Free Will and Indeterminism: Robert Kane’s Libertarianism
The consensus among free will theorists is that an agent can will freely to f without presently being able to form another volition. Frankfurt cases have helped to secure this agreement.1 It is still an open question, however, whether someone could be willing freely to f if there was nothing that she could have done to keep from forming that volition, her character having been determined
.Robert Kane believes that to be exercising a free will, an agent must have had an alternative at various times in her life, if not the present moment. Thus, he maintains that free will entails indeterminism: in order to will freely, an agent must have taken some decisions for which there were no sufficient conditions.2 That is, she must have had an alternative to forming any character trait that determines her decisions. His philosophical task, then, is to show that an act may be free, in the sense that it is something for which its agent is morally responsible, even though it was not guaranteed by antecedent mental processes. He must explain how the taking of a decision could be free while not the necessary outcome of the reasoning from which it issues.
One should think, however, that I am responsible for choosing, e.g., to support a certain cause only if that (mental) act is secured by the exercise of my reasoning ability. Moreover, having weighed the cause’s pros and cons as I did, unless I was rationally bound to decide in its favor the outcome of this reasoning process seems inexplicable and, thus, not something for which I should be held accountable: praised or blamed. My decision must be rational if praiseworthy or blameworthy, but (unless I am in a situation like that of Buridan’s ass) how could it be rational if the reasons motivating it do not ‘rule out’ the opposite choice? A deliberative mechanism that could arrive at more than one conclusion upon processing a given set of considerations appears erratic. (Think here of Peter Van Inwagen’s reluctant thief, alternating between deciding to steal and deciding against stealing, each choice the culmination of the same sequence of neural events unfolding according to the same laws.3) Its exercise, thus, would fail to insure a rational result redounding to my credit or discredit. Kane, therefore, faces a dilemma: either some mental actions are undetermined, in which cases the control and rationality requirements of free agency are not satisfied, or a free agent’s decision is always determined and explicable in terms of reasons that render irrational all but one course of action, in which case the alternative rational possibilities requirement of libertarian free will cannot be met.4 Alternatively, Kane must respond to the following chain argument:
In this paper, I examine Professor Kane’s most recent response to the above claim: that indeterminism would make for "lucky," inexplicable mental actions rather than freely willed decisions (known in the free will literature as the "Mind" objection). By developing the above dilemma, I shall challenge his account of indeterministic mental action as well as critique his notion of a "value experiment," which is supposed to help account for the rationality of undetermined choices.
Kane’s Response
Unlike many libertarians,6 Kane does not hold that all freely willed actions must be physically and rationally undetermined: have reasonable alternatives that could have been performed given the same antecedent events and laws of nature. He allows that responsible behavior often issues from a fully formed will: "character and motives" that necessitate (in both senses) a certain course of conduct.7 In such a case, acting ‘out of character’ would be a matter of luck. Thus, an agent whose will is fully formed cannot act freely without lacking alternative (rational) possibilities, that is, without being determined to act as she does. Her previous performance of undetermined acts that formed her will, however, is required for the conduct such a will yields to be free. If she never had had the possibility of developing a different character (say, if she had grown up in a Skinnerian society, where irreversible programming is the norm) she could not be morally responsible for its manifestations later in life.8 Kane calls the putatively undetermined actions involved in character development "self-forming actions" (SFAs). An SFA must be performed under conditions that allow for "plural rationality": its agent must be able to do (at least) two things, each one of which is supported by good reasons.9 Thus, Kane’s thesis may be restated as follows: a free will requires the performance of some SFAs.
Of SFAs, Kane says further that they are the outcome of the resolution of an "inner conflict," an agent’s struggle with competing desires.10 Such a conflict could involve deciding between business and pleasure, duty and self-interest, or long-range and short-term goals. To perform an SFA is to overcome the resistance provided by one competing desire so as to attempt to secure the object of the other. Thus, it requires an effort of will- specifically, an exercise of one’s moral or practical reasoning skills- to determine which one of the competing desires ought to become one’s motive, that is, reason for acting.11 Being caught up in such a trial, an agent will be uncertain as to how she will act; only her efforts at resolving the crisis can make clear what is to be done.12 The psychological discomfort inevitably accompanying indecision would be the price for failing to achieve closure. Thus, a conflict of this type yields an impetus to the performance of an SFA.
Physiologically speaking, the indecisiveness preceding an SFA is supposed to be a function of neural paths, transmitting data in the form of electrochemical impulses, being connected. Such connectivity allows the impulses coursing through one circuit to effect the transmission of data through another.13 In the case of conflicting desires, the impulses traversing one path are prevented from reaching the "activation threshold" that would be the choice to act upon one of the desires due to the influence of activity in a connected circuit, which is the attempt to realize the activation threshold that would be choice to act upon the other desire, inhibited itself by the former set of events.14 In other words, sans the interference of the other transmission, each data stream would produce the decision to seek a certain object of desire. As things stand, there is only irresoluteness. Breaking the impasse would mean that a successful effort had been mounted to halt one of the attempts to achieve activation threshold, allowing the other desire to become a motive.
How can an SFA and its alternate both be rational? If I eventually choose to make one competing desire my motive, how could the decision to act from the other have also been explicable? To suppose that it is not possible for an SFA to be rationally undetermined, according to Kane, is to hold the implausible thesis that all good reasons for acting must be conclusive: rule out alternatives as irrational.15 An SFA’s agent was faced with conflicting desires neither one of whose objects was more appealing than the other’s: her unformed will guaranteed as much. (If she had not been, so that one of the desires (along with her values and beliefs) had clearly dictated what rationally ought to be done, the conditions for performing an SFA would not have been met.) Thus, it would have been consistent with the demands of rationality for her to have made either one of the desires her motive. That is, she had a "good (but not decisive or conclusive)" reason to pursue the object she eventually sought as well as the object her shunning of which resolved the conflict in her will.16 Her choice, according to Kane, was "the initiation of a ‘value experiment’": a decision to embark upon a certain course of action sans conclusive reasons for doing so, being willing to be held accountable for one’s intended conduct and its consequences even while awaiting its justification.17 Given the multiplicity of rational possibilities confronting her, an SFA’s agent must conduct a value experiment to determine how to settle the conflict between her desires. (Kane compares her to a novelist deciding the fate of one of his characters.18) A successful outcome will mean that the character trait her resolution has engendered fosters the realization of her idea of the good life.19
Finally, it should not be supposed that the effort to perform an SFA, a "self-forming willing," must itself be freely willed, as this move would invite the very regress that positing SFAs is meant to avert, its purpose being to establish ultimate responsibility. The choice that is the outcome of the effort to resolve a conflicted will is freely willed, but not the effort itself.
20 An SFA’s agent is supposed to lack a self/character to which responsibility could be attributed. The performance of SFAs is required to develop a free will. Thus, it is open to Kane to treat the question 'Was the effort behind the performance of an SFA freely willed?' as a category mistake, insisting that this question should only be raised in connection with SFAs themselves and the actions of an agent whose will is fully formed. That is, he could plausibly maintain that it really doesn't matter how the motivational conflict and resulting effort of the will preceding an SFA arises. The important thing is that its outcome is the undetermined effect of that effort. Being forced to decide between two mutually exclusive options is consistent with being responsible for willing to exercise one rather than the other: we manage to develop free wills despite being "thrown into the world," to borrow Heidegger's phrase.The performance of an SFA, then, appears neither lucky nor inexplicable. Its agent’s efforts secured the outcome of the conflict preceding it (selection of its motive); her reasons, although inconclusive, render it understandable, whether or not it is justified being contingent upon ‘how things turn out’.
Two Dilemmas
Although Professor Kane’s account is the best-developed position on free will in contemporary metaphysics, I believe that it fails to demonstrate how an SFA would be both controlled and rational. Kane, as I see it, is faced with two dilemmas, which I shall elaborate in the remaining pages. To begin with, drawing on recent work in neurobiology, he assumes that the neural events leading up to the performance of an SFA leave it undetermined.21 But it is not clear to me that the outcome of a conflict between neural processes is anything more than unforeseeable, delayed but inevitable. After all, not all scientists are ‘sold on’ the idea of quantum indeterminacies.22 Moreover, even if they were to occur, they would not entail indeterminacy at the level of neuronal events, let alone in the operation of cognitive mechanisms. Instead of saying that either current could have reached activation threshold, why not maintain that the agent’s effort overdetermined the result? Alternatively, we could see the effort of will preceding an SFA as the deterministic exercise of an ability designed to resolve neural conflicts.
Regarding the first possibility, it could be maintained that the discomfort of the subject of a neural conflict summons forth an effort of will that merely accelerates the process leading to its inevitable resolution. Sans such an attempt, had the conflict’s subject been more patient, more willing to persist in a state of irresolution, the same choice would have been made, albeit at a later date. Indeed, one often seems content to let a conflict amongst one’s desires resolve itself, eventually taking a decision despite a lack of intellectual effort. It is not clear that an SFA’s agent could not have done likewise. And if she could have achieved the same result by doing nothing, the claim that she is morally responsible, because her effort was required to break the impasse in her will, would no longer hold.
On the other hand, suppose that an intellectual effort is needed to resolve a conflicted will. Why should we think that the exertion called for here would not yield a determined result? It may be thought of as the exercising of an ability, governed by deterministic laws: any given desire conflict having only one possible outcome once the neural mechanism underlying this ability becomes operative. (To make a "Taoist effort," which Kane believes is effective in resolving a conflicted will,23 one would need to learn to relax in the face of irresolution, that is, develop an ability.) Natural selection, it seems, would facilitate the development of a mechanism to restore the well being of an organism whose neural circuitry allows it to fall into a state of thermodynamic disequilibrium. However, it does not seem that a mechanism designed to allow one of two mutually interfering neural circuits to achieve activation threshold could do so without making it impossible for one of the circuits to realize this end, that is, by eliminating its activation potential. The conflict would seem to be unsettled until the mechanism acts to reduce the comparative amperage of the current travelling along one of the circuits. Thereafter, there would be no possibility of the correlated desire becoming a motive. Thus, an agent, who has made an effort of will sufficient to resolve a conflict between her desires, henceforth appears incapable of choosing either one as her motive. If, contrarily, the circuit that failed to reach activation threshold really could have done so (without there having been an alteration in the past or the laws of nature), one has to wonder about the role of said mechanism, how it works. It does not seem to make a difference in the selection of the motive for an SFA. In a word, it appears to lack control over the process.
Perhaps, though, an indeterminate effort could resolve a conflict of the will. The operation of the mechanism involved therein may not guarantee its outcome while still playing some role in how the matter is settled. In that case, we must revisit, I believe, the issue of whether or not an SFA’s agent is morally responsible. For an effort that did not prove determinative would not appear to redound to its agent’s credit. Consider an analog. As a battle rages in a civil war, the federal army’s general sends in her crack regiment, which she had been holding in reserve. These troops display their usual valor, but the battle’s outcome ‘hangs in the balance’ for several hours following their entry into the fighting. After the rebel’s resistance is finally broken, observers on both sides agree that the battle could have gone either way right up until its final stages. Under the circumstances, then, it would be inappropriate to credit the general for having secured the victory. After all, her deployment of the regiment did not prove decisive; it failed to secure its intended result- making it impossible for the rebels to continue fighting. It is not yet clear if anyone in particular should be singled out for praise, though the winning side as a whole merits acclaim.
Things appear no different with an SFA’s agent, as described by Kane. Her effort preceded the resolution of her conflict; but, since it did not guarantee its outcome, leaving open the possibility of the alternative, it is hard to see why she should be held accountable. If she decided to do ‘the right thing’, it is not because her effort disinclined her to act wrongly, for that remained an option right up until the moment of choice. It seems that nothing made her decide to act morally. (Indeed since her decision is not a manifestation of her character or abilities, which are as yet unformed, why should it not be regarded as morally equivalent to a toddler’s doings?) The question, then, is what is the connection between her effort and the ensuing choice that engenders moral responsibility on her part? Kane says that through her exertions she has "overcome temptation," but that is hard to reconcile with her giving in to it being a possibility in spite of her effort.24 How can she be praiseworthy if she has not presently eliminated her desire to do wrong?
Kane adduces two cases supporting the view that an agent may be morally responsible for an undetermined act that she performs. In one of them, an assassin’s bullet that actually found its mark might have gone astray had her arm twitched as she was firing her rifle due to undetermined events in her central nervous system. Despite the fact that the outcome of her attempt was undetermined, the assassin is morally responsible for killing the intended victim, according to Kane, "because she intentionally and voluntarily succeeded in doing what she was trying to do."25 There is an important difference, however, between this case and the performance of an SFA. An SFA must remained undetermined right up until the moment of its performance, whereas the assassination was not necessarily uncertain beyond the point at which the possible twitch failed to occur. The spasming that might have taken place stood merely as a potential obstacle to the realization of her goal, a contingency on the order of the possible intervention of the police. Given that it did not occur, we can say that the assassin’s exercise of her marksmanship guaranteed the killing of the intended victim. That is why it makes sense to hold her morally responsible: we can see how she effected the desired end. In the case of an SFA’s performance, however, the result is not secured by anything that its agent does, making it difficult to assign moral responsibility.
In response, Richard Reilly cites with approval Kane’s claim that an agent who does not "disown" her decision must be held responsible.26 But other theories making "endorsement" a criterion of free will have required modification in light of counterexamples involving deception or ignorance. With that in mind, I wonder whether an agent informed of the contrary behavior of a Lewisian counterpart would be willing to accept responsibility for what she did? No longer under the impression that her effort guaranteed the conflict’s resolution, realizing that it could have led to a different outcome, would she not cease to identify with her choice, seeing it as bearing no unique relation to her exertions? Would it be appropriate for the members of the regiment in my example to continue regarding themselves as having ‘carried the day’ after hearing the observers report that the issue remained in doubt long after they entered the fighting?
Thus, Kane is faced with the following constructive dilemma. Either the effort preceding an SFA guarantees it or it does not. If it does, then the alternative possibility requirement of libertarian free will cannot be met. If it does not, then its agent does not appear morally responsible.
A second constructive dilemma, I believe, is entailed by Kane’s attempt to demonstrate that an SFA’s agent would have been acting rationally even if she had made the desire that she rejected her motive. The basic question is: how can a conflict between desires be rationally settled without one of them achieving ascendancy as a reason for action? It seems that either an SFA’s agent merely possesses the liberty of indifference, which is not what Kane is attempting to show,27 or one of her opposing desires eventually appears as a better reason, culminating her deliberation. But if the latter is true, then it is hard to see how future events could bear upon the question of whether or not her decision was the product of a reliable mechanism and, thus, rational. Given the information available to her at the time it was taken, it will remain either a wise choice or an instance of poor judgment. What is to come may prove advantageous or unfavorable to continuing on the path chosen; but it cannot retroactively make embarking upon it misguided, if it was originally justified, or rational, if it then lacked support. (A failed scientific experiment does not discredit the reasoning behind the hypothesis tested; a successful one does not augment its strength.) Nor could the results of a value experiment turn one of the desires into the better motive were neither one a conclusive reason at the time it was initiated, as required by Kane. The unforeseeable, to borrow from Kant,28 can only make for one’s moral or practical "luck;" it could not determine the worth of one’s reasoning.
Thus, the second constructive dilemma facing Kane is as follows. Either the desires of an SFA’s agent are equal as reasons or they are not. If they are, then she decides with the liberty of indifference, no subsequent events justifying her choice. If they are not, then the condition of "plural rationality" is not satisfied, the conflict between them being resolved by the object of one emerging as more attractive, in her best interest.
Kane responds here by denying the first conditional. He holds that an SFA’s agent’s reasons are incommensurable, as they issue from "different aspects of (her)self and (her) personality": there is no objective measure by which to determine their comparative worth.29 Thus, neither can take precedence unless and until said agent, having made an effort of will to see how best to realize her nascent idea of a good life, places one above the other by initiating a value experiment. In settling the crisis this way, however, she does not reject the alternative course of action as irrational: there remains an equally compelling reason to exercise this option, thus dual rationality is preserved. In other words, it seems that an SFA’s agent cannot act akraticly, for do that she would have to do what she herself deems irrational.
But all desires, it seems to me, can be subjectively, if not objectively, ranked. Realizing that one’s choices build one’s character, one can ask oneself: ‘what kind of person would I most like to become, someone who would not allow self-interest to get in the way of doing the ‘right thing’ or someone who acts morally only when it is convenient?’ Granted that there is no agreed upon objective measure- the Platonic Good- by which to settle a conflict between considerations of different types, appealing to a subjective standard when faced with such a dilemma would not necessarily make one’s decision irrational. An agent could posit a Nietzschian ubermensch as an ideal to guide her SFWs. When forced to choose between moral and prudential considerations she could ask herself, e.g., whom would I rather be like, Socrates or Thrasymachus? Or she might simply ask herself: what would the person I most admire do? Were she to employ this method of volitional conflict resolution, she would be making reasons of different types commensurable.
If an agent were unable to bring herself to deliberatively prioritize her seemingly incommensurable reasons in this way or any other, her subsequent decision would not initiate a value experiment. Rather, it would reveal her tentatively established "plan of life," to borrow John Rawls’ phrase. The solution to her volitional crisis would be implicit in beliefs and values that already make one course of action more attractive than another: it is the decision she would make were she to undertake the task of deliberation. Moreover, the act of choosing cannot itself order her preferences, as Kane maintains,30 if the latter is to explain the former: her explanation would be circular were she to maintain both that she found a certain course of action more attractive than the alternatives because she chose that course of action and that she chose that course of action because she found it more attractive than the alternatives.
The main problem with the doctrine of plural rationality is that an unexercised option should be ‘ruled out’ by one’s reason for acting differently. That is why an adequate account of an agent’s conduct will be a reason that explains not only why she did what she did, but also why she did what she did instead of the alternative. It seems important to know why an agent did one thing rather than another. The reason of an SFA’s agent, as described by Kane, can provide only the first half of such an account. Were it to explain why alternatives were rejected, they would become irrational: an SFA would then be only "one-way" rational. The following case, I believe, will cinch this point. Faced with a threat to his team’s lead in the ninth inning of an important game, the manager brings on his closer, who retires the side to earn a save. That situation seems to repeat itself in all of its particulars in the following day’s game. But this time, the manager calls on a different pitcher, who gives up the winning run. Facing reporters following the loss, the manager is asked to explain the change in tactics. It seems that unless he can point to some perceived difference between the situations (e.g., the closer appeared tired while warming-up) he has acted irrationally: the epistemological analog of the ethical principle ‘like cases should be treated alike’ entails as much.
Is the standard of rationality presupposed here too strong? Is it possible to act rationally sans a compelling reason? It has been suggested that akratic actions, the reasons for which not only do not rule out alternatives but are weaker than the support they enjoy, are rational "insofar as they are performed for reasons in light of which they are intelligible."31 But, whereas an akratic agent can explain her failure to perform the alternative to her act by citing her akrasia, an SFA's agent cannot say why she decided against her other option without making it appear less rational than the one she actually exercised: Kane will not permit her to say that she had a better reason for her actual course of conduct. Or, to put it another way, the akratic agent can concede that she acted against her better judgment; an SFA's agent must maintain that had she had she acted differently her behavior would have been no more or less rational than her actual conduct. Thus, even if, in response to the possibility of rational, akratic conduct, we lower our standard of rationality so that a good reason for an action only makes alternatives less rational rather than irrational, SFAs appear irrational.
The suggestion has also been made that a rational agent could "set aside … what she judges (to be) rationally best" in order to advance a more compelling interest.32 Unlike with Kane’s description of an SFA, it is conceded here that one reason has emerged as conclusive. Nevertheless, the rationality of acting from the rejected reason is supposed to be preserved by the possibility of its being the content of a higher-order preference for the object of what has been deemed (at a lower level of reflection) an irrational desire, which would serve as "justification" for or make "understandable" the attempt at its realization. For example, a parent with careerist tendencies might (second-order) prefer "all things considered" staying home with her children to working at the office yet still be able to (third-order) prefer working at the office to (second-order) preferring "all things considered" staying home with her children to working at the office. If this view were defensible, then the requirement of dual rationality could be met (albeit not in the way Kane suggests): the woman in the example would have a good reason to remain at home- her second-order preference- were she to go to the office and a good reason to go to the office- her (unformed but formable) third-order preference- were she to stay at home. But it is difficult to understand how an agent can act rationally after having decided against doing what she herself believes is rational. If an agent does not really believe, all things considered, that rationality dictates her conduct, then either she has chosen to defy reason or she acts akraticly. In neither event does she act rationally.
Conclusion
Professor Kane maintains that to have a free will one must have performed undetermined acts at some point in one’s life. While he makes a strong case in support of this intuition, citing concerns over autonomy or the "origination" of one’s will,33 it remains to be seen how an act to which there are alternatives would satisfy the control and rationality requirements of free will. Thus, it behooves free will theorists to try dispensing with Kane’s assumption. I have argued elsewhere that possession of a free will is not a matter of having or having had alternative possibilities but of a certain faculty being unimpaired: specifically, those capacities/neural mechanisms having to do with decision making and self-evaluation.34 So long as an agent’s exercise of these capacities, her will, is unaffected by unhealthy conditions such as mental illness, brain washing, intoxication, deception, or extreme fear, she wills freely, even if she never has had an alternative to performing any of her actions. Notwithstanding the fact that her character is the inevitable outcome of her heredity and upbringing, as entailed by causal determinism, an unimpaired will allows it to freely manifest itself, yield controlled and rational behavior. The advocate of such a view, as many have noted since Thomas Hobbes first advanced it, must address Kane’s autonomy concerns. That is, she must defend her intuition that there is a significant difference between being subjected to the will impairing conditions noted above and having one’s choices entailed by facts that obtained prior to one’s birth and the laws of nature (or God’s knowledge, for that matter): is the capacity for self-expression sufficient for free will, regardless of the self’s origin?35 Despite this nagging question, Hobbesian compatibilism emerges as a viable alternative to Kane’s libertarianism in light of the dilemmas posed above. They show that indeterministic agency would be neither controlled nor rational. Thus, it cannot be the source of a free will. If our wills are free, it is not because Nature is so generous as to afford an individual more than one possible fate. It is because a person is endowed with capacities whose exercise makes her life an achievement rather than a fait accompli.
Notes