic process hits the decision-making
node, the agent will choose to steal Ann's car. And, so, I believe that we should not claim that the agent's
decision in Mele and Robb's case is brought about indeterministically in the right sort of way to allow for
the agent to be morally responsible for his choice and, so, Mele and Robb's case does not satisfy the
conditions of an FR scenario and, so, is not a counterexample to PAP.


16 It should be noted that others have criticized Mele and Robb's case as well: See Pereboom, 2001, pp. 14-
15; See also Kane, 2003; See also Widerker, 2000, pp. 183-185.









have done anything other than he did, but (2) the blockage is such that it in no way

affects what the agent actually does and, so, (3) the agent is morally responsible for what

he does. In "Moral Responsibility and Unavoidable Action," David Hunt suggests such a

blockage case that is based upon Locke's locked door example:

The original "Frankfurt scenario" is of course John Locke's famous example in
which "a man be carried whilst fast asleep into a room where is a person he longs
to see and speak with, and be there locked fast in, beyond his power to get out; he
awakes and is glad to find himself in so desirable company, which he stays
willingly in, i.e. prefers his stay to going away.""1 In this case the man is morally
responsible for what he does even though he cannot do otherwise .. But the
condition that makes for unavoidability in this example is not counterfactual in
nature: the door is actually locked; it doesn't lock only when someone approaches
the door and tries to leave. What makes the locked door compatible with the man's
moral responsibility is simply that it is not among the conditions actually leading
the man to stay in the room .. Locke's example does provide some initial
encouragement that the unavoidability essential to a Frankfurt scenario does not
have to rest on a counterfactual device.

Thus encouraged Imagine then a device that blocks neural pathways rather than
doorways The mechanism is not intervening directly in the series itself; it is
allowing the series to unfold on its own, but simply blocking all alternatives to the
series. Of course it can't block alternatives in response to the way the series is
unfolding, because then the blockage would be coming too late to have any effect
on the avoidability or unavoidability of Jones's actions. Instead the mechanism
blocks alternatives in advance, but owning to a fantastic coincidence the pathways
it blocks just happen to be all the ones that will be unactualized in any case, while
the single pathway that remains unblocked is precisely the route the man's thoughts
would be following anyways (if all neural pathways were unblocked). Under these
conditions, the man appears to remain responsible for his thoughts and actions ...
18

It is clear that Hunt takes his case to satisfy all three conditions of an FR scenario: (1) the

agent cannot do otherwise than he does (this is ensured by the mechanism, which,

coincidentally, blocks all unactualized pathways before the actual pathway is actualized),



1 Hunt's footnote reads: John Locke, An Essay Concerning Human Understanding Bk.II, ch. XXI, 10. Of
course Locke has his own agenda here, which is not entirely the same as Frankfurt's.


18 Hunt, 2000, pp. 217-218.









(2) what brings about the agent's inability to do otherwise in no way influences what he

does (because the operation of the mechanism does not affect the actual pathway), and

(3) the agent is morally responsible for what he does. Robert Kane has challenged the

intuition elicited by Hunt's blockage case, and Pereboom elaborates upon Kane's

challenge, arguing that our intuition that (3) the agent is morally responsible in this case

is not clearly correct, because it is far from clear whether it is reasonable to claim, given

the circumstances, that the agent's choice was brought about indeterministically. Kane's

challenge to such blockage cases is the following:

In [a case in which every other alternative is blocked except the agent's making A
at t], of course there are no alternative possibilities left to the agent; every one is
blocked except the agent's choosing A at t. But now we seem to have determinism
pure and simple. By implanting the mechanism in this fashion, a controller would
have predetermined exactly what the agent would do (and when); and, as a
consequence, the controller, not the agent, would be ultimately responsible for the
outcome. Blockage by a controller that rules out alternative possibilities is simply
predestination; and on my view at least, predestination runs afoul of ultimate
responsibility.19

Kane's claim here is that given the method of the mechanism, even though it only

coincidentally blocks the pathways that are unactualized, since the mechanism does this

before the actualized pathway is actualized, the blockage amounts to predestination of the

agent's choice. If an agent's choice being predestined means, as it means according to

Kane, that the agent is not ultimately responsible for such a choice, then, the presence of

the blockage removes the agent's responsibility. Kane's objection to blockage cases,

then, is that blocking alternative pathways prior to the actualization of the pathway that

leads to the agent's choice amounts to determining the agent's choice, in which case, the


19Kane, 2000, p. 162.









third condition of an FR scenario would not be met by such blockage cases (without

begging the question).

In Living Without Free Will, Pereboom strengthens this criticism by first

suggesting that the plausibility of claiming that the agent in a blockage case is morally

responsible is due to the fact that the presence of the blockage does not have any part in

the actual causal history of the agent's choice it is purely coincidental that just those

pathways were blocked that were unactualized (and would have been unactualized even if

the blockages were not present). Consider, though, the following two situations: In

situation A the agent makes choice x with no blockages or anything else unusual present

(and so could have at that time chosen differently), while in situation B the agent makes

choice x with the same causal history as in situation A, but in situation B the blockages

are in place in the way that Hunt suggests ahead of time such that coincidentally those

pathways other than the actualized pathway are blocked.20 Concerning situations A and B,

Pereboom says that "cases of the above sort might be misleading just because it is natural

to assume that the actual causal history of an event is essentially the same in each, given

that the only difference between them is a restriction that would seem to have no effect




20 1 have paraphrased the relevant aspects of Situation A (the first case) and Situation B (the second case),
the cases Pereboom uses at this point of the discussion. Situation A and Situation B are given on p. 16 of
[Pereboom, 2001] as follows: "Situation A: Ms. Scarlet deliberately decides to kill Colonel Mustard at t1,
and there are no factors beyond her control that deterministically produce her choice. When she chooses to
kill the Colonel, she could have chosen not to kill him. There are no causal factors that would prevent her
from not making the choice to kill Colonel Mustard. Situation B: Ms. Scarlet's choice to kill Colonel
Mustard has precisely the same actual causal history as in A. But before she even started to think about
killing Colonel Mustard, a neurophysiologist has blocked all the neural pathways not used in Situation A,
so that no neural pathway other than the one employed in that situation could be used. Let us supposed that
it is causally determined that she remain a living agent, and if she remains a living agent, some neural
pathway has to be used. Thus every alternative for Ms. Scarlet is blocked except the one that realizes her
choice to kill the Colonel. But the blockage does not affect the actual causal history of Ms. Scarlet's choice,
because the blocked pathways would have remained dormant."









on the event."21 Consider, though, the following two-situation case discussed by

Pereboom22, which seems to show that two situations with the (seemingly) same actual

causal history may differ in that the one with (something relevantly similar to) a blockage

may be essentially determined to proceed as it does:

Imagine a universe correctly described by Epicurean physics: At the most
fundamental level all that exists is atoms and the frictionless void, and there is a
determinate downward direction in which all atoms naturally fall except if they
undergo uncaused swerves.

Situation C: A spherical atom is falling downward through space, with a certain
velocity and acceleration. Its actual causal history is indeterministic because at any
time the atom can be subject to an uncaused swerve. Suppose that the atom can
swerve in any direction other than upwards. In actual fact, from ti to t2 it does not
swerve.

A counterfactual situation diverges from C only by virtue of a device that
eliminates alternative possibilities and all differences thereby entailed:

Situation D: The case is identical to C, except that the atom is falling downward
through a straight and vertically oriented tube whose interior surface is made of
frictionless material, and whose interior is precisely wide enough to accommodate
the atom. The atom would not have swerved during this time interval, and the
trajectory, velocity, and acceleration of the atom from t, to t2 are precisely what
they are in C.23

So, what is the lesson from Situations C and D? One might initially want to claim that the

causal histories in these two situations are identical between ti and t2, because the same

thing happens to the particle during this time interval in both situations. This intuition,

though, is questionable because it might be that the presence of the tube in Situation D

makes the causal history of the particle in Situation D essentially deterministic. If the

tube prevents the particle from swerving, and the possibility of swerving is what

21 Pereboom, 2001, p. 17.

22 Pereboom notes on p. 17 of Pereboom, 2001 that this two-situation case is "modeled on a reflection of
Hunt's" from Hunt's personal correspondence with Fischer, cited on pp. 119-120 of Fischer, 1999.


23 Pereboom, 2001, p. 17.









constitutes the indeterminate nature of the particle's movement, the tube precludes the

movement of the particle from being indeterminate in nature; if the tube precludes the

movement of the particle from being indeterministic, then the tube makes it the case that

the particle's movement is essentially determined and, so, makes the particle's causal

history deterministic in Situation D. In order for it to be the case that the movement of the

particle through the tube between times ti and t2 is indeterministic, it must be the case

that at ti, t2, and any moments between (if there are any) the world could have been just

as it was up until that moment, and yet the particle could have swerved (rather than not

swerved, which is what it actually did). Given the placement of the tube, though, the

particle could not have swerved and, so, it movement between ti and t2 is not

indeterministic. This would show that an object in two situations, here C and D, can

(seemingly) have the same thing happen to it in both situations, and yet in one situation,

C, what happens to the object is indeterministic, whereas in the other, D, the presence of

a blocking mechanism (here, a tube) makes it the case that what happens to the object is

determined. If this is the correct thing to say about situations C and D, this lesson applies

to Hunt's blockage case in the following way: the intuition that the agent is morally

responsible in Hunt's blockage case is taken to be reasonable, given that the actual causal

history of the agent's choice is not affected by the presence of the blockages. The lesson

from situations C and D is that the presence of such a blockage may affect the causal

history (making it deterministic), even though the agent's choice would have been the

same had the blockage not been present. Concerning what the lesson from situations C

and D shows us about blockage cases, Pereboom says the following:

Sympathy for Frankfurt-style arguments is generated by the sense that moral
responsibility is very much a function of the features of the actual causal history of









an action, to which restrictions that exist but would seem to play no actual causal
role are irrelevant. However, in a scenario in which such restrictions, despite initial
appearances, could be relevant to the nature of the actual causal history of an action
after all, one's intuitions about whether the agent is morally responsible might
become unstable. My own view is not that actual causal histories in blockage cases
are clearly deterministic, but only that these considerations show that they may be.
This type of problem should make one less confident when evaluating these
difficult kinds of Frankfurt-style cases.24

Internal-Sign Cases

Unlike No-Prior-Sign examples and Blockage examples, Internal-Sign examples do

not attempt to eliminate a sign whose presence indicates that the agent will make the

desired choice, C, and whose absence is sufficient for the intervention of the device to

produce choice C. Rather, Internal-Sign examples are set up so that there is a sign of this

sort, but, rather than occurring prior to the agent's action (i.e., rather than being a prior

sign), the sign is internal to the event of the agent's choice/action. Such a case is given by

Eleonore Stump in "Alternative Possibilities and Moral Responsibility: The Flicker of

Freedom":

Suppose that a neurosurgeon Grey wants his patient Jones to vote for Republicans
in the upcoming election. Grey has a neuroscope which lets him both observe and
bring about neural firings which correlate with acts of will on Jones's part. Through
his neuroscope, Grey ascertains that every time Jones wills to vote for Republican
candidates, that act of will correlates with the completion of a sequence of neural
firings in Jones's brain that always includes, near its beginning, the firing of
neurons a, b, c (call this neural sequence "R"). On the other hand, Jones's willing to
vote for Democratic candidates is correlated with the completion of a different
neural sequence that always includes, near its beginning, the firings of neurons x, y,
z, none of which is the same as those in neural sequence R (call this neural
sequence "D"). For simplicity's sake, suppose that neither neural sequence R nor
neural sequence D is also correlated with any further set of mental acts. Again, for
simplicity's sake, suppose that Jones's only relevant options are an act of will to
vote for Republicans or an act of will to vote for Democrats.

Then Grey can tune his neuroscope accordingly. Whenever the neuroscope detects
the firing of x, y, and z, the initial neurons of neural sequence D, the neuroscope


24 Pereboom, 2001, p.18.









immediately disrupts the neural sequence, so that it isn't brought to completion.
The neuroscope then activates the coercive neurological mechanism which fires the
neurons of neural sequence R, thereby bringing it about that Jones wills to vote for
Republicans. But if the neuroscope detects the firing of a, b, and c, the initial
neurons in neural sequence R, which is correlated with the act of will to vote for
Republicans, then the neuroscope does not interrupt that neural sequence. It doesn't
activate the coercive neurological mechanism, and neural sequence R continues,
culminating in Jones's willing to vote for Republicans, without Jones's being
caused to will this way by Grey.

And suppose that [in this case] Grey does not act to bring about neural sequence R,
but that Jones wills to vote for Republicans, without Grey's coercing him to do so.25

In setting up this case, Stump makes two presuppositions. The first presupposition

is that there is a correlation between the mind and the brain, that mental events are

somehow correlated with neural events. The nature of this correlation is left vague so that

various different theories of mind/brain interaction can be accommodated by this

Frankfurt-style case. This presupposition leaves it open whether each mental event is a

temporally extended event that occurs over the time period from the beginning to the end

of the corresponding neural event (in which case, the mental event does not occur unless

the entire neural sequence occurs), or whether each mental event occurs after the

completion of the corresponding neural event (in which case the mental event does not

occur until the entire neural sequence occurs, when the last neuron firing of the neural

event is completed). The second presupposition is that the correlation between a mental

event and the neural event it is correlated with is a one-many relation; each mental event

(say, a choice) is correlated to a neural event that consists of many neural firings.26


25 Stump, 1999a, pp. 303-305.

26 "When I suddenly recognize my daughter's face across a crowded room, that one mental act of
recognition, which feels sudden, or even instantaneous, is correlated with many neural firings as
information from the retina is sent through the optic nerve, relayed through the lateral geniculate nucleus of
the thalamus, processed in various parts of the occipital cortex, which take account of figure, motion,
orientation in space, and color, and then processed further in cortical association areas. Only when the
whole sequence of neural firings is complete, do I have the mental act of recognizing my daughter.










It seems to me that Stump's Frankfurt-style case is in no better shape than Mele and

Robb's, and is faced with a similar problem (refer back to footnote 20). In Stump's case,

it must be the case that there is indeterminacy in the event of the agent's choice being

made such that this indeterminacy is the right kind of indeterminacy, meaning that (at

least partly) in virtue of his choice being undetermined in this way the agent can be

morally responsible for his choice. We can see that it must be the case that the

indeterminacy in Stump's case lies in the initiation of the neural sequence that correlates

with the mental event 'choosing to vote Republican' (this neural sequence begins with

firings a, b, c), or in the initiation of the neural sequence that correlates with the mental

event 'choosing to vote Democrat' (this neural sequence begins with firings x, y, z).2

Stump wants to claim (see footnote 30) that unless a neural event is brought to

completion, nothing mentally significant has occurred (i.e., if a neural event is terminated

prior to its completion, no mental event occurs at all); so, for example, the firings of a, b,

c alone are not mentally significant, because these firings are only the beginning of a

neural event. If the indeterminacy occurs in the initiation of the neural sequence, and

nothing mentally significant occurs until after the completion of the neural sequence, then




Whenever neural firings are associated with an act of will, I take it that in this case, as in all others, the
correlation between the mental act and the firing of the relevant neurons is a one-many relation" [Stump,
1999a, p. 306].
27 Stump intends for the indeterminacy to be located in the initiation of the neural sequence, as is noted in a
later essay of hers: "So, as I originally presented the case, the initiating cause of the neural sequence in [the
case] whatever exactly that is is itself indeterministic and to be understood in a way that doesn't
preclude the indeterministic firing of the initial neurons of the sequence" [Stump, 2003, p. 142]. It must be
that the indeterminacy in Stump's case is at no point later than the initiation of the neural sequence. It
cannot be the case, for example, that the indeterminacy occurs after neurons a, b, c have fired, because then
it would be the case that neurons a, b, c fire (in which case the device does not intervene) and yet a
different neural event occurs, one other than that which correlates with the mental event 'choosing to vote
Republican' (in which case the agent could choose otherwise, even though the intervening device is in
place).









the indeterminacy is located prior to the occurrence of anything mentally significant.28

For an agent's choice to be undetermined in the right kind of way, to allow for the agent

to be morally responsible for that choice, it must be the case that up until the moment that

choice is made, the past (which in Stump's case would include every neuron firing up

until the last one of the neural sequence, everything up until something mentally

significant occurs) could have been just as it was and yet it be possible for the agent to

choose otherwise. This is not so in Stump's case, because once firings a, b, c occurs, the

rest of the sequence will be completed and the mental event 'choosing to vote

Republican' will occur. So, in Stump's Frankfurt-style case the agent's choice is not

undetermined in the right way to allow for the agent to be morally responsible for this

choice and, so, the conditions of an FR scenario are not met and, so, Stump's case is not a

counterexample to PAP.

Necessary-Condition Cases

Necessary-Condition Cases make use of a device that relies on a prior sign in order

to guarantee that the agent cannot choose otherwise than he does, but this prior sign is

merely a necessary condition for the agent doing otherwise, rather than a sufficient

condition of the agent doing otherwise. In a Necessary-Condition example, the agent

(indeterministically) makes choice C, but in order for him to fail to make choice C he



28 Suppose that Stump were to claim that the beginning of a neural sequence, for example firings a, b, c, is
mentally significant and, so, the indeterminacy lies in the initiation of something mentally significant. This
claim, though, would be problematic as well. If the beginning of a neural sequence is mentally significant,
then the agent in Stump's case would have an alternative possibility: firings x, y, z could have happened,
which would be the mentally significant beginning of the mental event 'choosing to vote Democrat'
(perhaps the mental event in this case would be 'trying to vote Democrat'), in which case the device would
intervene and bring about firings a, b, c and the mental event 'choosing to vote Republican' would also
occur. Since there is this alternative possibility under the stipulation that the neural firings at the beginning
of a neural sequence are mentally significant, under this stipulation Stump's case would not satisfy the
conditions of an FR scenario and, so, would not be a counterexample to PAP.









would have has to (indeterministically) choose/do A. If the agent chooses/does A, though,

the device would have intervened and brought about choice C. A Necessary-Condition

example is given by Derk Pereboom, in its most revised version in "Defending Hard

Incompatibilism."

Tax Evasion (2): Joe is considering whether to claim a tax deduction for the
substantial local registration fee that he paid when he bought a house. He knows
that claiming the deduction is illegal, that he probably won't be caught, and that if
he is, he can convincingly plead ignorance. Suppose he has a very powerful but not
always overriding desire to advance his self-interest regardless of the cost to others,
and no matter whether advancing his self-interest involves illegal activity.
Crucially, his psychology is such that the only way in this situation he could fail to
choose to evade taxes is for moral reasons. His psychology is not, for example,
such that he could fail to choose to evade taxes for no reason or simply on a whim.
In addition, it is causally necessary for his failing to choose to evade taxes in this
situation that he attain a certain level of attentiveness to these moral reasons. He
can secure this level of attentiveness voluntarily. However, his attaining this level
of attentiveness is not causally sufficient for his failing to choose to evade taxes. If
he were to attain this level of attentiveness, Joe could, with his libertarian free will,
either choose to evade taxes or refrain from so choosing (without the intervener's
device in place). More generally, Joe is a libertarian free agent. But to ensure that
he chooses to evade taxes, a neuroscientist now implants a device, which, were it to
sense the requisite level of attentiveness, would electronically stimulate his brain so
that he would choose to evade taxes. In actual fact, he does not attain this level of
attentiveness, and he chooses to evade taxes while the device remains idle.29

In Tax Evasion 2, Joe (1) cannot choose otherwise than he does, choose to evade

taxes, because if he met a necessary condition of choosing otherwise attained a certain

level of attentiveness to his moral reasons the device would intervene and bring about

the choice to evade taxes, (2) what brings about the situation described in (1), the device,

in no way affects what Joe actually does, and (3) Joe is morally responsible for what he

does because he did so on his own, was not determined to do so, etc.

Some proponents of PAP, though, may challenge (1) and claim that Joe does have a

relevant alternative open to him: he can voluntarily choose to attain the requisite level of


29 Pereboom, 2005, p. 231-232.









attentiveness to moral reasons for not evading taxes. His ability to so choose is a morally

relevant alternative because had Joe chosen to reach the requisite level of attentiveness to

moral reasons, the device would have intervened and forced Joe to choose to evade taxes.

Under these circumstances, though, Joe would not be morally responsible for choosing to

evade taxes, because this choice would be the product of the device and not made freely

by Joe. Joe's ability to choose to attain this requisite level of attentiveness to moral

reasons, then, is a relevant alternative one that explains why Joe is morally responsible

for his actual choice to evade taxes because in this counterfactual scenario Joe is not

morally responsible for choosing to evade taxes, while in the actual scenario he is

morally responsible for choosing to evade taxes; thus, Joe has an alternative open to him

in which he is not responsible for what in the actual scenario he chooses to do (i.e., he has

an alternative open to him in which he is not blameworthy for what he does).30

Pereboom, though, argues that having an alternative in which the agent is not

blameworthy for what he does is not the right kind of alternative possibility that could

explain why the agent is morally responsible for what he does; Pereboom argues that the

alternative possibility that the objectors are referring to is not a robust alternate

possibility and, so, his Tax Evasion 2 does satisfy the first condition of an FR scenario.

According to Pereboom, in order for an alternative possibility to be robust it must satisfy

the following conditions:

Robustness: For an alternative possibility to be relevant per se to explaining an
agent's moral responsibility for an action it must satisfy the following
characterization: she could have willed something other than what she actually


30 Otsuka, 1998 and, Wyma,1997.









willed such that she understood that by willing it she would thereby have been
precluded from the moral responsibility she actually has for the action.31

So, according to Pereboom, in order for an alternative possibility to be the kind of

possibility that is relevant to explaining an agent's moral responsibility for doing R, the

alternative possibility must satisfy the two following conditions: (i) the agent could have

chosen something, Q, such that Q is different from R AND the choosing of Q would

thereby preclude the agent of the moral responsibility she has for what she actually

chooses, R, and (ii) the agent understood that had she chosen Q, she would thereby be

precluded from the moral responsibility she has for doing R. In Tax Evasion 2, the

alternative possibility that is claimed by the objector to be morally relevant (the

possibility of choosing to reach the requisite level of attentiveness, which would cause

the device to intervene and produce the choice to evade taxes), while it satisfies condition

(i) of Pereboom's definition of 'robustness,' does not satisfy condition (ii). Had Joe

chosen to reach the requisite level of attentiveness to moral reasons, this choice would be

different from his actual choices (Joe actually chooses to not pay attention to moral

reasons and, so, chooses to evade taxes) and choosing to reach the requisite level of

attentiveness would preclude him from the moral responsibility he has for his actual

choices (as explained above, had Joe chosen to reach the requisite level of attentiveness,

the device would have caused him to choose to evade taxes and Joe would not be morally

responsible for this choice). This possibility of choosing to reach the requisite level of

attentiveness, though, according to Pereboom, does not meet condition (ii) above:

But, Joe does not have an alternative possibility available to him that is robust.
First, he does not even believe that if he had achieved the requisite level of
attentiveness he would thereby have been precluded from responsibility for


31 Pereboom, 2005, p. 230.









deciding to evade taxes. For he believes that achieving this level of attentiveness is
compatible with his never refraining from making this decision, or being seriously
inclined so to refrain, and deciding to evade taxes instead. In addition, Joe does not
know enough to understand that voluntarily achieving the requisite attentiveness
would preclude him from responsibility for choosing to evade taxes. True, were he
voluntarily to achieve this attentiveness, the intervention would take place, and he
would not then have been responsible for this choice. Nevertheless, Joe does not
understand that the intervention would then take place, or that as a consequence of
this intervention he would be precluded from responsibility for choosing to evade
taxes. Hence, no robust alternative possibility is available to him. Still, Joe is
morally responsible for deciding to evade taxes.32

Pereboom, then, maintains that Tax Evasion 2 meets the first condition of an FR

scenario and, so, (since it meets the second and third conditions as well) is a

counterexample to PAP. But, we can see that Pereboom's argument for the claim that Joe

has no robust possibilities is dependent upon the correctness of his definition of

'robustness.' Pereboom's argument for the claim that Joe has no robust alternative

possibilities in Tax Evasion 2 is just the following: the alternative possibility open to Joe

does not meet my (Pereboom's) conditions on robustness, and, so, Joe has no robust

alternative possibilities. Specifically, the alternative possibility open to Joe in Tax

Evasion 2 does not meet what I will call Pereboom's epistemic condition on robustness,

condition (ii) of the definition of 'robustness': in order for an alternative possibility to be

robust the agent must have understood that by making a choice other than the one he

actually makes he would thereby be precluded from the moral responsibility he has for

what he actually does.

In the following section, I will argue that Pereboom's epistemic condition on

robustness is not correct. While I will argue that there is some kind of an epistemic

condition that must be included in a definition of 'robustness,' I will argue that


32 Pereboom, 2005, pp. 232-233.






48


Pereboom's epistemic condition demands too much of the agent in order for the

alternative possibility open to him to be relevant to explaining the moral responsibility he

has for what he actually does. If the definition of 'robustness' that I will propose is

correct, the alternative open to Joe in Tax Evasion 2 is a robust alternative possibility. If,

then, the definition of 'robustness' that I will propose is correct, Pereboom's Tax Evasion

2 will not meet the first condition of an FR scenario and, so, will not be a counterexample

to PAP.















THE DEFINITION OF 'ROBUSTNESS'

In this section, I will argue that Pereboom's definition of 'robustness' is not correct,

and given what I take is a more accurate account of 'robustness,' Pereboom's Frankfurt-

style case, Tax Evasion 2, will leave the agent with robust alternative possibilities and,

so, will not be a counterexample to PAP. If Pereboom's Tax Evasion 2 is not a

counterexample to PAP, then the proponent of Frankfurt-style cases will either need to

formulate a Necessary-Condition Frankfurt-style case that does meet all three conditions

of an FR scenario, or will need to formulate a No-Prior-Sign, a Blockage, or an Internal-

Sign example that meets the challenges previously made to such cases. To show that

Pereboom's definition of 'robustness,' specifically his epistemic condition, is not correct,

I will first look at his argument to support his epistemic condition, this argument being

based upon our intuitions about a particular case. I will argue that our intuitions about this

case do not need to be explained by positing Pereboom's epistemic condition on

'robustness,' but, rather, are better explained by the epistemic conditions) that I will

suggest. I will spend the remainder of this essay discussing the conditions) that I

propose, and how given the definition of 'robustness' that I propose, Pereboom's Tax

Evasion 2 is not a counterexample to PAP. Before taking a look at Pereboom's argument

for his epistemic condition on 'robustness,' I will briefly discuss the criteria of adequacy

for (or, what we are looking for in) a definition of 'robustness.'

The term 'robustness' is a technical term that picks out a certain feature (or, a set of

features) of some, but not all, of the counterfactual/alternative possibilities that we









believe are open to agents. We will say that an agent has a robust alternative possibility if

and only if there is a counterfactual sequence open to him that has this certain feature (or

set of features). In order for a definition of 'robustness' to be an adequate definition, it

must spell out the necessary and sufficient conditions for when some agent has a robust

alternative possibility, and the definition must do this in a way that accords with our

intuitions about particular cases. Intuitively, it seems to me that an agent has a robust

alternative possibility in those cases in which the addition or removal of this possibility

affects our judgment of the agent's moral responsibility. It seems to me that our intuitions

concerning when some alternative possibility is robust are guided, then, in the following

way: An agent has a robust alternative possibility to choosing/doing x in those cases in

which the removal of that alternative possibility would change our judgment of him from

morally responsible for choosing/doing x to not morally responsible for choosing/doing x;

and, an alternative possibility to x is robust in those cases in which the addition of that

alternative possibility would change our judgment of an agent from not morally

responsible for choosing/doing x to morally responsible for choosing/doing x.

It is the job of one who proposes a definition of 'robustness' to give the necessary

and sufficient conditions that must be met in order for an agent to have a robust

alternative possibility, and the satisfaction of these conditions should occur when and

only when our intuition, guided in the way outlined above, tells us that the agent has an

alternative open to him that affects our judgment of his moral responsibility. In what

follows, I will argue that Pereboom's argument for his definition of 'robustness' is

flawed, and I will propose a definition of 'robustness' that will do a better job of

explaining our intuitions about an agent's moral responsibility in particular cases.









In arguing for his epistemic condition on robustness, Pereboom gives a case in

which (a) (the agent presumably knows that) there was an alternative open to the agent,

that he could have chosen otherwise, and (b) it is clear that the agent is morally

responsible, and (c) the agent does not understand that if he had chosen otherwise he

would thereby be precluded from the moral responsibility he has for what he does.

Roughly, Pereboom's argument for his epistemic condition on robustness is the

following: in this case, (a) the agent's ability to choose otherwise is not relevant to (b) the

agent's moral responsibility for what he does because (c) the agent does not understand

that by choosing otherwise he would thereby be precluded from the responsibility he has

for what he does:

Robustness also has an epistemic dimension, and it is important that it be made
explicit in the characterization of this notion. Imagine that the only way in which
Jones could have voluntarily avoided deciding to kill Smith is by taking a sip from
his coffee cup prior to making the decision, and this is only because it was
poisoned so that taking a sip would have killed him instantly. Suppose that Jones
does not understand that this action would preclude his deciding to kill, because he
has no idea that the coffee is poisoned. In this situation, Jones could have
voluntarily behaved in such a manner that would have precluded the action for
which he was in fact blameworthy, as a result of which he would have avoided the
moral responsibility he actually has. But whether he could have voluntarily taken
the sip from the coffee cup, not understanding that it would render him blameless in
this way, is irrelevant qua alternative possibility to explaining why he is morally
responsible for deciding to kill. Despite the fact that Jones could have voluntarily
taken a sip from his coffee cup, and doing so would have rendered him not morally
responsible for deciding to kill, this alternative possibility is nevertheless
insufficiently robust to have an important role in grounding the agent's moral
responsibility.1

So, Jones (b) is morally responsible for choosing to kill Smith, despite the fact that

(a) Jones could have voluntarily taken a sip of poisoned coffee, which would thereby

preclude him from the moral responsibility he has for killing Smith (had Jones


1 Pereboom, 2005, p. 230.









voluntarily chosen to take a sip of poisoned coffee, he would die and, so, he would not

kill Smith and, so, he would not be morally responsible for killing Smith). The reason

that (a) Jones's having the ability to choose otherwise to choose to take a sip of coffee -

is irrelevant to explaining (b) Jones's moral responsibility for what he does is because (c)

Jones did not understand that had he chosen otherwise he would be precluded from the

moral responsibility he has for killing Smith. Jones does not know that his coffee is

poisoned and, so, does not understand that by taking a sip he would thereby be precluded

from the moral responsibility he has for his choice to kill Smith. What this suggests,

according to Pereboom, is that in order for an alternative possibility to be relevant to

explaining an agent's moral responsibility, it must be the case that the agent understands

that by choosing otherwise he would thereby be precluded from the responsibility he has

for what he does; in order for an alternative possibility to be robust, it must be the case,

according to Pereboom, that the agent understands that he could choose otherwise and

understands that by choosing otherwise he would thereby be precluded from the moral

responsibility he has for what he actually does.

It seems to me that Pereboom's reasoning described above is flawed in that (c), that

the agent does not understand that by choosing otherwise he would thereby be precluded

from the moral responsibility he has for what he does, is not the only feature of the case

that could be appealed to in order to explain why (a) the agent's ability to choose

otherwise is irrelevant to (b) the agent's moral responsibility. There is another feature of

this case given above, a feature that better explains why (a) the agent's ability to choose

otherwise is irrelevant to (b) the agent's moral responsibility: (d) the agent does not have

an alternative (to what we can assume he is being claimed morally responsible for,









choosing to kill and killing Smith) open to him such that he believes that as a result of

choosing otherwise, the likelihood will be different (than if he did not choose otherwise)

that he choose to kill and kill Smith. I believe that feature (d) of the case more plausibly

explains why (a) Jones's ability to choose otherwise is irrelevant to (b) his moral

responsibility for choosing to kill and killing Smith. It is true that by choosing to take a

sip of coffee, Jones would not choose to kill Smith (and would not kill Smith), but Jones

does not believe that this is the case, nor should we say he has any good reason to believe

so. This explains why Jones's ability to choose otherwise is irrelevant to his moral

responsibility for what he chooses and does. Given this analysis of Pereboom's poisoned

coffee case, I propose the following definition of 'robustness':

Robustness: an agent has a robust possibility, y, to x at tn (where tn is not necessarily
the time x was chosen/done) if and only if (i) had he done y at tn he would not have
done x and, (j) he believes that he could have chosen and done otherwise, y, at tn and,
(g) the agent believes that he can choose/doy at tn such that the likelihood that he
choose/do x is different than it would be if he does not choose/doy at tn.2
So, the reason that Jones's ability to choose to take a sip of coffee is irrelevant to

his moral responsibility for choosing to kill and killing Smith is because Jones does not

believe that taking a sip will change the likelihood that he chooses to kill and kills Smith.

According to the proponent of PAP, then, I am suggesting that unless Jones has an

alternative to killing Smith that meets the criteria given above, Jones is not morally

responsible for killing Smith in the poisoned coffee case.





2 Recall that the reformulated version of PAP is as follows: having a robust alternative possibility to
choosing/doing x is necessary for an agent to be morally responsible for choosing/doing x. With this
definition of 'robustness,' then, what I am suggesting is that the intuition behind PAP really amounts to the
following: an agent is morally responsible for choosing/doing x at tl only if he could have either (a)
chosen/done otherwise at t1 or, (b) chosen/done otherwise, y, at some time prior to tl, where y satisfies my
definition of 'robustness.'









Let's test this definition against our intuition, guided in the way previously

outlined, about Jones's moral responsibility in a slightly modified version of the poisoned

coffee case. In Poisoned Coffee Case 2, Jones chooses to not and does not take a sip of

coffee and then chooses to kill and does kill Smith. Unbeknownst to Jones, though, his

coffee is poisoned and, so, had he taken a sip he would have died instantly. Suppose,

though, that Jones could not have done anything other than choose to not and not take a

sip of coffee and then choose to kill and kill Smith. Suppose that, unbeknownst to Jones,

he has a mental illness such that he always acts on the first course of action that comes to

his mind (perhaps he has some kind of attention deficit disorder that prevents him from

going through an adequate deliberative process of weighing reasons for and against

various courses of action), and in this case the first and only course of action he considers

is to not take a sip of coffee and then go kill Smith. Suppose also that, unbeknownst to

Jones, there is a device in his brain that is indeterministically stimulating his brain in such

a way that the device is indeterministically supplying his brain with the first course of

action that he considers, and, so, as a result of the combination of this device and Jones's

mental illness, Jones cannot do anything other than carry out the course of action

supplied to him by the device. In Poisoned Coffee Case 2, the device supplied Jones with

the choice to not take a sip of coffee and the choice to kill Smith and, so, Jones follows

through with this course of action and does not take a sip of coffee and does kill Smith. It

seems plausible that in Poisoned Coffee Case 2 Jones is not responsible for choosing to

kill Smith.

Now consider Poisoned Coffee Case 3, which is just like Poisoned Coffee Case 2,

except that the device is not present in Poisoned Coffee Case 3. In Poisoned Coffee Case









3 Jones could have taken a sip of poisoned coffee had he chosen to do so, i.e., had this

been the first course of action he considered (in which case, he would not choose to kill

and would not kill Smith, because he would have died instantly). In Poisoned Coffee

Case 3, it is also the case that Jones believes that taking a sip of coffee will make it less

likely that he chooses to kill and kills Smith (i.e., suppose Jones has an alternative

possibility that satisfies my proposed definition of 'robustness'); perhaps Jones believes

this based on past experiences with drinking coffee and making important decisions (such

as the decision to kill someone) and past experience with drinking coffee and shooting -

Jones has previously taken a sip of coffee before practicing his shooting and this has in

the past reduced his accuracy. (It is still the case, though, that Jones, due to his mental

illness, can only act on the first course of action he considers, where in this case the first

course of action he considers is not taking a sip of coffee and then killing Smith). In

Poisoned Coffee Case 3, the case in which Jones has an alternative possibility that

satisfies my proposed definition of 'robustness,' I believe that the intuition will be that

Jones is responsible for choosing to kill Smith whereas he is not responsible for doing

this in Poisoned Coffee Case 2, the case in which he had no such alternative. That Jones

has an alternative possibility such that he believes that by choosing otherwise he will be

less likely to do something something that he knows is wrong, choosing to kill Smith -

affects our judgment about whether or not Jones is morally responsible for what he does.

If I am right that one will have the intuition that Jones is morally responsible for choosing

to kill and killing Smith in Poisoned Coffee Case 3, and I am right that one will have the

intuition that he is not morally responsible for this in Poisoned Coffee Case 2, and the

difference between these two cases is that in Poisoned Coffee Case 2 Jones does not and









in Poisoned Coffee Case 3 Jones does have an alternative open to him that satisfies my

proposed definition of 'robustness,' then the addition of such an alternative to a case

affects our judgment of an agent's moral responsibility in the outlined above (i.e., the

addition of an alternative that satisfies my definition of 'robustness' changes our

judgment from not morally responsible for choosing and doing x, to morally responsible

for choosing and doing x).

One might object that in Poisoned Coffee Case 3, it is still the case that Jones has a

mental illness that makes it the case that he can only consider one course of action at a

time. One might claim that, despite having a robust alternative possibility, since Jones has

this mental illness, it is unjustifiable to hold him morally responsible for what he does.

Since Jones is not morally responsible in the case in which he is given an alternative

possibility that satisfies my definition of 'robustness,' my definition of 'robustness' is not

correct.

I believe that this objection is mistaken, and that there is no reason that an agent

with a mental illness cannot be held morally responsible for choices and actions when

that agent has an alternate possibility to what he does and the agent meets a certain

epistemic condition concerning the alternatives) open to him. If an agent could have

chosen otherwise and believes that he could do (could have done) so such that choosing

otherwise will (would have) make (made) it less likely that he kill another person, it is

reasonable to hold him responsible for killing that person. In Poisoned Coffee Case 3,

even though Jones has a mental illness, this mental illness does not prevent Jones from

choosing otherwise; in Poisoned Coffee Case 3, Jones can choose otherwise than to not

take a sip of coffee and to kill Smith. It could have been the case that the first course of









action Jones considered was to take a sip of coffee and then not kill Smith, in which case

Jones would have taken a sip of poisoned coffee and then died (and, so, not killed Smith).

It seems, then, that an agent that has a mental illness will only be exempt from moral

responsibility for doing something because of that mental illness, if it is the case that his

mental illness either prevents him from doing anything other than what he does, or if it

prevents the agent from having the belief associated with condition (g) of my definition

of 'robustness.' The mental illness Jones has in Poisoned Coffee Case 3 neither prevents

him from doing anything other than what he does, nor does it prevent him from having

the belief associated with condition (g) of my definition of 'robustness' and, so, Jones's

mental illness will not exempt him from moral responsibility for choosing to kill and

killing Smith.

So, in Poisoned Coffee Case 3, Jones is morally responsible for choosing to not

take a sip of coffee, choosing to kill Smith, and killing Smith, because he has a robust

alternative possibility to choosing and doing these things: he could have chosen to take a

sip of coffee such that (j) if he chose to take a sip of coffee he would not have chosen to

not take a sip of coffee, chosen to kill Smith, or killed Smith, and (g) he believed that

choosing to take a sip of coffee would make it less likely that he choose to kill Smith and

kill Smith. Notice, though, that according to Pereboom's definition of 'robustness' Jones

would not have a robust alternate possibility to choosing to kill and killing Smith in

Poisoned Coffee Case 3. Recall that according to Pereboom, in order to have a robust

alternative possibility to what an agent does, that agent must be able to choose otherwise

such that he understands that by choosing otherwise he would be precluded from the

responsibility he has for what he actually does. So, in Poisoned Coffee Case 3, in order









for Jones's ability to take a sip of coffee to be a robust alternate possibility to choosing to

kill and killing Smith, according to Pereboom, Jones would need to understand that by

taking a sip of coffee he would be precluded from the moral responsibility he has for

choosing to kill and killing Smith. Jones does not, though, understand this in Poisoned

Coffee Case 3, because Jones believes only that the likelihood of choosing to kill and

killing Smith is reduced if he takes a sip of coffee. In Poisoned Coffee Case 3 Jones

believes that if he takes a sip of coffee, it is still possible he will choose to kill and kill

Smith. So, since Jones does not understand that by taking a sip of coffee he will be

precluded from the moral responsibility he has for choosing to kill and killing Smith,

Jones does not, according to Pereboom's definition of 'robustness,' have a robust

alternative possibility. If, then, one agrees that Jones has an alternate possibility that is

relevant to his moral responsibility for choosing to kill and for killing Smith in Poisoned

Coffee Case 3, one cannot accept Pereboom's definition of 'robustness.' I believe that

one will have the intuition that Jones has an alternative open to him in Poisoned Coffee

Case 3 that is relevant to his moral responsibility for choosing to kill and for killing

Smith and, so, one will see that my proposed definition of 'robustness' is more accurate

than Pereboom's definition of 'robustness.'

Given the arguments presented above for the definition of 'robustness' I am

proposing, we can see how Pereboom's Tax Evasion 2 Frankfurt-style case leaves the

agent with a robust alternative possibility and, so, is not a counterexample to PAP

because it does not satisfy the first condition on an FR scenario. As given, though,

Pereboom's Tax Evasion 2 is underdescribed and, so, I will make a stipulation about the

agent's psychology and show that, given this stipulation, the agent has a robust









alternative possibility. Recall that in Tax Evasion 2, Joe chooses to not pay attention to

his moral reasons for not evading taxes and, so, chooses to and does evade his taxes. The

only way that Joe could have not evaded his taxes was by choosing to pay sufficient

attention to his moral reasons for not evading. Achieving this requisite level of

attentiveness to his moral reasons is necessary but not sufficient for his choosing to not

evade taxes; had he reached this requisite level of attentiveness (without the device being

in place) he could either choose to evade or fail to choose to evade. It is the case, though,

that a device is in place such that if Joe reaches the requisite level of attentiveness to his

moral reasons for not evading, the device will intervene and bring about the choice to

evade and Joe would evade his taxes. Joe actually chooses to not pay attention to his

moral reasons and, so, chooses to evade and does evade, even though Joe could have

chosen to pay attention to his moral reasons for not evading taxes. Recall that at this point

Pereboom argues that this alternative open to Joe the ability to choose to pay sufficient

attention to moral reasons is not a robust alternative to what he actually does. Given the

following reasonable supposition I will make about Joe's psychology, if the definition of

'robustness' I have proposed is correct, then Joe's available option of choosing to pay

sufficient attention to his moral reasons for not evading taxes is a robust alternative

possibility to his choosing to not pay attention to moral reasons: Joe believes that he

could have chosen otherwise chosen to pay sufficient attention to his moral reasons -

such that if he chose to pay sufficient attention to his moral reasons, he would have been

less likely to evade his taxes. I will call this version of Tax Evasion 2, in which Jones has

a robust alternate possibility to choosing to not pay attention to moral reasons, Tax

Evasion 21. In Tax Evasion 2' Joe's ability to choose to pay sufficient attention to his









moral reasons is an alternative to what he actually chooses, the availability of which is

relevant to explaining his moral responsibility for choosing to not pay attention to his

moral reasons.

So, in Tax Evasion 2, the defender of PAP can argue that Joe is morally responsible

for choosing to not pay attention to his moral reasons (on the supposition that Tax

Evasion 2 is just like Tax Evasion 211'), because he had a robust alternate possibility to

doing so. Perhaps, though, this is all Joe is morally responsible for in Tax Evasion 2;

perhaps Joe is not morally responsible for choosing to evade and for evading his taxes.

After reading Tax Evasion 2, I believe one has the intuition that Joe is responsible for

something, and the most obvious candidate for this is the action he performs evading

his taxes. It seems to me, though, that it is open to the proponent of PAP to claim that this

is not what Joe is morally responsible for in Tax Evasion 2, and that he is only morally

responsible for his choice to not pay attention to moral reasons, in which case Tax

Evasion 2 is not a counterexample to PAP because Joe does have a robust alternative

possibility to what he is morally responsible for, choosing to not pay attention to his

moral reasons.

What, though, would be the case if Joe does not have such a belief, if Joe does not

believe that choosing to pay sufficient attention to his moral reasons will make it less

likely that he choose to evade taxes? It seems very hard to imagine that a person could

not have any beliefs concerning how the consideration of moral reasons will affect one's

choices and behavior. Consider case Tax Evasion 21', in which we suppose that Joe does

not have any such beliefs that he does not believe that paying sufficient attention to his

moral reason will in any way change the likelihood that he will evade taxes would we









still want to claim that Joe is morally responsible for anything he does? If we would want

to claim that Joe is morally responsible for choosing to not pay attention to his moral

reasons and for choosing to evade and evading his taxes, despite the fact that he does not,

according the definition of 'robustness' I have proposed, have a robust alternate

possibility to choosing/doing as such, then either my definition of 'robustness' or PAP (or

both) would be false. I think, though, that both my definition of 'robustness' and PAP can

be upheld in the case at hand, the case in which Joe's alternative possibility the ability

to choose to pay sufficient attention to his moral reasons does not satisfy my definition

of 'robustness.'

Tax Evasion 2~' is just like the original poisoned coffee case given by Pereboom,

even though it is very difficult for us to imagine that any person could hold the choice of

whether or not to consider moral reasons as on par (with respect to the effect that such

consideration will have on one's future choices and actions) with the consideration of

whether or not to take a sip of coffee. If we can picture that, with respect to the effect that

the agent believes that the consideration of whether or not to x will have on his future

choices and actions, Joe believes that the choice of whether or not to pay sufficient

attention to his moral reasons is on par with the choice of whether or not to take a sip of

coffee (this is what Joe believes in Tax Evasion 2~1'), I believe the intuition that Joe is

morally responsible for choosing to not pay sufficient attention to his moral reasons,

choosing to evade, and evading taxes will fade away. In what follows, I will argue that in

Tax Evasion 2-~', Joe is neither responsible for choosing to not pay sufficient attention to

his moral reasons, nor for choosing to evade and evading his taxes.









Consider Tax Evasion 2* which is set up just like Tax Evasion 2, except that Joe*

will choose to evade and will evade taxes only unless he chooses to and takes a sip of

coffee; if Joe* does choose to and take a sip of coffee, under normal circumstance, he

could then choose whether or not to evade taxes. Suppose, though, that there is a device

such that if Joe* will choose to take a sip of coffee, the device will prevent him from

making this choice and will bring about the choice to not take a sip. Joe*, then, cannot

choose to do anything but not take a sip of coffee and evade his taxes. Suppose further

that Joe* does not believe that the decision of whether or not to take a sip of coffee will

in any way affect the likelihood of making any future choice or performing any future

action. Suppose that for his own reasons, Joe* does not take a sip of coffee and chooses

to evade and does evade his taxes. In Tax Evasion 2*, I cannot see what could justify the

claim that he is morally responsible for not taking a sip of coffee. Taking or not taking a

sip of coffee is not normally the kind of activity that one is morally responsible for,

unless of course, the agent has some kind of a belief associated with the effects of taking

or not taking a sip on his behavior (and, Joe* does not have any beliefs like this in Tax

Evasion 2*). If Joe* is not morally responsible for not taking a sip of coffee in Tax

Evasion 2*, and if the decision of whether or not to take a sip of coffee is considered by

the agent as on par (with respect to the affect on one's future choices and behavior) with

the decision of whether or not to consider moral reasons, then Joe is not morally

responsible for choosing to not pay attention to his moral reasons in Tax Evasion 211, the

case in which he does not believe that paying attention to his moral reasons will affect the

likelihood that he evade taxes. So, if Joe did not have a robust, according to my definition

of 'robust,' alternative to choosing to not pay attention to his moral reasons (as he would









not in Tax Evasion 2 '), he would not be morally responsible for choosing to not pay

attention to his moral reasons.

A defender of Frankfurt-style cases might object at this point by claiming that the

previous argument ha suggested a Frankfurt-style case that satisfies all three conditions of

an FR scenario: Tax Evasion 2-'. In Tax Evasion 2 ', Joe does not have a robust

alternate possibility to choosing to evade and evading his taxes (given the placement of

the device Joe cannot do anything but choose to evade and evade taxes), what makes it

the case that he cannot do otherwise does not bring about his actual choice/action (in Tax

Evasion 2 ', it is still the case that Joe chooses to not pay attention to his moral reasons

and, so, chooses to evade and evades his taxes, and he chooses/does so without the

intervention of the device), and it is plausible to claim that Joe is morally responsible for

choosing to evade and for evading taxes since he did so indeterministically (just like Joe

did so indeterministically in Tax Evasion 2).

I do not, though, believe that Tax Evasion 2 is a case in which the indeterminacy

in Joe's choice to evade taxes is the right kind of indeterminacy to allow for moral

responsibility for so choosing and doing. We must remember what role Tax Evasion 2 (or

any Necessary-Condition Frankfurt-style case) is playing in the debate. Recall that

Pereboom's Tax Evasion 2, and any Necessary-Condition example for that matter (for

example, Tax Evasion 2 1), is a response on behalf of the proponent of Frankfurt-style

cases to the second horn of the dilemma defense of PAP. The challenge set forth in the

second horn of the dilemma is to give a case in which the agent's choice is in some

relevant way undetermined and yet the agent has no robust alternative possibilities. The

whole motivation for challenging the proponent of Frankfurt-style cases in this way









comes from the Incompatibilist who claims that in order for an agent to be morally

responsible for what he chooses or does, his choice must be in some relevant way

undetermined. The Incompatibilist who makes such a challenge, though, will not be

satisfied with the presentation of an agent whose choice is undetermined in just any way.

For example, consider a case in which an agent's choice is undetermined in just the

following way: a particular agent who is 21 years old will at age 24 choose to go to

medical school only unless at age 22 he chooses that red is his favorite color. Suppose

that the agent does choose at age 24 to go to medical school. The Incompatibilist, though,

will not accept that the kind of indeterminacy present in this case is the right kind of

indeterminacy to allow for moral responsibility, and I believe that the Incompatibilist will

not accept this kind of indeterminacy for the following reason: the agent has no reason to

believe that there is any connection between his choice at age 22 of a favorite color and

his choice at age 24 of what career to pursue. In order for the indeterminacy present in a

case to be the kind of indeterminacy required for an agent to be morally responsible for

some choice C, it must be the case that whatever explains the indeterminacy (here, the

ability to choose or not choose red as his favorite color) is in some way relevant to the

undetermined choice for which the agent is claimed to be morally responsible (here, the

choice to go to medical school). Contrast the kind of indeterminacy present in the case

just given with the indeterminacy present in Pereboom's Tax Evasion 2 case. In

Pereboom's case, what explains the fact that Joe's choice concerning whether to evade

taxes is undetermined is that he has the option of (voluntarily) paying attention to moral

reasons for and against doing so. The reason, I believe, that this kind of indeterminacy

will satisfy the Incompatibilist proponent of the dilemma defense of PAP is because the









weighing of moral reasons for and against doing something is taken to be (assumed to be)

relevant to the undetermined choice Joe is claimed to be morally responsible for,

choosing to evade taxes. Although the kind of indeterminacy present in Tax Evasion 2 is

the right kind to allow for the agent to be morally responsible for what he does, it is not

necessarily the case that this kind of indeterminacy suffices for moral responsibility and,

so, one can still argue that Joe is not morally responsible for choosing to evade and for

evading taxes in Pereboom's Tax Evasion 2. (It is still the case, then, that it is open to the

proponent of PAP to argue that Joe is not morally responsible for choosing to evade and

for evading taxes in Tax Evasion 2, and that what he is morally responsible for is

choosing to not pay sufficient attention to moral reasons, something he has a robust

alternative to choosing/doing.).

The indeterminacy in Tax Evasion 2 1', the case in which Joe has no belief

concerning the effect, on his future choices and actions, of paying sufficient attention to

his moral reasons is like the indeterminacy in the case of the agent who will choose at age

24 to go to medical school only unless he chooses at age 22 that red is has favorite color.

In the case in which Joe has no belief concerning the effect, on his future choices and

actions, of paying sufficient attention to his moral reasons, what explains that his choice

to evade taxes is undetermined is that he could fail to so choose if he chooses to pay

sufficient attention (if the device were not in place). But, if his belief concerning the

effect, on his future choices and actions, of considering moral reasons is on par with his

belief concerning the effect, on his future choices and actions, of considering whether or

not to take a sip of coffee, then the indeterminacy present in Tax Evasion 2- 'is just like

the indeterminacy present in the case of the student who will choose at age 24 to go to









medical school only unless he chooses at age 22 that red is his favorite color. If it is the

case, then, that Joe does not believe that paying attention to moral reasons will affect the

likelihood of choosing to evade and evading taxes, there will not be the right kind of

indeterminacy present in the case to allow for moral responsibility. Tax Evasion 2~-' (i.e.,

the case in which Joe does not have an alternative open to him that satisfies my proposed

definition of 'robustness'), then, is not a case that is a response to the dilemma defense of

PAP and, so, will not help the proponent of Frankfurt-style cases argue his way out of the

dilemma.

In Tax Evasion 2~', then, Joe is neither morally responsible for choosing to not pay

sufficient attention to moral reasons, nor is he morally responsible for choosing to evade

and for evading taxes. That Joe has no robust alternative in Tax Evasion 2~' to so

choosing/doing according to my proposed definition of 'robustness,' then, is neither

problematic for my proposed definition of 'robustness,' nor for PAP.

On the one hand, then, we can see that if Joe has an alternative possibility open to

him that satisfies my proposed definition of 'robustness,' (the version called Tax Evasion

2'W) Tax Evasion 2 will not satisfy the first condition of an FR scenario (the agent must

have no robust alternative possibilities to what he actually chooses to do) and, so, will not

be a counterexample to PAP. On the other hand, if the alternative possibility open to Joe

does not meet my proposed definition of 'robustness,' (the version called Tax Evasion

2~-') Tax Evasion 2 will not satisfy the third condition of an FR scenario (the agent must

be morally responsible for his actual choice) and, so, will not be a counterexample to

PAP. I conclude, then, that Pereboom's Frankfurt-style case does not satisfy all three

conditions of an FR scenario and, so, is not a counterexample to PAP.









The discussion of section three leads me to conclude that the proponents of

Frankfurt-style cases have yet to give a case that is a counterexample to PAP because

they have yet to give a case that satisfies all three conditions of an FR scenario. It is open

to the proponent of Frankfurt-style cases to support Pereboom's definition of 'robustness'

or some other definition that shows that Tax Evasion 2 does not include such alternatives

open to the agent. The proponent of Frankfurt-style cases could also support one of the

three other styles of cases the No-Prior-Sign example, the Blockage example, or the

Internal-Sign example by showing that something is mistaken with the criticisms of

such cases presented in the second section of this essay. It is also open to the proponent

of Frankfurt-style cases that finds the definition of 'robustness' I have here proposed

attractive to give a Necessary-Condition example in which the agent's choice is

undetermined in the right kind of way and yet the agent has no robust alternative

possibilities given the definition of 'robustness' that I have proposed.

In this essay I have argued that the proponent of Frankfurt-style cases has yet to

give a case that satisfies all three conditions of an FR scenario and, so, has yet to give a

case that is a counterexample to PAP. I have argued this by discussion a dilemma that is

posed to the proponent of Frankfurt-style cases, the dilemma defense of PAP. I have

argued that the proponent of Frankfurt-style cases has not adequately responded to the

first horn of the dilemma; if it is supposed that the prior sign is sufficient for the agent's

choice in a Frankfurt-style case, the use of a Frankfurt-style case to support the

Compatibilist's conclusion is unnecessary and, so, this first horn is problematic if the

Compatibilist claims that Frankfurt-style cases go some way to advancing the debate,

between Compatibilists and Incompatibilists, concerning moral responsibility. Although









the problem with the supposition that the prior sign suffices for the agent's choice has, up

until now, been posed as a problem of begging the question against the Incompatibilist, I

have argued that this is not the real problem with the supposition but that, regardless, the

proponent of Frankfurt-style cases (Fischer) has not adequately responded to the real

problem posed by the supposition of the first horn of the dilemma defense of PAP.

I have also argued that the proponent of Frankfurt-style cases has not adequately

responded to the second horn of the dilemma defense of PAP. I have done this by

discussing the four types of cases that proponents of Frankfurt-style cases have used to

respond to this horn of the dilemma, showing that neither strategy succeeds in satisfying

all three conditions of an FR scenario and, so, neither strategy succeeds in giving a

counterexample to PAP. I have paid special attention to the Necessary-Condition

strategy, used by Derk Pereboom, showing that the proponent of PAP can use my

proposed definition of 'robustness' to argue that Pereboom's Tax Evasion 2 is not a

counterexample to PAP. I have gone some way to showing that my proposed definition

of 'robustness' is more accurate than Pereboom's definition of 'robustness,' and, since

little attention has been paid in the literature to attempting to define this term, I have

(hopefully) helped us to understand more about the Principle of Alternate Possibilities

and when it is satisfied.















LIST OF REFERENCES


John Martin Fischer, The Metaphysics of Free Will: An Essay on Control (Cambridge,
MA.: Blackwell, 1994).

John Martin Fischer, "Recent Work on Moral Responsibility," Ethics 110 (1999): 93-139.

John Martin Fischer, "Frankfurt-type Examples and Semi-Compatibilism," in The Oxford
Handbook of Free Will, ed. Robert Kane, ed. (New York: Oxford University Press,
2002), 281-308.

John Martin Fischer, "Free Will and Moral Responsibility," in The Oxford Handbook of
Ethical Theory, ed. David Copp (New York: Oxford University Press, 2006), 321-
354.

Harry G. Frankfurt, "Alternate Possibilities and Moral Responsibility," Journal of
Philosophy 66 (1969): 829-839.

Harry G. Frankfurt, "Freedom of the Will and the Concept of a Person," Journal of
Philosophy 68 (1971): 5-20.

Stewart Goetz, "Stumping for Widerker," Faith and Philosophy 16 (1999): 83-89.

Stewart Goetz, "Frankfurt-style Counterexamples and Begging the Question," in Midwest
Studies in Philosophy Volume XXIX: Free Will and Moral Responsibility, eds. Peter
A. French and Howard K. Wettstein, guest ed. John Martin Fischer (Boston:
Blackwell Publishing, 2005), 83-105.

David Hunt, "Moral Responsibility and Unavoidable Action," Philosophical Studies 97
(2000): 195-227.

Robert Kane, Free Will and Values (Albany: State University of New York Press, 1985).

Robert Kane, The Significance of Free Will (New York: Oxford University Press, 1996).

Robert Kane, "Responses to Bernard Berofsky, John Martin Fischer and Galen
Strawson," Philosophy and Phenomenological Research 60 (2000): 157-167.

Robert Kane, "Responsibility, Indeterminism and Frankfurt-style Cases: A Reply to Mele
and Robb," in Moral Responsibility and Alternate Possibilities, eds. David
Widerker and Michael McKenna (Burlington, VT: Ashgate, 2003), 91-105.






70


Michael McKenna, "Alternative Possibilities and the Failure of the Counterexample
Strategy," Journal of Social Philosophy 28 (1997): 71-85.

Alfred R. Mele and David Robb, "Rescuing Frankfurt-Style Cases," The Philosophical
Review 107 (1998): 97-112.

Michael Otsuka, "Incompatibilism and the Avoidability of Blame," Ethics 108 (1998):
685-701.

Derk Pereboom, Living Without Free Will (New York: Cambridge University Press,
2001).

Derk Pereboom, "Defending Hard Incompatibilism" in Midwest Studies in Philosophy
Volume XAXX: Free Will and Moral Responsibility, eds. Peter A. French and
Howard K. Wettstein, guest ed. John Martin Fischer (Boston: Blackwell
Publishing, 2005), 228-247.

P.F. Strawson, "Freedom and Resentment," in Free Will, ed. Gary Watson (New York:
Oxford University Press, 1982), 59-80.

Eleonore Stump, "Alternative Possibilities and Moral Responsibility: The Flicker of
Freedom," The Journal of Ethics 3 (1999a): 299-324.

Eleonore Stump, "Dust, Determinism, and Frankfurt: A Reply to Goetz," Faith and
Philosophy 16 (1999b): 413-422.

Eleonore Stump, "Moral Responsibility without Alternative Possibilities," in Moral
Responsibility and Alternative Possibilities, eds. David Widerker and Michael
McKenna (Burlington, VT: Ashgate, 2003), 139-158.

Peter van Inwagen, An Essay on Free Will (New York: Oxford University Press, 1983).

David Widerker, "Libertarianism and Frankfurt's Attack on the Principle of Alternate
Possibilities," Philosophical Review 104 (1995): 247-261.

David Widerker, "Frankfurt's Attack on the Principle of Alternative Possibilities: A
Further Look," Philosophical Perspectives 14 (2000): 181-201.

David Widerker, "Responsibility and Frankfurt-type Examples," in The Oxford
Handbook of Free Will, ed. Robert Kane (New York: Oxford University Press,
2002) 323-336.

David Widerker and Michael McKenna, eds., Moral Responsibility and Alternative
Possibilities, (Burlington, VT: Ashgate, 2003).

Keith Wyma, "Moral Responsibility and Leeway for Action," American Philosophical
Quarterly 34 (1997): 57-70.















BIOGRAPHICAL SKETCH

Jaclyn Genchi was born in Palm Beach Gardens, Florida, and went to elementary,

junior high, and high school in the surrounding areas. She attended Suncoast Community

High School, where she received an International Baccalaureate degree. Jaclyn went to

Elon College, in Elon, North Carolina, for her first year of undergraduate education, and

then transferred to the University of Florida to complete a Bachelor of Arts in philosophy

in May 2004. After graduating with her Master of Arts in philosophy in August 2006, she

will start law school at Duke Law as a IL in the graduating class of 2009.