The Atlantic: Free will is an illusion, but we need to keep that illusion

May 23, 2016 • 2:00 pm

Yes, I know I’m writing about two Atlantic pieces in one day, but so be it: such are the laws of physics. The second piece, much better than the article on FGM, is an essay by Stephen Cave, “There’s no such thing as free will but we’re better off believing it anyway.” I’ll try to be brief, as the piece is long. The salient points:

  • Libertarian free will [the “we could have chosen otherwise” form] is dead, or at least dying among intellectuals. That means that determinism reigns (Cave doesn’t mention quantum mechanics), and that at any one time we can make only one choice.
  • But if we really realized we don’t have free will of that sort, we’d behave badly. Cave cites the study of Vohs and Schooler (not noting that that study wasn’t repeatable), but also other studies showing that individuals who believe in free will are better workers than those who don’t. I haven’t read those studies, and thus don’t know if they’re flawed, but of course there may be unexamined variables that explain this correlation.
  • Therefore, we need to maintain the illusion that we have libertarian free will, or at least some kind of free will. Otherwise society will crumble. 

Cave offers two solutions to keep the illusion.

One is “illusionism” as adumbrated by Israeli philosophy professor Saul Smilansky: “the belief that free will is indeed an illusion, but one that society must defend.” Smilansky believes this illusion is vital because discarding libertarian free will—the notion that we can really “choose otherwise”—undermines both praise and blame. How can we hold people responsible for their good or bad deeds if we don’t think they could have chosen to be good or bad?

My response is that we can still hold people responsible, but not morally responsible. We should realize that an individual who does something bad should be punished, but punished to keep them out of society, to rehabilitate them, and to deter others. In that sense, an individual who does something should be held responsible, but we should also realize that that individual didn’t have a choice. To me, that realization, which you get only if you accept determinism, is the upside of discarding libertarian free will. And individuals who do something good should be praised, for that not only prompts them to do more good, but shows others that if they also do good, they’ll also get goodies.

It’s impossible to act as if you’re a robot, so deeply ingrained in our brains is the idea that we’re conscious agents who can choose freely. But remembering that individuals don’t freely choose to do good or bad is a healthy thing to keep in mind, if for no other reason than it improves the way we dispense justice. We can be determinists in how we run society while still victims of our illusions that we’re free agents.

Lest you say (and some of you will) that “we already fully incorporate determinism into our justice system,” my answer is, “No we don’t.” I got an email this morning (sadly, I trashed it), by some person who, after reading one of my essays on free will and criminality, thought that I was dead wrong—that we need to punish people even more severely for what they do. Needless to say, that would be a disaster for the criminal justice system in America.

Cave’s other solution is this: we should become compatibilists and conceive of free will as Dan Dennett does: the fact that the human brain, although its actions be determined, requires a complex nexus of inputs and brain processing to give an “output”—a choice. While data show that this form of “free will” isn’t the one that most people hold (and certainly isn’t the view that most religionists hold), it’s basically a semantic relabeling of libertarian free will.

If you want to say that we’re free because we have complex computers in our heads (and, of course, so do many animals), then go for it. What’s important to me is not how you define free will, but to always remember that determinism (absent any quantum effects) holds: that at any time we could not have chosen otherwise. And to author Cave (though perhaps not to all readers) this leads to a sea change in society:

Waller’s definition of free will is in keeping with how a lot of ordinary people see it. One 2010 study found that people mostly thought of free will in terms of following their desires, free of coercion (such as someone holding a gun to your head). As long as we continue to believe in this kind of practical free will, that should be enough to preserve the sorts of ideals and ethical standards examined by Vohs and Baumeister.

Yet Waller’s account of free will still leads to a very different view of justice and responsibility than most people hold today. No one has caused himself: No one chose his genes or the environment into which he was born. Therefore no one bears ultimate responsibility for who he is and what he does. Waller told me he supported the sentiment of Barack Obama’s 2012 “You didn’t build that” speech, in which the president called attention to the external factors that help bring about success. He was also not surprised that it drew such a sharp reaction from those who want to believe that they were the sole architects of their achievements. But he argues that we must accept that life outcomes are determined by disparities in nature and nurture, “so we can take practical measures to remedy misfortune and help everyone to fulfill their potential.”

Well, Cave gives no data to show that Waller’s (and Dennett’s) definition of free will is the one most people hold; in fact, the study of Sarkissian et al. shows otherwise: surveys in four countries showed that most people hold a libertarian view of free will (see below). The figure below shows the proportion of people surveyed in four countries who think the universe is not deterministic: that human decision making is not governed by the history of the individual. Between 65% and 85% of people are pure libertarian free-willers.


But again, never mind. The second paragraph above is the telling one: “No one has caused himself: no one chose his genes or the environment into which he was born.” (Women are left out here, but they’re of course they’re subject to the same constraints!). And that does lead to a very different view of justice and responsibility than the one we hold today. Yes, the American justice system does exculpate those who weren’t thought to have a choice, but nobody, and no criminal, had a choice! That is a radical view for most people, but it happens to be true.

 Waller’s last thought above is equally true: determinism leads to the notion that “life outcomes are determined by disparities in nature and nurture.” I can’t emphasize that strongly enough, for it completely overturns the “just world” philosophy that pervades conservatism: that people are responsible for what happens to them, and should be treated accordingly. If you’re poor, that’s your fault; if you’re a criminal, that’s your fault, too. Those who make a lot of money deserved it.

In the end, what you label “free will” doesn’t matter to me so much as people’s need to accept the determinism that science tells us is true. For only when we embrace that can we begin to walk the path of treating people fairly.

h/t: John O’Neill

159 thoughts on “The Atlantic: Free will is an illusion, but we need to keep that illusion

  1. I’ve never understood the critics’ claim that responsibility, as a social concept, must follow free will out the window if it goes. Even if a brain tumor “makes” someone do something, it doesn’t change the fact that they did it. Non sequitur.

  2. But how can we defend the “illusion” of free will without either lying to others or deluding ourselves?

    And when we say that people who don’t believe in free will act badly do we mean they won’t have any choice but to act badly once they fid out they don’t have free will? Because if they have a choice…

    1. Very good question. It either involves deceiving ourselves, which is impossible, or making the Little People argument, which is patronizing. It’s like saying we need to believe in God lest society fall apart. And, of course, that argument has been made. In that sense, and many others, free will is treated the same as God.

      1. Yep, its exactly like that. And I would wager a fair amount of donuts that – also like the ‘God or bad behavior’ argument – its empirically untrue when tested.

        Yeah I know some tiny studies with college kids in a lab setting show some local/immediate changes in behavior (more willingness to cheat or play hardball in game-theory type games). But I don’t necessarily think those are predictive of long-term social behavior. It seems to me one could do population surveys in combination with incarceration data to investigate with more confidence whether determinism as a belief is correlated with bad social acting…and I bet it isn’t.

      2. In fact this is how it is described by some partisans of “agent causation”. I think it was the (otherwise very capable, but partisan of libertarian free will) philosopher R. Chisholm who “concluded” that we are in possession of a property traditionally only associated with the divine. Shouldn’t one stop to think one when comes to that conclusion, and maybe consider this an (informal) _reductio_? 😉

    2. Our conscious mind seems to be able to be deceived by our unconscious instinctual self, or atleast unaware?

    3. In my experience deluding ourselves in these matters comes naturally and we cannot completely avoid it. I see it as a classical battle between my rationality and my emotions (or maybe better Daniel Kahneman’s system 1 and system 2), and as we all know, emotion wins most of the time.

      But on the good side: society will not fall apart because of my believes about freewill. I still feel morally responsible for my actions though I’m 100% sure it’s an illusion.

  3. I worry about the state of criminal justice systems in most of the world. That in the US is seemingly designed to create recidivism. Also being caught for one mistake can destroy your life, while another person committing the same crime but not caught can go on to be president.

    One fairly minor thing but one I think is indicative is how people who’ve been incarcerated cannot vote for life in the US in a majority of cases. It’s another way of ensuring they have no stake in society and further alienates them.

    1. I meant to add that when there is a greater acceptance of determinism, as there inevitably will be, it should see huge reform of the criminal justice system.

      1. I don’t think so, and I (still) don’t think that JAC has made his case for determinism -> prison reform. If we are deterministic organisms (DO’s for convenience), we could be DO’s that respond more effectively to punitive physical punishment just as easily as DO’s that respond more effectively to rehabilitation. “What empirically works to lower crime” is an entirely different question from “are we deterministic” and I don’t necessarily see why they would correlate. Being more on the liberal side of the political spectrum, I hope and expect rehabilitation will work better than more punitive measures. However I don’t think the determinism/nondeterminism question provides any relevant evidence or support for my position on prison reform.

        1. There’s already evidence that less punitive and more restorative methods of criminal treatment results in lower crime rates, if only because it doesn’t encourage recidivism. For instance, The Spirit Level offers statistical evidence in support of the empirical claim, with accompanying analysis.

        2. The reason I think acceptance will lead to prison reform is that the system won’t be so much about revenge as it looks to me it is now. If people accept that a person’s path in life inexorably led them to this point, the next step will be about what will make them change direction rather than taking vengeance.

          Also the conditions of many prisons makes them part of the punishment. The sentence is supposed to be depriving someone of their liberty (and in the US, maybe their life), but the reality is they’re placed in a Lord of the Flies environment, or worse. That destroys humanity.

    2. So my compatabilist brain, despite the science, is still making me think like a determinist and there is nothing I can do about it?

    3. ” That in the US is seemingly designed to create recidivism.”

      It sure does. It is so good at creating recidivism it’s hard not to come to the conclusion that recidivism has been consistently selected for for a long time, or that the system was purposely designed to create it.

      Local to me is a case that perfectly illustrates this. Recently a 12 year old child that goes to the same school my children attend has been arrested and sentenced to jail time (almost certainly a juvenile detention center, but still) for possession of pot. WTF are the chances that this punishment will prevent this kid from smoking pot in the future? And even if it does at what cost? The damage to this child caused by the punishment, the social stigma associated with it and the lack of types of intervention that will actually help improve the child’s life rather than fuck it up? The result is this child is being groomed to be a repeat offender known by all local law enforcement.

      And this is the norm. Hell, there are police officers stationed on school grounds. This also is the norm. And they aren’t there to protect against terrorist attacks, at least that isn’t the only reason. They are regularly used to discipline kids. Kids as young as 5 and 6 years old. As in come into the classroom, handcuff them and remove them from class. Because they are having a temper tantrum. How screwed up is that?

      1. “How screwed up is that?” VERY.. 5 and 6 yr olds? that’s bordering on insanity.

        I’ve heard of something similar being done to first time juvenile offenders. They ‘give’ them a night or a few hours in a remand cell to give them a taste of criminal life.
        I cannot give you ‘success’ rates but i believe it has an impact on some kids.
        Their free will and liberty gets a jolt and with a little emotional tagging that’s enough to change the behaviour, at least for now.
        I would say the reverse could happen depending on the child’s personality and environment as in, harden their resolve and give the parents, police a big one/two finger salute.
        I’m my own child and all of you are going to know it!

      2. Absolutely outrageous. How enough people think this is a good idea for it to happen boggles the mind. I even think the perp walk thing for adults is counter-productive, but to subject kids to that is unbelievable. And you’re right, this reaction is just going to make any behaviour issues worse. The poor kids in these situations haven’t got a chance – their futures are being determined in front of us.

  4. The legal system needs a change of mind set, from a criminal must be punished, to a mind set of, the public must be protected.

    1. The public should be protected. Must is almost a normative claim charged as a declaration.

      Religion, for example, should be irradiated from our species. I would never claim that it must.

  5. It’s impossible to act as if you’re a robot, so deeply ingrained in our brains is the idea that we’re conscious agents who can choose freely.

    Is that really so? We can simply act as though we are deterministic robots who are not consciously aware of our decision-making brain (which is true, none of us are consciously aware of stuff whizzing around our 10^14 neural connections).

    I’ve never had any problem acting like that! How am I not acting like that?

    1. I wonder if this has been tested in any Milgram-type experiments. Are determinists more likely to follow harmful instructions? Some social science theory suggests that people are more likely to object when they are reminded they have a personal choice, or when they see others object. To many people, “determinism” says you don’t have an authentic choice in what you do; perhaps people would feel a lesser sense of autonomy if reminded of that.

        1. That’s not the question though. The question is how peoples’ behavior correlate with their beliefs about free will. Milgram-like experiments are all about uncritical submission to authority, so I wonder if that’s been tested while controlling for beliefs about free will.

          1. I suspect you wouldn’t see a change in behaviour over the long term because the sense of agency is powerful and even though we understand we don’t have free will, we can’t help but act as though we do. I often think of the scene in the Matrix when Cypher is asking to be reinserted:

            I know steak doesn’t exist. I know that when I put it in my mouth, the Matrix is telling my brain that it is juicy and delicious.

            Moreover, we have other parts of our brain stopping us from doing bad things (empathy, theory of mind). So, even if we want to, our minds will prevent us from going thermonuclear (at least those of us with the standard brain hardware and software).

            1. “the sense of agency is powerful and even though we understand we don’t have free will, we can’t help but act as though we do.” In social contexts, the Milgram experiments show the exact opposite. Most people have an overriding impulse to abdicate their agency and simply follow instructions. They literally act like robots. Some recent experiments hint that peoples’ brain may not evaluate those instructions before acting, so people literally act like robots. But voluntary action involves more processing and is connected with “a greater sense of responsibility.” What we want is for people to think critically about their orders and refuse to comply if they are instructed to violate ethics or law. So, does believing I’m a robot make me less likely to critically evaluate my instructions? I don’t know the answer, but it would be no surprise if determinist thinking makes people more compliant.

              1. I don’t think the Milgram Experiments at all disagree with what I wrote. The fact that people tend to obey authority figures doesn’t interfere with a sense of agency; it simply means they obey and they do so thinking they choose to do so.

              2. The theory is that people “switch off” their agency in certain situations, and there’s a fair amount of study on what can make people re-engage their independent judgement. The EEGs seem to indicate that people don’t contemplate their actions as much while following orders, and they don’t feel as though they are making choices. They are just acting out someone else’s ideas. It’s not precisely the same concept as determinism, but there is some psychological cross-talk between determinism and ones sense of agency or autonomy.

              3. But they still think they have agency. They think they are choosing to shut off their agency but in reality there is no agency. I think I’m typing this response and thinking about what to say and choosing how I write it but really there is no me doing that and there is no agency.

              4. I don’t know what they’re thinking or feeling below the surface, but in this situation most people traditionally say, “I had no choice, I was following orders.” Only a small fraction of people challenge the experimenter when they ordered to continue shocking a person who is begging for it to stop. Presumably the majority of people would not autonomously choose to behave like that. There is clearly a physical, measurable difference between the activity of merely acting out commands, vs reflecting critically on those commands and deciding whether it would be wrong to comply. The latter is what I would call “agency”, and it doesn’t contradict determinism. Some research has hinted that people tend to have less “will power” when they contemplate determinism, such as less ability to resist a cigarette. It’s a psychology finding so the results are a bit tenuous, but I wonder if what people really understand about “free will” overlaps with their conceptions of power and independence, and if that would show up in a compliance experiment.

              5. I fail to see how any of this is relevant to a person’s sense of agency. “I was only following orders” still has a doer. The person doesn’t say “something took over my will”.

              6. That’s exactly what they are saying. They literally say, “I didn’t have a choice, I did what I was told.” That means the same thing as “something took over my will.” People stop willing in the presence of authority, and you can see it in their brains.

              7. No, people never did “will” but hey think they did and then did not.

              8. Oh and the instruction to do something from a person in authority is the input to the brain to make someone obey. It’s a good example of determinism so I’m not sure what you’re arguing here.

              9. First of all, I’m not making an argument here so much as posing a question, which sounds like it’s still open. Second, since I’m a compatibilist I’m not arguing against determinism, so you can stop trying to convince me that brains are deterministic. I am suggesting that “agency” or “will” have meanings that can coexist with determinism, and perhaps this example of uncritical compliance vs reflective volition offers some support for a compatibilist view, where “free will” is often identified with self-directed action as opposed to a coerced or externally compelled action.

              10. And this is why I agree with Jerry here – the arguing about free will isn’t useful. The real thing we need to convince the majority of people, who are dualists, is the reality of determinism.

              11. The real thing we need to convince the majority of people, who are dualists, is the reality of determinism.”
                I totally agree.
                The failure to believe in determinism is a denial of the value of science and of the value of evidence- a denial that leads to all sorts of human folly.
                And perhaps the LAST thing we determinists should be propagating as a certainty is that we have no free will. Given the strength of the supposed “illusion” of free will, the person considering the our argument for determinism can only but think “these determinists even say that I don’t even make my own decisions, which I KNOW I do. Therefor determinism must be totally illogical rubbish”

              12. So, does believing I’m a robot make me less likely to critically evaluate my instructions?

                For what it’s worth, I’ve been long convinced I’m a neural network program operating in a brain in a human body – aka a robot – and I haven’t noticed any corresponding drop in my critical faculties any more than I’ve noticed a drop in quality in my colour vision or writing ability. I still remain wary of rhetorical questions with leading answers, for one thing.

                I think the bigger problem is that our stereotyped idea of a robot is stuck in “mindless factory automaton” mode. People convinced of the higher, fundamentally different aspects of humanity tend not to be receptive to any evidence that suggests human cognition is anything involving cold, hard brain circuitry.

      1. It has been researched (Kathleen Vohs, Jonathan Schooler). But it couldn’t be replicated and it’s conclusions where heavily debated.

  6. When someone says, “We don’t have free will, therefore we should do X, Y and Z” I don’t get that. What am I missing?

    1. Like the Master Oogway said: “Quit, don’t quit… Noodles, don’t noodles”

      The universe does not care. In the face of determinism, these truths still remain:

      What Matters Is What Matters To People.


      We Can Always Do Better.

      Two of Sean Carroll’s recent considerations in The Big Picture.

    2. It’s basically like saying, “Based on our understanding of how the world works, it makes more sense, logically, to behave like THIS rather than like THAT.” The hope is that by saying that and putting that thought out in to the world, you can influence other minds to accept it and change their behavior. Hopefully, the end result is to make the world a better place.

      I don’t know if that clears anything up for you. It all comes down to inputs and outputs.

      1. Like Jon Snow, I know nothing. Doesn’t ‘influencing other minds’ imply those minds have some freedom to choose?

    3. Donmac,

      I don’t believe you are missing anything – quite the opposite: I think you are detecting a real problem.

      1. pacopicopiedra,

        “It’s basically like saying, “Based on our understanding of how the world works, it makes more sense, logically, to behave like THIS rather than like THAT.””

        The only possible force such logic can have to motivate anyone’s behavior is if there is a goal in mind – i.e. the goal of “making the world a better place” or whatever. So the logic is “IF you want to achieve that goal, THEN you should take THIS action rather than THAT action.’

        In other words, it assumes we actually could choose either option.

        But that makes no sense if one starts with the position such choice is only an illusion – hence we really don’t have the ability to do either THIS or THAT.

        So the contradiction Donmac points out remains. And this problem is not addressed by saying we can influence each other’s behavior; we want to be actually giving coherent arguments, speaking truths, to influence behavior.

  7. Yes, the American justice system does exculpate those who weren’t thought to have a choice, but nobody, and no criminal, had a choice!

    My problem with this one-size-fits-all approach to justice is that it elides an important distinction between two broad classes of bad behavior. On the one hand, there are bad acts committed by people capable of competent moral reasoning who know they’re doing wrong but for one reason or another think they can get away with it. On the other hand, there are acts committed by people whose moral judgment is impaired in some way that makes them incapable of appreciating the bad consequences of their actions.

    These different categories of bad behavior call for different social responses. Deliberate miscreants need to be taught that their bad choices have consequences, while those incapable of learning such lessons need to be protected from making bad choices in the first place. Treating people fairly, it seems to me, demands that we recognize such distinctions.

    1. To me, understanding the consequences of your actions is the key issue. When an animal, child or severely mentally disabled person commits harm, it should be treated differently than when a mentally competent adult commits harm. The law emphasizes the word “knowingly.” I am not sure this is different from “morally responsibility” since, to me, morality is about having a due concern for the effects of one’s actions on others. I am quite willing to discard the contentious term “free will.”

      1. Although I am troubled by the case of an amygdala tumor victim who knowingly commits harm. There is a sense in which he lacks “free will” because he cannot control his emotions and actions. I am thinking of Charles Whitman.

    2. Yes, of course I agree with Gregory K. I didn’t say that everyone should be treated equally, nor do I think I’ve ever said that. What I’m saying is that the notion that a criminal had no choice should always be kept in mind when figuring out with to do with him. Of course the amount of sequestration, the kind of rehabilitation, and the need for deterrence based on the criminal with vary with the person.

  8. Apologies for the length of this comment. I continue to be puzzled by your posts on freedom and determinism. Maybe you could help me to see how your view is supposed to go?

    You write: “What’s important to me is not how you define free will, but to always remember that determinism (absent any quantum effects) holds: that at any time we could not have chosen otherwise.”

    I know that many, many compatibilists have pushed you on this, but let me try again. It’s not obvious that your conclusion here (that we could not have chosen otherwise) follows from your premiss (that determinism holds). I’m not sure how you mean to formulate the argument, but on some natural formulations, the argument is invalid. Here is such a natural reconstruction, followed by an explanation of why it is invalid:

    [P1] Necessarily, if the laws and initial conditions are such and so, then we choose and act in this way.

    [P2] The laws and initial conditions are such and so.

    [C1] Necessarily, we choose and act in this way.

    Or, using a basic fact about how modal operators interact with negations, we could express [C1] as:

    It is not the case that it is possible that we choose and act otherwise than in this way. I.e., we could not have chosen or acted otherwise.

    What follows is that we do, in fact, choose and act in this way. But the further modal claim that we *could not* have done otherwise does not follow. What follows from the premisses is that we *do* choose and act in this way. That is, in the actual world, we choose and act this way. But the conclusion requires more: that we choose and act this way in *every* possible world. But since the premiss [P2] does not ensure that every world shares our laws and initial conditions, it does not follow that we choose and act in this way in every possible world. More on this after an analogous (obviously incorrect) argument.

    Formally, the argument above is equivalent to the following, obviously erroneous argument:

    [P3] Every cat is a mammal.

    [P4] Garfield is a cat.

    [C2] Everything is a mammal.

    The problem in the obviously-wrong argument is that there are things for which the antecedent in [P3] — that x is a cat — does not hold. The analogous problem in the argument against the possibility of choosing or acting otherwise is that there are possible worlds that either do not share our laws or do not share our initial conditions (or both). For example, there are possible worlds where determinism is true — a fact about the laws — but where the initial conditions are different from the ones in the actual world. Supposing such a world has a “me” in it at all, it could easily be the case that he chooses and acts differently than I do in the actual world. After all, the worlds are very different places! But plausibly, when one says, “We could not have chosen or acted otherwise,” one is saying that there is *no possible world* in which we choose or act differently than we do in this world.

    Perhaps you mean to say that in every world that shares our laws and initial conditions, we act just as we act in the actual world. That looks okay. But two things. First, it’s at best unclear that saying, “In every world that shares our laws and initial conditions, we act just as we act in the actual world,” is equivalent to saying, “We could not have done otherwise.” Second, it’s at best non-obvious that there are any possible worlds that share both our laws and our initial conditions and that are distinct from the actual world. (In fact, I’m not sure anyone working on the metaphysics of laws of nature or of possible worlds would endorse the view that there are non-actual possible worlds that share all of our laws and also share our initial conditions.) But if there are no possible worlds distinct from the actual world that share the actual world’s laws and initial conditions, then to say, “In every world that shares our laws and initial conditions, we act just as we act in the actual world,” is just to say, “We act as we actually do.” But literally no one disputes that claim.

    Perhaps you mean something else? I’m just not sure what that something else might be. I would genuinely like to understand what you mean to be saying. But I think that so far, I don’t.

    1. Helpful questions and comments, Jonathan. It’s useful to note as well that the chart that Prof. Coyne puts up to suggest that most people agree with his understanding of free will (and hence compatibilists are changing the subject) suffers from similar problems in the way it presents determinism. The responses presented in the chart are to the question: Do you think we live in universe A (“deterministic”) or B (“indeterministic”), following this closing contrasting description of the universes:
      In Universe A, “each decision has to happen the way that it does. By contrast, in Universe B, decisions are not completely caused by the past, and each human decision does not have to
      happen the way that it does.”

      1. That’s not problematic as most people don’t have an inkling that quantum mechanics may throw a small wrench into behavioral determinism. That graph certainly shows that most people believe in a form of libertarian free will, regardless of quantum mechanics, and this is in direct conflict with those, like you, who claim that most people are compatibilists: i.e. they accept determinism and some deterministic-compatible version of free will.

        1. “That graph certainly shows that most people believe in a form of libertarian free will, “

          It still doesn’t seem a settled subject.

          And, again, I think we have to end up getting clearer on whether the people who seem to hold a libertarian idea of free will, and even those doing the studies, are actually coherent on the subject. In other words, whether they are mixing up people’s theory of free will with a definition of free will. They are very importantly not the same thing, but easy to confuse.

          Just like morality.

          A lot of people think they are getting their morality from the bible. But we atheists often point out this is a confusion about how they actually think about morality – that is, they are actually bringing their own morality to the bible and using it to cherry-pick the bible.

          Further, if you ask many religious people – e.g. Christians – they will say morality has a supernatural basis, morality comes from God and if that isn’t the case, then morality doesn’t exist.

          But when you ask why they think this, you get replies such as “Well, if there is no God then what reason would I have to not rape or kill, or to do good acts?”

          In which case, it can be pointed out that what they ACTUALLY think morality is based upon are reasons for doing good. That’s what they really want. It’s just that they think God supplies those reasons for doing good. They just have a bad theory for where the reasons for moral actions come from.

          In other words, they are actually blind to some of their own assumptions about morality.
          And it’s a very important distinction to pick up because IF you stick only with the superficial take on the belief of the religious person, you’d be saying “Since God doesn’t exist, morality doesn’t exist.”
          But understanding that what they are actually concerned about is the existence of reasons for actions – that THAT is the ACTUAL basis, what is important, to their concept of morality…then you can go on to explain “See…what you actually care about is still there. It was obscured by your own
          naive explanation for it, but so long as reasons for moral action is what you care about, then we can supply the ACTUAL, real world reasons for acting morality. Morality doesn’t go away: it’s better supported.”

          And most compatibilists argue the same for free will. It’s a better account for the freedoms we really do have, the freedom that is actually important to most people in everyday affairs.

        2. Again, I’m puzzled. I don’t see how what I’m saying or what Eddy is pointing to about Sarkissian’s study has anything to do with quantum mechanics. My point is that there is an easy conflation between determinism and fatalism, and as a result, people often don’t think clearly about the implications of determinism. Eddy is pointing out (I think) that the way Sarkissian’s study is written runs together fatalism and determinism in a similar way. So, it’s not clear what to take away from their studies.

          Now, I tend to think that most people are libertarians about free will (sorry Eddy, you’re going to have to keep working to bring me around). But that they are libertarians is not demonstrated by the chart you show from Sarkissian et al., even if their studies worked as intended. To see whether most people are libertarians, we need to look at the [Incompatibilist, Indeterminist] box in the contingency table, since libertarianism about free will is the conjunction of the belief that we do have free will with the belief that free will is incompatible with determinism.

          Sadly, Sarkissian et al. do not actually write down the contingency tables for their studies, but we can get an approximate reconstruction based on their report that 61% of those who responded as determinists also responded as compatibilists (see page 353 of their paper). Based on that, here are reconstructions of the contingency tables for the U.S. respondents and for the Hong Kong respondents. (In order to increase the chance that the formatting comes across, I’ll use some dots as spacers. “Det” means “Deterministic”; “Comp means “Compatibilist.”)

          U.S. Respondents:





          Hong Kong Respondents:





          So assuming that the study questions get at what Sarkissian et al. wanted to get at, what we see here is that about 68% of people in the U.S. are libertarians about free will and 49% of people in Hong Kong are libertarians about free will (still the modal view, but not the majority of people in Hong Kong).

          There is a further question about how people would respond were they to be convinced that determinism is true. Sarkissian et al. say (and I agree) that “our results suggest that if people are persuaded that the universe is deterministic, they will not end up concluding that human beings are never morally responsible … [rather] they will simply conclude that moral responsibility is compatible with determinism” (353).

          But that said, I’m not sure we should take our cues about what to think about free will from what ordinary people think. (This is not to say that there aren’t interesting philosophical implications of the experimental results.) I assume you agree on that, since you think that most people are libertarians, while you also think that people do not have free will.

    2. As I mention in all these contexts: what are the possible worlds here?

      (Also, the universe does not *have* initial conditions, only parts of it do, say the local hubble volume.)

      1. I’m not sure if this is helpful, but briefly:
        The possible worlds are not like the multi-verse worlds posited by some recent theories of quantum physics. They are just meant to be a way of thinking about possibility (and necessity) claims. So, think about a lightning bolt hitting a particular tree in a forest at a particular time T. Could it have hit another tree? [Is that possible?] It depends what we mean by ‘could have’ and ‘possible’. It seems pretty intuitive that the answer is ‘yes, the lightning could have hit another tree.’ Coyne suggests that if determinism is true, maybe the answer should be no–nothing could happen differently. Jonathan and I are suggesting that one way of understanding that the answer is yes, even if determinism is true, is that there are ‘possible worlds’ (roughly, very similar situations to the actual one) that have slight differences that explain why the lightning hits a different tree.

        In any case, we (including scientists) need a way to talk about possibilities even if we think determinism is true. Whatever way we use (e.g., to talk about the possibility of lightning hitting different trees, etc.) may be appropriate to use about people making different choices. Even if determinism is true and I chose the soup at time T, I could have chosen the salad at time T (if so, some conditions would have to be different).

        1. Right, but in my view possible worlds doesn’t actually answer the question: Lewis’ possible worlds, even in their “ersatz” versions, are *logical* possibilities (though determined through “semantic ascent”). I.e., formalize the state (in what reference frame??) of the universe at a given time and then it is consistent with this description that the next state be *this* description. So we can say, loosely, that the second state is “possible”.

          (I’ve never understood any notion of logical possibility other than that applied to propositions.)

          As for *other* notions of possibility, if determinism is correct (in its traditional sense) there is only *one* nomological possibility. Epistemic possibility: well, that depends on whether you take that as an individual or social notion.

      2. The question, “What is a possible world?” is very fraught. The basic idea is to start by thinking about how things are in the universe (or maybe the multiverse, if there is one). Take that whole thing. Everything that exists or happens. That’s one possible world. Alternative possible worlds are just like this one except that some facts are different. Some are different in trivial ways (worlds where one atom is displaced by one micron for one picosecond); others are different in radical ways (worlds where there is no matter or where the laws really are Newtonian or …). As you can see from the SEP article on possible worlds, if you’re interested (, there is disagreement about whether we should think of possible worlds as concrete or as abstractions (maybe like big fictional stories) or in some other way. Typically, thinking about possible worlds is motivated by some pragmatic argument: possible worlds do a ton of work in metaphysics, philosophy of language, philosophy of mind, and even philosophy of science. For my own part, I would prefer to replace talk of possible worlds with some other mechanism for talking about possibility. But I’m not entirely sure that any alternative will work as well.

        Thanks for the correction on initial conditions. I’m curious: Do you think anything hangs on it? Can we just substitute “initial conditions of a very large portion of the universe — a portion in which we live” everywhere that I’ve talking about the universe having initial conditions and get something sensible?

        1. [I should add to my above remarks: I don’t mean to toot my own horn. But I did do an MA and some further work, the former actually in metaphysics, so if it helps I do know the literature ;).] I just don’t think “possible worlds” work to understand the problems it supposedly does. And this is *independent* of whether you buy Lewis’ stuff or not.

          What notion of possibility? Constrained by laws (not by law statements!), e.g., as in vol. 3 of Bunge’s _Treatise_. That might be too strong, for some philosophical tastes, but I think it is the only one we have for the present purpose outside of “logically possible” which in my view immediately becomes Boolos’ provability calculus or something similar.

          As for initial conditions, maybe, just substitute the Hubble volume correction, since if Vic Stenger and others are right, the big bang does “erase”.

          1. Formally, though, the question “what notion of possibility” just boils down to figuring out what the accessibility relation looks like or what the generic structure is with respect to the space of possible worlds. In standard treatments, anyway, you’re still going to have possible worlds regardless of whether you’re dealing with logical possibility, metaphysical possibility, physical possibility, or epistemic possibility.

            You said in reply to Eddy that if determinism is true, then there is only one nomological possibility. But that is only true if you take the initial conditions — perhaps defined piecemeal over portions of the universe — to be one of the laws. If the laws only describe transitions of state, then there are *many* nomologically possible worlds distinct from the actual world.

            1. Right, one possibility relative to a given state. This assumes, as most discussions do, that the laws are “memoryless” – or alternatively that the state includes the history, if one prefers.

              I agree that formally one just picks the accessibility relation and the set of worlds oneself. (Think of Kripke’s remark about possible worlds being stipulated.)

              But these two items represent all the metaphysical work that needs to be done!

  9. The notion that punishment for the purpose of deterrence is justified under the “no free will” doctrine is incoherent.

    1. How so? If your decisions are determined by your life history and that history includes awareness that certain actions have certain negative consequences, that will weigh against your deciding to act in that way.


      1. “How so? If your decisions are determined by your life history and that history includes awareness that certain actions have certain negative consequences, that will weigh against your deciding to act in that way.”

        Under the “no free will” doctrine (and I’m not making a value judgement here) no one ever decides anything. That’s why the notion of deterrence is incoherent.

        1. I don’t see that at all. Humans respond to information, including that in the form of threats of unpleasant consequences. This response is a reaction to the environment. Just like blowing hot air on a spider will make it move to cool off.

          Free will is unnecessary for the affect to occur.

          1. Deterrence presupposes free will. I don’t see any way around that. Now if by “deterrence” you mean something like “setting the conditions for minimizing crime”, OK, but that relieves the word of doing any work.

            1. “Deterrence presupposes free will.”

              I don’t think that it true. Deterrence presupposes consequences. We can have deterrence without the philosophical baggage of free will.

              1. Language is rife with the philosophical baggage of free will. “Deterrence” is a particularly striking example.

              2. Language is rife with all kinds of baggage. That’s interesting but, I think not relevant. We aren’t slaves of language and we don’t need free will to produce deterance. We only need to be able to recognize the relationship between action and consequence.

            2. Deterrence doesn’t presuppose free will at all. I’m puzzed at that assertion! We are evolved beings that act in such a way as to minimize pain and stress. If you find out that you’ll not get penalized if you run a red light, you’ll run them. If you find out you’ll get a $5 ticket, you’ll run them a bit less. If you know you’ll get a $100 ticket you’ll hardly run them at all. That doesn’t presuppose any kind of free will; just the ability of an evolved organism to associate different inputs with different levels of pain.

              1. Not only that, but you could probably train an ant or an even simpler organism to behave in a particular way using a system of deterrence. I know some people would say an ant has free will, but not many.

              2. Training an ant (or anything else) with some sort of conditioning is rather different from persuading potential criminals not to commit crimes under threat of imprisonment or death, don’t you think? When you can get an ant to behave in a particular way by punishing other ants, get back to me.

              3. Does killing ants (or some other insects) release a pheromone that deters other ants (or other other insects)?

                So capital punishment would work …


              4. I don’t quite agree with how Stephen Barnard
                seems to be arguing the contradiction between incompatibiism and deterrence. Or at least it doesn’t seem spelled out exactly.

                I would put it something like this:

                There are many ways to deter behavior, and many do seem to presuppose free will in the sense of assuming one can do otherwise.

                You can try to deter someone by appealing to her reason: “X action vs Y action really makes more sense – given your personal, or moral goals.” But this assumes some robust sense of “could do otherwise” that incompatibilism seems to deny.

                Punishment-style deterrence also seems to presume one could do otherwise, since the one punishing is doing this on the assumption that “to do otherwise” – i.e. to not punish – will not meet a goal. Plus it presumes the same on behalf of the people the punishment is supposed to influence: it presume they could be influenced to choose otherwise.

              5. Another factor to consider is the mind’s ability to generate its own “world” of images; one that is perceived AS real and one that can have as much of an influence on our decisions and behaviors as the real world outside us can. A particular person might not be “deterred” from running a red light, even if the fine were $5,000 dollars, if they have an internal, mental world that tells them, “The cops won’t catch ME!” In fact, to hold on to this type of “false-reality”, to one extent or another, is necessary before one engages in any kind of illegal or dangerous activity, even though the individual may be unaware of it or even denies it.

              6. When someone contemplates the consequences of a criminal act, balances the potential risks and rewards, and decides not to commit the crime, that’s deterrence. You might think the conscious decision is an illusion, and you may be right, but the vast majority of people see it as an act of free will.

              7. “You might think the conscious decision is an illusion, and you may be right, but the vast majority of people see it as an act of free will.”

                Sadly (or luckily?) reality isn’t up for a vote. The vast majority of people are wrong about a very large number of things. They think, for example, that angels exist. The minority view is, I think you’ll agree, closer to the truth.

              8. The vast majority of language speakers determine the meaning of a word like “deterrence”.

              9. Assuming the democracy of word definitions, you’ll need more than one cherry-picked one.

                Google tells us this: “discourage (someone) from doing something, typically by instilling doubt or fear of the consequences.”

       “1. to discourage or restrain from acting or proceeding:
                The large dog deterred trespassers.
                2. to prevent; check; arrest:
                timber treated with creosote to deter rot.”

                “: to cause (someone) to decide not to do something
                : to prevent (something) from happening ”

                The Free Dictionary:
                “1. To prevent or discourage from acting, as by means of fear or doubt
                2. To prevent or discourage (an action or behavior).
                3. To make less likely or prevent from happening”

                Cambridge Online:
                “to prevent or discourage someone from doing something”

                Oxford Dictionary: “Discourage (someone) from doing something, typically by instilling doubt or fear of the consequences:
                only a health problem would deter him from seeking re-election.”

                None of these definitions (what popped up when I asked the Interweb what the word means) include the concept of free will. So if we’re going to use dictionary definitions as proxies for a popular vote I’m going to continue to insist that there’s no need to insert “free will” into a conversation about deterrence. The two words serve no purpose and simply confuse things.

      2. The absence of free will does not provide justification for any type of punishment. We need a moral theory in order to say that something is “justified,” and determinism doesn’t necessarily compel a unique moral theory. I gather that most people here subscribe to some type of utilitarianism, but that’s not predicated on determinism.

  10. I think it was Dennett that likened free will to things like money or the law. They’re social constructs in the purest sense of the phrase, but constructs that society needs in order to exist.

    So maybe not so much social constructs as they are social pillars or pre-constructs or whatever.

    1. I accept that it’s determinism all the way down, but I don’t accept ‘illusionism’ or ‘compatibilism’ (for certain types of compatibilism). However when humans live in complex ‘troop’ societies our behaviour is heavily conditioned by living in that society. So we have various sorts of ‘signalling’ to manipulate the thoughts of others. Punishing crime is signalling. Identity signalling is another, as is virtue signalling and (in religious times) sanctity signalling. Perhaps ‘free will’ is just ‘agency’ signalling, including ‘signalling’ to ourselves.

      I’d be interested to find out (tricky experimental design required) if herd animals have a strong sense of self or ‘free will’.

  11. Jerry, you wrote “We should realize that an individual who does something bad should be punished, but punished to keep them out of society, to rehabilitate them, and to deter others.”

    Interesting. If what I do is predetermined, how will knowing a bad act’s likely consequences stop me from doing it?

    1. Because knowledge of such things forms part of the environment that alters the way a person’s brain works.

    2. Because, knowing the consequences of a bad act are part of what makes the predetermination. If awareness of the consequences are not enough to stop you from committing the bad act then other factors, other inputs, had stronger sway than that knowledge. Those factors need to be determined and accounted for and possibly adjusted, all while you are removed from society, in order for you to return to society with new inputs that won’t deliver a similar outcome as before.

      I think.

      I think I’m finally beginning to understand it. At least I hope I am.

      1. Yes, I think I’m beginning to get it too, and I like your summary.

        Your mention of adjustment reminded me of The Adjustment Bureau, which has “The Chairman” (AKA god) supervising the adjusting.

    3. You’re confusing “determined” and “pre-determined.” Your action is not written by the fates into the Book of Life, an unalterable destiny. It is determined by your genes and environment and the environment is ever changing. Being presented with the knowledge that a particular action will result in a particular punishment, may be enough to alter your trajectory.

  12. The author of this blog regularly confuses me with his articles about free will. It seems that I need a clearer distinction between agency and free will. The two seem related but one is spilled over by theological concepts whilst the other is quite mechanistic. I feel that if we come more appreciative of what agency entails (the ability to make choices on the sum of experiences and learning), the arguments in favor of a system of rehabilitation seem much more acceptable.

    My point is, often within these articles there’s only ripping on the concept of free will and the people over at Templeton, an ever lasting glory to neuroscientists that operate within a rather Cartesian frame, and a lack of a clear explanation of what is really going on then. Stating that life is just a chain of events is too general and, within its counter argument, almost makes it looks like the act of learning is rather futile to begin with since there’s already a fate decided for you by the chain of events happened in the past. At least, for me that is.

    I guess I will need to dig into Denett’s essays on intentionality to get a deeper understanding of it.

    1. Don’t confuse determinism with predeterminism. Determinism is not fatalism but simply the concept of one thing following another. You are made up, to oversimplify things, of your genetics and memories at any point in time. Inputs influence you (an organism of genetics and memories). If someone slaps you all of a sudden, do you kill that person, do nothing, slap them back? It all depends on who you are at the time you get slapped.

      Are their consequences to slapping a person back or killing them? Do these consequences matter to you? Can your opinion on these things change? All this influences how you react at any given time. Maybe you try to kill the person but after spending time in rehabilitation, realize that your decision was a bad one. If you go back in time, you have no choice but to try to kill the person again because you don’t have that rehabilitation input yet. You are a different person at that point in time.

      The input into your brain of “not to kill” does not require free will to change how you feel and change how you behave if you find yourself in the same situation in the future.

      Determinism, which compatibility and non-compatibilists both agree about, is really the more important discussion and more important argument than free will because it’s the understanding of determinism that influences how we handle the mentally ill, the criminal minded and should lead to reform of our justice systems, which suppose people can choose differently and must be punished for bad choices (vs rehabilitating or where rehabilitation is not possible, keeping away from society).

      1. Apologies for “their” and my iPad autocorrected “compatibilists” to “compatibility”.

      2. Determinism, defined as “one things following another”, is a simple statement of Newtonian cause and effect. In the 2014 Edge question on what scientific ideas should be retired, our host proposed “free will”, but another proposal was “cause and effect.” It’s possible that cause-and-effect will prove to be an inadequate narrative for fully explaining reality. Indeed, we now know that time and space are themselves illusory to some extent. If compatibility sets and determinists agree on one thing, it’s that libertarian free will is a poor explanation of human actions. But that doesn’t mean we are compelled to fully endorse determinism; it may not be the last word on how things really work.

        1. No, it isn’t as simple as Newtonian cause and effect and that’s why I purposely didn’t use that term. There could even be quantum influences we are not aware of but quantum mechanics is deterministic as well.

          Because we can’t predict or know all the variables, doesn’t mean determinism doesn’t exist.

          1. Determinism exists an idea. Hard determinists would have us make big changes to our intuitions about justice and responsibility based on that idea. All I’m saying is that I can be skeptical of both the libertarian and determinist positions. I can think about justice and responsibility without reference to that metaphysical dilemma. If we build theories of morality and justice that don’t hinge on metaphysical propositions, then I think they’ll be stronger theories.

            1. The problem is our justice systems do take a philosophical position. They are based on the notion that a person could have chosen otherwise and therefore should be punished, even killed for their actions because they are bad for making a bad choice.

              1. Justice systems are comprised of many people with varied philosophical views. Those who write and interpret laws and policy do tend to focus heavily on deterrence. Indeed, the concept of punishments in proportion to a crime’s severity is part of a framework for generalized deterrence. The idea is that if I see you being punished severely for an infamous crime, it makes me less inclined to commit any crimes. This is fully compatible with determinism and it’s a common outlook in criminal justice. If anything, I think determinism and utilitarianism have been dominant in legal thinking for over a century. On the other hand, virtually every moral theory (including those connected with libertarian free will) has a wing that opposes retributive justice. Among scholars, retributivism is rare. So I think the position of determinism is not necessary for the conclusions you want to advocate, and I think it adds more confusion than it resolves.

              2. Justice systems aren’t comprised of people. They are based on ideas. In our case, the justice system is based on the idea that once is criminally responsible and could choose otherwise. This is classic dualism and it is scientifically untrue.

                There are some instances of determinism like pleading insanity but otherwise, the system is dominated by the desire to punish more than anything else. It’s why the US still executes people.

              3. If you go to court, you’ll see a lot of people there. In legislatures, in prisons, in police departments, in law schools and criminology departments; lots and lots of people with individual ideas about things. The gospel of determinism has been known for centuries and probably isn’t going to have a seismic shift on how they think about justice. For the most part they are concerned with deterrence. Some of them worry about rehabilitation, but retribution is a more complex subject. As I said, many libertarians are opposed to retributive justice, and not every determinist is opposed to retributive justice. Determinism just isn’t a magic key that answers hard social questions.

              4. Because lots of people have ideas about things, as you speculate, doesn’t negate the nature of our justice system as predominately retributive.

              5. It is retributive, in the sense that we place a high value on the rights of the innocent. We allow probably-guilty people to go free all the time because that’s considered better than punishing a possibly-innocent person, because we don’t want to punish someone who doesn’t deserve it. That’s the other side of retributive thinking. But the justice system doesn’t reflect a single philosophy. In practice, we tolerate a low (not zero) rate of innocent convictions because it serves the goal of general deterrence. Here’s a little snippet on the topic from a European project on traffic safety:

                “General deterrence can be defined as the impact of the threat of legal punishment on the public at large… The general assumption underlying police enforcement is that it should primarily aim at general deterrence, which is first and foremost achieved by increasing the subjective risk of apprehension.”

                For general deterrence, public perception and compliance is more important than meting out “true” justice, and that’s the pragmatic focus of most wings of the justice system.

        2. Odd how that works, since there is a renaissance in understanding causality! In fact, I have been doing a literature review for the local CFI group on the subject and even I was amazed at how much is being done.

    2. For freewill you need agency but it is not enough.

      Especially compatibilist’s use their own definitions of freewill so yes it is very confusing. And without proper definitions what we are talking about, discussions make no sense.

      I use this simple and effective definition of freewill:

      “The unique ability of persons to exercise control over their conduct in the manner necessary for moral responsibility”.

      We need of course clarification what moral responsibility means:

      “A person who is a morally responsible agent is not merely a person who is able to do moral right or wrong. Beyond this, she is accountable for her morally significant conduct. Hence, she is, when fitting, an apt target of moral praise or blame, as well as reward or punishment. And typically, free will is understood as a necessary condition of moral responsibility since it would seem unreasonable to say of a person that she deserves blame and punishment for her conduct if it turned out that she was not at any point in time in control of it.”.

      It’s from “”.

      Sam Harris has written a nice book on freewill and explains why it doesn’t exist.

      1. Sam’s book is heavily contested, and Sam’s view has been debated heavily on WEIT.

        How useful is moral agency if we can’t distinguish between a morally responsible agent and an agent that doesn’t act responsibly? We learn what we ought to do often by being told what we ought to have done, when we acted without moral responsibility. Praise and blame are tools
        for molding moral behavior.

  13. One does not have a choice but to cope with the illusion of having choice and going about the business of living in the everyday world…thus have we been determined!!

  14. To paraphrase an old (most likely apocryphal) quote about evolution: “If is IS true that we have no will, let us pray that, at least, that it not become generally known.”

    This because of all the people who, not understanding “free” will (or the lack of it) in the first place, would leap to use its “absence” as an excuse for the bad behavior they wanted to exhibit in the first place.

  15. I think the view expressed in this post is really similar to Smilansky. One of Smilansky’s papers argues for “funishment”, the idea that determinist punishment compels us to provide prisoners with a very high quality of life, since they don’t “deserve” to be there, and we should compensate them for what we’ve taken from them. It leads me to think of examples of people who are unlikely to be rehabilitated, like the worst pedophiles. Should we treat their prisons as “sanctuaries”, the way we might relocate a problem grisly bear to a safe place where it can indulge its instincts.

    On a different angle: if the chief goal of justice is deterrence, then one might conclude that its important to visibly punish someone, not necessarily to punish the person who actually did it. Right now a lot of people are arguing that we can deter rape by ensuring someone is punished, even though the person’s real guilt or innocence is inconclusive. At universities we’re now told to use a “preponderance of evidence” standard (50% likelihood); some would say that an accusation alone is sufficient to meet this standard, if the defendant can produce no counterevidence. Should we care about the small percentage of innocent men who will be convicted as rapists?

    For another example, Trump wants to go after the families of terrorists. Almost everyone (on the left, at least) thinks that would be wrong. It could work to deter some terrorists. So why is it wrong? In these examples it seems, to me, that we can’t really set aside the concept of moral responsibility; those pedophiles don’t deserve to have a great life at our expense, and innocent people don’t deserve to be harmed for the greater good.

  16. I am reading Carroll’s “The Big Picture” right now, and find it extremely excellent. He argues “our best theories of people, presented without reference to underlying particles and forces, leave plenty of room for human choice.” (32) I find his argument far more compelling than the determinist view.

    1. Compatibilists like Carroll call it emergence. No free willers call it illusion. IMO, neither provides a drop-the-mic argument as to why their description is right.

      Then again, if this question has puzzled great minds for two millenia, it would be surprising if we know the answer now.

    2. I agree that it is a very good book (I’m reading it now). Not sure I agree about the human choice part. But I’m still mid-book and must be patient….

    3. “I find his argument far more compelling than the determinist view.”

      I haven’t read the book, but I’d be incredibly surprised if Carroll is not a determinist. Can anyone else clarify this?

  17. My boss thought that evolution was the big bang.
    He makes four times as much as I do. When he told me everyone has free will, I told him the burden of proof comes from the one making a positive claim. He told me he would need time to respond, I told him to take at least two days for rebuttal. It is now over two weeks and he has nothing to say.

    1. I smiled a bit at the wage difference. I’ve reported to a lot of doofuses in my day who made a lot of money.

  18. Suppose you knock over a wineglass, whether accidentally, in a fit of rage, or by flubbing a parlor trick. It doesn’t matter. Does a microscopic quantum mechanical description of the motion of the atoms in your arm shed any light on the event?

  19. It’s kind or dorky to bring up dictionary definitions, but here’s the first one I came across:

    Deter: to cause (someone) to decide not to do something

    … which sums up the meaning of deter for me.

    Will you grant, at least, that that definition is in conflict with determinism?

    1. No, because “to cause (someone) to decide” is a rather good definition of determinism!

      (Emphasis on cause rather than decide)

    2. Actually, I think the words “to decide” are completely superfluous in that definition. So I’ll only grant that they obscure the actual meaning of the word “deter”.

  20. I read that Atlantic article when Sam Harris tweeted it. Nice to see it get comment here.

    I really don’t care for the “illusionism” approach, for I think similar reasons most of us here would reject it. (We want to be talking about the truth).

    Jerry wrote: “My response is that we can still hold people responsible, but not morally responsible.”

    I’m very sympathetic to that line of reasoning. It seems to me that a truly deep “blameworthiness -buck-stops-here” type of moral responsibility is very problematic, and it’s best if we recognize this.

    At the same time, it seems to me we have to be careful not to undermine morality: I think we all agree morality still applies whether determinism is true or not, and it does seem odd to be considered moral agents…but never morally responsible.

    The studies that purportedly show that lowering belief in free will is associated with lowered sense of moral responsibility – if sound – undermines how we really don’t want to be making a mistake on this subject.
    To me this is what can happen when you aren’t careful how you speak about determinism, and the baby slips out with the bathwater. If you want to emphasize the lack of moral responsibility for past actions by saying “People really didn’t have a choice” then the logic of determinism projects forward as wall that “we really don’t have a choice” and then (if those studies are right) you risk people acting in a confused manner, sort of fatalistic in terms of their moral responsibility.

    We need a way of talking about the reality of determinism – and all the environment/biological influences on behavior – and the lack of “ultimate libertarian freedom”….while NOT making the mistake of undermining the very real choices we have to make. That is we need to retain a robust sense of “having a choice” so as not to invoke a false sense of fatalism.

    And, to no surprise, I think compabibilism still does this best. (As I’ve argued before, both Sam Harris and Jerry have rightly argued that determinism does not entail fatalism…but then slip into ways of talking about determinism that DO invoke a form of fatalism).

  21. My take on illusionism would be that I find it preferable to know and believe what is true, may the chips fall where they will.

    That being out of the way…

    My response is that we can still hold people responsible, but not morally responsible.

    If the consequence is jail time / a reward if somebody is morally responsible, and jail time / a reward if somebody is responsible but not morally responsible, and if we still distinguish between fully culpable and insane / too immature / not fully in control as we do now (“Yes, of course I agree with Gregory K. I didn’t say that everyone should be treated equally, nor do I think I’ve ever said that.”), then I just don’t see the point of the whole discussion. Yes, there are people who argue for vengence-based punishments, but they are not the compatibilists who generally comment here, or Dennett for that matter.

    it’s basically a semantic relabeling of libertarian free will.

    I am fairly sure that when people invented the word free will, they did so to differentiate between doing what I want and being tricked or forced into doing something that I do not really want. In other words, out of my own free will is the Germanic synonym of Latin-based voluntary. That is also how my native language treats the issue. And that difference is an empirically demonstrable reality, and determinism is irrelevant to whether I was able to act on my innate preferences or not.

    so deeply ingrained in our brains is the idea that we’re conscious agents who can choose freely.

    Same here: This is not an idea somebody came up with in their armchair, it is an empirical observation. I am conscious when I am not asleep or knocked out; that is what conscious means, that is how it was originally defined, that is the difference our ancestors invented it to describe. I can choose freely unless you trick me or hold a gun to my head or suchlike; that is what choosing freely means, that is how it was originally defined, that is the difference our ancestors invented it to describe.

    What’s important to me is not how you define free will, but to always remember that determinism (absent any quantum effects) holds … people’s need to accept the determinism that science tells us is true.

    I always find that a bit puzzling, because quantum effects do exist, and thus determinism can probably be considered empirically disproved. The really important point may then be that randomness doesn’t give us libertarian free will either; the whole concept was incoherent and ill-conceived from the get-go.

    1. My response is that we can still hold people responsible, but not morally responsible.

      How about if this was re-worded to:

      My response is that we can still hold people responsible for their behaviour as agents, but judge their behaviour according to their circumstances as far as we can determine them.

      In the end the judicial system is an attempt to reduce personal affront and bias in judging behaviour – and sometimes works.

      1. Well yes, but isn’t that what all civilised countries already try to do, even without any judge or lawyer having gone through courses in neuroscience and philosophy of the mind? A six year old who steals from a shop is treated differently than a thirty year old kleptomaniac, and both are treated differently than a thirty year old neurotypical person. Telling people not to say free will will neither change that nor bring it about.

  22. “That is a radical view for most people, but it happens to be true.”
    Whose truth its that? Is that a scientifically proven truth? Can science actually ultimately ‘prove’ anything? (see below)

    “. . . accept the determinism that science tells us is true.”
    I was under the impression that science doesn’t really tell us anything is true, but rather that it seems very likely to be true (or false) given our present knowledge. This knowledge can, at any given time, be revised or dispensed with entirely in view of new findings,

    It’s fine if you want to propagate your view and do so with verve, but when you invoke scientific evidence, which is always subject to revision, to underpin your arguments, coupling it with the word ‘truth’, it sticks in my craw, even though I basically agree with many of the things you say. You say, with apparent certainty, “it happens to be true,” and I can’t help thinking that that is a very odd statement coming from a scientist. Perhaps I am being a semantic nitpicker here, but hey, you’re trying to get a philosophical message across, and semantics become very important in that context.

  23. Just skimmed the Sarkissian etc al paper that Jerry refers to (its sort of related to work so I can sort of justify it), and its a poll of 231 English or Spanish speaking undergrads in 4 countries.

    In all 4 the majority went for dualist free will (“indeterminist” in the picture), and also incompatibilism (similar percentages).

    The authors note that those who voted for “determinism” (no dualist free will) also tended to vote for compatibilism (still fully morally responsible). Annoyingly, they only show their two questions individually, they don’t actually give the full table of both results for all 4 countries.

    Summary: most people said that we do have dualist free will, but if we didn’t then we wouldn’t have moral responsibility, or that we don’t have dualist free will, but still have moral responsibility.

  24. Life’s an illusion, love is the dream
    But I don’t know what it is
    Everyone’s saying things to me
    But I know it’s okay, okay

    Everybody’s happy nowadays

  25. This is my first comment here, so I’d just like to quickly say how much I’m enjoying this blog.

    Now to my point: Of course we have free will! You mention quantum mechanical effects twice but don’t follow through the implications of that. Indeterministic quantum mechanical effects certainly operate at a sufficiently macroscopic level to affect the operation of the human brain to the extent of influencing the outcome of decision making processes.

    1. You are dead certain about things that nobody knows for sure. There are arguments about the macroscopic effects of quantum-mechanical phenomena on behavior. If you have data that there are, produce it. You’re apparently the only person in the world who is certain about this.

      We’re waiting. And, of course, since you don’t seem to have followed the discussion, “free will” has nothing to do with the vagaries of quantum mechanics, because we can’t choose to produce those effects.

      1. To address your second paragraph first, (but ignoring your snarky “we’re waiting”) I’m not sure I follow your point – if free will refers only to decision-making processes that are under 100% conscious control, then we have it by definition. If internal or external indeterministic processes are admitted, then does that make our will less free than before? In one way, it does because that decision is then less than 100% subject to our conscious control, in another way, it allows hypothetically hypothetical decision taken under identical circumstances to come out differently, which is another perfectly good definition of free will.

        I will cite:

        The Free Will Theorem

        Tha Strong Free Will Theorem

      1. Worse, a “roulette wheel” (or dice, etc.) So that’s not a “variety of free will worth wanting”.

        I once encountered a very religious physicist who thought that (like K. Miller) quantum randomness was actually god or (libertarian) free willed humans making choices. I wondered how this was compatible with the Born rule, but …

  26. In assessing the societal benefits of embracing Determinism, upgrades to our Justice system is usually the first point made. I think this a huge mistake. The way I see it, the changes Determinism would make to our existing Political beliefs would be radically changed for the better and would have far more consequences to everyone’s daily life than issues of crime & punishment.

  27. I cannot see how any scientifically oriented person who believes that free will does not exist, and even that the sense of it is nothing but an illusion can argue that there is any sane reason to use this sense of illusion to help establish or regulate human behaviours. Science demands truth and the dispelling of all incorrect illusion. Therefor we must teach everyone that they are indeed no more than robots and should only be valued as such. Along these lines then, since our behaviours are not attributable to our own moral responsibility but merely to our “robotic responsibility” I see no reason whatsoever to put any value on the individual mind itself or have any consideration for the person who is a “misprogrammed robot” – i.e. one guilty of behaviour deemed to be a crime. It MUST and it should be a Clockwork Orange type of total reprogramming that we impose on that criminal robot. Also, penalties used as a deterrent for criminal behaviour should also severe rather than mitigated – in a true robot society.
    Of course I myself need not even consider the necessary (and rather horrific) implications of this Clockwork Orange Incompatibilist world …as I myself am a compatibilist.

  28. Good thoughts Jerry.

    As a fellow free will skeptic and disillusionist (meaning I think it’s important that we educate people on the lack of “could have done, of one’s own accord, otherwise” free will), here is my position on free will illusionism (and that Atlantic article):

    I also take issue with compatibilist “semantic shift solutions” that cause problems and bypass important issues in the process.


    1. I also take issue with compatibilist “semantic shift solutions” that cause problems and bypass important issues in the process.”
      I totally disagree.
      Rejecting compatibilists redefinition of free will is done merely a straw man debating exercise used by incomptibilists…. arguing against a obsolete dualist definition of free-will that no compatibilst actually holds. It is a tactic of avoiding real debate on the subject.

      Furthermore, the rejection of such redefinitions in science totally and utterly wrong-minded. Otherwise, for example, we would have to dispense with the term gravity. Redefinition, as a clearer understanding of phenomena developed, is done ALL the time in science. For example, quoting Rossdale and Cox in their communication “Terminology: a mark of scientific progress” – “Scientific terminology must be as exact as is possible within the state of knowledge available.”

      1. @howiekornstein

        >>”Rejecting compatibilists redefinition of free will is done merely a straw man debating exercise used by incomptibilists…. arguing against a obsolete dualist definition of free-will that no compatibilst actually holds. It is a tactic of avoiding real debate on the subject.”<<

        The incompatibilist sees the exact opposite…that it is quite often compatibilists who argue against a strawman version of the incompatibilist:

        That being said, it is sometimes the case that incompatibilist will argue against a strawman as well.

        More often, however, the hard incompatibilist **agrees** that philosophical compatibilist versions of free will "exists" (as they defined) but that such semantic shifts cause way too much confusions and problems for the common layperson. It is like redefining "god" as "the universe" and then suggesting that people will not affirm their "conscious deity" ideas around you telling them "god exists".

        And the incompatibilist version is not "obsolete" when the majority of common laypersons have either incoherent libertarian notions of free will, the incoherent variety of compatibilist notions (rather than a more coherent philosophical compatibilist notion) or both, per various studies and per what is intuitively obvious.

        Redefining terms causes various problems which can be noted here:

        Most of science and philosophy is not about word revisionism, and for good reason. That being said, there is often ambiguity already and it is best for everyone to define terms.

        Sure, you can define any word the way you like…but we need to be careful not to do it at the expense of other understandings, for example, what it means that we couldn't have, of our own accord, done otherwise, etc.

        1. More often, however, the hard incompatibilist **agrees** that philosophical compatibilist versions of free will “exists” (as they defined) but that such semantic shifts cause way too much confusions and problems for the common layperson.”

          Goodness -this sort of stance certainly qualifies as yet another “little people” sort of argument – this being that what compatibilists say is actually TRUE and that compatibilists really have an appropriate description of the free-will phenomenon as it does exist, but since the little people use an unscientific definition we must discount ALL definitions of this phenomenon, even reject the phenomenon itself.
          This is absurd. The “little people”, even reasonably scientifically interested “little people”, have many inaccurate or confused understanding of most of science – especially quantum physics, relativity, cosmology, biology, cognitive science etc. Terms used in these disciplines are totally misinterpreted by little people. Therefore gravity does not exist, evolution does not exist, space does not exist, genetic inheritance does not exist, matter does not exist etc. etc. Now I assume that you accept that despite the little peoples lack of good definition of these phenomenon, that the phenomena do exist. So why discount free will? Surly your selective view in this matter is nothing but straw man argumentation.
          As has been highlighted in many of our debates here on WEIT, what we need to do is educate the little people about DETERMINISM and not confuse them with our own quite abstract debate on the free will controversy.

          1. ps:

            Although I can appreciate the effort you have made to express your view on Free Will, I seriously disagree with you inference that this provides any avenue of providing a path for the “betterment of mankind”.
            Two points here:
            1) Very strong argument can be made that rejecting free will is a path to seriously degrade the accepted norms of moral humanism. This is NOT just about how people might behave if they did not believe in free will. If you want to move the argument onto this vein i would be most happy to oblige.
            2) The very statement of your title shows that you have an agenda. A moral/political/ethical agenda I would infer. I would point out that people who let agendas of this sort into their motivations when doing science are apt to err in their conclusions. That is why we all reject Templeton, isn’t it?

            1. 1) I disagree, and that is why I provided the first link:

              2)If by “agenda” you mean a concern over what is true…then yes, the truth is my “agenda”. But I do not place agendas before the philosophy / reasoning – I place them afterward. Does Coyne have an agenda by explaining why evolution is true? Yep…to educate people on why evolution is true and out of their faulty beliefs in creationism, etc.

              So yes, I have an agenda based on reasoning, NOT reasoning based on an agenda. An important distinction.

              Templeton’s agenda precedes the process. My process builds my agenda, and if something changes, so does my agenda.

          2. I think terms are defined by what the majority of people think. This is why pantheism that defines “god” as “the universe” is as problematic as compatibilism. Such semantic shifts allow the more incoherent notion of the terms to go unabated.

            The fact of the matter is, people have both incoherent compatibilist notions and libertarian notions of free will…*not* the philosophical compatibilist notion. Compatibilism is word revisionism. We might as well call the solar system “geocentric” and just redefine what geocentric means to align with our scientific understanding of planetes revolving around the sun. Rather, we don’t do this. We say geocentricism does not exist, and create a new word (heliocentric). This is far more clear. Compatibilist definitions of free will can be far mor clear without ever using the term “free will which contains all of the incoherent baggage…similar to the way the word “god” contains baggage.

            And no, “free will” is not a scientific term defined via a scientific assessment- so your analogy is off the mark.

            1. And one last point, most people would not say that they know “quantum physics, relativity, cosmology, biology, cognitive science etc.”…and would probably refer to scientists. Most people, however, do say that they A) have free will, B) people are morally responsible in the “just desert” sense, and C) could have, of their own accord, done otherwise, and so on.

            2. No…. Free will is a functional attribute, NOT the explanation of how the attribute comes about. Explanations CAN change while the functional attribute remains quite the same. Explanations improve as human knowledge advances, as has occurred with a compatibilist EXPLANATION of the free will function. This type of progress in explanation occurs with philosophy, history, economics etc etc as just as it does with science

              1. I’d have no problem with the term “free will” changing if the vast majority understood that shift, but the fact of the matter is, the “functional attributes” of “free will” that people believe they possess are of the harmful, incoherent variety….not of the variety that the philosophical compatibilist proposes. And most philosophical compatibilist bypass the fact that people hold to these incoherent notions…in turn allowing them to go on unabated. This is completely different than “advancing knowledge”. The fact of the matter is, we can educate people on all of those “attributes” that the philosophical compatibilist suggests without ever using the term “free will” for it which contains all of the extra baggage. Again, should we convert the word “god” like pantheists would do? Shouldn’t we recognize the confusions that would (and do) occur in doing so? Just look at how many creationists point to Einstein as a theist to support their position…even though he didn’t use the word in that anthropomorphic sense.

                Progress is abandoning (in terms of suggesting they exist) really poor words like “god” and “free will”…not letting them live on in some other form that will only create confusions, ambiguities and problems with historical demarcations of thought. These are not words worth trying to save.

                But we both have our position through a process that lacks free will…so I understand why you think the way you do. 😉

              2. I just cannot accept any of your points of argument. This is again a repeat of the endless “definition games” that we all seem to play in these free will debates. You say you would accept an alternative term for that which presently describes compatibilist free will, to avoid any misinterpretation by the “little people”. Suppose then I make up such a new term, say “decisional agency” which replaces “free will” as the descriptor of the function of the self in its facility to make decisions and bear ultimate responsibility for the decisions made. You would deny this newly defined property even exists, would you not?
                So why then even bother to change terms, it is just a waste of time for us both?

              3. Yes, the debate between compatibilist and incompatibilists is one of a semantic disagreement.

                Regarding the term, most free will skeptics do not deny that we are decision-making machines, or even that we deliberate between options.

                We often don’t have a problem with terms such as “decision”, “agency”, or even “will” or “choice” for that matter…but would take issue with “free decision”, or “free agency”, or “free will”, or “free choice” that qualify these things in confusing ways:


                The free will skepticism I and many others address is not, for example, of the fatalist, futile, or defeatist variety. We think consciousness exists, and conscious thoughts, decisions, and actions play key causal roles in the output of the future. These are indeed things that “exist”.

                The problem comes when people think they could have, of their own accord, done otherwise – in turn allowing for responsibility in the strong (non-utilitarian) sense to be re-enforced:


                Various studies show that the free will abilities that people intuitively feel they and others possess extend to far more *incoherent* properties than simply being an “agent” that “makes decisions”. This is what should not be re-enforced through semantic-shifting of such problematic words. The term “decisional agency” is far less problematic (though I think there are better words than “agency” which is often ambiguous itself).

                Anyway, you get the idea. Peace. 😉

Leave a Reply