A new paper says that belief in free will makes people more likely to criticize unethical acts and favor strong punishment

People who are determinists but are also compatibilists—that is, they believe that our actions are dictated by the laws of physics, and we can’t “will” them beyond that, but that we can still conceive of some form of free will—often justify their re-framing of “free will” on utilitarian grounds. That is, although we may not have “contracausal” free will (the kind that says that, at any given moment, we could have decided to do two or more different things), it’s still important for society to believe in some form of free will, for such a belief acts as an important social glue. Without a belief in agency, these folks say (Dan Dennett is one of them), society will fall apart, as everyone will just lie in bed and do nothing because determinism makes them nihilists. That argument resembles the old Argument from Utility for God; even if there’s no god, it’s good for society to believe in one, for it helps people behave morally. It’s also been called the “Little People” argument: although I know the truth, we need to let the Little People have their belief in contracausal free will.

Both arguments are wrong. Even though we feel and act as if we have contracausal free will, we can still accept rationally that we don’t, and make the social reforms mandated by that realization—especially reform of the judicial system. We don’t need Little People arguments to have a good society.

One of the famous papers used to justify compatibilism was published by Kathleen Vohs and Jonathan Schooler in Psychological Science, “The value of believing in free will: Encouraging a belief in determinism increases cheating.” But that paper is problematic. Besides its design flaws (i.e., “cheating” was tested shortly after students read passages either promoting or denigrating free will, with no long-term monitoring of behavior), it’s also failed to be replicated at least twice (see here,  here and here). And there’s at least one paper showing that accepting determinism makes you more empathic and less vindictive, which isn’t that surprising if you don’t think people are able to “decide’ whether to do good or bad things.

My own view is that while these studies are interesting, we need to concentrate less on testing the effects of free will on people’s views or on confecting new semantic arguments for free will, and more on working out the consequences and implications of determinism for society, for we know that determinism, (or, if you include quantum effects, “naturalism”) happens to be true, and we should accept it and work from there.

But the papers keep coming out. There’s a new one in PNAS by Nathan D. Martin, Davide Rigoni, and Kathleen Vohs (the same Vohs) that examines the effects of believing in free will on attitudes towards unethical acts and the need for “strong punishment”. (Reference and free link below, be sure you have the legal “Unpaywall” extension.)

I’ll try to be brief. Martin et al. used existing data from the World Values Survey, which asks questions not only about free will (the effect of “fate”), but whether certain actions are considered unethical and whether “criminals should be severely punished” (how severely is not specified). They also correlated the results with the nature of the country surveyed (65,111 people were surveyed from 46 countries): was the country one with a high “corruption index”, and was it a democracy?  They then did a massive correlational analysis to see what effects various demographic variables, as well as belief in contracausal free will, had on attitudes towards ethics and criminal punishment.

Here are the data, with the left and right sides showing the regression of either intolerance for unethicality (left side) or support for strong criminal punishment (right side) on some variables; the bottom part of the graph (“B”) is better as it takes care of other variables correlated with the two items of interest:


Considering the regression coefficients (β) and the probabilities of getting them by chance alone (those with asterisks are considered statistically significant), you find these main results:

a.) Belief in free will is positively and significantly correlated with both intolerance for unethicality and support for criminal punishment; the association is stronger for the latter. This isn’t surprising: if you think people can choose to behave good or badly, then you would be more prone to pass moral judgment of them (be intolerant of their acts) and support punishment for them. (The authors note from other studies that the punishment is often seen as “retributive”: not to deter others or sequester criminals for “curing” and removal from society, but simply because criminals deserve to be punished because they made the wrong choice.

The regression for disapprobation of unethical acts, is pretty small, and of borderline significance. In fact, factors like gender, age, education, and religion are at least as strong. (As you’d expect, religious belief is positively associated with both disapprobation of unethical acts and calls for severe punishment; that’s likely because the Abrahamic religions are explicitly dualistic, favoring contracausal free will.)

b.) The relationship between belief in free will and support for criminal punishment holds across all societies, but the relationship for “intolerance of unethicality” holds only in countries that have “institutional integrity”—those where corruption is uncommon. Examples of high-integrity societies (determined in other analyses) include Chile, Japan, and Spain; medium-integrity societies are Poland, South Korea, and Trinidad and Tobago; and low-integrity societies are Iran, Moldova, and Rwanda.

Again, this is unsurprising; in all societies, people who accept contracausal free will should call for retributive punishment; but you can understand why those who live in corrupt societies might go a bit easier on those who commit unethical acts, for that’s how you get along there. As the authors note:

Institutional integrity reflects the extent to which countries’ public sectors are free of corruption and maintain strong, transparent governance. Among residents of countries with average to high institutional integrity, stronger free will beliefs predicted stronger intolerance of unethical behavior. However, in countries with widespread corruption and lax governance, whether people slightly or strongly endorsed free will beliefs was decoupled from their attitudes toward unethical action. In these countries, unethical behavior could be attributed to external circumstances or viewed as a rational strategy rather than a reflection of moral character.

c.) The authors construe this as strong support for the Vohs and Schooler paper and its conclusion that belief in free will makes you less likely to cheat.  While the present paper doesn’t really say anything about people’s beliefs and tendency to act unethically, it does show a connection between free will and how you judge the behavior of other people. And since it’s based on a huge sample of people from many countries, it shows that a connection between free will and punishment or disapprobation holds pretty widely, especially in societies of “integrity.” I don’t doubt their conclusion, as it would seem to follow naturally from realizing that contracausal free will makes you see other people’s actions as a result of free choice, and thinking that they could have done otherwise.

What’s the upshot? The authors, mercifully, don’t even come close to saying that we should believe in free will because it’s good for society. All they do is point this out:

Belief in free will might seem esoteric, unworthy of scientific study, or academic (in the pejorative sense). It is not. Attesting to its widespread impact, our global analysis found that the more that people endorsed notions supporting free will beliefs, the harsher their attitudes toward wrong-doing and wrong-doers, with one notable exception. For residents of countries with corrupt and ineffectual public sectors, free will beliefs did not bear on judgments of unethical actions but nevertheless predicted preferences that criminals receive harsh punishments. The influence of free will beliefs in people around the world, along with the moderating influence of countries’ institutional integrity, provides evidence that seeing one’s own and others’ actions as reflecting personal choice, accountability, and self-determination can broadly affect moral attitudes and judgments.

Well, I have no problem with that; it’s a statement of fact. But what you do with that fact is disputable. Seeing that the “free will” associated with moral judgment is true contracausal free will, we already know that that belief is simply wrong. And I feel that basing social policy on something that’s known to be wrong is, in the long run, bad.

Naturalism and physical causation of behavior are true, and we must accept that. When we do, then we can deal with its supposed inimical effects. I for one think that a view of determinism (or naturalism) behind people’s behavior will be good for societies in the long run. I’m a naturalist and determinist, and I don’t lie in bed every day wondering what is the point of getting up. And yes, I feel as if I make free choices. But in my heart (or brain) I know that I don’t. This realization, to me, is one that can have immense and salubrious effects on the criminal justice system, and Robert Sapolsky, who’s way smarter than I, agrees. But I’ve already discussed that, and won’t go into it here.

h/t: Karl


Martin, N. D., D. Rigoni, and K. D. Vohs. 2017. Free will beliefs predict attitudes toward unethical behavior and criminal punishment. Proceedings of the National Academy of Sciences 114:7325-7330.


  1. Gregory Kusnick
    Posted July 16, 2017 at 11:03 am | Permalink

    It’s also been called the “Little People” argument: although I know the truth, we need to let the Little People have their belief in contracausal free will.

    This is certainly not Dennett’s view. He thinks people should be educated about a naturalistic view of agency, rather than being told (falsely, in his opinion) that agency can’t exist “because physics”. Nowhere (not even in the Erasmus essay) does he advocate the promotion of contracausal free will to keep the Little People in line.

    • Posted July 16, 2017 at 11:11 am | Permalink

      I didn’t say it was; I said that Dennett said that belief in agency is necessary for society. Have a look at the quotes at the link:

      There is—and has always been—an arms race between persuaders and their targets or intended victims, and folklore is full of tales of innocents being taken in by the blandishments of sharp talkers. This folklore is part of the defense we pass on to our children, so that they will become adept at guarding against it. We don’t want our children to become puppets! If neuroscientists are saying that it is no use—we are all already puppets, controlled by the environment, they are making a big, and potentially harmful, mistake.

      . . . the deep conviction Erasmus and I share: we both believe that the doctrine that free will is an illusion is likely to have profoundly unfortunate social consequences if not rebutted forcefully.

      Well, most people conceive of free will as contracausal free will, so this is pretty damn close to the Little People’s argument. Dennett spends a lot more time writing about how we can have a form of free will (two books) than about how we should deal societally with determinism.

      But granted: I should have said that the Little People argument runs that people should be able to have a belief in either contracausal free will or some kind of “free will.”

      • Gregory Kusnick
        Posted July 16, 2017 at 12:55 pm | Permalink

        If we observe that people make worse decisions if they think their decisions don’t matter, would that be a Little People argument? That’s the harm Dennett is worried about. What false belief is he promoting by expressing that concern?

    • Vaal
      Posted July 16, 2017 at 11:56 am | Permalink

      Agreed, Gregory. Even if it were settled most people think free will is contra-causal (hardly settled), Dennett spends much of his philosophical time rebutting the very idea of contra-causal free will – to the “little people” public as much as to other philosophers (re, his books).

      He’s trying to correct a mistake, promote what he thinks is true, not a “necessary falsehood.” He simply isn’t make a Little People argument, full stop.

    • Posted July 16, 2017 at 7:11 pm | Permalink

      Also, it’s not Dennett’s view that “everyone will just lie in bed and do nothing because determinism makes them nihilists.” Rather, incompatibilism is the most dispensable thing that makes them nihilists. And Dennett wants to dispense with it.

  2. Tom
    Posted July 16, 2017 at 11:39 am | Permalink

    Could “free will” simply be the capacity to believe?
    People change their beliefs for all sorts of reasons, something strikes a chord and a new belief is generated. But is this a conscious decision or has the inner world of neurons merely found something a little more “exciting” that generates new associations and ways of preserving the well being and even the life?

    • Kevin
      Posted July 16, 2017 at 10:55 pm | Permalink

      Belief is probably a necessary ingredient.

  3. rickflick
    Posted July 16, 2017 at 11:53 am | Permalink

    When I access the paper, I get this error:

    “The Unpaywall extension couldn’t find any legal open-access version of this article.”

    Has anyone else gotten this message?

    • Michael Fisher
      Posted July 16, 2017 at 12:05 pm | Permalink


      Yes – same message
      [unpaywall 1.5 on firefox]

      • Posted July 16, 2017 at 12:07 pm | Permalink

        It might work better with Chrome. I accessed it this morning. If you want the paper, just email me.

  4. Heather Hastie
    Posted July 16, 2017 at 12:12 pm | Permalink

    For the last few years I’ve been one off the people surveyed annually in the NZ Attitudes and Values Survey. NZ is also currently 1= in the world in the International Transparency Index (lack of corruption) with Denmark. Also, I’m a determinist.

    I see the results slightly differently. I’m intolerant of corruption because that’s a value of the society I live in. There is no need nor excuse for it here. Institutions, including the Police, can largely be trusted. We grow up in that atmosphere.

    I’m also opposed to harsh punishments. That’s not just because I’m a determinist. Finding out about determinism simply supported the views I already had: that harsh punishments didn’t work and are a waste of resources. Long prison sentences are also not common in NZ, so it’s a societal value, just like lack of corruption, though there are always people calling for harsher penalties.

    Countries that do not have long sentences and where people are generally less vindictive towards criminals, like Norway, are not overrun by criminal behaviour.

    I think the conclusion drawn from the data has a lot to do with the attitudes of the people analyzing it.

    • rickflick
      Posted July 16, 2017 at 12:29 pm | Permalink

      “Countries that do not have long sentences and where people are generally less vindictive towards criminals,…”

      In the US, I’d say people are very vindictive. In part perhaps because of religious fundamentalism, and perhaps also because we are still a bit of a frontier country where only 150 years ago the local sheriff would round up a posse comitatus and hang the rustlers.

      • Heather Hastie
        Posted July 16, 2017 at 1:11 pm | Permalink

        It’s only a hundred years ago that my grandfather and great-grandfather were carving their farms out of the bush after emigrating from England.

        We’re an even younger country than you. There were no people here at all until c. 1300 CE, and the first European didn’t land until 1769. Our nation’s founding document, The Treaty of Waitangi, is dated 1840.

        We were originally a British colony of course, so our law enforcement has always been very similar to theirs. Our police still don’t carry guns. Sheriffs and posses aren’t part of our lore.

        • Colin McLachlan
          Posted July 17, 2017 at 4:37 am | Permalink

          You raise an interesting point: why did NZ not have a “wild West”? You could perhaps work up a post on that subject in Heather’s Homilies 🙂

          • Posted July 17, 2017 at 1:02 pm | Permalink

            Or, why was Canada *less* wild?

            Perhaps it was the nominal unified government (even if it was a colonial power)?

  5. Posted July 16, 2017 at 1:24 pm | Permalink

    Thanks for reporting on this paper! I will need to spend some time scrutinizing it myself, but I just want to say that drawing any substantive conclusions from these types of purely exploratory, correlational analyses is very difficult. And that’s especially the case when working with hierarchical data, as the authors do here with individuals nested within countries.

    Such research can certainly be interesting, but it really irritates me how it is published in marquee venues like PNAS alongside far more rigorous science.

    • Posted July 16, 2017 at 7:16 pm | Permalink

      I agree about correlational analyses; thanks for pointing that out. It’s very hard to correct for potential confounding variables with such an approach.

  6. BJ
    Posted July 16, 2017 at 1:48 pm | Permalink

    First, an important mistake in your opening to point a: “Belief in free will is positively and significantly correlated with both intolerance for ethicality and support for criminal punishment; the association is stronger for the latter.”

    I believe you meant to write intolerance for UNethicality, not ethicality.

    Anyway, it’s important to have a society that is intolerant of unethicality. If determinism is correct (and even if it isn’t), then it seems to me such intolerance is an input that is more likely to make average people less likely to act unethically. It also seems to me that the same holds true for punishment of unethical actions.

    Also, as I said in another thread about determinism and its relationship with justice system reforms (specifically, reforming our system of punishment), the only change I can see happening if we all accept determinism is better treatment of prisoners, which is something that should be supported regardless. If significant prison sentences for criminality exist as an input, then it is more likely to deter criminal behavior than very light sentences. Many democratic societies already have rehabilitative, rather than retributive, justice systems. The US should have such a system too, and it doesn’t really matter whether determinism is correct or not. Therefore, it seems a belief in free will is still an important social glue, and that it isn’t likely to have an enormous effect on how justice is carried out in terms of sentencing and deterrence.

    • Posted July 16, 2017 at 3:58 pm | Permalink


    • Kevin
      Posted July 17, 2017 at 10:06 am | Permalink

      Conceptually I am locked into my own world where the concept of free will has no bearing why someone would, for example, be a retributivist.

      The best bet is to simply tell people demanding retribution for harm others inflicted is not the best answer. Period.

      Maybe at some point later we can convert people to determinism, or at the least compatibilism, but better to endorse criminal practices which have determinism in mind.

  7. Gregory Kusnick
    Posted July 16, 2017 at 2:10 pm | Permalink

    we know that determinism, (or, if you include quantum effects, “naturalism”) happens to be true

    I’m all for framing the argument in terms of naturalism to arrive at a truer view of the world. But note that this generalization requires us to jettison the claim that we couldn’t have done otherwise in identical circumstances. Naturalism per se — i.e. the claim that our decisions have physical causes — doesn’t preclude the possibility that the same physical causes could produce a different outcome. “Couldn’t have done otherwise” requires a commitment not just to naturalism, but to a particular kind of physical causation.

    • Posted July 16, 2017 at 4:01 pm | Permalink

      In the identical physical situation, only quantum mechanics could create a different behavior. Are you asserting explicitly that quantum mechanics can affect behavior that way on a macro level? If not, what basis do you have for your last claim?

      • Gregory Kusnick
        Posted July 16, 2017 at 4:37 pm | Permalink

        We know that quantum mechanics affects the distribution of galaxies on a cosmic scale.

        At human scale, here is one of many quantum random number generators available on the web. After viewing that page, your brain state will have been determined at least in part by quantum processes. This is just one of many ways that quantum events can enter our awareness and affect our brain state; in fact we’re surrounded by them: the random flickering of fluorescent tubes, the irregular beeps of smoke detectors, the graininess of digital photos taken in conditions of very low light. Since we agree that brain state determines behavior, it seems undeniable that quantum events can and occasionally do play a role in that determination.

        The larger point is that we needn’t make an ideological (and probably counterfactual) commitment to “couldn’t have done otherwise” in order recognize that our behavior has physical causes. In my view, arguments for a naturalistic worldview are stronger if they avoid the pitfalls of “can’t do otherwise” determinism.

        • Posted July 16, 2017 at 8:14 pm | Permalink

          If determinism is false, then you perhaps could have done otherwise in an actual situation, but this wouldn’t make your behavior more up to you than under determinism since it isn’t you but a random fluctuation that determines the outcome. So this sort of indeterministic “could have done otherwise” doesn’t confer a contra-causal free will that makes us more deserving of punishment than we would under determinism.

          • Posted July 17, 2017 at 5:00 am | Permalink

            I’m talking about dualism here: what most people think of as free will. I’ve written a lot about “random fluctuations” (presumably you mean quantum fluctuations) and how they don’t give you a free will; you’ve apparently not read that.

            • Posted July 17, 2017 at 6:39 am | Permalink

              I was responding to Gregory, and yes I remember you making this point. It always has to be pointed out that indeterminism can’t add to agency and responsibility, although it does complicate the otherwise simple picture of cause and effect.

          • Gregory Kusnick
            Posted July 17, 2017 at 11:13 am | Permalink

            Tom, of course physical indeterminism can’t give us unphysical powers to violate causality. So presumably you would agree with me then that the question of whether more than one outcome is physically possible in a given situation is irrelevant to the question of whether, under naturalism, there’s any meaningful sense in which we can be said to have agency. Which in turn implies that incompatibilists are barking up the wrong tree in placing so much emphasis on “couldn’t have done otherwise”.

            If you’re suggesting that “couldn’t have done otherwise” should be read as something like “couldn’t have chosen otherwise”, then this strikes me as question-begging, since it assumes a priorithat naturalistic choice isn’t possible.

  8. Ken
    Posted July 16, 2017 at 5:42 pm | Permalink


  9. Posted July 16, 2017 at 8:24 pm | Permalink

    The authors conclude:

    “The influence of free will beliefs in people around the world, along with the moderating influence of countries’ institutional integrity, provides evidence that seeing one’s own and others’ actions as reflecting personal choice, accountability, and self-determination can broadly affect moral attitudes and judgments.

    Haven’t been able to access the paper as yet using the unlock feature. But here’s a question, since you’ve read it:

    When the authors refer to belief in free will, do they specify that it was *contra-causal* free will that was asked about in the survey?

    • Posted July 17, 2017 at 4:59 am | Permalink


    • Eddy Nahmias
      Posted July 17, 2017 at 4:54 pm | Permalink

      Well, here are the two questions (from the World Values Survey) that indicated people’s degree of belief in free will:

      1. Some people feel they have completely free choice and control over their lives, while other people feel that what they do has no real effect on what happens to them. Please use this scale where 1 means “none at all” and 10 means “a great deal” to indicate how much freedom of choice and control you feel you have over the way your life turns out.

      2. Some people believe that individuals can decide their own destiny, while others think that it is impossible to escape a predetermined fate. Please tell me which comes closest to your view on this scale on which 1 means “everything in life is determined by fate,” and 10 means that “people shape their fate themselves.”

      I am not sure that answering higher on these scales indicates greater belief in *contra-causal* free will and vice versa. Even if one doesn’t believe in that type of free will, one might be up and down the scale, depending on how much (compatibilist-style) control one thinks one has (or how much one doesn’t believe in fate). And to some extent, people who do believe in contra-causal free will might respond up and down the scale, depending on how much they think the desired outcomes of their free choices are thwarted by external factors beyond their control (such as poverty or bad luck, etc.).

      In general, the scales used to measure “free will” beliefs use much more “metaphysically neutral” terms, such as “choice,” “control,” or “freedom”, and sometimes problematic statements of “determinism”, e.g., “such things as beliefs, desires, and values don’t in fact guide decisions.”

  10. Rosmarie Maran
    Posted July 17, 2017 at 7:57 am | Permalink

    One should consider that, if determinism should be true and effective, it would have to be 100% determinism, not 9,99999…%
    Which means that it would be anyway not possible to discern whether determinism reigns or not, whether the future is closed or open. Plus, if determinism, meaning that the future is 100% determined and hence closed,is true, then “causality” also looses its meaning, as nothing would be caused by anything, as everything would have been fixed all along and nothing could have “caused” it to happen differently.

%d bloggers like this: