Short take on a book: Dennett vs. Caruso in “Just Deserts”

April 19, 2021 • 10:15 am

This is not a book for everyone, for it’s rather hard-core philosophy (albeit written in an accessible way), and is about one question: do we have free will or not? Since a lot of us have engaged in free-will debates here over the years, it’s appropriate for many of us. I’m really glad I read it.

And so to the Rumble in the Ivory Tower:

In one corner is Gregg Caruso, described on his page as “Professor of Philosophy at SUNY Corning, Visiting Fellow at the New College of the Humanities (NCH London), and Honorary Professor of Philosophy at Macquarie University. He is also Co-Director of the Justice Without Retribution Network housed at the University of Aberdeen School of Law.”

In the other corner is Dan Dennett, whom most of us know; he’s “the co-director of the Center for Cognitive Studies and the Austin B. Fletcher Professor of Philosophy at Tufts University.”

Both men have published extensively on free will. Caruso is a self-described “free will skeptic”; he thinks that because none of us can control our actions in a way that would change what we do at any given moment, we are not morally responsible for our acts, though we are “answerably responsible” or “causally responsible”. That is, if we do something good or bad, then we must be held accountable by society for our act in some way. Caruso adheres to a “pubic health” model of punishment: if you transgress, you are quarantined for possible cure and to keep you from hurting other people. You are not quarantined to deter others, as we don’t do that with carriers of infectious diseases. Ergo Gregg doesn’t see deterrence as a valid reason for “punishment” (or “quarantine”).  Caruso also sees no concept of “free will” that makes any sense, much less the historical one of “dualistic” free will—the one in which at any time we could have willed our choices and behaviors to be other than what we chose.

Dennett, like Caruso, is a determinist, agreeing that at any moment we have no free choice about what we do. However, he believes in a form of free will different from the traditional one; a form that, he argues, is the only kind of free will worth wanting. He thus sees his form of free will as compatible with determinism, so he’s a “compatibilist.”

What is Dennett’s form of free will? For him “freedom” consists of what we do when we’re members of the “Moral Agents Club”: that group of citizens who have been properly brought up and are responsive to reason and guidance by other responsible people. So for Dan, though free will isn’t “free” in the traditional sense, he sees it as “the concept of responsible, reliable self-control.”  In other words, people do things—make “choices”, if you will—that conform to the strictures of society. And so Dan says members of the Club have “moral responsibility.”

The screenshot below links to the Amazon order site.

I’ll briefly describe the Battle of the Heavyweights. You already know whose side I’m on! But let me say first that I greatly enjoyed the book, as it shows two top-notch philosophers arguing about a topic dear to my heart, and although the back and forth is civil (it’s a conversation, with each person writing between a paragraph and a few pages before the other person responds), it’s also hard-nosed, with each man querying and parrying the other, trying to find holes in their defense. 

As the title says, the argument is about “Just Deserts”, which to Dan means that people deserve to be praised or blamed for their actions because those actions are taken in a state of moral responsibility. Gregg sees no real reason for people to deserve their praise or blame, and so praise and blame must be allotted according to whether these actions help society or not (with some limitations). Blame should be limited, though, as it’s not really deserved; and “quarantine” rather than moral shaming is the best way to proceed.

In general, both Caruso and Dennett are consequentialists: they think the system of reward and (especially) punishment are largely justified by the consequences these systems have on society. To Dan, punishment is warranted by its effect on sequestering bad people and preventing them from hurting others, by its ability to help effect reformation of the criminal (if that’s possible), and to deter others from committing similar acts. Caruso, however, differs from both Dan and me in arguing that deterrence should not be a goal of punishment, because it uses people as means to control other people’s behavior, which he sees as fundamentally immoral. For example, one might say that in Dan (and my) society, even if someone is innocent of a bad deed that’s been committed, you might want to frame them to deter others from doing that deed. But that doesn’t seem right, does it? My answer would be that the consequences of punishing the innocent would be detrimental in general. But perhaps they need not be! The issue of deterrence is one I’m still thinking over.

So what is the difference between Dan’s views and Gregg’s? Gregg in fact spends almost all his time trying to answer that question, and he presses Dan on whether Dennett’s views are retributivist (which both men abhor: punishing someone simply to get back at them for bad deeds). But Dan sometimes comes close to saying that with his view of “moral responsibility”. At one point, frustrated by Dan’s apparent rapid changes of view during the conversation, Gregg compares Dan to a slippery eel. (There are moments of palpable frustration like this, though both guys behave civilly, like members of Dan’s Moral Agents Club.)

In the end, I would say Gregg won, simply because Dan doesn’t seem to make a good case for people deserving the punishment or praise they get just because they’re member of the “Moral Agents Club”. As Gregg (and I) have pointed out before, you have no choice about whether you’re a member of the Moral Agents Club: circumstances beyond your control have determined whether you are responsive to reasons and adhere to the social contract that makes you “morally responsible.” You might not have had the right upbringing, for instance.  Both Dan and Gregg agree, though, that strenuous prison reform is needed, and largely along similar lines. So to me, the debate either comes down to a difference in semantics or to an opacity of views on Dennett’s part that makes parsing his ideas very difficult.

But it’s great to see these two intellectual heavyweights slug it out. There are no knockouts, but I judge Caruso the winner on points.  And I have to do some thinking about deterrence. Right now I still think that deterrence is a valid aim of punishment.

Regardless of whether you’re a compatibilist or a free-will skeptic (or somewhere in the middle), this book will stimulate your thinking. Do read it if you’re interested in the free-will debate that’s occupied so much of our time. And I really do wish that we could have more debates like this: real back-and-forth conversations in more or less real time. That’s one reason I’m debating Adam Gopnik on whether science or its methods are the only way of gaining knowledge.

Oh, and after you read the book, you can vote on who you think made the best arguments; the voting site is here. Do not vote unless you’ve read the book!

36 thoughts on “Short take on a book: Dennett vs. Caruso in “Just Deserts”

  1. “Dan doesn’t seem to make a good case for people deserving the punishment or praise they get”

    The fact that you go ahead and praise them both for doing a good job with the topic is an example of what gnaws at those who are not quite willing to dismiss the entire concept of people deserving praise. Did they do well? Does that matter? Is your argument praise worthy?

    1. Good point. The objection is that consequentialist accounts such as Dan’s don’t give moral-realist accounts of “deserve” or “morally good” or whatever. That’s right, they don’t, there is no such thing as objective, moral-realist morality.

      All of these are pragmatic concepts that we use, as social animals, to influence each other. Why do we praise or condemn people? To influence their behaviour? When do they “deserve” praise or condemnation? When we want to influence their behaviour. That’s all there is to it. But we couldn’t live as social animals without those concepts.

    1. I have not read their debate book yet, but have read earlier works by both authors. I read your review Ed Gibney and I have two issues with it.

      1. You overstate “our shared capacity for learning” so much that it gets close to essentialism. Your claim that “Wherever we find examples of impairment in learning, or poor exposure to ‘good’ or ‘bad’ culture, we lower the judgments of moral responsibility for that person.” is not true. Many currently fare much worse than others in life due to being unlucky enough to have some learning capacity or other to a lesser degree that their peers. Myths about “self-made men” etcetera sprout from deeper myths about free-will based desert that are still very widespread.

      2. Regarding the manipulation argument. Your reply that “The evil neurosurgeons cannot be taught a lesson by punishing their puppets” is confusing. For how is that fact an objection? The genes and the laws of physics *also* cannot be taught a lesson by punishing their puppets. Symmetry thus remains. In both cases it is *unfair* to claim that the puppet morally *deserves* punishment for its unavoidable behaviour. (Acknowledging that unfairness still leaves open the further question if deterrence punishment, cleansed of any whiff of desert, is what we all things considered have strongest reason to pursue even when such a policy contains some unfairness.)

      1. Thanks for reading and responding, “determined”. (I wonder which side of this debate you fall on.) I’m happy to respond to the issues you raised.

        1. No essentialist thinking here. I regularly cite Dennett’s paper about “Darwin and the Overdue Demise of Essentialism”. (https://ase.tufts.edu/cogstud/dennett/papers/Demise_of_Essentialism.pdf) If you think humans don’t have neuroplasticity as a rule, you are ignoring basic facts of biology. As for your quibble about “Wherever we find examples of impairment…” I will grant that if you emphasise “wherever” then that sentence is too strong. But if you strengthen “impairment” the sentence is fine. We do/should/would lower judgments of moral responsibility for infants, Alzheimer’s, Down Syndrome, and children raised by wolves. That’s my point so I believe it stands.

        2a. “Genes and laws of physics” are at different levels of consideration than evil neurosurgeons. Would you say a car can’t drive because steel can’t drive? That would be an asymmetry. (And a category error.) I’m comparing people to people. I’m comparing free agent actors to puppet-masters. I don’t think that’s too confusing.

        2b. You are taking determinism too far when you say behaviour is *unavoidable*. That is fatalism. Read the book for more on Dennett’s “familiar concept of desert” which is all I would also advocate by saying someone deserves praise / blame / punishment / reward. It basically just means they are the locus of control for the behaviour that occurred and is being judged.

        2c. Dan also has a nice argument for deterrence based solely on consequentialist considerations. The rules of society are set in place for everyone. The universal application of those rules is what acts as the deterrence, not any single example of punishment. None of that requires whiffs of retributivism.

        By the way, I’m not convinced calling all of this “free will” is the right outcome. But until a good replacement comes along, I see why Dan and other compatibilists say it is better than telling everyone we don’t have any free will. There are important ideas, both right and wrong, that are bound up in the term as it stands.

        1. Thanks for discussing.

          1. “We do/should/would lower judgments of moral responsibility for [1] infants, [2] Alzheimer’s, [3] Down Syndrome, and [4] children raised by wolves.” Going that way with the sentence makes it true about what we currently do but at the cost of becoming insufficient to answer my complaint: Many people (billions!) not in groups 1-4 currently fare much worse than others in life due to being unlucky enough to have some learning capacity or other to a lesser degree that their peers. We do far too little about that. Dennett’s type of view, and yours too it seems, shrouds the unfairness of that in myths and jargon about “free will”, “shared capacity” and such.

          2a. You still misunderstand the manipulation argument. It draws vivid attention to empirical similarities in the cases namely that the puppet’s current behaviour has causes beyond its control. That sameness of situation supports a sameness in moral conclusion: in none of the cases can desert apply to the puppet. The “neurosurgeon’s cannot be taught a lesson …” bit doesn’t work as objection because (A) it doesn’t differentiate the cases (in none of them can the neurosurgeon – or any other external cause of the puppet’s behaviour – be taught a lesson) and (B) the reply is irrelevant anyway since lesson-teaching is not what the argument is about in the first place. Have you read Pereboom’s original formulation of the manipulation argument? It is well worth the time.

          2b. ‘Unavoidable’ in the sense that a behaviour that in fact occured in a particular situation was determined to happen in that way at that time, ultimately by the laws of physics. That’s not fatalism.

          2c. “None of that requires whiffs of retributivism.” … but big gulps of complacency about harmful unfairness (above). Dennett previous work on this topic has a long bad history of that. He has backtracked some of it in the last decade under debate pressure from free will skeptics like Harris and Waller. I’d be happily surprised if he backtracks even more in this tango with Caruso.

        2. Great discussion. I agree about your comment RE: neuroplasticity. I’m wondering… I know there have been loads of psych studies using genetically identical mice/rats. Have there been any studies using these types of lab animals that might support determinism?

  2. I’m reading the book but, so far, I’m a bit frustrated because they seem unable to focus down on a few differences. In other words, they seem to be mostly talking past each other, though in an interesting and enjoyable way. As requested, I will abstain from voting until I finish it. I guess I better get my butt in gear.

    1. Hi Paul, I would love to hear more about how you think they are talking past each-other, what are some examples of this? Really great observation!

      1. Actually, I spelled it out a bit more in another comment on this page. Here’s what I said:

        Dennett (rightly, IMHO) identifies control as the central difference between their points of view. Both accept that the control we think we have derives from the laws of physics via a causal chain. Caruso thinks that means we don’t have control but Dennett maintains we (most of us, anyway) have the control we need to make decisions and moral judgements. They never really resolve this here.

        Dennett is saying that we have all the free will we need. This is a somewhat glib way of acknowledging that each and every influence that goes into a decision we make has a causal chain going back to the laws of physics and the Big Bang but that it just doesn’t matter to how free will works. As I see it, all Incompatibilists disagree with this fundamental idea. Caruso dances around this difference but never seems to come to grips with it. I suppose it is just as much Dennett’s fault for not focusing the discussion on this issue. Anyway, that’s how I read it.

  3. I’ve not read the book, but:

    Caruso … thinks that because none of us can control our actions in a way that would change what we do at any given moment, we are not morally responsible for our acts, …

    Followed by:

    Caruso, however, differs from both Dan and me in arguing that deterrence should not be a goal of punishment, because it uses people as means to control other people’s behavior, which he sees as fundamentally immoral.

    Isn’t that self defeating, to do away with morality, and then argue for a position on moral grounds?

    And yes, of course we try to control other people’s behaviour. That’s what social interactions are all about! Also:

    … we are not morally responsible for our acts, though we are “answerably responsible” or “causally responsible”. That is, if we do something good or bad, then we must be held accountable by society for our act in some way.

    But that’s all that “moral” responsibility is, that others want to hold you accountable. I guess you might want more to it than that if you were a theologian, but really “morality” is just a pragmatic matter of approving or disapproving of modes of behaviour as a means of influencing how often they occur.

      1. If we interpret the latter as:

        “you did the right thing [the thing that conforms to approved social norms] when, had circumstances and influences been slightly different, you might have done the wrong thing [something contrary to societal norms], and the point of me saying this is in order to add to the influences reinforcing societal norms, thus reducing the prevalence of the wrong thing.”

        … then the two are pretty much the same.

    1. “And yes, of course we try to control other people’s behaviour. That’s what social interactions are all about! ”

      Yes, it is true that we try to control other people’s behaviour. However, the argument is about whether it is legitimate to instrumentalize people (i.e. use them as tools) to control other people’s behaviour.

  4. I’m not prepared to forsake deterrence as a legitimate penological goal, but I will confess that I’ve long thought it to be in tension with the tenets of secular humanism — viz., that one ought not to use the incarceration of our fellow human beings as a tool to “send a message” (particularly a message that tends to be received through a glass darkly, if at all) to our general societal population.

  5. Interesting discussion! I disagree with Caruso about deterrence, if used wisely. The greater good wins out for my utilitarian tendencies. And I do think Dennett is onto something with the “Moral Agents Club” notion as a way of connecting individual determinism with the greater community that influences the “non-choices” that the individual makes. Perhaps a useful vehicle for improving those choices in the future.

  6. The case discussed here last week of the Mohammedan who had marijuana & murdered a poor woman – is that affected by this debate? The effect of drugs?

  7. I agree with Dennett though I found nothing in this book that came close to nudging me one way or the other. Dennett (rightly, IMHO) identifies control as the central difference between their points of view. Both accept that the control we think we have derives from the laws of physics via a causal chain. Caruso thinks that means we don’t have control but Dennett maintains we (most of us, anyway) have the control we need to make decisions and moral judgements. They never really resolve this here.

    I read books like this one and realize once again that although I like philosophy generally, I don’t at all enjoy the kind of discussion they had in this book. There’s just too much jargon. I understand why philosophers use it but, as in this book, it allows them to spend a lot of time arguing about the words to the expense of the concepts and making progress.

  8. I have read the book …click rom to read the review.

    Overall I don’t think there was a winner … ultimately it was battle that:
    “Dan and Greg in their book Just Deserts have decided to engage one another and their readers in the quagmire of perception.”

  9. you have no choice about whether you’re a member of the Moral Agents Club

    That’s not quite right. You have no choice about whether to enter the Moral Agents Club, would be close to the truth. (You have no choice about whether to start to enter, would be the more exact truth.) But you have a choice whether to exit. You could commit suicide. You could damage your brain extensively. Maybe you could deliberately drive yourself insane, or less deliberately, choose to remain in a situation that will eventually drive you insane. Of course, almost nobody wants to do these things, because meeting the qualifications of the Club comes with lots of (what most people see as) benefits.

    1. “But you have a choice whether to exit.”

      Well… you might exit that way. That does not mean you actually have a choice in the matter.

      1. I know there are arguments that purport to show that all “choices” are illusory. But those arguments are based on erroneous physics or bad logic. In the case of the Manipulation Argument, a bit of both.

        1. The point is that the possibility that your day could end in suicide is not support for free will (which I assume you mean by “choice”).

          1. My point is – as illustrated by the actions necessary to stay alive or to die – most people assess responsibility based on the agent’s abilities at the time, without a need to inquire about the agent’s childhood. If you think this is a mistake, you need some argument to show why such deep history is crucial – preferably one that doesn’t start from false premises. Your side is the one arguing for more extensive changes in society’s practices. At least practically speaking, that puts a greater burden of argument on you.

            (Intellectually speaking, it does so as well. But that point would take us into philosophy-of-language weeds, reflecting on how words – including “responsibility” – get their meanings.)

            1. I don’t think I have a “side” here, at least in the argument you seem to be trying to foist on me. All I’m point out is that there is a logical gap in your case. The fact of a possible future outcome does not necessarily mean choice exists. It only means that you don’t know the outcome ahead of time. You’re conflating the absence of certainty about the future with free will.

              1. I have failed to explain my point, but I will give it one more try, and you can have the last word. “My case” is not my case. Notice that I never used the term “free will” above. Instead it’s about responsibility, and Dennett’s Moral Agent Club. Members are considered responsible if they had a “choice” about something – you can put scare-quotes around “choice” if you like – and their action was morally significant. Nobody but philosophers (or those heavily influenced by them) thinks it is necessary to go into the deep childhood history of a moral agent if it is clear that they are currently at human-normal levels of intelligence and rationality. This is not offered as “proof” of moral responsibility, much less of free will. But it establishes a burden of argument regarding moral responsibility – at least, as a practical matter.

  10. All you need to do to make “compatibilism”compatible with hard determinism is to stop calling “freedom” to what simply is “responsiveness to reason”.

    In the physical world there are only deterministic laws and randomness. In the mental world, only reason or unreason (or less reason). There are degrees of reason, not degrees of freedom.

    1. And yet we use the concept “degrees of freedom” in science and engineering all the time, and we use “free” and “freedom” in everyday life. (A prisoner released from jail now has “freedom” — that is not asserting that he is no longer subject to the laws of physics, but is saying he can now wander around where he wishes to, he has greater “freedom”.)

      So the better way of adopting compatibilism is simply to interpret words like “freedom” and “choice” in line with what they do actually mean in everyday life.

      1. In philosophical discussions about free will, if you agree that what we have is actually “responsiveness to reason”, you should call it “responsiveness to reason”.

      2. I agree. No one is disputing (or should dispute) that our choices are based on brain activity (rather than ghosts in the machine), which is subject to the laws of physics- whether deterministic or not- and ultimately rooted in genes and environment. But there is still a world of difference between choices that are made voluntarily (‘freely’)- that is, in accordance with our desires (giving someone money as a gift because we want to) and those that are made under duress- that is, contrary to what we desire or under pressure from external agents or forces (giving money to a thug at gunpoint).

        Compatibilism recognizes this meaningful distinction- one we routinely make in everyday life, as reflected in our language and social interactions. So, of course, there is ‘free will’, but it’s not something that is exempt from the laws of physics.

        And I agree again with Coel about “morality” being ‘just a pragmatic matter of approving or disapproving of modes of behaviour as a means of influencing how often they occur.’ Just think about training your dog (good dog/bad dog!). We call things or people ‘good’ if we approve of them, and ‘bad’ if we do not. Apart from that, there is no such thing as objective morality.

  11. “As Gregg (and I) have pointed out before, you have no choice about whether you’re a member of the Moral Agents Club: circumstances beyond your control have determined whether you are responsive to reasons and adhere to the social contract that makes you ‘morally responsible.’ You might not have had the right upbringing, for instance.”

    Right. Some folks end up morally competent (know right from wrong, are basically rational) but also morally flawed: they habitually or occasionally do the wrong thing. Are they blameworthy for ending up that way? Dennett says yes since most biologically normal humans develop self-making capacities and can be held responsible – blamed – for not exercising them correctly. Even those growing up in tough environments have a good chance, he says, to eventually possess sufficient competence for correct self-formation: “…it is worth reminding ourselves that in some cases – maybe most cases – the very hardships and injustices and assaults they endured hastened their achievement of self-control and responsibility.” (p. 74) Data from many sources flatly contradict this claim: trauma, violence, and abuse in childhood and adolescence often compromise the brain-based capacities for impulse control while modeling anti-social behavioral styles often adopted by victims as coping strategies. Are such individuals really to blame for their deficits in self-control and for having imbibed defective moral norms? I don’t think so, and we can see here an instance of how Dennett routinely downplays or ignores the deterministic causal story of individual development in order to defend desert. Too bad. That said, I think he’s right that punishment as deterrence is sometimes an unfortunate necessity. It can work for those who have something to lose, who are aware of the penalties, and have the capacity and opportunity to reflectively consider the possible consequences of wrongdoing, e.g., potential white collar criminals.

    Just Deserts is reviewed at https://naturalism.org/resources/book-reviews/responsibility-in-question-caruso-and-dennetts-just-deserts

    1. Dennett understands that society decides the criteria by which its citizens are to be held morally responsible. It’s a subject for discussion and decision. The jury is out on how much abuse or neglect children must undergo before they are not to be held responsible for their actions. You and Dennett may disagree on precisely where to draw the line but this doesn’t have much to do with the free will discussion.

  12. Deterrence works, it makes society work. Nowhere in the world is there a nation, a tribe, that can do without a punitive system, nor is it known that such a thing ever existed in times past.
    As a social tool, deterrence works, of course, at the expense of the individual who is imprisoned and segregated for the rule-breaking he has committed, even though he was not could not have done otherwise than break the rule.
    But is this segregation of the condemned individual really so bad?
    A biting dog is segregated in the shelter, although it cannot do anything for its actions.
    A dog infected with rabies is killed even though it is not responsible for its infection and its danger to others.
    An attacker who threatens to kill people may be killed by the police in a dangerous situation, although the attacker could not have acted otherwise.

    As long as there is no substitute for the tool of deterrence, all societies will continue to use it. This does not mean that there will be no approaches of modification; the more prosperous and the more enlightened a country is in terms of scientific knowledge, the more modern the penal system will be ( as in Norway for instance).

Leave a Reply to Ed Gibney Cancel reply