The Free Will Wars
We’ve been recently advised by philosopher Dan Dennett not to tell folks they don’t have free will. Of course we have it, he says, and to say otherwise risks wholesale cultural demoralization. After all, some psychology experiments show that, when informed that there’s no free will, people start behaving badly: cheating more, giving conspecifics bigger doses of hot sauce (a measure of aggression), and generally slipping down the slope of irresponsibility into unchecked mayhem. After all, if we don’t have free will we can’t be held accountable, no one’s to blame, and everything’s permitted, right?
But of course we have to pause here and ask the tedious but necessary question of what we’re talking about when denying free will. What do we mean by “free will”? It turns out there’s considerable and sometimes vociferous disagreement in the philo-scientific community on what it means, and what it should mean – a disagreement that reflects variability in ordinary folks’ conception of free will.
Some, like Dennett, when defending free will are referring to our capacity for voluntary, reasons-guided choice-making and behavior. We are free when acting sanely on our own and perhaps others’ behalf, so long as we’re not subject to coercion or manipulation. Such free will – being in sane control of one’s actions – is compatible with determinism. We’d be acting freely in this sense even if it’s the case that our behavior is fully a function of antecedent causes, including our neurally-realized beliefs and desires. Even if, in actual situations (as opposed to counterfactual), we couldn’t have done otherwise due to iron-clad determinism, our actions would still be up to us: we’d be controlling them, not someone else. Since we’re usually in control of our behavior, to tell folks they don’t have free will in this compatibilist sense would be false. So don’t do that.
Others, like scientists Anthony Cashmore, Jerry Coyne and Sam Harris, when denying free will are usually referring to a purportedly contra-causal capacity for choice and action, such that we could have done otherwise in a way that’s deeply up to us as a situation played out. Such freedom, incompatible with determinism, obviously makes the human agent more of an originator – somewhat of a first cause – since choices and actions aren’t fully traceable to antecedent conditions. An exact replay of the situation up to the moment of choice might have turned out otherwise.
However, there’s little evidence to support the claim we are free in this libertarian, contra-causal sense. Libertarians such as philosophers Robert Kane and Mark Balaguer, and neuroscientist Peter Tse, have floated various proposals for how this sort of free will might work. But it hasn’t yet been made clear to most philosophers’ and scientists’ satisfaction how indeterminism makes an agent more responsible – more of an originator – than under determinism. Given indeterminism, things might have turned out otherwise, but breaks in causal regularities in generating options to choose from can’t be credited to the agent. Moreover, we don’t want the agent’s choice among options to involve much randomness, since that would undermine responsibility by sidelining the agent’s character and intentions as primary determinants of the choice. What we can call agent determinism is necessary for control and responsibility. So, if a contra-causal swerve helps to determine a choice, that choice can’t more reflect the agent’s character or motives, so doesn’t make the choice more her doing in any morally significant respect than under determinism, even though she could, and might, have done otherwise. The upshot, it seems to me, is that it’s perfectly ok to tell folks they don’t have free will in this purportedly responsibility-enhancing sense, since there’s no good reason to suppose they do, even if determinism is false. So go for it. But we have to make sure they understand that they can still, and will, be held responsible, since being held responsible is crucial in helping people do the right thing.
We can thus see that the free will wars – disputes about whether or not we should go around denying free will, and what free will really is – are a function of differing definitions. If you’re referring to our capacity for voluntary choice-making that gives us rational control over our behavior, and that makes us responsible, then it would be wrong to deny that. If, on the other hand, you’re referring to a contra-causal capacity that supposedly makes us more responsible than what deterministic voluntary action affords, then it would be wrong not to deny that, at least on the assumption that we want a well-informed public. So the first order of business when discussing free will is to make clear what you’re talking about, then make your point.
The No-No of Denying Proximate Control
In his Big Think video, Dennett relates the fictional tale of man who is treated for obsessive compulsive disorder by having a microchip implanted in his brain (there are such chips, apparently). But after the surgery, he’s told by the well-respected neurosurgeon that his behavior will be monitored and controlled, 24/7, by her clinical team. She tells him, in so many words, that he no longer has free will, even though he might feel like he has it. Although the team exerts no such control (the neurosurgeon is simply playing a rather bad practical joke on the poor fellow), he leaves her office believing that he’s no longer in charge of his behavior; someone else is. Acting under this misinformation, he behaves irresponsibly and ends up in trouble with the law. The point of this cautionary tale, according to Dennett, is that neuroscientists (and more generally, scientists) should stop telling people they don’t have free will, lest folks believe they aren’t responsible and start misbehaving. The free will at issue is the compatibilist free will of being in robust proximate (not ultimate or contra-causal) control of one’s actions.
In USA Today, biologist Jerry Coyne debunks free will, but in the contra-causal, libertarian sense: he says there’s no reason to believe we would do otherwise if we replayed an actual situation, all factors set the same. Quite right, on the assumption of determinism (and indeterminism doesn’t help, see above and note 3). But in critiquing libertarian free will, he seems to conclude that we’re not in control and don’t really make choices:
You may feel like you've made choices, but in reality your decision to read this piece, and whether to have eggs or pancakes, was determined long before you were aware of it — perhaps even before you woke up today. And your ‘will’ had no part in that decision. So it is with all of our other choices: not one of them results from a free and conscious decision on our part. There is no freedom of choice, no free will. And those New Year's resolutions you made? You had no choice about making them, and you'll have no choice about whether you keep them.
But of course we do make choices – decide between alternatives as determined by our beliefs and desires – that’s what the brain is for. It’s just that our wills and choices are constrained by the laws of physics and biology and behavior and whatever other causal regularities end up being true about human beings. Coyne overlooks the agent as an identifiable and potent causal factor in explaining behavior, assigning all control, all the decision-making, to physical laws and antecedent conditions. And leaving out the agent as a source of control is exactly what Dennett is inveighing against. Of course, if you define choices as requiring contra-causal free will, then no one ever makes them. But then you’d need another word for the very real, agent-deterministic process of selecting among alternatives that we go through dozens of times a day.
Like Coyne, in his book Free Will Sam Harris rightly targets contra-causal free will: the idea that “each of us could have behaved differently than we did in the past.” But he also seeks to debunk the idea that “we are the conscious source of most of our thoughts and actions in the present,” which he thinks is central to popular conception of free will (p. 6). Although the causal contribution to behavior of consciousness per se (phenomenal experience, qualitative states such as sensations and emotions) is contested, there’s no dispute that the neural processes associated with consciousness are critical in behavior control. The question, then, is whether we are something other than the neural processes that could be the source of our thoughts and actions, such that our neural processes are putting us out of a job. And of course for naturalist-physicalists like Harris, the answer is no, there’s no immaterial self that’s in competition with, or that could be disempowered by, the brain.
But if this is true, it’s puzzling when Harris says that “your brain has already determined what you will do,” as if there’s a significant difference between you and your brain (p. 9). Such a locution suggests that we are being pushed around by neural processes, passive witnesses to what the brain cooks up. Or, as he puts it, “If you don’t know what your soul [or your brain] is going to do next, you are not in control” (p. 12). But if there is no immaterial self that could be left out of control, then in fact we are in control, as persons physically constituted by brains and bodies. So it isn’t that we just consciously feel that we are authors of our thoughts and actions, but are mistaken in that judgment (p. 26), rather we actually are such authors – after all, no one else is producing them. It’s just that consciousness per se might not be the controller we commonly suppose it is. On a naturalistic view of human persons, I am a perfectly real originator of thoughts and actions, even if I’m completely the product of factors I ultimately didn’t choose, and even if “I, as the conscious witness of my experience, am not the deep cause of [my behavior]” (p. 44). Although the jury is very much out, the conscious “I” and the rest of our experience could simply be an entailment of our being representationally complex behavior control systems.
So in Coyne and Harris we have two instances of scientists engaged (Dennett isn’t, particularly) in the important project of debunking the myth of libertarian free will – ultimate, contra-causal control – but in so doing they verge on denying the reality of robust proximate control. Biologist Anthony Cashmore is a third; he overlooks agent determinism in an otherwise excellent 2010 essay for the National Academy of Sciences that challenges contra-causal free will, see here.
These instances show that Dennett’s tale of the nefarious neurosurgeon is well taken, but the lesson isn’t that we shouldn’t deny free will, if by that we mean the libertarian variety. After all, the scientists in their labs aren’t out to irresponsibly prove and disseminate the idea we don’t control our behavior – that the brain is left out of the loop. Rather they are elucidating the neural mechanisms of such control, and in the process debunking the notion that there exists a contra-causal controller. So the lesson is: when naturalizing the self, be careful not to give the impression that lack of libertarian freedom in any way disempowers us, or renders us unaccountable for our actions, or makes us “biochemical puppets” (Harris’s phrase in Free Will, p. 47). Promulgating such false conclusions might indeed demoralize and destabilize. There is much to be gained in getting clear about who we actually are as agents – connection, compassion, and control – but we don’t want to leave out the agent in the process. As I like to put it, don’t forget about me.
What to Tell People about Free Will
So what should we tell people about free will? Here are some suggestions for engaging in the free will debate which are neutral with respect to any particular view:
- Acknowledge that “free will” is not univocal, such that trying to establish or enforce its “real” meaning might be a fool’s errand.
- Define what you mean by free will up front. What capacities are you referring to?
- State clearly the nature and capacities of human agents as you see them.
- Draw conclusions about the existence (or not) of free will based on the second and third items.
- Draw conclusions about what, if anything, needs to change in light of the fact that we do or don’t have free will in the sense defined. This is why the debate about free will matters.
As might be seen, we could leave out talk of free will and simply talk about our view of human capacities and what the substantive practical and moral implications are of that view, for instance in criminal and social justice, or for our conceptions of credit and blame. But since “free will” and its varying definitions seem here to stay, short hand for various constellations of capacities, we perforce will end up talking about it. When we do, we must be clear, and not mislead folks (this is now my view talking) by claiming that 1) the ultimate, contra-causal control of libertarianism is a live naturalistic possibility – there’s no good evidence for that, or 2) we aren’t robust proximate controllers in our own right – there’s plenty of evidence we are. Not falling into error on either side, we walk the middle way of naturalism.
 For a discussion of Vohs & Schooler’s cheating experiment, see here. As it turns out, this study was one of many psychology experiments which when replicated failed to produce the original findings, see here. Schooler reacts here.
 There’s a burgeoning literature of experimental philosophy on what the folk believe about free will; see for instance Deery et al., The Free-Will Intuitions Scale and the question of natural compatibilism, Philosophical Psychology, Volume 28, Issue 6, 2015, and Rose, et al., Neuroscientific Prediction and the Intrusion of Intuitive Metaphysics, Cognitive Science, 2015.
 Further observations on conveying, responsibly, the radical disutility of libertarianism:
If you tell people they couldn’t have done otherwise in an actual situation, this might strike them as saying they don’t have control, that we are victims of determinism. If we can have done only what we did, given the circumstances as they were, how can we be free? The answer is that, should determinism be the case, we are still free in the sense that what we do, most of the time, is a function of our beliefs, desires, and intentions, so is up to us, not someone else. Even though things couldn’t have turned out otherwise, most of the time we are in control of our behavior, free from coercion or manipulation by others.
Some folks, perhaps many (and all libertarian philosophers), might bridle at this suggestion, saying that real control requires that I could have done otherwise in a situation, and perhaps would have done otherwise, given the exact conditions in play. This gets us back to the problems with libertarian, contra-causal free will mentioned in the main text. The problem is that such additional control, something that could override all the causal factors in play, or decide among them in choosing a course of action, would have to involve a control element other than my very own character and inclinations. Such libertarian control, therefore, would be arbitrary with respect to me and my motives, so would do me no earthly good in terms of guiding effective action. I would have no basis on which to exert it, no motive in service to which the final libertarian, indeterministic choice is made. So I rationally don’t want that sort of extra control. The control I rationally want is for my actions to reflect - to be caused by - me, and I already have that control under determinism, what we called agent determinism. And, as seen in the main text, any random factor introduced in the control process can’t be credited to me, so can’t increase my responsibility.
The upshot is that the additional control some folks think we need to be free and responsible, such that we could have done otherwise in an actual situation in a way that’s deeply up to us, either subverts agent determinism or fails to enhance control or responsibility. If determinism is false, we perhaps would have done otherwise as the situation played out, but not in a way that augments agent determinism in giving us more power or control. On this see also The Flaw of Fatalism.
 About the causal role of consciousness, see Experience and Autonomy: Why Consciousness Does and Doesn’t Matter.