The Complex Tapestry of Free Will: A Philosophical Odyssey

This book is a comprehensive presentation of Robert Kane’s important views about free will and moral responsibility. Kane is a proponent of libertarianism about free will (and moral responsibility), the conjunction of two claims: (1) the incompatibility of causal determinism with free will/moral responsibility, and (2) our freedom and moral responsibility. Some incompatibilists, such as Carl Ginet and Peter van Inwagen (among others), focus primarily on establishing (1) but do not pay so much attention to (2). Kane is distinctive in giving both core claims of libertarianism considerable attention.In defending (1), Ginet and van Inwagen emphasize what van Inwagen dubbed “the Consequence Argument,” which employs primarily principles expressing the fixity of the past and natural laws. Kane accepts these principles but does not explicitly employ this argument as the driver of his incompatibilism, although I would claim the intuitive ideas and principles involved in the Consequence Argument do play a role in Kane’s argumentation. Similarly, while some incompatibilists, especially about causal determinism and moral responsibility, such as Derk Pereboom, emphasize Manipulation Arguments, Kane does not explicitly invoke them as central to his rejection of (1), although, again, I would note that manipulation concerns do play a role in Kane’s critiques of compatibilism. Kane’s distinctive arguments for (1) rely primarily on consideration of examples and more general kinds of circumstances and facts about human nature and behavior, and the application of inference to the best explanation.In his defense of (2), Kane offers a detailed account of indeterministic free will. He avoids appeal to “extravagant” and “obscure” elements in this account, arguing that it is entirely consistent with naturalism and science. It is not committed to any sort of dualism about the relationship between the mind and body, nor to different “realms” or “perspectives,” as in Kantian views about agency. Kane also presents critiques of other approaches to these topics (compatibilist, incompatibilist, and skeptical) and defenses of his views against objections offered over the years. These further clarify and sharpen both Kane’s defense of incompatibilism and his account of indeterministic free will/moral responsibility. Kane’s critiques and ruminations throughout the book are insightful and illuminating.Kane is one of the most important and influential contemporary philosophers working on this cluster of topics of ancient as well as contemporary interest. Unfortunately, he passed away a few months before the publication of this book. Bob was a generous and supportive mentor and a sympathetic interlocutor, even with those with whom he had fundamental disagreements—a role model for writing clearly and doing philosophy in a collegial way.I confess: I am one of those philosophers with whom Kane has had fundamental disagreements, and I will sketch some of them here, even as I worry that it may seem a bit ungracious. I’ll start with Kane’s defense of incompatibilism (1), then turn to the development of his particular account of indeterministic freedom, as it relates to (2).(1) The Defense of IncompatibilismA central feature of Kane’s overall theory of free will and moral responsibility is his notion of “self-forming actions (SFAs)” (12–13). He observes that many of the seemingly intractable debates (including much of the contemporary discussion) are misguided in focusing solely (or primarily) on the notion that moral responsibility requires alternative possibilities, the AP condition. Rather, Kane holds that there is a deeper and more important requirement for moral responsibility, “ultimate responsibility,” the UR condition (11–14). He contends that a necessary condition of our moral responsibility is that we are the ultimate sources of our behavior (in the first instance, our will or choices) in the way captured by UR: to be ultimately responsible for an action, an agent must be responsible to some degree for anything that is a sufficient reason (a sufficient condition, case, or motive) for the action’s occurring (11).Kane believes that UR implies that, if we are to legitimately attribute moral responsibility to an agent for an action, we need to trace back to some point at which there is indeterminism. The indeterminism enables SFAs. For Kane, it is a mistake to focus solely on what he calls “freedom of action.” Rather, a more basic (though obviously related) issue is “freedom of the will,” which requires an SFA (13–15). Crucially, Kane takes the claim that SFAs require alternative possibilities (freedom to will/choose otherwise) and causal indeterminism to be intuitive. I don’t, however, see an argument for it, apart from the idea that genuine “depth” of human moral responsibility and value depends on such freedom.The problem is that all Kane is entitled to, based simply on intuition, is that there be an exercise of “freedom” or “free will” at some point along the action-generating path, but it is contentious (and not decided by intuition alone) whether this is alternative-possibilities freedom or an actual-sequence freedom. Harry Frankfurt (1969, 1971) distinguished between freedom to do/choose otherwise and acting freely. This distinction, or something like it, goes back at least to Chrysippus and is noted by philosophers such as Bramhall and Locke, as well many contemporary philosophers (in addition to Frankfurt). I contend that, at this point in the dialectic, that is, the analysis of the intuitive and pre-theoretic idea of an “SFA” (a character-forming choice that helps to ground moral responsibility), Kane is not entitled to the idea that SFAs require alternative-possibilities, rather than actual-sequence, freedom.One could put it this way. Kane and I agree that moral responsibility for an action requires an exercise (or, at least, possession) of freedom along the path to the action, but we disagree about the kind of freedom: alternative-possibilities or actual-sequence. We all want a deep interpretation of the free will required for moral responsibility, but it is not justified, at an intuitive level of argumentation, to assume that actual-sequence freedom cannot be deep. If, as I contend, there is nothing in our intuitive and pre-theoretic idea of an SFA that points to alternative-possibilities, rather than actual-sequence, freedom, much of Kane’s argumentation becomes problematic. Throughout his various arguments for incompatibilism, he does not give sufficient weight to the possibility of actual-sequence freedom. (He does offer such an account of voluntariness, but not of freedom.)Consider his discussion of B. F. Skinner’s Walden Two (26–28). Frazier, one of the designers of the “utopian” community, behaviorally conditioned people (in a causally deterministic way) from childhood to want and choose only what they could have and do (26). Frazier described the community as “the freest place on earth, the freest place imaginable” (26). Of course, Kane thinks this is false; the people are free only in a superficial sense (“freedom of action”) but not in a deeper sense (“freedom of the will”). He holds that this deeper sense requires an SFA, and thus indeterminism. Kane writes: “…the worry about the freedom implicated in moral responsibility grows to the extent that the conditioning may be completely determining …” (28).I agree that B. F. Skinner’s community is not the freest place on earth. It is more like a “gangster’s paradise,” as in the song sung by Coolio and L.V. I deny, however, that the only way to capture this “deeper” kind of freedom is to accept AP and UR. Actual-sequence freedom can fill the bill. There are various plausible accounts of such freedom, and my point here does not depend on accepting any particular account of actual-sequence freedom, only that such freedom exists (and can be separated, at least analytically, from alternative-possibilities freedom).I adopt a “guidance control” model of actual-sequence freedom. On my view, an agent acts freely only if she exhibits guidance control—that is, she chooses/acts from her own appropriately reasons-responsive mechanism (Fischer 1994; Fischer and Ravizza 1998). The people in Walden Two lack the capacity for guidance control and thus lack actual-sequence freedom. They do not act freely. They lack not merely the freedom to want and do what they do not in fact want or do (perhaps as the result of deterministic conditioning by other people or other clandestine compulsion). It is deeper than this: one must choose and act from one’s own reasons-responsive mechanism. The people arguably fail to act from their own mechanisms, and those mechanisms may fail to be reasons-responsive in the relevant sense.I contend (although I cannot argue for this here) that guidance control is compatible with causal determinism (Fischer and Ravizza 1998). Thus, even if alternative-possibilities freedom (freedom to choose and do otherwise) is incompatible with causal determinism, actual-sequence freedom need not be. We should not then infer that the problematic status of the denizens of Walden Two must be explained by causal determination qua causal determination; it might be explained by the kind of determination that precludes guidance control.Kane writes that some compatibilists would insist that the difference between deterministic control by others (such as Frazier) and such control by natural processes of heredity and environment (and “ordinary” human interaction) is relevant here. He claims that these compatibilists would resist attributing responsibility to the citizens of Walden Two, but not necessarily to any agent causally determined to choose and act as she does. He finds this implausible: Why is there a difference based solely on these sorts of facts about the distal initiation of the sequences (agential and nonagential) (36–38)?A compatibilist, however, need not hold that the distinction between “controlling agents” and “natural processes” (as specified above) at the beginning of the sequence in itself makes (or explains) the relevant difference. Whereas the first kind of path arguably lacks guidance control, the second kind may contain an instance of it. A compatibilist can thus agree with Kane about the people in the community of Walden Two but reasonably resist the extrapolation to a general thesis of incompatibilism.Kane discusses an actual case of a young man who assaulted and raped a young girl. Kane’s initial reaction was filled with anger and resentment toward him. He writes: “But as I listened daily to the testimony of how the young man came to have the mean character and perverse motives he did have—a sordid story of parental rejection, sexual abuse, bad role models, …some of my resentment toward the young man decreased” (35). Kane, however, resisted transferring all the responsibility to others (the parents, in particular). He attributes this hesitation to his questioning whether the young man’s behavior was causally determined by his adverse early circumstances. Kane contends that the degree of transference (I’ll stick with this term of Kane’s, rather than transferal, which I think preferable) of responsibility depends on the degree of probability of the behavior that issues from the defendant’s upbringing, which was out of his control. In the extreme, causal determination expunges all moral responsibility (or, alternatively, renders a complete transference appropriate). The basic idea is that if causal determinism were to obtain, the young man had no alternatives available to him, and thus it would be unfair to hold him responsible.I contend that this is not the only (or best) explanation of the phenomenon of transference of responsibility. As his formative background renders it more and more probable that the young man will commit the assault and rape, it becomes more likely that he did not act freely (or exhibit guidance control) in behaving as he did. We thus have an explanation of the transference of responsibility that invokes a deep notion of freedom that is compatible with causal determinism. This explanation does not put any weight on the distinction between distal initiation by a clandestine controller and such causation by “ordinary natural processes.” On this approach, it is not a matter of whether the individual could “rise above” his formative circumstances but whether, given those circumstances, she acts freely/exhibits guidance control.Consider, also, Kane’s analysis of the (in)famous Frankfurt-Style Cases (FSCs) (46–57). He is willing to concede that there are FSCs in which the agent appears to be morally responsible for her action but could not have done otherwise. But he contends that this conclusion is drawn too hastily, because Kane’s notion of moral responsibility (“ultimate responsibility”) requires an SFA somewhere along the path to the choice and action in question. So the FSCs do not show that moral responsibility (ultimate responsibility, which is the gateway to blame, punishment, and so forth) requires no alternative possibilities, and thus their potency in defense of compatibilism is etiolated.I reply as above. I agree that moral responsibility requires freedom along the path (proximal or distal) to the relevant choice and action, but it cannot simply be assumed, without argumentation, that the freedom in question is alternative-possibilities freedom. This assumption would be particularly egregious in an evaluation of the FSCs and their relationship to the alternative-possibilities requirement for moral responsibility. Further, I see no reason why there can’t be “Frankfurt-style counterfactual interveners” all the way back, so to speak, thereby preemptively overdetermining all prior choices and actions along the path to the choice whose responsibility status is under consideration. Kane’s response seems to involve complexities of putative quantum mechanics in the brain, and it is not easy to grasp. At the least, I do not find it convincing.I cannot engage with all the argumentation involved in Kane’s resourceful defense of incompatibilism. In my opinion, a key problem throughout is that he relies on an analysis of the required way of forming one’s character (via SFAs) that implies alternative possibilities, but this is not justified simply by intuitive considerations. All that is justified is a requirement of freedom, but this in itself does not decide the issue as between alternative-possibilities and actual-sequence freedom. In seeking to explain various features of our responsibility practices, Kane fails to take actual-sequence freedom seriously. He assumes that all such freedom, or perhaps actual-sequence resources, including his “teleological guidance control” (81–82), available to a compatibilist, are “superficial” (or, at best, relevant to voluntariness, but not freedom). This is incorrect.In addition to moral responsibility, Kane contends that a principle analogous to UR (with its requirement of indeterminism) is a necessary condition for the development of fundamental features of the self (what he calls “the dialectic of selfhood”), conferring “greater value on …goods …such as creativity, autonomy, self-creation, desert, …dignity, respect, and others” (284). I agree that exercises of metaphysical freedom render such goods immeasurably greater in value (if it is not a requirement of their having value, or existing, at all). The reader will not be surprised to know that I would insist, however, that it does not follow that only alternative-possibilities freedom can do the trick.There is, then, a problem at the very heart of Kane’s critique of compatibilism, which is central to his overall approach to free will and moral responsibility. This problem, if it is not fatal, at least points to a significant incompleteness in Kane’s argumentation. Put a bit less charitably, the use of UR, as Kane it, is to explain the intuitive about moral responsibility and the of so central to human as we and of this book and throughout his Kane does not argue against compatibilism but considerable attention to an indeterministic (and thus account of free will. Peter van Inwagen that he must choose between the and the and that he the It is to Kane’s that he to the Kane holds that our freedom and moral responsibility back to SFAs at in the path to the choices and actions under consideration. He an insightful and account of which he contends take place in of between that point in different and or and and so In such on Kane’s view, one and this to a distinctive of indeterminism in the brain, by quantum mechanics The choice under such circumstances is a result of an of the will. the one choice rather than the agent an action, she and acts and and doing what she to, as a result of her (in the all Kane contends that the choice and action are not or the result of or They are free in the sense that moral is the sketch of a and account of freedom that Kane in The of Kane takes is that, as the agent with the she seems to be to choose (and and also to choose (and although she she can’t do In Kane of this But to that the agent is to do two different she she can’t do is arguably to in the SFAs that are the of moral responsibility, and Kane has that this is thus offers a account of SFAs in This is an In a for an SFA, the agent may have the to an to choose and also the to an to choose to the of the of the relevant is more of Kane’s is that we the of will to choose as over a of with a of to choose by (or no choice at all). at first As I it, the the agent the to to choose and engage in to choose over with or no choice at lack the to the in any and I note that I find it to It is that the account avoids the that and how it with Kane’s requirement of (the for moral responsibility especially not because the alternative sequence might issue in It has a more and one is in of the for the to an problem for Kane’s account of freedom (in any of its and It makes our free will and moral responsibility and on views in and Kane holds not only that indeterminism is in the but that it is of a quantum by in like this is required for his and it is that this is the way the (although not that the are causally hold that the but not necessarily in this way. The is out (and I them to render a theory implies that, if causal determinism were (or at least the were or indeterministic but not in the way required by the we would not be free and morally These basic facts about free will and moral on a they would on a kind of of about the status of the fundamental of and about the of our concede that our freedom and responsibility should depend on facts about our and capacity for I am a although my account of guidance control does not require For if it were that we are not at the of for them and (as to Peter would the that we are free and morally Similarly, if it were that we are not at all in the extreme, were in our by we were and we are to choose and as we as in a by Peter van this would result in my the views that we have free will and moral but which I do not think it even plausible to that our moral responsibility (and free should depend on whether the fundamental of nature are or whether the is indeterministic in responsibility should not be so that it depends on this kind of It should not on this sort of (or we the view, so basic to human as we it, that we morally or punishment, if we were that the of have rather than I think in is the rather that Kane accepts (and has a a of and of have Kane as an and he does in fact (and that only (and also of but he makes it in the book that he is an (although of a different sort than those on He contends that this of his overall it to two for the agent problem and the problem of causal is to however, how is to these I will to the problem of or causal a for causal in general action, justified and so Some have to the problem with only have simply an notion of causal The more approach is not obviously to and the is, not or does in to an theory of free action, as in Kane’s in this I see As I it, the idea is (or might that the agent by and the of the Kane’s that she has the to start and is This would seem to be in any case, to to the that an Kane to as the to exercise “teleological guidance the in such a way as to of the causation as it The sequence in an action, including and so is a occurring a of not an then, can at the beginning causal along the one with an to the kind of causation along the path to the why not simply require and it a here seems to be a I noted that my critiques might be but this is not a It is not even an as I have no Kane’s ideas will on and philosophers for as they have for a engage in and even is not or even It is a of deep respect, the sort I have for Robert am for insightful by and Derk I am to detailed from who from many to for which entirely and solely