Tag Archives: Philosophy

Free Will Part 3: Moral Responsibility and Luck

Taking into account what I’ve said about Harris on free will, I want to look at what I take to be a credible threat not to free will, but to moral responsibility; this is the problem of luck. I will try to offer some reasons for why moral responsibility can be preserved even while taking into account the extent to which luck undermines control and moral responsibility.

Aside from the belief that libertarian free will entails mind/body dualism and contra-causal powers, the most pervasive confusion gripping this area is the conflation of free will and moral responsibility. To Harris’s credit, he does see how they come apart, and offers us a consequentialist based notion of moral responsibility that he claims is sufficient to preserve a criminal justice system and holding each other accountable. Although seeming to contradict his whole book, this tension resolves once we understand that there are a plurality of justifications for incarceration. Most frequently appealed to and understood is retribution. Retribution is the practice of bringing about an intrinsic good by harming another to the extent that they are responsible for a proportional harm. Less commonly, we restrict the freedom of or directly harm others for reasons such as rehabilitation or protection of the public. I agree with Harris that this is all we really need to preserve our norms of quarantining and penalizing the more dangerous and less trustworthy members of society. This shouldn’t be foreign to anyone, given that this is exactly how we justify the harm parents and educators bring on young children, and to a lesser extent, teenagers and young adults. It’s only at a particular, but indeterminate stage of development that the robust sense of retribution seems appropriate, where you are punished precisely because you deserve it.

Theories of moral responsibility can broadly be distinguished as forward-looking and backward-looking. The former justifies both holding people accountable and society’s response (the enhancing  and restricting of the population’s freedom) based on their ability to avoid harms and produce goods. A desert based form of moral responsibility is committed to the idea that were justified in holding people responsible because they’ve demonstrated themselves as good or bad in a way that is under their control and exemplifies their genuine character. Harris doesn’t think anyone really is deserving of reactive attitudes (e.g. praise and blame)  because he takes desert to require the sort of control he thinks is impossible.

I share some of Harris’s skepticism over our capacity to exemplify a desert based notion of moral responsibility. There is a something inarguably true that our character/nature is not of our own making. We don’t have control over our parents, genes, culture, history, upbringing; all of which shape and constrain our mental and moral capacities. This is the problem of luck. Luck being just whatever is causally and morally relevant to your behavior that you don’t have control over. Almost all of the causal conditions behind our capacities and dispositions are outside of our control, and given that these are the principal sources of our action, how can we be held responsible for them?

I can see four strategies for reconciling constitutive luck and desert:

1. Argue that there are control-compensating features intrinsic to human agency  

2. Reject moral responsibility altogether

3. Revise moral responsibility by substituting desert with an alternative justification

4. Qualify our notion of desert

(1) is the route of most free will affirmers, be they compatiblists and incompatiblists, (2) is endorsed by free will skeptics and hard determinists, (3) is accepted by  Harris, Pereboom, and Vargas,(4) is my own proposal.

The qualification I have in mind consists of making explicit what one is responsible for and what one is deserving of, but first I’d like to clarify the folk conception of desert. The basic conditions for desert are free will and moral competence. Roughly speaking, moral competence consists in an agents capacity to understand and react to reasons (moral and otherwise). Moral competence explains how we respond differently to agents that interfere with our desires. We typically don’t react to a violent dog the same way we do to a violent person, or a drunk bar attendant, or an overly critical coworker. We expect these different classes of agents to be more or less (or not at all) understanding of why their behavior is unacceptable. This is grounded in their ability to understand moral reasons plus the availability of such reasons. For example, this is exactly what is expressed when someone says “they should have known better” in justifying their indignation at another’s behavior.

The central problem with desert is that we can only be justifiably blamed for what we can control, but much of our behavior is a result of our character, which we don’t control. We are born with a set of genetic dispositions, some of which are inexorably expressed, others can be nurtured or stifled. The extent and direction of these latter dispositions is largely dependent on our environment. We have little to no control over these influences, but especially as a child; and this is exactly when the majority of our character development takes place. For the most part, these factors are going to determine whether we’re hot-headed or easy going, depressive or cheerful, ambitious or content, curious or uninterested, stoic or sensitive, prone to addiction or not, etc; all of which can influence our moral competence  

The challenge is whether it even makes sense to blame someone in light of these facts. I think the majority of this problem resolves itself once we understand desert as being explicitly restricted to the aspects of behavior that morally competent agents have control over. This sounds obvious, but most of our retributive practices ignore this, leaving its justification  under-articulated. The result is the fusion of metaphysical freedom and moral competence, leading to the incoherent notion of having metaphysical freedom over one’s moral competence. This seems to be what is going on in the variety of retribution behind our criminal justice system. The attitude behind sentencing is often that the guilty individual choose to become the sort of person who is disposed to steal, kill, abuse, lie, etc. Their behavior expresses who they really are, which they alone determined.  

While acknowledging this as impossible, I think we can preserve a desert based conception of moral responsibility. Praise or blame being justifiable in instances where a morally competent, metaphysically free agent is caught in a torn condition and chooses to act either in pursuit of what is just or what is merely prudent. Consider the following example. Dave is selling his truck out of desperation. He recently accrued severe gambling losses and will suffer worse if he doesn’t immediately settle his debt. A friend of his informed him of an especially credulous, undiscerning, and moderately wealthy coworker that’s been looking for a similar truck. He’s told that he is nearly guaranteed a sale, even while exaggerating the vehicles worth up to 50%. Also of note is that Dave was raised by gambling addicts that actively encouraged him to gamble. He started young and continued to do so until he fell into a set of drug habits. Eventually his health was in danger to the extent that he was desperate to do anything he could to distract himself from drugs. Finding gambling especially effective, he spent most of his free time at the casino, but having an addictive personality, was unable to stop himself. Finally, when showing the truck to his friends coworker, he’s met with enthusiastic response he expected. As tempted as Dave is to sell the truck at an exaggerated price, this is in conflict with his central desire to do what is right. He finds himself in the torn condition of being equally motivated to what is right and what is expedient. In the end, he sells his truck for far more than it’s worth.

I think it’s fair to say that Dave is blameworthy for his action because he choose to pursue what was convenient while knowing it was wrong and having the capacity to do otherwise. Though as we saw, Dave’s nature was largely shaped by features outside of his control. There were elements of luck involved in his nature, his circumstances, and the success of his actions. Since you can only be justifiably held responsible for what you can control, this severely limits the extent of his blameworthiness. There was constitutive luck involved in his genes and upbringing, circumstantial luck in the opportunity to sell his truck, and consequential luck in his loses and buyers acceptance. If any of these features would have been otherwise, he would very likely never scammed the man. Knowledge of these facts should incline us to restrict the extent of our reproach. In the same way we restrain our blame when dealing with children and the mentally impaired, we ought to modify our reactive attitudes in light of the way luck functions in life of all moral agents.

Blameworthiness can be undermined in a variety of ways. Luck is just one. It will be useful to situate luck within the general structure of limiting blame (1). Take some seemingly blameworthy act: out of nowhere, Mac pushes Jack. We naturally ask why he did it. There’s three strategies available to get Mac off the hook. That is, he can be justified, exempted, or excused, all of which can be mitigating or exculpatory. Justification means Mac had morally compensating reasons for pushing Jack. Perhaps if Mac hadn’t pushed Jack, he would have been hit by a brick. Alternatively, Mac could be criminally insane. Mac sometimes lashes out in a way that he can’t control. This would render Mac exempt from blame because he is not equipped with sufficient moral competence and self-control to function as the proper target of praise and blame. Finally, maybe Mac was pushed by Zach, so he couldn’t help but push Jack. In this case, Mac is excused from blame because didn’t have the opportunity control his action in this precise circumstance. These three are examples of exculpatory pleas. They are sufficient to wholly relieve one of blame.

However it doesn’t always work this way. Often when investigating a harm, we’ll discover  context that partially relieves one of blame. For example, if Mac was being intentionally intimidated and threatened by Jack, we’ll find Mac not quite as at fault for pushing him. Since Jack hadn’t yet physically harmed him, Mac wasn’t fully justified in pushing, but perhaps he was to some degree. Next, assume that Mac has a severe anger problem. It doesn’t take much to upset him, and he only has so much control over this. Mac’s limited emotional stability ought to limit the degree to which we hold him responsible. This doesn’t mean he’s not sufficiently morally competent; it just entails that is somewhat diminished in self-control. So when Mac pushes Jack as the result of Jack teasing Mac, we might be not so severe in our condemnation of him. Finally, assume Mac is typically good-natured and exhibits self control, although recently he has been working two jobs (both of which he might lose), getting little sleep, as well as having severe marriage and financial problems. Jack tells a tasteless joke at Mac’s expense. Upon hearing the joke, Mac pushes Jack. Though it’s difficult to hold Mac entirely responsible given the number stressors, he has an excuse for his behavior (to an extent). In these three cases we have instances of discovering information that partially diminishes the extent of Mac’s blameworthiness. A partial justification, a mitigating exemption, or a mitigating excuse.

To review, one’s blame can either be exonerated or mitigated. This can either happen by way of having sufficient warrant to harm (justification), lacking the capacity to grasp and apply moral reasons and to govern one’s behavior in light of such reasons (exemption), or being deprived of the opportunity to properly exercise one’s capacities (excuse).. Whereas Harris seems to think of constitutive luck as exonerating, I take it to function as a universal mitigating exemption. Assuming we sometimes have the opportunity to do otherwise and possess the requisite moral competence, we can still be blameworthy for our actions, but this must be measured against the degree to which we lack control.

This mitigation applies as much to praise as it does to blame, but that’s not nearly as significant. What’s important is proportioning the extent to which we hold people blameworthy to the degree that they exercise control. On any plausible theory of free will and moral responsibility, whether it entails basic desert or not, the current criminal justice system is unjustified. This simply follows from the fact that we don’t include all of the possible mitigating and exonerating features in our assessments. In order to do so, one would have to have perfect moral knowledge as well as access to all of the facts relevant to the how control and luck have factored into anyone’s behavior. This is impossible. For this reason we can’t structure our criminal justice system around people possessing free will in the sense required for basic desert. This doesn’t mean we won’t have one. We simply must appeal to alternative justifications when dealing with those who harm (i.e. protection of the public, rehabilitation), which if performed consistently will inevitably lead to more humane treatment of the incarcerated.

The upshot of all this is that we can still make sense of basic desert in light of the ubiquity of luck, but just not without qualification. No one is praiseworthy or blameworthy simplictor , but this remains consistent with being praiseworthy or blameworthy in the deontic sense. This conception of desert has the advantage of preserving the notion that we exemplify the control necessary for moral responsibility, as well as making us sensitive to the relevant factors that are not in our control.  However this case is dependent on us sometimes being able to make metaphysically free decisions. If Harris turns out to be right that we don’t have this capacity, then I think we would have sufficient reason to conclude that humans are never morally responsible in the basic desert sense.

  1. Franklin, Christopher Evan. ‘A Theory of the Normative Force of Pleas’Philosophical Studies 163:2, (2013): 479-502.

Free Will Part 2: Sam Harris and The Coherence of Free Will


Sam Harris is likely the most influential contemporary figure discussing free will. He attempts to demonstrate not only that we don’t have free will, but that the very notion is incoherent, and therefore impossible. Harris delivers a handful of arguments, each of which he takes to be sufficient to undermine possibility of free will. Although Harris has been largely ignored by professional philosophers, his level of impact warrants a response. Here I show that his arguments either fail outright, or don’t establish what they intend to. I will close by briefly discussing the issues of psychological determination and the self.

A central problem subsuming Harris’s critique is his failure to define his terms, which is shown in his equivocation over the terms  “free” and “will”.

Generic Free Will: “The popular conception of free will seems to rest on two assumptions: (1) that each of us could have behaved differently than we did in the past, and (2) that we are the conscious source of most of our thoughts and actions in the present” 

Perfect Control: “Consider what it would take to actually have free will. You would need to be aware of all the factors that determine your thoughts and actions, and you would need to have complete control over those factors”

I want to contrast these claims with my preferred account,

Free will: The will is the capacity to deliberate, make decisions, and to translate those decisions into action, and exercise of will is free when the agent could have done otherwise in a way that is under the control of the agent. Where an agent is anything that is capable of making decisions.

Although standard, GFW is still too strong. A free exercise of will just requires that our conscious intentions be a unique source of our thoughts and behavior, not necessarily the source. For example, the quarks, molecules, and organs that comprise us will be among the sources (causal conditions) of our actions, but their inclusion in no way undermines our intentional contribution. Further, not every decision must satisfy these conditions, for it could just be that we only sometimes act freely, even if most of our decisions aren’t free. With that said, much of his work is spent arguing against the latter (PC) while claiming victory over the former (GFW). The second conception is clearly impossible, so a demonstration of its incoherence can hardly be taken as a serious contribution.

The first of Harris’s arguments demonstrates the problem,

                                            The Incoherence Argument

“Free will is actually more than an illusion (or less), in that it cannot be made conceptually coherent. Either our wills are determined by prior causes and we are not responsible for them, or they are the product of chance and we are not responsible for them”

Taking this claim apart, we’ll see it only undermines PC, not GFW. It’s not clear what Harris means by “our wills”. This could refer to our character, our deliberative capacity, or any particular decision. However regardless of interpretation, this claim is false. If will refers to our character then it just doesn’t follow that we’re not responsible for them because some of those prior causes could be our own free decisions, which would make us partially responsible for our character. However, if ‘free’ means ‘having complete control over’ and ‘will’ refers to our psychological dispositions, then no one has free will. Neither definition, however, addresses LFW. The proponent of LFW means something different by those terms. Next, if our will refers to our capacity to make decisions in the first place, then it is certainly true that we don’t have free will because no one decides to be the sort of being that can make decisions. Though again, this isn’t what is captured by GFW.

Finally, our will could be any particular decision. The problem here is that the determined/random binary is misleading. It’s true that an event is either determined (causally necessitated) by antecedent conditions or indeterminate (not causally necessitated), but it does not follow that events that aren’t causally necessitated are random. For example, if i’m deliberating on whether I want to read or watch tv, and I don’t have any reason to prefer one over the other, so I just choose one, my decision isn’t random in a control-underming sense. My decision is adequately non-random in virtue of its being rationally authored, and can be made intelligible by citing the reasons I weighed against the competing option.

One might protest that the decision merely seems free, but really there must have been an unconscious neurological process that was the sufficient condition for my choosing one over the other. First, my point here is merely to dislodge worries about incoherence by illustrating how causal determination/indetermination doesn’t of itself render the notion of free will unintelligible, not to argue that we have free will. Second, one would need an argument that all brain activity is necessarily deterministic.

Here Harris thinks the issue is settled

                                                The Empirical Argument

“We know that determinism, in every sense relevant to human behavior, is true…One fact now seems indisputable: Some moments before you are aware of what you will do next—a time in which you subjectively appear to have complete freedom to behave however you please— your brain has already determined what you will do. You then become conscious of this “decision” and believe that you are in the process of making it.”

In support of this he cites the study performed by Benjamin Libet, where they were able to detect brain activity informative of the subjects behavior prior to when the subject reported deciding; which they took to mean that the “decision” was made prior to the subject’s intention to perform the behavior. In the experiment, Libet asked the participants to spontaneously flex their wrist while watching a clock, after which they would go back and pick out which point the clock was on when they decided to flex. It turned out that the average reading of the readiness potential (the relevant brain activity) started about 550 milliseconds before the flex, but the intention occured only about 200th milliseconds before the behavior. They took this to mean that the decision was settled unconsciously, then generalized this to all decision making, meaning we never act freely

Alfred Mele summarizes the argument:

      1. The participants in Libet’s experiments don’t make conscious decisions to flex. (Instead, they make their decisions unconsciously about half a second before the muscle burst and become conscious of them about a third of a second later).
     2. So people probably never make conscious decisions to do things.
3.An action is a free action only if it flows from a consciously made decision to perform it.
4. Conclusion: there are probably no free actions (Mele, 23-24)

This argument is superior in that it’s relevant to GFW, but it still fails. First, the experiment can’t establish that conscious intentions aren’t effective. The readiness potential was only recorded after the person flexed their wrist, which means it doesn’t tell us if the readiness potential can occur with no behavior. That is, it could be that the readiness potential is merely a necessary, but not sufficient condition for flexing. In which case, it’s possible that the decision was made later; perhaps even at the 200 millisecond, right where the intention is located.

Fortunately, Libet did another experiment in which he requested the participants to intend to flex, but then not. The results matched the previous experiment, in that the readiness potential would rise prior to the intention, but would drop after 150 to 250 milliseconds. Libet took this to mean that the decision to flex was still unconsciously decided when the readiness potential began to rise, but in addition we have a conscious “veto-power”. The problem is that the participants never really intended to flex; they intended to prepare to flex and then not. Given this, it’s more likely that the earlier activity correlates with preparing to intend to flex, rather than the intention itself (Mele 12).

In another experiment where the subjects were instructed to click a mouse upon hearing a tone, it took on average 231 milliseconds for the muscle burst to to occur after the tone played.Which is exactly what would we should expect if our intentions are causally relevant (the extra 31 milliseconds would be the time it takes to register the sound) (Mele, 21) Combining these two studies, we have the readiness potential without the intention and action, and the intentional action without the extended preconscious activity. This gives us reason to think that the Libet studies don’t demonstrate that ineffectiveness of conscious intentions. However even if Libet did confirm that the intention to flex was settled unconsciously, this gives us no reason to think all decision making is like this. The subjects were specifically instructed to not consciously think about when they were going to flex. These sort of conditions leave little room for conscious reasoning to play a role. So even if the decision to flex was settled unconsciously, that doesn’t mean that conscious reasoning is never causally relevant.

Similar problems show up in the other studies Harris cites. In one done with an fMRI machine, where “The experimenters found two brain regions that contained information about which button subjects would press a full 7 to 10 seconds before the decision was consciously made”. First, they were only able to do so with 60% accuracy, which isn’t much better than chance (Mele, 27). Also, the finding is consistent with the idea that what they were tracking is merely an unconscious bias, which is compatible with our actions still being up to us. Finally, he mentions an experiment where they could predict where the subject would report when they felt the urge to move with up to 80% accuracy. First, although eighty percent is high, it still doesn’t entail determinism. Additionally, like the Libet experiments, brain activity could be tracking something other than than the decision, such as the preparation to move. The urges also could be unconsciously determined without themselves determining the intention and behavior. Possibly the subject still could consciously choose to act on them or not (Mele, 32-33). And this is exactly what some of the participants reported to have done in the 2nd Libet study. Finally, as in with the 1st Libet study, we have reason to be skeptical these findings generalize into all decision making, especially where conscious deliberation appears to be playing a more prominent role.

Even if these studies fail to demonstrate the ineffectiveness of conscious intentions, he’s convinced that if we had ideal neuro-imaging technology, we could perfectly predict every single thought and action that a person might perform. However this is just to assert neuro-biological determinism. If neuroscience ever does come to construct such a device or something near enough, I think everyone ought to accept that free will is an illusion. Until then, I think the free will affirmer is justified in not being moved by the supposed threat posed by neuroscience.

                                              The Explanatory Argument

The Incoherence Argument and the Empirical Argument are the heart of Harris’s case, but they don’t exhaust it. In addition, Harris occasionally performs an experiment on his audience that he claims to show that we don’t exemplify the freedom we take ourselves to have. He requests us to think of a city, any city. That’s it. He then asks us for an explanation for why we choose the city we did, and upon realizing our inability, claims this to illustrate that we merely witness the deliberation and selection of our decisions. We lack control both over the candidates for selection and the selection itself, given that we can’t even say why it was selected. Harris then rightly asserts that we could gesture at some sort of story for why we chose what we did, but it would almost certainly be a post hoc confabulation.  

Although not obvious, it’s easy to see the problem. Imagine I were to ask you to choose among a set of options, but not to choose one for any particular reason, and then paraded it over you that you couldn’t tell me why you choose the option you did. You would be justifiably frustrated by the implicit conflict. The problem is the demand for a contrastive explanation in a situation where none should be expected. Contrastive explanations take the form of explaining why X rather than Y, where X and Y are both possible, but some additional factor tips the scales in favor of one over the other. The problem is that contrastive explanations are deterministic explanations because the additional factor in conjunction with the prior state necessitates the future state, but free actions aren’t deterministic, so one shouldn’t be expected to deliver one.

One might object that this leaves our decisions inexplicable, and that an action without any explanation can hardly be free. Although a free action cannot be without explanation, not all explanation is contrastive. I can run my own version of Harris’s thought experiment. I start by taking a moment to consider cities: Seattle, New York, Omaha, Portland, Boston… After a minute, I pick Boston. Upon request, I explain why I picked Boston. First, I wanted to comply because it’s necessary to address the point . Next, Boston was among the options that happened to occur to me. Further, it wasn’t one I immediately rejected (I might have some unfavorable association with a particular city). Finally faced with a number of options where I have no overriding reason to prefer one, but motivated by time, I pick one.

I think this line of reasoning suffices as an explanation despite its being a non-contrastive one. Harris might press the issue further, “I see what you mean, but I want to know why you picked that city rather than any other”. I would respond that based on the criteria given, there’s no answer I can give, I just picked. Where he might reply “Yes, but now don’t you see my point, there is an explanation for why you picked Boston, and you’re not even aware of it; where’s the freedom in that?”

The relevant sense of freedom resides in the space of actions unconstrained by the request. To push this as insufficient merely begs the question, in that it assumes that a decision that can’t be given a sufficient reason undermines the control of that action. By sufficient reason I just mean a motivation that entails that a person makes a particular decision. And a decision that is necessarily implied by a motivation isn’t free in the relevant sense. Harris wants to say that my decision to pick Boston is unconsciously settled in a way that undermines my control, but we need not see it that way. Rather it could be that I consciously give permission to my unconscious to decide for me because I’ve exhausted the limits of what conscious deliberation can do. As long as the unconscious efficient cause is constrained by my conscious reasoning, there’s little reason to see it as a threat to my control.

Alternatively, you could take the explanation I gave at face value. That is, in addition to my reasons and desires, the efficient cause is just me. This proposal is just the agent-causal theory of action. Harris would likely consider agent causation untenable given its inconsistency with reductionism. Reductionism is the thesis that some seemingly distinct entity can be identified with some lower level entity, once properly understood; where X reduces to Y just if X is nothing but Y. For example, a pile of dirt reduces to a collection of rocks, minerals, dust, etc; whatever causal powers the pile of dirt has (e.g. fire extinguishment), is merely an aggregate of the powers of the molecules that make it up. So the pile of dirt is nothing over and above the collection that makes it up. An irreducible/emergent substance on the other hand is one that makes an addition to the inventory of what exists (i.e. a mere collection of things doesn’t make a new thing), which is picked out by its causal powers being irreducible to its component parts. For example, a water molecule being irreducible because it exemplifies propensities (to extinguish fire, or dissolve a solute) and liabilities (to freeze at 32°) that aren’t shared by Hydrogen (a highly flammable substance) and Oxygen(freezes at -361.8°F). In the case of agents, substances that can make decisions, they would have irreducible powers to sometimes freely to decide among a plurality of options. Meaning the nature of the agent’s parts and environments would not be sufficient to settle what happens. The agent itself would settle what happens, even if its severely constrained by it’s own nature and environment.

It’s by no means obvious that reality includes emergent substances. It could be that the world wholly consists of fundamental particles and their space-time locations, where the higher level entities we refer to in everyday discourse (e.g. doors, cats, gold, etc) turn out to be little more than mental conventions, and by that very fact, reducible to arrangmentments of those fundamental particles (or some alternative). I don’t think this is the case, but I want to highlight the controversial nature of the issue at hand, and to emphasise that the metaphysical assumptions we hold will frame our evaluation.  

Going back, there are three possible readings of the decision to pick Boston; two freedom-friendly readings (one reductionist, one non-reductionist), and one not. Harris’s position is that that our decisions are always going to be determined by something that isn’t up to us, and so like in any instance, the reason I can’t articulate precisely why I picked Boston over the alternatives is because it wasn’t up to me. There was an unconscious determinate, which I had no control over. However as I explained, this is just an interpretation. A fellow reductionist could explain the inarticulable aspect of the explanation as the granting of permission to the unconscious, but I had conscious control over that, so the essential sense of it being my decision is preserved. Finally, if agent causation is true, then it’s completely legitimate to take our self-reports of “just choosing”  in these sort of cases at face-value. That is, me and my reasons and desires jointly cause my action. In instances of motivational underdetermination, the efficient cause is just going to be my free exercise of one of my unconstrained powers.

I want to note that nothing about my description of free will and agent causation has involved anything to do with dualism or materialism. This is because those issues are irrelevant. I agree with Harris that dualism does nothing to assist the possibility of free will. Everything about my case is consistent with naturalism. More radically, I would extend agent causation to much of the animal kingdom. So when Harris asks “The indeterminacy specific to quantum mechanics offers no foothold: If my brain is a quantum computer, the brain of a fly is likely to be a quantum

computer, too. Do flies enjoy free will?”, I can say that if humans do, then likely so do flies.  My account doesn’t appeal to quantum mechanics, but I also don’t see it as a problem; its influence on the brain is an open scientific question. However, further than that, agent-causation entails a theory of causality which largely side-steps the determinism/indeterminism issue by understanding causes as substances with particular dispositions, not events. My proposal is that the irreducibility of particular substances can be of assistance because it explains how wholes can be causally superordinate to their parts.

                                               Psychological Determination

You might have noticed that my explanation of free decisions is reliant on motivational underdetermination, but what about cases where I do have an overriding reason to prefer a course of action? Assuming one is mentally stable and adequately rational, in those instances you are in some sense determined to attempt the course of action you do, and therefore are not free in one respect. Is this a problem? Yes and no. For the most part, this is exactly how we want our mental life to be. We want to pursue the courses of action, that all things considered, appears to be the best thing to do, and we want to do so for the reasons we claim to have.

The threat of this control undermining feature depends on how often and in what ways we are determined by our motivations. In the extreme, one might object that we always pursue what we’re most motivated to do, and therefore never act freely. This may sound almost self-evident, but it assumes that we always have sufficient motivation for whatever we do. The central problem around this issue is its dependence on our self-reports. The psychological determinist will almost always claim that their experience of their own decisions fits the description of always being sufficiently motivated, though myself and others don’t see it the same way. For example, I often find myself in situations like the ones described above, where I don’t have a sufficient reason to do anything in particular. The psychological determinist might find this unintelligible; taking this to mean that I either do things that I don’t want to do, with no ulterior motive, or I do things for no reason at at all.

I have a reason, but not always a sufficient reason for every feature of my decision. This can take the form of a sufficient reason being unavailable or being equally motivated to pursue two or more incompatible courses of actions (torn conditions). For example, let’s say I’m listlessly bored and trying to think of something worth doing. I consider reading, listening to music, watching a lecture, or going for a walk. None of them sound more appealing than any of the others, but all of them sound better than what I’m doing; so being sufficiently motivated to leave my current circumstance, I pick one of them. Alternatively, in a torn condition, I’m positively drawn to two or more incompatible courses of action. Say I have to pick between pizza and pasta, or between staying home and going out. Both sound equally appealing, but I can’t do both. Ultimately, I have to choose. If agent causation is true, then your reasons and desires are not sufficient for your decision (in these sort of scenarios), what makes the difference is you, which is exactly what we want.
It’s also important to note that the extent to which a sufficient motivation determines your behavior depends on how specific it is. If i’m sufficiently motivated to become a lawyer within the next five years, then a large number of possibilities are closed off to me, such as moving to Antarctica or getting my Phd in chemistry, but that goal doesn’t always determine what im going to do at any moment. The same thing happens at every level of specificity. Let’s say, I’m torn between going out and staying in, and I decide to go out. My friend then gives me a number of bars to choose from; three of them sound terrible, but I’m indifferent to two of them, so I choose one of them. Once at the bar, I’m torn between two different hefeweizens, but I don’t want to hold up the line, so I quickly select one. At each of these junctions, either by indifference or tearing, I am afforded the psychological space for a metaphysically free choice. Even if a lot of the actions in between were determined by my mental states, not all of my actions were. We have to accept that our choices are largely bounded by our goals and desires, but I think these limitations still leave ample room for frequent metaphysically free decisions.

                                                                The Self

The concept of the self is rightly met with suspicion by much of the scientific and philosophical community, which is a sentiment that Harris shares. More than that, Harris outright denies the existence of the self. I largely agree with Harris’s dismissal of the self as he describes it. This is where there is a central ego that inhabits the body, specifically in the head and between the eyes. I think this cartesian sense of the self is untenable. In addition to the lack of evidence for a homunculus in the brain, there is the regress problem of the seeming need to posit a homunculus in the homunculus to make sense of its selfhood. Although alternatively, we could think of the self as being the whole irreducible substance. On this account, selves would just be substances with intentional states. There is no special part of substance that is the real self; the self is just the whole substance.

This issue is relevant to free will insofar as one could claim that there is no free will because there is no one to have it. Again, I want to note how influential ones general metaphysics is going to be when assessing this issue. There are careful arguments in favor of denying the existence of the self, but there are also good arguments on the other side. Harris might have reasons to deny the self as I described it, but it is by no means obviously false, and is sufficient to supply the sense of self necessary for free will.


Looking back, Harris gives us four reasons to reject free will:

1. Defining free will as the impossible power of perfect control.

2. Asserting that all events, including decisions, are either determined or random; neither of which is inconsistent with being freely performed.

3. The demonstration of the ineffectivity of conscious intentions by neuroscience.

4. Pointing out our own inability to sufficiently explain our own decisions.

I have attempted to show that these strategies are unsuccessful, at least as argued by Harris. Other philosophers have developed more sophisticated, but similar lines of reason, so I don’t want to say that his skepticism of free will is without warrant. Of the four offered by Harris, I take the neuroscientific angle to be the best, but at this point free will remains an open scientific question. My hope is for the reader to leave this piece appreciating that the free will issue has to be understood in the context of the relevant philosophical issues. Until we have a better understanding of the nature of the self, causation, motivation, reduction/emergence, explanation, and the brain, we will remain in the dark about the extent of our own control.

                                                          Work Cited

  1. S. HARRIS, Free Will, Free Press, New York 2012
  2. A. MELE, Free: Why Science Hasn’t Disproved Free Will. OUP 2014

Hedonism and Cardinal Utility

Recently, I have decided to embrace preference-based theories of welfare over hedonism. I had two reasons for doing so. First, due to the difficulty of assigning cardinal utilities to mental states. Second, because I no longer feel as though there are any grounds for claiming that preferences that do not supervene over mental states are irrational—and if they are not irrational, then we ought to respect them when it comes to moral decisions. In this post, I will expand upon the first point.

Before talking about ethics, it would be beneficial to briefly review the decision theory necessary for understanding the notion of cardinal utility. For more detailed discussions of these topics, see Schwarz or Peterson (do note that the former is open source). Put simply, according to orthodox decision theory, every individual has a utility function, which intuitively measures “how good” outcomes are for that individual (according to their preferences—we’re not yet concerned with making any substantive claims on what those preferences should be). Formally, the utility function is a function from the set of possible outcomes to the real numbers; outcomes with higher numbers assigned to them are preferred by the agent. The utility function does not just describe the individual’s preference ordering; it also captures the magnitudes of their preferences. Let’s say A, B, and C are outcomes, and say A has utility 0 and C has utility 1.  Suppose further that I prefer C to B and B to A. Then the utility of B should be somewhere between 0 and 1. But it matters where between 0 and 1 the utility of B is—the mere ordering of these numbers is not all that matters. If the utility of B is, say, 0.6, then in some sense my preference for B over A is “stronger” than my preference for C over B. If the utility of B were 0.4, this would not be so. Because relative magnitudes between preferences matter, we call the assigned utilities cardinal utilities. This is distinct from mere ordinal utilities, which only capture the order of one’s preferences.

According to expected utility theory, if an agent is faced with a choice between two gambles, she should choose the one with a higher expected utility. A gamble is an assignment of probabilities to possible outcomes. If the possible outcomes of a gamble G are o_1, o_2, ..., o_n, each of which will yield with respective probabilities p_1, p_2, ..., p_n, then the expected utility of G is given by \mathrm{EU}(G) = \sum_{i=1}^n u(o_i)\cdot p_i, where u is the utility function assigning real numbers to outcomes. As an example, consider the above situation with outcomes A, B, and C which have respective utilities 0, 0.6, and 1. Let the gamble G1 denote getting outcome B with certainty, and let G2 denote the gamble which gives A with 70% probability and C otherwise. G1 has expected utility equal to 0.6. G2 has expected utility 0*0.7+1*0.3 = 0.3. G1 has a higher expected utility, so I should choose G1. Expected utility can be thought of as the average yield of a gamble; if I were to partake in G2 over and over, on average I would get 0.3 utility. Notice that the assigned cardinal utilities are crucial to what decision I make: if the utility of B were 0.2, my preference ordering of outcomes would be the same, but I would have to choose G2. So, knowing how my preferences order the outcomes isn’t enough to tell me how I should make choices under conditions of uncertainty.

So far, I have only described cardinal utilities as representing the “strength” of a preference. However, this is quite a weak basis for expected utility theory. For sure, people do often have some sense of the relative magnitudes of their preferences (this used to be a standard way of defining cardinal utility—see Harsanyi 1955). But basing cardinal utilities on this intuitive sense would not be anywhere near precise enough to serve as a foundation for expected utility theory. One would not be blamed for concluding that there is no fact of the matter regarding the relative magnitudes of preferences—that the only real facts are about one’s preference ordering of outcomes. In fact, this position, called ordinalism, was argued for in the beginning of the 20th century (Schwarz). Schwarz calls this the “ordinalist challenge”: to give a basis for assigning cardinal utilities to preferences.

How, then, can the expected utility theorist respond? On what basis can we actually compare the relative strengths of preferences, in a way that doesn’t just rely on some vague intuition? The answer is to look at the agent’s preferences under conditions of uncertainty, and to try to use these preferences to define their utility function. Here’s a rough example: as above, I prefer B to A and C to B. Let Gp denote the gamble which yields A with probability p and C otherwise. I thus prefer G0 to B and B to G1, since G0 just always yields C and G1 always yields A. Moreover, in some sense, the “value” of the gamble Gp seems to increase continuously with p. It stands to reason, then, that there must be a unique value of p such that I’m indifferent between B and Gp—if I start at Gp where p = 0 and increase p, eventually the value of Gp will have to “cross” that of B. Let’s say this value for p ends up being 0.6. The expected utility of Gp is just p; since I’m indifferent between B and Gp where p=0.6, we should take the utility of B to be 0.6. This method can be used to assign cardinal utilities to all outcomes.

Two things should be addressed. First, the fact that I was only able to assign a utility to B because I had already assigned utilities to A and C. How, then, am I supposed to get these utilities? As it turns out, you can just choose them arbitrarily. Choosing the utilities of A and C just specifies my choice of the units with which we measure utility. This is similar to measuring temperature, for example. It doesn’t matter what number you assign to any particular temperature; what matters is how these numbers relate to one another. I can choose any numbers I like for two temperatures—say, the freezing and boiling points of water—and I get a valid way to measure temperature just as long as all other temperatures are consistent with the first two. Likewise, I may arbitrarily choose A to have utility 0 and C to have utility 1, just as long as all other utilities are consistent with this choice.

The second thing to address is that the above reasoning likely seems circular. The original challenge was to assign numbers to outcomes so that we can tell agents that they should maximize expected utility. I showed a method for assigning such numbers by reasoning backwards, and assigning the numbers which would make the agent’s preferences consistent with expected utility theory. Is this not circular? A utility function is said to represent an agent’s preferences provided that she prefers a gamble G1 to G2 iff the expected utility of G1 is greater than that of G2, with respect to that utility function. The expected utility theorist merely wants to say that there exists some utility function that represents a given agent’s preferences—if the agent has rational preferences. A representation theorem is a theorem which guarantees the existence of a utility function representing an agent’s preferences, provided that her preferences satisfy certain axioms. Arguing that agents should be expected utility maximizers, then, comes down to arguing that their preferences should satisfy the axioms of some representation theorem. This eliminates any circularity. One example of such an axiom is transitivity: if you prefer G1 to G2 and G2 to G3, then you should prefer G1 to G3. Another example is continuity: if you prefer the outcome A to B and B to C, then there should be some probability p such that you’re indifferent between B and a gamble that yields A with probability p and C otherwise. These two axioms are found in von Neumann and Morgenstern’s representation theorem, which defines cardinal utility by looking at preferences between gambles. Another is Savage’s representation theorem, which not only provides a utility function but also a probability function assigning probabilities to events.


Preference utilitarianism holds that total (or average, but that’s beside the point) utility should be maximized, where an agent’s utility function is just constructed from their preferences as above—or, from what their preferences would be if they were rational (see Railton 1986 for a discussion how this notion of welfare can work). It should be noted that although decision theory generally does not place substantive constraints upon agents’ preferences (merely formal constraints, given by axioms such as transitivity and continuity), preference utilitarians may hold that there are reasonableness constraints which should be taken into account (Harsanyi 1977). These may take the form of substantive constraints on the content of preferences.

Hedonistic utilitarianism disagrees with preference utilitarianism in how it defines an individual’s welfare. According to preference utilitarianism (as I have described it), an individual’s welfare is measured by the utility function that represents an idealized version of that individual. According to hedonism, cardinal utility is an objective property of mental states—some mental states just are intrinsically better to experience than others, and by a certain amount. In referring to the utilities of mental states as objective, I mean this: whether a mental state in better than another, and by how much, only depends on properties of the two mental states, and not explicitly on any other factors, such as the individual’s desires. In particular, these utilities are the same for different individuals. An individual’s welfare at a certain time, according to hedonism, is thus given by the utility of the mental state they are experiencing at that time. Hedonistic utilitarians seek to maximize the sum (or average) of these utilities over all sentient beings.

The fact that for the hedonist cardinal utility is taken to be an objective property of mental states, and not a mere result of an individual’s subjective preferences between those mental states, is crucial. This is also the aspect of hedonism which, in my view, makes it untenable. There are no such facts intrinsic to mental states, and even if there were, they would be in principle impossible to know, even approximately. Any attempt to define the cardinal utility of mental states in terms of something knowable (whether observed from the outside or by introspection) inevitably leads to a notion of utility too subjective to be suitable for hedonism.

Effectively, I am making the claim that hedonists do not have any way to meet the ordinalist challenge described above, whereas preference utilitarians do. A natural proposal is for the hedonist to respond to this challenge the same way expected utility theorists do. Recall that in decision theory, an individual’s utility function is determined by their preferences under uncertainty. A hedonist may likewise attempt to base the cardinal utility function on an individual’s preferences between mental states under uncertainty—call this the “preference approach”. As a matter of fact, I think this is the only option for hedonists. The older approach, in which agents simply judge the relative strengths of preferences of mental states, is basically equivalent to this approach (although it is less precise). This is because, if the individual in question is rational and therefore cognizant of the fact that they should be an expected utility maximizer, judgements about the relative size of utility increments are equivalent to preferences between gambles whose outcomes are mental states. Thus it suffices to focus on the preference approach in judging whether hedonism can meet the ordinalist challenge.

There are two natural ways hedonists can use the preference approach as a basis for assigning cardinal utilities to mental states. First, they can define the utility of mental states according to agents’ preferences between mental states under uncertainty. Second, they can claim that utility is a more fundamental property of mental states, of which preferences are an approximate expression. The first does not seem to work, because different agents may disagree about which gambles are better than others, even if they agree on how they order mental states. Since utility is just defined using preferences, we have no basis by which to resolve this disagreement1. It follows that cardinal utility would not end up being an objective property of mental states. Thus, I imagine most hedonistic utilitarians would prefer the second strategy, and that’s what I’ll focus on for most of the rest of this post. My problem with this position is that, if it is true, then it is in principle impossible to measure the posited properties of mental states, even approximately. This is because there would be no reason to think that preferences between mental states track the actual goodness of those mental states. Thus, the utilities of mental states can vary while one’s preferences between mental states change to “cancel out” the changes in utilities, thus causing these changes to be unobservable. I hope my argument will also motivate the stronger claim that there are no such facts about mental states, but as we will see, this stronger claim is not necessary to refute hedonism.

If hedonists want to claim that our preferences track the objective utilities of mental states (at least approximately), they need to provide some explanation as to why this is so. For example, they can argue that having preferences track the value of our mental states is evolutionarily beneficial. Put very simplistically, the reason why some mental states are preferable to others is to motivate us to act in ways that are beneficial for our survival—the good mental states result from beneficial behaviors, thus incentivizing us to behave that way. For example, getting punched is more painful than getting poked because it’s more important to avoid the former than the latter, since it causes more damage to one’s body. But this can be achieved just as well if an individual’s preferences are “out of sync” with the objective value of that individual’s mental states. This is because any “error” in how objectively good or bad a mental state is can be corrected by another “error” in the agent’s preferences between mental states. We, then, cannot expect the utilities of mental states to be tracked by anything observable, or anything available to introspection.

Here’s an explicit example to illustrate the above point. Let’s say that it’s “twice as important” to avoid getting punched than it is to avoid getting poked (relative to the status quo), in the sense that it is optimal that we are indifferent between getting poked and a 50% chance of getting punched. How is it that our biology makes us exhibit this behavior? According to the hedonist’s story, this is accomplished by having some resulting mental states feel better or worse than others. Say M1 is the mental state had after getting pinched, and M2 after getting punched. Assuming the status quo has utility 0, agents will exhibit the desired behavior if i) the utility of M2 is twice that of M1 (with both being negative), and ii) agents’ preferences are represented by the utilities of these mental states. Thus, we have an explanation for why our preferences should track the objective utilities of mental states: if they didn’t, then in situations like the above, agents would not exhibit optimal behavior.

But there’s a flaw in the above explanation. There are many ways we can get agents to have the “right” behavior when it comes to getting punched and poked. For one, we could do as above, and assign M1 and M2 utilities of, say, -1 and -2, and have agents’ preferences represented by these utilities. But we could also do the following: assign M1 and M2 utilities of, say, -1 and -4, and have agents’ preferences fail to reflect these utilities, and instead acting as though M2 had utility -2. Agents like these can be thought of as experiencing more pain when punched than those in the first case (or perhaps less pain when poked, or some combination of these), but also dispreferring larger amounts of pain to a lesser degree, so that their behavior is the same.

More generally, say Smith is rational according to hedonism, so that the utilities of her mental states represent her preferences. Let M be the set of possible mental states, and f: M \to \mathbb R the function which assigns the correct utilities to mental states. Now suppose Jones differs from Smith in the following two ways. First, in situations where Smith would experience a mental state with utility u, Jones experiences a mental state with utility u^3. Second, Jones’ preferences are not represented by f, but rather by f^{1/3}. Jones is irrational, according to hedonism: she doesn’t take more “extreme” outcomes as seriously as she should. But outcomes are in general more extreme for her than for Smith, so that the two end up having the same behavior.

Given that Smith and Jones will exhibit the same behavior in any given situation, why would evolution make us like Smith (as is required for hedonists) instead of making us like Jones? It seems that the only causal role played by the utility of mental states is acting in conjunction with one’s preferences in order to produce certain behaviors. But as we saw above, one’s utilities and preferences between utilities are underdetermined by behavior. In other words: if I observe that someone prefers a state A to a state B by a factor of two (relative to the status quo), this can be explained in two ways. First, by saying that the mental state associated with A is actually twice as good as that associated with B and that the intrinsic goodness of these mental states are reflected by her preferences. Second, by assigning arbitrary utilities to the mental states associated with A and B and stipulating that the agent’s preferences are not represented by these utilities. We have no reason to think the former is happening instead of the latter.

Obviously the above evolutionary story is simplistic, but it illustrates my point. Moreover, I think the same reasoning can be used to undermine any story the hedonist tries to give for why our preferences track objective utilities.

I would go so far as to say that this sort of underdetermination shows that mental states have no such properties. No doubt, this sort of argument will remind some of arguments made against phenomenal properties of consciousness more generally, such as Dennett’s attacks on the concept of qualia using the considerations like inverted spectrum thought experiment (Dennett 1988)2. The idea is that, assuming color qualia exist, there is in principle no way to tell whether yours are different from mine—that is, no way to tell whether you actually see what I call “blueness” when we look at something green, and vice versa. Dennett uses considerations like that to argue that the concept of qualia doesn’t refer to anything real. I’m basically telling a similar story about utility, replacing the inverted color spectrum with an “inversion” of preference and utility.

I happen to agree with Dennett’s conclusion, but no doubt many people find this too counter-intuitive, and still assert that there is a fact of the matter about whether your color spectrum looks the same as mine, even if it is in principle impossible to find out. Should these individuals reject my argument here on a similar basis? I think there’s good reason to accept my argument as damaging to hedonism, even if one isn’t convinced by eliminativism about qualia. Because even if one still wants to assert that mental states have objective utilities, I have shown that these utilities are unsuitable to serve as the basis for a normative theory. If hedonism is to be nontrivial, it has to tell us what to do in some situations where ordinal values of outcomes are not sufficient to make a decision. Hedonism tells us to base these decisions off objective cardinal utilities of mental states. But, as I have shown, there is no basis by which to resolve disagreements about these objective utilities—thus, agents do not have access to the information required to make (even approximately) correct decisions according to hedonism. In other words: even if one wants to assert that mental states have objective utilities, one must admit that it is impossible to use them as a basis for decision making. Thus hedonism fails to be actionable in any cases which require comparisons of cardinal utilities. Imagine I had a normative theory which treated as absolutely crucial interpersonal comparisons of color qualia—the theory says you should do one thing if our color qualia are the same, and another thing if our spectra are inverted with respect to one another. Even if you’re committed to believing in color qualia and that there is a fact of the matter as to whether our spectra are inverted, such a theory should nevertheless be seen as defective, as it requires us to act according to information that we in principle cannot have access to. But hedonism is just as bad. It is in principle impossible to know whether we are like Jones or like Smith, but hedonism prescribes different actions in each case.

To summarize the above point: in order to assign cardinal utilities to mental states, hedonists appeal to our intuition that some preferences between mental states are stronger in magnitude than others. But all that is immediately apparent is that these preferences track a subjective attitude towards gambles between mental states, not recognition of some objective property of them. Hedonists have to provide some justification for the claim that these objective properties exist, and why our attitudes track them. I have shown that even if they do exist, they are causally inert (because any change in objective utility can be “cancelled out” by a change in preferences between objective utilities, keeping everything observable constant), so there is no reason to expect our preferences to track them, even approximately. This undermines any story the hedonist provides for why our preferences track objective utilities of mental states.

There’s one more point I should address. Above, I stipulated changes in mental states without corresponding changes in agents’ preferences. That is, I assumed that it is possible for individuals’ preferences between mental states to not be represented by the utilities of those mental states. Could the hedonist then claim that it is in principle impossible for this to happen? Lazari-Radek and Singer in passing make the ordinal version of this claim in their book when discussing future-Tuesday-indifference. When considering the rationality of someone who doesn’t care about anything taking place on future Tuesdays, they stress the importance that this only applies to future Tuesdays by saying “If I am now experiencing a sensation that I have no desire to stop, then what I am feeling could not be agony” (p. 46) So, they’re basically claiming that it is in principle impossible to be wrong about whether agony is worse than the status quo—if someone thinks a feeling is better than the status quo, then it wasn’t agony to begin with.

Could hedonists make the same claim, but about cardinal utilities? That is, could hedonists claim that it is impossible for agents to really disagree (to a large extent) about the values of gambles between mental states, and argue that there would have to be some difference between the mental states in questions to account for any disagreement? This view is really problematic. Say I’m trying to figure out my preferences between gambles G1 and G2. At a first pass, my choice as to which I prefer will depend on three variables: properties of G1, properties of G2, and the properties of myself when I am making this decision. To use this defense, then, the hedonist needs to claim that the outcome of this decision process actually doesn’t actually vary with the third variable. This is a substantive claim which would need justification. Moreover, it seems to be clearly false. One cannot just magically output their preferences given G1 and G2; the brain has to have some process by which the outcome is determined given the necessary information about G1 and G2. But then a person could just be re-wired so that this process works differently, and the outcome is reversed even when G1 and G2 stay the same. Maybe such a re-wired individual would necessarily be irrational, but they would still be a counterexample to the claim that preferences in principle must reflect objective properties of mental states.


All of the points I’ve made warrant further elaboration, but this is getting a bit long for a blog post, so I’ll conclude here. To summarize, I disagree with hedonism on the basis that there isn’t any suitable basis for assigning cardinal utilities to mental states, even granted the existence of objective ordinal values of mental states. Any attempt to define the utilities of mental states in terms of preferences (or, equivalently, introspection about those mental states) leads to a notion of utility too subjective to be suitable as a basis for hedonism. Any attempt to claim cardinal utility as an objective property of mental states leads to skepticism about utility, because there is no way to derive the utilities of mental states unless we assume our preferences between them are (at least approximately) rational, and there is no basis for making such an assumption.

I welcome any feedback. I’m not horribly well-read, so it’s not unlikely that someone has made arguments similar to the above. If they have, let me know.



  1. When talking about disagreements about gambles whose outcomes are mental states, I rely on the assumption that interpersonal comparisons of mental states are possible. This assumption is innocent here, since hedonistic utilitarians assume such comparisons are possible anyway.
  2. In fact, Dennett briefly argued against the idea of pleasure and suffering as intrinsic qualities of mental states in Consciousness Explained.



Schwarz, Wolfgang. “Belief, Desire, and Rational Choice”, 2017.

Peterson, Martin. “An Introduction to Decision Theory”, 2009.

Harsanyi, John. “Cardinal Welfare, Individualist Ethics, and Interpersonal Comparisons of Utility.” 1955

Dennett, Daniel. “Quining Qualia”, 1988

Lazari-Radek and Singer. “The Point of View of the Universe”, 2016.

Railton, Peter. “Facts and Values”, 1986

Free Will Part 1: Agent Causation & Metaphysical Naturalism

It’s no longer uncommon for free will to be met with suspicion. This suspicion is even greater when it comes to libertarian free will, and overwhelming regarding agent causation. This belief is largely arrived at via the notion that agent causation or even free will in general is inconsistent with Metaphysical Naturalism. This attitude is mistaken. Here I propose to show that even an agent causal account of action is consistent with Naturalism, which implies that free will in general is. I’ll close by arguing that at least some people are justified in believing in free will.


Metaphysical Naturalism (MN) is a meta-philosophical position regarding the fundamental nature of Being, the world, etc. What it entails is largely debated, but I will be using two definitions that are generally accepted.

MN1: Everything that exists is natural. There are no supernatural entities or forces.

MN2: Reality is exhausted by space-time and its contents, or an ensemble of space-time manifolds.

MN1 is the most common version, but it’s largely uninformative because “natural” is left unaddressed. We’re merely left with picking out paradigmatic supernatural entities/forces such as ghosts, gods, magic, and the like, and asserting that nothing of the sort obtains. I prefer M2, but I think assuming the truth of either one of them is sufficient for what I hope to demonstrate.

Free Will

To understand why people assume agent causation is inconsistent with MN, we have to clarify what free will is. First, the will is the capacity to deliberate, make decisions, and translate those decisions into action (Franklin, 2015). I take the folk conception of free will to mean that persons are sometimes able to exercise their will such that they could have done otherwise. That is, at least some decisions aren’t necessitated by their nature and/or environment

More clearly, an action is free only if it satisfies the following conditions:

Sourcehood: The agent is the actual source of ones action (e.g. no manipulation).

Intelligibility: The agent performs actions for reasons that are understood by the agent (e.g. a spontaneous jerk isn’t a free action).

Leeway: The agent is able to refrain from performing the action.

It’s often assumed that naturalism entails determinism, and that determinism is in conflict with the leeway condition, and by this very fact naturalism is in conflict with free will. But this entailment does not hold. There’s nothing about naturalism itself that implies that all causal relations are determinate (necessitated by the relevant antecedent conditions). All that’s required of causality on MN is that nature is causally continuous. Which means that there is only one metaphysical causal kind within the world (i.e. Dualism is false), and that there aren’t external non-natural causal forces affecting the natural world. For these would almost be by definition supernatural. Further, contemporary physics already admits indeterminism in at least six interpretations of quantum mechanics (three remain agnostic, and four are explicitly deterministic). So if one is going to reject free will in virtue of MN, it can’t be because of MN entailing determinism. One might object that indeterministic events don’t take place in higher-level settings, such as the firing of a neuron, so a naturalistic interpretation of human behavior will be deterministic. First, there’s nothing about naturalism in itself that requires this. Second, whether some events in the brain operate indeterministically is an empirical thesis that remains to be settled, and there are already models of how this might work (Tse, 2014; Franklin, 2013; Weber, 2005)

Given what has been outlined above, we can make sense of an event causal libertarian account of free will fitting within MN. In these sorts of instances, one’s mental states cause one to act but in such a way that you could have done otherwise. That is, the features of yourself that cause the action wouldn’t necessitate the action. You could have refrained or performed an altogether different action. It’s also helpful to note that this model fits nicely with the reductive account of mind, where any token mental state is identical to a particular brain state. Most philosophers specializing in free will recognize event causal libertarianism as a possibility worth considering, even if they remain skeptical of its reality (Balauger, 2004, 2010).

Agent Causation & Substance Causation

This charitable tone tends to drop once agent causation is proposed. This is typically followed by accusations of anti-scientific and “spooky” metaphysics. This is primarily grounded in the assumption that agent causation implies substance dualism. They can’t imagine what this agent could be besides a disembodied mind that interacts with the body. I think the agent causal picture people have in mind is much like how Kant thought freedom of the will worked. Essentially, the physical world that we experience is fully deterministic. Everything runs like clockwork with the exception of human action. In addition to bodies, persons are also noumenal selves that transcend the empirical world, making sovereign unconstrained choices each time they deliberate and act. So on this picture, the world consists of two different sorts of causes, natural events and agents. Given this sort of description, it’s of little surprise that so few philosophers take agent causation seriously.

Before we contrast the previous description with how agent causation has been recently updated, it will be useful to offer a brief description of what event causation is supposed to be. Event causation essentially involves some complex state of affairs or process causing another. For example, a heart pumping causes the movement of blood or a brick being thrown causes the window’s shattering. Further, the way these events unfold are explained by whatever laws of nature happen to obtain, be they deterministic or probabilistic. Causation cashed out as event relations can either be understood as ontologically primitive or reducible to something more basic such as facts concerning the global spatiotemporal arrangement of fundamental natural properties or sequential regularity.

Timothy O’Conner offers two similar, but philosophically distinct analyses of causation which clearly sketch the relevant difference between event and agent causation (O’Conner, 2014):

Event causal analysis: “The having of a power P by object O1 at time t produces effect E in object O2.”

Agent Causal analysis: “Object O1 produces effect E, doing so in virtue of having power P at time t.”

In the first case it is the “possessing a power”, an event, which is the cause of the effect; in the second it is the object. What’s of crucial importance here is that the agent causal analysis isn’t actually just one of agent causation, but is of the more general theory of substance causation. Substance causation is just the theory that substances or objects are what cause effects. So on this account, it’s not the throwing of the brick that causes the window to shatter; properly speaking, it’s the brick. Now this might sound absurd; how could the throwing of the brick not be a cause of the windows breaking? The absurdity drops once we consider the thrower. Really, the thrower and the brick jointly cause the windows shattering, where the throwing is a manifestation of a power possessed by the thrower. Powers theory is crucial to any plausible theory of substance causation. It’s not merely the object in itself that causes the effect, but the nature of the object that is constituted by the powers it possesses.

Most of the mysteriousness of agent causation disappears once we understand it as a species of substance causation. So take any ordinary substance, a rock, an electron, a water molecule, etc; any time any substance causes an effect on another substance, we have an instance of substance causation. What distinguishes agent causation from ordinary instances of substance causation is that there is an intention behind it. This entails that agent causation is fairly common place within the animal kingdom, which itself is good reason to believe that agent causation is consistent with naturalism.

A robust defense of substance causation is beyond the scope of this paper, but I can briefly sketch some reasons for accepting it. One is the numerous problems with alternative theories of causation. The constant conjunction or sequential regularity theory is currently one of the most popular and has been since Hume proposed it. On this account, for x to cause y is just for it to be the case that every time x occurs, y occurs. So on this view there is no intrinsic or necessary connection between the fire and the smoke that follows; this is just the way the universe happens to unfold. A contentious assumption on this theory is that all instances of causality are temporarily ordered. But we can make sense of non-temporal causation such as two cards propping each other up or a ball making an impression on a pillow that it’s been resting on for eternity(i.e. there was no prior time where ball was not affecting the pillow).

The other popular account reduces causation to counterfactual dependence, which is something like this,

“1) If A had not occurred, B would not have occurred.

2) If A had occurred, B would have occurred.

3) A and B both occurred. “ (Scholastic Metaphysics, pg. 60).

So the throwing of the brick causes the window breaking because if you remove the throwing of the brick then the breaking would not have happened. One problem with counterfactual dependence is the infinite number acts of omission that are involved in any causal sequence. So my successfully walking across the street was dependent on not being crushed by an elephant, not being transported, the earth not blowing up, etc. Another issue that’s applicable to both theories is that both of them seem to get the dependence relation wrong. It’s because of causation that there is constant conjunction and counter factual dependence. They are symptomatic of causation.

Next, here is a simple argument in favor of substance causation: (Whittle, 2016)

1. Some actual substances possess causal powers.

2. If a substance possesses a causal power, then it is efficacious.

3. If a substance is efficacious, then it can be a cause.

4. Some actual substances’ causal powers are manifested.

5. Therefore, some actual substances are causes.

The only premise I can imagine being rejected is (1). On the face of it, this might sound absurd; as if it means that nothing has the power to do anything. Though really the individual who rejects causal powers would have alternative explanations for why things do what they do. A not uncommon answer is that we only need appeal to the laws of nature to understand and explain how events unfold. This is problematic. On one hand, if you take the laws of nature just to be descriptions of regularity, then the laws themselves don’t do any explanatory work. On the other hand, if you take the laws of nature to be something that dictates and enforces the activity of things from the outside, then you’ve committed yourself to a form of platonism, where naturalism must be rejected. Finally, you can take the laws themselves to be the causal implications of the intrinsic natures that the substances possess, and in that case we’re back to powers theory.

Metaphysical Irreducibility

One might object to my earlier claim that agent causation is fairly common place because in reality there are no agents, merely matter in motion or atoms in the void.This is where the possible reducibility of macro-level objects becomes an issue. So a largely reductionist metaphysics will hold that much of what we consider ordinary objects are nothing over and above their parts. So what they are is wholly reducible to a set of fundamental constituents and relations. Another way to think of about this is that if we were to take an inventory of everything that really exists, much of what we take to exist would turn out to not. At its most extreme, the reductionist thesis holds that there’s nothing over above quarks, bosons, or whatever a complete theoretical physics takes to be fundamental. Ordinary objects will be described as simples (indivisible physical objects) arranged in a particular way. So to be a cat is just to be simples arranged cat-wise.

If one were both a reductionist and a substance causation theorist, then one could rightfully reject agent causation because there would be no agents in the relevant sense. In order for agent causation to obtain, the agent has to be a unique substance that’s not merely the sum of its parts. If agent causation were true, then agents would be irreducible substances whose persistence conditions are picked out by their higher-level causal powers(e.g. Purposiveness, narrativity, & self-reflection). That is, we are unique irreducible substances because we possess capacities that aren’t exemplified by our constituents. The constituents have come together in the right way; they are not merely a collection of them. A unique form is exemplified that puts constraints on the activity of its lower-level constituents. Which is an example of top-down causation if anything is. On reductionist substance causation, the lower level substances do all of the causal work.

A possible strategy for motivating a non-reductionist account mirrors the demystifying of agent-causation. That is, if irreducible objects aren’t special cases that are essentially restricted to persons, then there’s less reason to be suspicious of irreducibility in general. This does not mean that I think that all ordinary objects are irreducible substances. I take objects of artifice to be clearly reducible to their chemical constituents. So houses, cars, computers, tools, etc are reducible to their constituent parts. Edward Feser offers a clear description of the distinction I have in mind,

The basic idea is that a natural object is one whose characteristic behavior – the ways in which it manifests either stability or changes of various sorts – derives from something intrinsic to it. A nonnatural object is one which does not have such an intrinsic principle of its characteristic behavior; only the natural objects out of which it is made have such a principle. We can illustrate the distinction with a simple example. A liana vine – the kind of vine Tarzan likes to swing on – is a natural object. A hammock that Tarzan might construct from living liana vines is a kind of artifact, and not a natural object. The parts of the liana vine have an inherent tendency to function together to allow the liana to exhibit the growth patterns it does, to take in water and nutrients, and so forth. By contrast, the parts of the hammock – the liana vines themselves – have no inherent tendency to function together as a hammock. Rather, they must be arranged by Tarzan to do so, and left to their own devices – that is to say, without pruning, occasional rearrangement, and the like they will tend to grow the way they otherwise would have had Tarzan not interfered with them, including in ways that will impede their performance as a hammock. Their natural tendency is to be liana-like and not hammock-like; the hammock-like function they perform after Tarzan ties them together is extrinsic or imposed from outside, while the liana-like functions are intrinsic to them” (Scholastic Metaphysics, pg. 182)

I don’t commit myself to the idea that all natural particulars are irreducible or simple (without parts) or that only objects of human construction are reducible. For example, a rock made of limestone would reduce to a collection calcium carbonate, that may or may not have an irreducible intrinsic nature. The correct account of reduction/non-reduction relation is a severely under-explored issue in metaphysics. The hope here is merely that this example is useful in communicating an idea of what an irreducible relation/substance is supposed to be.

Final Arguments

Before summing up the arguments, it’ll be useful to explain what sort of advantage an agent causal account of freedom has over an event causal one. It stems from what’s called the “disappearing agent” objection to event causal libertarianism. The idea is that on the event causal analysis the agent-involving events (the particular mental states, preferences, reasons, etc) that non-deterministically cause the decision don’t actually settle which option is selected. The leeway condition is satisfied in that we could roll back the event and you could have otherwise but you, yourself don’t actually choose it. Your agent-involving states merely constrain which options are possible for you. Where it goes from there is a matter of luck. This can be thought of as claiming that an event causal view doesn’t satisfy the sourcehood condition for free will. The events, which do the work, merely flow through you, but you don’t really settle which option occurs. Agent causal theories have the advantage of saying that you certainly do play an explanatory role.

With this work behind us, we can abridge the essential story into a few brief arguments.

1. Substance Causation is consistent Naturalism.

2. The metaphysical irreducibility of certain substances (persons among them) is consistent with Naturalism.

3. If (1 & 2), then agent-causation is consistent with Naturalism.

4. Therefore, Agent Causation is consistent with Naturalism.

I think 1 and 2 are fairly straightforward in that nothing about my description of them implied that they transcend space and time, and 3 isn’t much more than the definition of agent causation.


1. The leeway condition is consistent with Naturalism (i.e. Nothing about naturalism implies that all causation is deterministic or that all causally relevant neural sequences are deterministic).

2. The sourcehood condition is consistent with Naturalism (since the most demanding form of satisfying it (agent causation) is consistent with Naturalism).

3. The intelligibility condition is consistent with Naturalism (I can’t say much more than I’d be completely puzzled if someone denied this, beyond maybe saying that all of our reasons for action are post hoc confabulations).

4. If (1,2 & 3), then Free Will is consistent with Naturalism (A priori true).

5. Therefore, Free Will is consistent with Naturalism.


1. Substance causation is a plausible theory of causation.

2. The irreducibility of certain biological substances is not implausible.

3. Indeterminism is plausible.

4. If (1,2, & 3), then free will is plausible.

5. We’re justified in holding independently plausible positions if they cohere with our background beliefs*.

6. Therefore, at least some people are justified in believing in free will.

Plausible: A position is plausible just in case it is coherent, contains sophisticated arguments or evidence in favor of it (ones that are aware of and address the relevant issues and objections that might undermine it) and is void of any obvious insurmountable objections.

*Epistemic axiom: We’re justified in believing what seems to be true unless we have sufficient reason to think it’s false.

*Phenomenological claim: Some of our decisions seem to be free, to at least some of us.

Without question, this is the weakest of the arguments I’ve offered. Plausibility is context dependent, which means many will find this unconvincing. Some of the most obvious candidates are committed reductionists, scientismists, eliminativists, determinists, and event causal theorists. Though this is not my target audience. My hope is that fence sitters, or anyone who’s just generally skeptical yet open to free will and agent causation might be persuaded to take the position seriously. No one should be moved to believe in free will merely based on what I’ve offered here, but it might be sufficient to motivate some to re-assess their position.

                                                           Works Cited:

Balaguer, Mark. Free Will as an Open Scientific Problem. MIT Press, 2010

Balaguer, Mark. A Coherent, Naturalistic, and Plausible Formulation of Libertarian Free Will. Noûs, Vol. 38, No.3 (Sep., 2004), pp. 379-406

Feser, Edward, “Scholastic Metaphysics: A Contemporary Introduction”, Editiones Scholasticae, 2014

Franklin, Christopher Evan, “Agent-Causation, Explanation, and Akrasia: A Reply to Levy’s Hard Luck”, Criminal Law and Philosophy 9:4, (2015): 753-770.

Franklin, Christopher Evan, The Scientific Plausibility of Libertarianism’, Free Will and Moral Responsibility, eds. Ishtiyaque Haji and Justin Caouette. Newcastle upon Tyne: Cambridge Scholars Publishing (2013): 123-141.

O’Conner, Timothy. “Free Will and Metaphysics,” in David Palmer, ed., in Libertarian Free Will (ed. D. Palmer, Oxford), 2014

Tse, Peter Ulric, Neural Basis of Free Will: Criterial Causation, MIT Press, 2013, 456pp.

Webber, Marcel. Indeterminism in Neurobiology. Philosophy of Science, Vol. 72, No. 5, Proceedings of the 2004 Biennial Meeting of The Philosophy of Science AssociationPart I: Contributed PapersEdited by Miriam Solomon (December 2005), pp. 663-674

Whittle, A. (2016). A Defence of Substance Causation. Journal of the American Philosophical Association , 2(1), 1-20. DOI: 10.1017/apa.2016.1