-- The Determinism and Freedom Philosophy Website --

It is sometimes supposed that philosophers are independent thinkers, even pioneers or revolutionaries. They think their own thoughts, each in his own ivory tower, sometimes hers. The reality is wonderfully different. Philosophers exist in herds, blind movements, mutual-admiration societies, troupes of bright and dim graduates of the more select universities, circles of natural affinity, mobs and other groups. Richard Double is among philosophers who are a little different. If you read his book The Non-Reality of Free Will, you will not be reading the same old stuff. You may not agree with it, but you won't be dozing off. If you read his new piece here, you will also be getting outside the orthodoxies of the determinism and freedom debate. It begins with his own clear abstract or summary.

ABSTRACT The following is a criticism designed to apply to most libertarian free will theorists. I argue that most libertarians hold three beliefs that jointly show them to be unsympathetic or hard-hearted to persons whom they hold morally responsible: that persons are morally responsible only because they make libertarian choices, that we should hold persons responsible, and that we lack epistemic justification for thinking persons make such choices. Softhearted persons who held these three beliefs would espouse hard determinism, which exonerates all persons of moral responsibility, or, at least, would not espouse libertarianism. I do not address the view held by some libertarians that we do have epistemic justification for thinking that persons make libertarian choices, a minority position that I believe cannot be sustained.

I propose four premises to argue that most libertarians are hard-hearted (unsympathetic, not morally conscientious). (1) Libertarians believe that we may hold persons morally responsible only if they exercise libertarian free will. (2) Libertarians believe that we should hold persons morally responsible. (3) Most libertarians believe that we have scant epistemic justification that persons have such free will. I believe, but shall not argue in this paper, that libertarians who believe they have epistemic justification for libertarian free will are mistaken.1 I add a fourth premise describing how soft-hearted persons would respond to accepting the first three premises, and conclude that those libertarians who do not believe they have epistemic justification that persons make free choices are hard-hearted. In the first section I present my argument. In section two I consider three implausible replies libertarians may make to my argument. In section three I examine three replies that are not available to libertarians.

1. The Moral Problem with Libertarianism

Libertarians believe that persons choose freely in such a way as to make them morally responsible at least some significant amount of the time, and that persons are free and morally responsible only because they make undetermined choices. Whatever variety of undetermined choices they postulate, few libertarians purport to provide epistemic justification that persons actually make such choices. Traditionally, the existence of agents in agent-cause theories has always been treated conjecturally, as evinced in the writings of Roderick Chisholm (1976), Richard Taylor (1966) and Peter van Inwagen (1983). More recent agent-cause advocates such as William Rowe (1995), Timothy O’Connor (1995A), and Randolph Clarke (1995) give more details, but provide little reason to suppose that such choices really occur. Robert Kane’s event-cause account (1996) sketches how choices made at times of conflicting motives might coincide with the amplification of quantum indeterminism in persons’ brains, but Kane provides little reason to believe that such choices occur. Mark Balaguer (1999) argues that we have as much reason for believing choices are undetermined as we do for believing they are determined, because nobody knows how the brain works. But his argument has to discount the inductive evidence that every other macroscopic event we understand seems to be causally determined by previous events. Even if one accepts Balaguer’s point, it provides no more reason to think that brains make undermined choices than we have reason to think we make determined choices.

Immanuel Kant proclaims that we can have no epistemic justification for believing that persons make libertarian choices, but recommends that we postulate on faith alone the existence of trans-empirical selves ‘in’ a noumenal world who (that?) make such choices. William James (1962) provides a will-to-believe type of argument for believing that persons make libertarian choices, but his reasoning, like his argument regarding belief in God, is designed to operate in lieu of epistemic justification, not to provide it.

Libertarians’ confessed lack of epistemic justification that persons make libertarian choices raises the issue of moral hardness. By the libertarians’ own lights, are they being morally conscientious when they (1) adopt libertarianism as a theory or (2) put their belief in libertarianism into practice in their own treatment of persons? Because the former question is intellectual, it might seem that the issue of moral hardness pertains only to the latter, which deals with action. But the theory of libertarianism supports the practice of holding persons morally responsible. If the practices sanctioned by libertarianism are morally objectionable, the charge of lack of moral conscientiousness seems to apply to libertarian theory. Moreover, a theory can be hard-hearted even if no one ever acts on it, such as Descartes’s denial of animal suffering or certain theodicy-arguments.

Holding persons morally responsible includes a wide range of positive and negative behaviors: expressed reactive attitudes, verbal recrimination, praise and blame, retributive punishment and just-deserts rewards, all the way to eternal torment in Hell and bliss in Heaven. Libertarians disagree among themselves over how much of that range moral responsibility includes. But because even the mildest of the adverse behaviors harms persons, libertarians use the assignment of moral responsibility as a justificatory mantra that turns otherwise immoral treatment into just-deserts goods. Whatever the degree of harshness of the assignments of moral responsibility they select, libertarians agree that to assign those adverse aspects of moral responsibility to persons who lack libertarian free will is to act wrongly. Such reasoning, no doubt, lies behind many libertarians’ expressed contempt for compatibilism. By the same reasoning, then, libertarians need to provide a moral justification for visiting these evils upon persons. By their own lights, if libertarians are to hold persons responsible while avoiding the charge of treating persons immorally, libertarians should provide epistemic justification that persons actually make libertarian choices. The mere hope will not suffice.

How much epistemic justification that free will exists must libertarians provide to avoid the accusation of endorsing unwarranted recrimination, blame, and retribution? I believe it requires a great deal, well over a 0.5 probability, which is more than any of the above-cited libertarians claim exists. If we have insufficient epistemic justification for believing that persons satisfy the conditions we deem necessary for them to deserve assignments of moral responsibility, then, unless there are overwhelming consequentialist reasons for doing so (I discuss this later), we have insufficient moral justification for those assignments. Consider a simple case stipulated so that there are no consequentialist reasons for the assignment of responsibility: If I lack evidence that you committed an arson, then I should not blame you for committing that arson—even if you did commit the arson.

Robert Kane (1985, 1996, and in correspondence) distances himself from other libertarians by pointing out that Kanian free choosers have only partial control over their choices and, hence, are only partially morally responsible for their actions. Because Kane’s theory makes indeterministic choices depend on and coincide with indeterministic (and indeterminate) quantum events, he acknowledges that Kanian free persons lack control over their choices before they make them. Kane (1999) disagrees with his critics over whether losing antecedent control means losing control per se, which Kane admits would eliminate responsibility. Regardless of this dispute, Kane gives a plausible reason why Kanian free persons would merit a less strident kind of moral responsibility than would non-materialist Kantian trans-empirical selves or Cartesian souls. The latter are postulated as dictators of choices irrespective of the influences from their bodies, including all genetic and environmental influences. Traditional libertarians will dislike this implication of Kane’s view precisely because lessened control fails to underpin the strong type of responsibility that libertarians historically have wished to assign. It would be difficult to over-emphasize the importance to libertarians of justifying the practices of holding persons morally responsible.

Despite the plausibility of the claim that partial control merits partial responsibility, Kane at most mitigates this paper’s objection without rebutting it. Regardless of the strength of moral responsibility they endorse, all libertarians are committed to exonerating all persons of all moral responsibility if they came to believe that persons are determined. All libertarians who confess that they lack epistemic justification that anybody chooses indeterministically should exonerate everyone of moral responsibility, whether the most blood-chilling type of blame or Kane’s milder variety.

2. Three Unsatisfactory Replies: Digging in One’s Heels, Conceptual Schemes, and the Pragmatic Value of Believing in Libertarianism

A first reply by libertarians might be to insist that moral responsibility requires libertarian free will and that nothing could ever make them surrender the belief that persons make libertarian choices, despite their lack of epistemic justification. One could put a happy gloss on this by seeing it as intransigence for a good cause, but this would be bad epistemology. Surely the belief that persons make libertarian choices—an a posteriori claim that is acknowledged as unknowable by almost everyone—is a paradigm of a defeasible claim. This is a strategically inapt place to dig in one’s heels.

A second reply might be that libertarianism, although not epistemically warranted in the face of deterministic skepticism, is acceptable because of its centrality to our conceptual scheme as persons.2 If we are to see ourselves as persons who deliberate about our actions, then we must see ourselves as making libertarian dual-choices, even if we have no knowledge that we are the libertarian deliberators we take ourselves to be. Thus, the following are synthetic a priori truths; if we take ourselves to be persons, then we must take ourselves to make libertarian choices.

In response, I see little merit in “conceptual frame” arguments to support substantive philosophical theories. I concede to the libertarians that we cannot deliberate about choices we believe are going to occur irrespective of our choosing them, but to my knowledge no compatibilist has ever maintained that determinism entails this absurdity. As to whether deliberation requires two-way, metaphysically open libertarian choices, I think it is clear that the only openness we need is an epistemological openness in which we do not know what we will choose until we choose it (Kapitan, 1986). Clearly, this is consistent with determinism.

As a counter-balance to the above libertarian argument, consider Peter Strawson’s (1963) famous argument for the irrelevance of determinism to moral responsibility based on his non-libertarian analysis that reactive attitudes would persist irrespective of what we learned about the determinism of human behavior. Even if Strawson were right about this, a highly questionable assumption, the fact that both compatibilists and libertarians can use conceptual scheme arguments to support their contradictory views suggests caution regarding such arguments. 

A third reply might defend the claim that persons make libertarian choices, despite its lack of epistemic credentials, by claiming that our believing in libertarian free will and moral responsibility has salutary moral effects, a claim Kant shares with James. Of course, this claim is debatable armchair psychology. To the best of my knowledge, psychologists have not studied the results of the belief and non-belief in moral responsibility, so the whole issue seems to me conjectural. But so long as we are conjecturing, perhaps there are moral advantages to rejecting moral responsibility, as Saul Smilansky (1994) avers, or even advantages that make doing so morally superior to libertarianism, as B. F. Skinner (1948) and Bruce Waller (1990) claim. Libertarians are not obviously right on this conjecture.

But let us concede libertarians the dubious psychological premise and see if it helps to rebut my objection. Any libertarian who stresses the pragmatic value of believing in libertarianism is committed to consequentialist cost-benefit analysis. So, even if the psychological speculation were true, the benefit of inculcating better moral behavior among the public at large must be weighed against the possible cost of the unfairness of assigning moral responsibility to persons who do not deserve it. Instead of reveling in the hope that we might have libertarian freedom, a compassionate incompatibilist would consider the injustice we do to persons when we retributively punish them under the banner of libertarian free will and then consider the probability that we actually have such freedom. To ignore the probability that we have libertarian freedom when we adopt libertarianism is as confused as ignoring the probability that we will win the sweepstakes because we are inveigled by the size of the prize. Moreover, the belief in moral responsibility may do considerable damage even if persons generally are morally responsible: for example, retributive punishment dispensed to individuals who are mistakenly thought to have committed crimes.

Therefore, libertarians who endorse this pragmatic argument, by their own premises, find themselves in an unenviable position. They are committed to believing that an improbable moral good (encouraging better moral behavior at large) outweighs the probable moral evil of inflicting negative reactive attitudes, blame, and retributive punishment on undeserving victims. In sum, by their own lights, libertarians should not use the presumed pragmatic advantage of believing in libertarianism to justify imputing libertarian choices to persons. The probable cost of the pragmatic strategy should be enough to make any rational, morally sensitive libertarian reconsider using the pragmatic argument.

I find it odd for libertarians to do two things: (1) to condemn behavior modification as the sole rationale for punishment, because, following Kant, doing so treats persons as mere means, thereby doing them a moral injustice. Kant writes:

Judicial punishment can never be used merely as a means to promote some other good for the criminal himself or for civil society, but instead it must in all cases be imposed on him only on the grounds that he has committed a crime. (Kant, 1965, 100)

(2) to impute to persons libertarian free will (without epistemic justification) in order to make them act better morally. Both strategies of behavior modification and imputing libertarian free will have consequentialist motivations, and if the former strategy uses persons, so does the latter. Worse, there is a second degree of using persons operative when libertarians endorse the pragmatic strategy. Libertarian fans of the pragmatic argument impute libertarian free will to persons not simply to improve humanity’s moral conduct, which could be done just as well by endorsing compatibilism’s route to moral responsibility. Libertarians wish to improve moral conduct while holding on to their incompatibilist conviction that only libertarian free will underpins moral responsibility.

If “using” seems too strong, I could rephrase the last point. Libertarians underestimate the likelihood that their theory (incompatibilism and the belief in undetermined choices) is false and then discount the harm done if that theory is accepted and is false. If libertarians were softhearted, they would not risk hurting persons by imputing unknowable libertarian free will (and moral responsibility) to them simply on the strength of their libertarian vision. Fallibilism about one’s views is a desirable quality in general, but it is morally obligatory when dogmatism has potentially harmful repercussions for persons.

I can see two ways for libertarians to respond to my objection to the pragmatic argument. The first appeals to some barbarous moral thinking under the guise of an unjustifiable metaphysical premise. The second leads directly to compatibilism.

The first response would go like this. Libertarians could reply that if persons lack libertarian free will, then none of us have moral worth, and, hence, it does not matter if libertarians mistakenly subject persons to blame and retributive punishment. Libertarians might as well assume that persons exercise libertarian free will, because if persons do not, nobody can treat anybody immorally. One cannot treat a being lacking in moral worth immorally. (Kant thought our duties to beasts were only indirect: One must not mistreat beasts only because it may lead to mistreating humans.) A zany variant on this would be to insist that because “ought” implies a libertarian “can,” if libertarianism is false, no one ought to do otherwise than they do, which includes refraining from assigning unwarranted moral responsibility. The premise behind this response would be question begging against compatibilists. Worse, this response is unspeakably hard-hearted, using an unjustifiable philosophical premise to support a cruel moral position.

For a second response to my objection, consider Daniel Dennett’s (1984, 164) suggestion in Elbow Room:

we simply hold persons responsible for their conduct (within limits we take care not to examine too closely). And we are rewarded for adopting this strategy by the higher proportion of “responsible” behavior we thereby inculcate (1984, 164).

Dennett’s cavalier position appears heartless until Dennett adds to it the compatibilist premise that determined persons can be morally responsible. Until Dennett does this, his recommendation not to look closely at the causation of human actions will seem to overlook what incompatibilists see as the gross injustice of holding determined persons morally responsible. Thus, to avoid my objection to the Kant/James pragmatic line, Dennett needs to claim that determined persons could be responsible. As a compatibilist, Dennett happily endorses this premise, although, Dennett’s idea of moral responsibility is a milder variety than most libertarians endorse. Unfortunately for libertarians, accepting this premise would be to surrender libertarianism for compatibilism.

3. Three Replies Not Available to Libertarians

The following are reasonable responses to my argument, but each is unavailable to libertarians. First, libertarians might treat their theories as hypothetical: “I believe that if persons are morally responsible, they need to make libertarian choices; but because I have no epistemic justification for the claim that anyone makes such choices, I shall not hold anyone morally responsible until I find some.” Someone who says this combines conviction about incompatibilism with a reasonable skepticism about the likelihood of libertarian metaphysics, thus avoiding the worst excesses of dogmatism and heartlessness. The problem for the libertarian is that adopting this agnostic position would be a tentative acceptance of hard determinism. This position is contrary to what most historical libertarians want: If one’s libertarian theory recommends that one act like a hard determinist, then on epistemic and practical grounds, one probably should be a hard determinist. Erstwhile libertarians who adopted this move would, of course, leave open the possibility of endorsing libertarianism should new evidence for libertarian choices emerge.

   One variation on this theme would be for the libertarians to refuse to endorse the negative aspects of moral responsibility, but continue to endorse the positive ones. This proposal reminds one of the compatibilist Susan Wolf’s (1980) proposal that blame requires libertarian free will, but praise does not. The problem with this modulated proposal is that it is not consistent with the traditional libertarian aim of vindicating the whole range of morally responsibility. It would be tantamount to accepting hard determinism for the part of moral responsibility that is most important to most libertarians and compatibilism for the rest.

   Another variation on this theme suggested by Lee McCracken in correspondence is that libertarians might concede that human beings could never know whether persons make libertarian choices, but God would. So, libertarians might deny persons the right to hold persons morally responsible, but grant it to God. There are three things to note about this riposte. First, the result would be that libertarians should act like hard determinists, as noted above. This applies with even more force against libertarians such as Kane (1996, 81-89) who allege that an extremely wide array of qualities require libertarian free choices: not simply free will and moral responsibility, but creativity, autonomy, individuality, dignity, and worthiness of love and friendship. The more extensive the qualities the libertarians make depend on libertarian free will, the more extensive the qualities libertarians should withdraw in light of the epistemic problem.

Second, if we do not have epistemic justification that persons have free will, how could we have epistemic justification that God knows they do? An omniscient God would know all truths, but the question at stake is precisely whether persons make libertarian choices. Postulating that God would know who makes libertarian choices if they do is no more epistemically justified than our belief that persons make libertarian choices.

Third, I find it a curious metaphilosophical motivation to endorse a theory for which one has little epistemic justification, not to improve humanity’s lot (as do Kant and James, whose pragmatic argument I have criticized), but to exonerate a postulated entity. The theist might reply that ‘vindicating God’s justice,’ to paraphrase Leibniz, does serve humanity or, at least, makes some theists happy. If it difficult to see, though, the practical benefit to persons of being assured that if God exists, and if He discovers that persons make libertarian choices, then His punishment will be just.

   A second libertarian response would be to become agnostic about incompatibilism to defend the belief in responsibility. Libertarians might say that the belief in responsibility is so important that even if persons cannot be shown to be morally responsible by incompatibilist lights, they are willing to reconsider those lights and adopt compatibilism, at least provisionally. This modest position looks much like Alfred Mele’s “agnostic autonomist” (1995), but it also is not libertarianism.

Finally, erstwhile libertarians might defend responsibility by saying that both determined choices and undetermined choices can be free. [In Double 1996 and 1999 I call this Free-Will-Either-Way Theory.] But because this view endorses compatibilism’s positive claim that one can be morally responsible for determined choices, it, too, is not libertarianism.

4. Conclusion

Libertarians can answer this paper’s objection if they (1) reject incompatibilism, or (2) quit holding persons morally responsible, or (3) provide epistemic justification that persons make enough libertarian choices to support the practices of expressing reactive attitudes, blaming, and punishing retributively. Because libertarians cannot do either of the first two and remain libertarians, and seem unable to do the third, the three strategies are unpromising. All in all, the wisest and kindest strategy for libertarians would be to let go of moral responsibility—for the time being, at least.3


1 Some famous libertarians have offered arguments for libertarian free choices, but I have never found one that provided epistemic justification. C. A. Campbell (1957, 168-70) thinks we can tell by introspection that we make uncaused choices, but his view is epistemically worthless. My intuitive ‘feeling’ that I can choose in multiple ways given my complete mental state immediately preceding my choice is explained just as well by the hypothesis that there are determinants of my choice of which I am unaware as it is by the postulation of libertarian dual choices (Double, 1991). Sartre proclaims that “man is freedom,” but one cannot settle a substantive metaphysical problem by stipulation. Van Inwagen (1983, 206-9) deduces that persons make libertarian choices from the premises that (1) incompatibilism is true, and (2) we know we are morally responsible for our actions. Needless to say, both premises are highly disputable.

2 I owe this suggestion to Kenton Machina.

3 I thank Mark Balaguer, Robert Kane, Lee McCracken, and Bruce Waller for helpful comments on previous drafts of this paper. Also helpful were the members of the Philosophy faculties of Illinois State University and Illinois Wesleyan University to whom I read a version of this paper. Finally, I am grateful to the editor of this journal, Quentin Smith, for correcting several mistakes and infelicities.


Balaguer, M. (1999). Philosophical Studies. [vol. 93], 189-211.
Campbell, C. A. (1957). On Selfhood and Godhood. New York: Macmillan.
Chisholm, R. (1976). Person and Object. LaSalle, IL: Open Court.
Clarke, R. (1995). "Toward a Credible Agent-Causal Account of Free Will." In O’Connor, 1995, 201-15.
Dennett, D. (1984). Elbow Room. Cambridge, MA: MIT Press.
Double, R. (1991). “Determinism and the Experience of Freedom.” Pacific Philosophical Quarterly. [vol.72], 1-8.
Double, R. (1996). Metaphilosophy and Free Will. New York: Oxford University Press.
Double, R. (1999). Beginning Philosophy. New York: Oxford University Press.
Fischer, J. M. (1994). The Metaphysics of Free Will. Oxford: Blackwell.
James, W. (1962). "The Dilemma of Determinism."  In Essays on Faith and Morals. New York: World Publishing Company, 145-83.
Kane, R. (1985). Free Will and Values. Albany: SUNY Press.
Kane, R. (1996).  The Significance of Free Will. New York: Oxford University Press.
Kane, R. (1999). “Responsibility, Luck, and Chance: Reflections on Free Will and Libertarianism.” Journal of Philosophy. [vol. 96], 217-40.
Kant, I. (1965). The Metaphysical Elements of Justice. Indianapolis: Bobbs-Merrill.
Kapitan, T. (1983). “Deliberation and the Presumption of Open Alternatives.” Philosophical Quarterly, [vol. 40], 230-51.
Mele, A. (1995). Autonomous Agents. New York: Oxford University Press.
O'Connor, T., ed. (1995). Agents, Causes, and Events. New York: Oxford University Press.
O’Connor, T. (1995A). "Agent Causation." In O'Connor, 1995, 173-200.
Rowe, W. (1995). “Two Concepts of Freedom.” In O’Connor, 1995, 151-71.
Skinner, B. F. (1948). Walden Two. New York: Macmillan.
Smilansky, S. (1994). "The Ethical Advantages of Hard Determinism." Philosophy and Phenomenological Research. [vol. 54], 355-63.
Strawson, G. (1986). Freedom and Belief. Oxford: Oxford University Press.
Strawson, P. (1962). “Freedom and Resentment.” In Watson, G., ed., Free Will. New York: Oxford University Press, 59-80.
Taylor, R. (1966). Action and Purpose. Englewood Cliffs, NJ: Prentice-Hall.
Van Inwagen, P. (1983). An Essay on Free Will. Oxford: Oxford University Press.
Waller, B. (1990). Freedom Without Responsibility. Philadelphia: Temple University Press.
Wolf, S. (1986). “Asymmetrical Freedom.” In Fischer, J. M., ed., Moral Responsibility, Ithaca: Cornell University Press, 225-40.


This article was first published in Philo, Vol. 5, No. 2, pp. 226-234.

HOME to Det & Free website front page
HOME to T.H. front page