Is Free Will a Necessary Fiction?

Saul Smilansky thinks that belief in contra-causal free will, which he concedes doesn't exist, is necessary to provide essential support for morality, meaning, and the worth of human beings. I argue that he is mistaken on all counts, and that we would be better off morally and existentially without believing the falsehood that we have free will.  Moreover, Smilansky's view could motivate systematic deception about our causal connection to nature, which is neither possible, necessary, nor desirable. Free will is not a necessary fiction, and making known the naturalistic truth about ourselves is a far better basis for human flourishing.

Some ideas, it is thought, are inherently dangerous. Some pose a threat because they are false, and by misleading us, encourage destructive behavior. Others are dangerous because although true, they might have dire consequences for those who fall under their sway. An example of the former is the belief that human beings are not a significant contributor to global warming, and that therefore there’s nothing we can or need do to avert a climate disaster over the next century. Believing such a falsehood will have manifestly destructive consequences for the planet.

Examples of what are considered to be the latter aren’t hard to find either. A favorite conservative theme is that telling children the truth about our country’s history, for instance that prostitution was widespread in mining towns during the gold rush, will undercut proper feelings of reverence and respect for the good old US of A. So it’s better to sanitize the past in carefully edited high school textbooks than risk demoralizing the next generation.

In his chapter, "Free Will, Fundamental Dualism, and the Centrality of Illusion" in Robert Kane’s The Oxford Handbook on Free Will (also online), philosopher Saul Smilansky argues that although we don’t have free will in the traditional libertarian, contra-causal sense, a widespread appreciation of this truth would constitute a dire threat to our moral commitments and practices. Disbelief in such free will is a literally dangerous idea of the second category. In order to protect our moral virtues, it's better to have the false belief that we have free will. Believing that we are ultimately responsible agents – god-like, miniature first causes who choose without being entirely determined to choose – is necessary to supply the requisite motivation to maintain a strong sense of ethical duty and responsibility. Moreover, the very possibility of finding meaning in our lives and truly valuing each other depends, says Smilansky, on not believing the causal, naturalistic truth about ourselves. Free will, therefore, is a necessary fiction, without which the entire social-psychological fabric of meaning and morality is at risk.[1]

I want to take exception to this claim, and show that we need not accept Smilansky’s argument from dire consequences. Instead, we can safely make known the naturalistic facts about what it is to be human, while keeping and indeed improving our moral commitments and practices. We don't need to be misled about our true nature, since a robust morality and quest for meaning are perfectly compatible with naturalism. Once we see that our moral inclinations stem from our shared biological human endowment, honed by culture, we don’t need the notion of ultimate responsibility or self-creation ex nihilo to back up our sense of ethical duty. In fact, the very belief in ultimate agency arguably creates attitudes and beliefs which reinforce some of the worst human tendencies, e.g., for retribution and self-aggrandizement. We would be morally better off without it. This is my argument from good consequences, which has the signal virtue of being aligned with the truth about our situation as both Smilansky and I (and most philosophers and scientists) see it.

As Smilansky acknowledges, the standard libertarian notion of strong, contra-causal free will, in which we are in some important respect the first cause of our characters and actions, cannot be sustained in the light of logic or scientific evidence: "The libertarian project was worthwhile attempting: it was supposed to allow a deep moral connection between a given act and the person, and yet not fall into being merely an unfolding of the arbitrarily given, whether determined or random. But it is not possible to find any way in which this can be done" (491).

One might wonder why, if we aren’t supernatural interveners in causal processes, this makes us "merely an unfolding of the arbitrarily given"?  From a naturalistic perspective, individuals aren’t just the unfolding of events, they remain centrally important features of our world which contribute essential elements of the causal story as it evolves. We couldn’t tell this story without reference to persons and their powers (just try!). True, inclusive naturalism sees each and every aspect of the person as arising within the causal matrix, but the individual doesn’t in the least disappear as a locus of proximate control and origination. To suppose otherwise is to buy into strong reductionism, the idea that because we are composed of sub-personal, biological and computational processes that are ultimately physical and causal, that therefore the person-level causality of character-based reasons and motives is somehow invalidated or made irrelevant when devising causal explanations.[2]

Compatibilism

Compatibilism, generally, is the thesis that a robust sense of moral responsibility, sufficiently robust to ground essential moral and criminal justice practices, is compatible with our being completely included in the natural causal order and therefore not having libertarian free will. Smilansky acknowledges that certain distinctions which obtain even in a completely deterministic world are necessary to make sense of morality. He defends what I think is a reasonable, naturalized sense of moral desert, based on the idea that some actions, even if they are ultimately determined by factors outside our control, are justifiably thought of as "up to us," compared to other actions which aren’t up to us. In most everyday contexts, the ultimate arbitrary "givenness" of our situation as seen from a cosmic perspective loses practical significance in comparison to our capacity for reflective, rational control – acting according to one’s character and wishes and in the light of reason. When we act under these conditions, behavior is indeed proximately up to us - our desires and rationality – as opposed to being coerced or the result of a mental defect. Agents deserve, naturalistically, the morally relevant responses to their acts (praise, blame, etc.) in the limited sense that such responses work to guide behavior toward good consequences in ways we deem appropriate.[3]  Credit and blame are accorded to enable people to live as "responsible beings," that is, beings which behave responsibly and ethically. Normal and widely accepted excusing conditions come into play to pick out those agents we don’t hold responsible: "situations in which they lacked the abilities, capacities, and opportunities to choose freely and therefore are not responsible in the compatibilist sense" (495). All this is the necessary, and I believe, sufficient basis for moral responsibility, conceived consequentially.

Note on compatibilism vs. neo-compatibilism

Many, perhaps most, contemporary compatibilists don't believe that naturalized desert is limited to a consequentialist understanding of morality.  That is, they believe that a specifically deontological justification for moral practices exists, for instance that would justify retributive punishment.  People deserve praise and blame in the sense originally connected with contra-causal free will, a sense that needn't be justified by appeal to consequences, but simply by virtue of the agent's blameworthy or praiseworthy act. True, we are fully caused to be who we are, says the compatibilist, but we are nevertheless deeply deserving, in a non-consequentialist sense, of praise and blame, or punishment and reward. It's our deontological duty, therefore, to mete out just deserts, independent of whatever outcomes, positive or negative, these might produce. The person should suffer, or feel pleasure, simply because she deserves it.

But why, one wonders? Under a libertarian, contra-causal conception of freedom, in which the self is not completely beholden to non-self factors, this strong sense of desert makes intuitive sense - after all, my choices are ultimately up to me.  But the naturalist compatibilist understands that the self ultimately owes everything to conditions it didn't choose, so why insist that naturalized desert encompass punishments and rewards that serve no behavior-guiding function? Traditional compatibilists often ignore the problem that our natural freedoms do not support moral responsibility practices that are normally thought to presuppose contra-causal free will, for instance retributive punishment. Therefore their claims to have naturalized moral responsibility ring hollow. Without exploring this question further here (see the papers at the Criminal Justice page), I'll use Owen Flanagan's term "neo-compatibilism" to refer to the naturalized, strictly consequentialist conception of moral responsibility that I think is justifiable, given that we are not self-made selves.[4] Neo-compatibilists recognize that if we take science seriously with regard to human nature, our notions of responsibility, agency, and even the aims of the law might change, so neo-compatibilism is revisionary in a way that mainstream compatibilism is not. Some of these revisions, tending in a humane, non-punitive direction, are discussed below in "The significance of naturalism."

Smilansky says that (what I will now call) neo-compatibilism is "grimly insufficient," since no one can be "ultimately in control" or "ultimately responsible." The causal perspective given by naturalism shows that it’s ultimately all luck, ultimately not one’s fault or to one’s credit what one becomes. True, but this fact doesn’t impugn neo-compatibilism, which can still distinguish between those who are and are not moral agents, and which can justify adequate (and humane) responses to criminality and moral infractions, all while accepting a naturalistic universal causality.[5]  The insufficiency Smilansky finds with neo-compatibilism (and its naturalistic basis) is a matter of judgments made from the libertarian perspective, it isn’t the insufficiency of a thoughtfully worked-out neo-compatibilist position on moral responsibility. Smilansky says that if we punish someone on neo-compatibilist grounds, for instance to deter, incapacitate, or rehabilitate, this is unjust since the person isn’t ultimately responsible for who they are and their actions, so it’s not really her fault (493). But why, if we agree that libertarian free will is incoherent, since no one could possibly meet the requirements of libertarian agency, does Smilansky judge neo-compatibilist notions of justice by such a standard? Why should intuitions about fault or responsibility or control, derived from an impossible contra-causal conception of the person, be thought reasonable criteria for assessing the fairness of neo-compatibilist justice? Such intuitions might be someone’s knee jerk response if they haven’t been exposed to fully naturalized versions of moral responsibility, but once the adequacy of naturalism for morality has been explained, we are not bound to use impossible supernaturalistic standards of agency to assess our moral practices.

Smilansky claims that neo-compatibilist grounds for moral worth and respect are "shallower" than libertarian grounds (493), but since the libertarian view is unreal, why should the neo-compatibilist view be judged shallow in comparison? It’s simply what’s possible, given the reality of things. It may be shallow compared to what we are used to thinking, but once we understand libertarian free will and agency don’t exist, then the charge of shallowness ceases to cut any ice against neo-compatibilism, especially once we see that it's perfectly adequate to ground nearly all our intuitions about the moral worth of persons. Libertarians and hard determinists may claim that "real responsibility" and "real agency" must involve contra-causal freedom. But, once it is seen that such freedom is an impossibility (which hard determinists agree is the case), then the responsibility and moral agency based on it can’t logically be characterized as real. After all, something real has the obligation to exist, or at least be a coherent possibility. Neo-compatibilist moral agency meets these requirements, while libertarian agency does not.

Smilansky charges neo-compatibilism with a "complacent compliance with the injustice of not acknowledging the lack of fairness and desert" (493). But again, the notions of fairness and desert here derive from the discredited libertarian perspective. Since Smilansky agrees that nothing could actually confer such fairness and desert, then it is unfair to find neo-compatibilism lacking on these grounds. True, we may hanker after such things, but having discovered they don’t exist, we can’t fault the truth of the matter for contradicting our hankerings. Instead, we should give them up. In rendering his judgment against neo-compatibilism, Smilansky is straddling two worlds, even while admitting one is a fantasy.

Similarly, he goes on to say that neo-compatibilist practices are "in one way unjust, owing to the absence of libertarian free will, which implies that our actions are on the ultimate level not up to us" (495). But why should such a sense of injustice be given any weight once we see that it’s simply not possible to have authored ourselves? It would only be unjust if libertarian agency were a live possibility, but it isn’t. Just practices reflect what’s "up to us" in the neo-compatibilist sense that Smilansky describes, a sense which is perfectly adequate to capture our intuitions about who can be held responsible and who can’t. The question is, why does Smilansky continue to hold onto to this sense of injustice, driven as it is by a definition of freedom that he acknowledges could never be fulfilled? Once we see that libertarian free will is impossible, the notion of justice based on it simply isn’t a fair criterion for assessing neo-compatibilist grounds for our practices.

The Significance of Naturalism

Smilansky says: "There is no reason to claim that the absence of libertarian free will is of no great moral significance and moreover to deny the fact that without libertarian free will even a vicious and compatibilistically-free criminal who is being punished is in some important sense a victim of his circumstances" (497). The absence of libertarian free will is, one might say, the presence of naturalism, since it is the naturalistic, scientific perspective which establishes the incoherence and empirical implausibility of libertarian, contra-causal, supernatural free will. Because it does just that, naturalism is of great moral significance, I believe, but not in the way Smilansky supposes.

First, simply because circumstances create us, we are not thereby victims of circumstance. We can only be victims of circumstance in this global sense by supposing that we might conceivably have been masters of circumstance in some global sense, i.e., by having libertarian free will. But as Smilansky concedes, naturalism forces us to abandon that perspective as wishful thinking, a traditional cultural myth, a false folk-metaphysics undone by science and logic. So the person being punished for his crimes under a neo-compatibilist regime is not a victim by virtue of naturalism or universal causality, so there is no intrinsic unfairness to such punishment, as Smilansky implies. Unfairness would only accrue if the person punished had been insane, coerced, or in some other sense didn’t have the capacity or opportunity to act responsibly.

The moral significance of naturalism - of the absence of libertarian free will - isn’t that it makes us all victims, but that it undercuts retributive and fawning attitudes based on the assumption of libertarian agency. Since we don’t create ourselves, we do not deserve praise and blame in the ultimate, traditional sense of being the uncaused originators of our actions. Retributive rage and the urge to punish harshly, e.g., by imposing the death penalty, permitting rape in prison, or withholding opportunities for rehabilitation, are directly linked to the assumption that persons are self-caused and could have done otherwise in the situations that shaped them. Similarly, the acquiescence to extreme social and economic inequalities is at least partially a function of the belief in the self-made man: those who succeed deserve unlimited rewards since their success is a matter of their self-chosen drive and acumen, not a matter of external conditions.   Likewise, those who fail deserve their impoverishment since they chose not to succeed. If, under naturalism, we dispense with the self-made self, whether criminal or exemplary, such attitudes lose their primary metaphysical basis, and so might wither in favor of more compassionate, cause-appreciating attitudes. This in turn may have a beneficial effect on criminal justice and other social policies, creating a more enlightened, less punitive culture.

This is why it’s puzzling that Smilansky says: "If we reflect upon the fact that many people are made to undergo acute misery while the fact that they have developed into criminals is ultimately beyond their control, it is hard to dismiss this matter in the way that compatibilists are wont to do" (497). As I hope I made clear in the paragraph above, and in much else at Naturalism.Org, it’s not at all the case that neo-compatibilists such as myself and Flanagan (and hard determinists and incompatibilists such as Honderich and Pereboom) dismiss the factors beyond an individual’s control that shape their lives, or the misery meted out by punishment, quite the opposite. Neo-compatibilists, perhaps more than libertarians, tend to be cognizant of the manifold causes that create individuals and societies.  And it is precisely the deep appreciation of such causes that motivates less punitive attitudes and practices which in turn can lessen the unnecessary toll of misery taken by, for instance, our criminal justice system.  Such insights constitute, partially, the practical moral significance of naturalism.

Smilansky goes on to say that "Similarly, any favorable compatibilist appreciation of persons is necessarily shallow for, in the end, it rests upon factors not under the person’s control. Any factor for which one is appreciated, praised, or even loved is ultimately one’s luck. That compatibilists are indifferent to such ultimate arbitrariness, shallowness and injustice is morally outrageous" (497). The answer here is that neo-compatibilist appreciation for persons isn’t shallow, since the deeper sort of appreciation based on a person’s capacity for self-choosing isn’t possible, given that this capacity doesn’t exist. Therefore, it isn’t morally outrageous to accept the fact that our appreciation of people’s virtues and faults is based on the existence of such characteristics, not on their being self-originated. As Pereboom points out, knowing that persons are ultimately caused in every respect doesn’t erase the moral and aesthetic distinctions we draw between individuals, even though it may undermine certain attitudes based in traditional notions of agency.[6]

The Risks of Naturalism

Because Smilansky thinks that neo-compatibilism is "grimly insufficient" to ground our moral practices, he says we must protect society against realizing the truth of what Daniel Dennett has called "creeping mechanism," that is, a fully inclusive naturalism that understands human agents as completely caused creatures. But remember that Smilansky thinks this insufficiency isn’t a matter of justification, it’s that he supposes neo-compatibilism lacks the necessary persuasive power that only the myth of libertarian agency can provide. As he puts it: "Yet while the justification for these values does not require libertarian free will, in practice they might be at risk were the lack of libertarian free will internalized" (498, emphasis added). In other words, while Smilansky and others in the educated, professional classes might be able to understand that neo-compatibilist justifications are all we need to ground moral practices, the great unwashed can never be brought safely to such a realization:

We often want a person to blame himself, feel guilty, and even see that he deserves to be punished. Such a person is not likely to do all this if he internalizes the ultimate hard determinist perspective, according to which in the actual world nothing else could in fact have occurred -- he could not strictly have done anything else except what he did do (499).

But of course, whether someone who adopts a completely causal view of his behavior is unlikely to feel that his punishment is appropriate is an empirical claim needing research, as are all the claims Smilansky makes concerning the dangers of understanding the impossibility of libertarian free will (or, more positively, accepting naturalism). Smilansky assumes that naturalism inevitably demoralizes, no matter what sort of education about causality and compatibilism we might provide. Such might be the case for some individuals. But the psychological responses and changes in attitudes that might be generated by encountering naturalism really are open questions, not to be prejudged without investigation. In my experience (hardly definitive, of course), once people see that universal causality does not constitute a blanket excuse, that is, that sufficient reasons exist apart from ultimate, retributive desert to impose sanctions on offenders, namely social protection, deterrence, rehabilitation, and restitution, then they quickly understand that not everything is permitted under naturalism. They might indeed feel that their punishment is deserved in the neo-compatibilist sense of being necessary for guiding good behavior (see Morse’s "Guiding Goodness") and for achieving other agreed-upon social goals. In any case, we need to actually find out what sorts of responses people have to naturalism, not imagine the worst and then restrict knowledge about it on that assumption.[7]

Smilansky says that the sentiment supporting common intuitions of justice should be "almost instinctive" (498). But of course it is instinctive, in fact it’s so hard-wired that it’s probable that ethical instincts drive the libertarian assumption of contra-causal agency, not the other way around. This suggests that we don’t need to protect such instincts from naturalism. Intuitions about justice and morality derive from deeply embedded emotional responses to injury, threat, cooperation, and mutual aid that most likely conferred survival advantages upon the individuals (and groups, perhaps) which manifested them.[8] This is why it is highly unlikely that naturalism will undermine such intuitions, although it may moderate retributive impulses commonly justified by free will. So despite Smilansky’s claims to the contrary, it seems unlikely that we need to continue inculcating the myth of contra-causal, supernatural agency to prevent wholesale demoralization.

Smilansky claims that understanding that we don’t have free will is "an obvious danger to our moral motivation," but this is true only if neo-compatibilist reasons for ethics aren’t appreciated, which they can be quite easily. It also suggests that our values only function from the standpoint of wanting and imposing strong, ultimate credit and blame as defined by supernaturalistic and dualistic views of agency, when in fact they derive from our natural inclinations to cherish and protect. So in order to "function adequately as moral agents," we don’t need contra-causal freedom, nor need we believe in it. A strong parallel exists between Smilansky’s claim and the claim that human beings need to believe in God to be good. But just as atheists are perfectly capable of goodness, and indeed aren’t handicapped by religious absolutism in achieving tolerance of other views, those who don’t believe in free will are perfectly capable of being moral agents. Thorough-going naturalists may even have an advantage in achieving certain moral virtues, since they are perhaps less susceptible to the radical individualism and egoism generated by belief in libertarian agency. And it should be noted that in cultures in which the notion of free will is unheard of or considerably muted (e.g., in some Eastern societies) moral codes nevertheless thrive apace.

In a similar vein, Smilansky says that "Illusory beliefs are in place concerning free will and moral responsibility, and the role they play is largely positive. Humanity is fortunately deceived on the free will issue, and this seems to be a condition of civilized morality and personal value" (500). The benefit he sees of maintaining the fiction of strong, originative free will is to sustain morality and meaning as they currently exist for perhaps the majority of citizens in Western society. But 1) Is the "civilized" status quo what we really want, given the psychological and social pathologies generated by assigning ultimate credit and blame, and the excessive rewards and punishments justified by libertarian free will?  2) Should we continue to ignore the causes of social and personal dysfunction, violence, and crime by continuing to believe that such evils ultimately derive from uncaused human choices?  3) Is the fiction of free will really necessary to rally support for perfectly viable neo-compatibilist moral and criminal justice practices that can be justified on their own terms, are consistent with our scientific understanding of the world, and which avoid the punitive excesses driven by the belief in libertarian agency?  4) Do we need the assumption that we are self-caused choosers in order to perceive ourselves and others as truly valuable? The answer to all these questions, I submit, is no. Instead, appreciating and accepting a naturalistic view of ourselves and our behavior while adopting an explicitly neo-compatibilist morality, far from demoralizing us, will be a vast improvement over the status quo, both for morals and meaning.[9]  This is my counter-proposal to Smilansky’s, which besides resulting in better, revised practices concerning punishment and social policy, has the additional virtue of avoiding the mass promulgation of falsehoods.

Orwell, or an Open Society?

In his paper (and in his book, Free Will and Illusion) Smilansky's illusionism about free will implies that we would be better off ignorant of naturalism.  It would be better for morality and meaning if scientific findings that show human beings are indeed caused creatures were not widely disseminated or understood. In fact, if we took it seriously, Smilansky’s illusionism might well motivate an Orwellian project of doublethink, in which an increasingly manifest truth – our causal interconnectedness with nature – is perpetually denied.

To make a convincing case that we need the free will illusion, Smilansky has to prove that people would indeed lose their moral compass, that under naturalism they would succumb to "pragmatic consequentialist temptations" (e.g., punishing the innocent), or be driven to "unprincipled nihilism" (501). But of course it would take considerable empirical research to show that such horrors would necessarily follow the dissemination of naturalism, instead of the consequences that I foresee – a lessening of ego-driven, punitive, and fawning attitudes, and an increase in knowledge about the actual causes of crime and social and personal dysfunction that we can use to reduce human suffering. Smilansky’s case for "motivated obscurity" re free will – that is, shielding people from naturalism – is not, as he thinks, prima facie, because people can understand compatibilist grounds for our criminal justice practices and they don’t necessarily fall into a moral panic when they discover they’ve never had libertarian free will, but still possess all their real causal powers.[10]

In short, it’s not at all clear that we need the illusion of free will for an "adequate moral and personal reality" (502). Free will of the contra-causal variety is not a necessary fiction.[11] Smilansky raises issues that suggest research is needed on beliefs about agency, responsibility, and the psychological consequences of being disabused about libertarian free will, research that is now ongoing. But must we remain deceived about the truth of our full causal connection to nature? Nothing suggests such deceit is necessary, possible, or desirable. The only, and best, way forward is in the light of our best collective understanding about who and what we ultimately are, that delivered by continuing exploration of how we fit into the natural scheme of things.

TWC, written July 20 2002 and revised November 2005

Notes

[1] Many readers will be familiar with Daniel Dennett’s Darwin’s Dangerous Idea, in which he makes vivid the danger of Darwinian, selectionist, bottom up processes ("cranes") to the assumption of mind-first creative processes ("skyhooks"). In his book Freedom Evolves (on free will, the ultimate skyhook, it would seem), Dennett raises the issue of the moral dangers of a belief in determinism, and consistent with his earlier work on free will, Elbow Room, seeks to show that such dangers are exaggerated. But he also recommends proceeding with all due caution in order to head off any fatal confusions about the implications of determinism. I whole-heartedly concur.

[2] For more on the case against strong reductionism, see the section titled "Strong Reductionism vs. Causal Explanation" at Neuroscience and the Human Spirit, and "Playing God, Carefully".

[3] See Stephen Morse, "Guiding Goodness", for a cogent defense of such a view in the context of criminal justice studies.

[4] Flanagan, in his marvelous book The Problem of the Soul, (reviewed here) points out that compatibilism necessarily redefines the tenable conceptions of agency, freedom, and responsibility: "The compatibilist, meanwhile, if he thinks free will is compatible with determinism, must have changed the subject. He cannot be saying that the Cartesian [contra-causal] conception of free will is compatible with determinism because, well, it isn’t. And indeed if one looks at the literature one will see that compatibilists invariably mean something different by free will than what the orthodox concept says it is" (The Problem of the Soul, p 126).

[5] See for instance my papers "Materialism and Morality" and "The Freedom of Susan Smith," my review of Michael Moore’s Placing Blame, and Stephen Morse’s papers "Guiding Goodness" and "Rationality and Responsibility".

[6] See Pereboom, Living Without Free Will, pp. 187-213.

[7] Research is underway that is designed to test Smilansky’s empirical claims about what people actually believe about free will.  See here, and secondarily here.

[8] See Flanagan’s chapter on "Ethics as Human Ecology" in The Problem of the Soul, and my review of Moore’s Placing Blame, the section entitled "The natural function of morality".

[9] See "Spirituality Without Faith" for my take on the quest for meaning carried out within naturalism.

[10] For reassurances about fatalism and other supposedly nefarious consequences of naturalism, see the Resources page.

[11] Others besides Smilansky have supposed free will is a necessary fiction, e.g., Stephen Pinker, Marvin Minsky, and those mentioned at "Free Will Is a Necessary Fiction". However, Smilansky's work on free will illusionism is by far the most philosophically significant, spanning 15 years and including many papers and his recent book, Free Will and Illusion. The latter is the longest and most carefully worked out defense of free will illusionism extant.