Do free will skeptics swallow their own medicine?

  • English

(Book review of Daniel C. Dennett and Gregg D. Caruso: Just deserts. Debating free will, Cambridge: Polity, 2021, 223 pp, $15.99 PB. A slightly shortened version was published in Metascience)

How can we ever hope to resolve philosophical questions with which some of our smartest thinkers have wrestled for centuries, with no agreed resolution in sight? Perhaps we should delegate two of our leading philosophers, each representing one of the opposing views, and allow them to fight it out in public. Not a quick duel on the stage, but a long and extensive dialogue, in which they tread carefully, step by step, resolving ambiguities and misunderstanding as they go along. That is exactly what Daniel Dennett and Gregg Caruso have tried to do with the question of free will. The result is Just Deserts, an in-depth dialogue that runs to over 200 pages. Caruso and Dennett represent the two main camps in the modern free-will debate: those who think that free will is compatible with determinism and those who do not. Although, despite their best efforts, neither philosopher succeeds in persuading the other, their book is exemplary as far as philosophical dialogues go. Both thinkers have respect for each other, they are civil yet passionate about their beliefs and do not pull any punches, and they make a genuine effort to understand how on earth the other side comes to hold the views they do. Their sense of bafflement with each other’s position is genuine and mutual.

To this day, even the very definition of “free will” remains contentious. Believers and skeptics both try to appropriate the concept, accusing each other of obfuscating and changing the subject. The debate in this book follows a recurring pattern, which shows how deep the conceptual quagmire runs. Time and again, Dennett offers a pragmatic and down-to-earth version of free will and related concepts, but Caruso insists that it fails to be this other, more basic, more absolute, intrinsic, metaphysical concept. And each time, Dennett responds that this other thing is a chimera, a will-o-the-wisp, a metaphysical phantom that we do not need anyhow. For instance, Dennett defends a homely and everyday version of just deserts (e.g., the punishment of fouls in sports), but Caruso objects that it is not basic desert. There is the free will that Dennett deems worth wanting, but then there is the elusive metaphysical freedom that is ruled out by determinism. Punishing wrongdoers is proper and necessary to uphold respect for the law, argues Dennett; but is it “intrinsically good” that they suffer, asks Caruso (42)? Rational agents are causally responsible for most of their actions, but are they ever ultimately responsible for anything they do (29)?

Disagree to agree

Ironically, in pretty much all of these cases, both Dennett and Caruso wholeheartedly agree that people do not have the more intrinsic, ultimate, fundamental, metaphysical version of freedom or responsibility. Except Dennett believes these notions to be hopelessly muddled and incoherent anyway, a form of “free will inflation” (Dennett, 2020), whereas Caruso thinks that we could (and should?) have had them but we do not, and this changes everything. For his part, Caruso also grants Dennett virtually all his pragmatic and practical distinctions, about degrees of autonomy and freedom, responsiveness to reasons, difference between adults and children or demented people. And yet, says Caruso, it’s not quite basic desert, it’s never ultimate control.

Dennett and Caruso end up not so much “agreeing to disagree”, but something more akin to “disagreeing to agree”. They agree about a lot of things, but they disagree about what this agreement entails. At several points, Caruso tries to entice his opponent to his side, arguing that Dennett’s position is “much closer to that of the free-will skeptic” than he acknowledges (27, 41, 194). And indeed, Dennett is equally skeptical about all the things Caruso is skeptical about. It is just that Dennett believes the skeptic is denying something that is not even worth denying, because it could not possibly exist.

But what difference does all this make to our morality? Thankfully, Just Deserts does not linger for too long on the threadbare metaphysical arguments about determinism and choice, but spends most of its pages on moral and legal implications. For both thinkers, a lot is at stake: Caruso believes that giving up on free will and moral responsibility would make for a much better world, while Dennett believes that it would be disastrous (26).

Legal reform

At first sight Caruso’s public-health quarantine model for criminal behavior (Caruso, 2021) looks quite revolutionary, since it not only discards the idea that criminals are ever morally responsible for their crimes, but also does away with the notion of punishment, a step which not all free will skeptics are willing to take. The only reason why the state may incarcerate or incapacitate criminals, according to Caruso, is out of self-defense, not because the incarcerated deserve it, and not even to deter further wrongdoing. By analogy, if we quarantine those infected with COVID-19, we intend no harm to them and are not even expressing moral disapproval.

Pushing back against Caruso, Dennett argues that such a model would always need a punitive stick behind the door. But Caruso ably defends his thesis that, strictly speaking, enforcement of quarantine does not necessitate punishment. It may just require more effective incapacitation, with as little inconvenience as possible. When a tiger escapes from the zoo, you do not punish it.  You just erect a higher fence. Likewise, although mandatory quarantines for COVID-19 are often enforced with punishments for non-compliance, such enforcement is not strictly necessary. If someone breaks his quarantine, try to reason with him. If he will not be persuaded, just lock him up in a hotel room. If he trashes the room and deliberately coughs in the face of the hotel staff bringing him food, push the food through a hatch. If he refuses to pay for the damage, too bad.

The question is not whether treating humans like zoo animals is theoretically possible, but whether even Caruso would do it if push comes to shove. Moreover, what if we discovered more effective ways for protecting society against criminality apart from incapacitation? In public health, quarantines are lifted as soon as the infection risk has subsided, since, as Caruso rightly points out, we are not punishing people for being infectious. At some point, Dennett brings up the thought experiment of the perfect anti-crime pill, which extinguishes all criminal tendencies with no side-effects. Caruso bravely bites the bullet: if such a pill were to exist, there would not be any reason for society to incapacitate or otherwise inconvenience murderers for even a single day. In such a desert-less utopia, all murderers would go scot-free; at least for one crime, after which their criminal tendencies would be extinguished. As Dennett says, however, in such a world our laws would neither command respect nor deserve it (124). 

The trouble is that the sheer radicalism of Caruso’s proposal is conveniently disguised by our lack of effective protection against recidivism. With the current state of technology, Caruso’s model would still include a form of incapacitation and other decidedly unpleasant measures — all of which would still feel like punishment even though Caruso can plausibly deny that they are. “Prisons” would still exist in all but name. And if incapacitation happened to deter crime, that is welcomed by Caruso as an unintended “natural side-effect” (147). Quite convenient!


Even more tellingly, Caruso seems to smuggle some desert-like notions through the back door by stipulating that we can only incapacitate people after they have committed a crime, not when they are only about to commit one (133), and that we are never allowed to frame an “innocent” person, even if this would protect the community against a grave threat: “it would be wrong to incapacitate an innocent person because that person, being innocent, is not a danger to society” (187). But the very notion of “innocence”, with which Caruso is much concerned (see also Caruso, 2021, section 8.4) does not make much sense in a world without just desert. After all, we also quarantine people who might be infected.  When you arrive at Auckland airport these days, you will not be treated with a presumption of epidemiological innocence.

If the only grounds for incapacitation are self-protection, then why should we not incapacitate those strongly suspected of having committed a crime, or for that matter, those who are about to commit a crime but have not done so yet, like in Steven Spielberg’s movie Minority Report? In some circumstances, someone who has already committed a crime might pose less of a threat to society than someone who has not actually broken the law yet. Compare the following cases: Walter is a peaceful and law-abiding man who one fateful night bludgeoned his wife’s lover in a fit of jealous rage when catching the two of them in flagrante delicto, and then was immediately stricken with remorse; Donald is a sadistic psychopath who has not murdered anyone yet but is just biding his time. Given that Caruso does not believe anyone deserves to be incapacitated in the first place — for him, it is purely a public health intervention — it is unclear why locking up Donald would be morally more objectionable than incarcerating Walter. Unless, that is, you believe in some sort of desert. It seems as if Caruso is recoiling from the radical implications of his own views.

Taking a dose of your own medicine

It is one thing to bite the bullet of a thought experiment about a non-existent legal system, but quite another to live up to the implications of a radical philosophy in real life. I found myself wondering: what exactly are skeptics like Caruso doing differently in their own lives, now that they no longer believe in moral responsibility? Has Caruso stopped praising people for their accomplishments, or blaming them for their mistakes? Or, alternatively, has he continued praising and blaming like the rest of us, only silently muttering to himself that of course no-one ever truly deserves any praise or damnation?

I’m not suggesting that free-will skepticism commits Caruso to the “objective attitude” towards human beings that Peter Strawson (1962) famously opposed to the “reactive attitudes” of resentment, gratitude, and indignation. Even in Caruso’s utopia, there is still a difference between agents who are accountable to reasons and those who aren’t. There are still moral emotions and sentiments.

But suppose that Dennett had swindled Caruso out of his half of the royalties for Just Deserts, tricking him into thinking that everything would be shared evenly. Imagine that, with a diabolical twist, Dennett had done so precisely because he reckoned that his co-author would not hold him morally responsible because he does not believe in moral responsibility anyway. Finally, thought Dennett, here is someone I can swindle and get away with it! Gregg would be disappointed and sad, of course, but he would understand that Dan, just like everybody else, is merely a puppet of causal factors over which he ultimately has no control whatsoever.

Or would he? I bet that Caruso would be just as resentful and angry as the rest of us, and would think that Dennett deserved public opprobrium for this dishonesty. At this point, Caruso might respond that he is only human and that it takes effort to truly wean yourself off the all-too-human belief in moral responsibility (50). But I suspect that it just cannot be done. Even after careful reflection, Caruso would still be making distinctions to which he is not entitled as a free-will skeptic, apportioning blame and credit just like we all do. At most, he might add a silent disclaimer, along the lines of “But of course Dan did not truly and intrinsically deserve it”. But what difference would that make in real life?

If my suspicion is correct that even the staunchest free will skeptic will merely talk the talk and never walk the walk, this may also assuage Dennett’s worries about free-will skepticism “return[ing] humanity to … Hobbes’s state of nature” (26). I think moral responsibility is too deeply ingrained in our thinking for anything like that to happen. When free-will skeptics take a break from their speculations and settle down to play a game of backgammon, they will be just as upset if someone cheats.


Caruso, G. D. (2021). Rejecting Retributivism: Free Will, Punishment, and Criminal Justice: Cambridge University Press.

Dennett, D. C. (2020). Herding Cats and Free Will Inflation. Romanell Lecture for American Philosophical Association.

Strawson, P. F. (1962/2008). Freedom and resentment and other essays. Routledge.