Sam Harris says the concept of free will is incoherent. Humans are not free and no sense can be given to the idea that we might be. There are good arguments in philosophical and scientific literature that call into question the ability of humans to make truly free choices. Those arguments generally are rigorous attempts to show that certain necessary conditions for free will can’t obtain or particular sufficient conditions don’t obtain. That is, they unpack a clear definition of what it might mean to be free and then attempt to show that nothing could or actually does fulfill the requirements of the definition. Sam Harris’ new book Free Will takes a somewhat unique, and I think ultimately inconclusive, approach. I will focus mainly on the first part of the book in which Harris lays out his philosophical case. The last part of the book is more about application and I agree with Harris that assuming his philosophical case works, his description of how such a situation would apply to the world seems largely correct.
In this section, I summarize what I think is going on in the book and provide a quick analysis of it for those who don’t want to wade through more dense material. For those interested in the fuller story, there’s more in the second section.
In this book, Harris argues that the concept of free will makes no sense and so those who believe they act freely and are responsible for those actions are being duped by their biology. When a person make a decision (and then acts on that decision), the thought “I choose a over b” appears in his mind in the same way a pain experience or a desire for chocolate might. The person doesn’t seem to have any control over where the thought comes from—it just appears “out of the darkness.” Further, the person, try as they might, can’t trace where the thought comes from. It’s genesis is a product of a complex nexus of biology and environment about which the person knows very little and does not control. If this accurately describes the situation, how can we say that such a person is free in any sense that would please those who claim the person is responsible for the decisions they make? No coherent answer can be given to this answer so free will must be a false idea.
But many of us appear to have a strong feeling that we are in control of the choices we make. We seem to think that when presented with options, we get to decide the way things turn out. Harris admits that this feeling is strong but when analyzed, it breaks down. Everything a person would need to really make a free choice—access to everything that gives rise to the choice and complete control over those things—doesn’t happen. But even if we did have everything Harris says we would need, we still could not claim to make our choices freely. This is because the choice still is the product of what is going on in our brain, influences from our upbringing, and our environment. Any control we would appear to have would still be the product of those things. Our brain and environment is involved in everything we do. Because of this, no account of freedom really makes sense. This is partly why, I think, he says the concept is incoherent.
So there’s a problem here. If Harris is right, there is no way to even describe what it actually would be like for an action to be free. But libertarians (people who believe some human actions truly are free) disagree. They argue that this mysterious “appearance” of the thought to choose one thing over another and this feeling that we do perform free acts is the essence of freedom. The thought is mysterious only if one assumes that there must be a story that involves other causes like brain events and environmental influences and the feeling that we are in control is the basis for believing that it is actually us that brings the thought into existence. The two taken together provide support for the idea that some acts are free in the sense that Harris says can’t be possible.
So Harris’ story involves an assumption that everything that happens in the world—including human action—is the product of other events that precede it. Since we don’t have know what those other events are and the cause of the thought involving a choice isn’t something we do—they just seem to appear—we aren’t free. According to libertarians, this assumption is incorrect and our experience of being free along with the fact that we don’t have access to other causes that describe how our thought to choose one thing over another provides a least some reason to think some of what we do is the product of a genuinely free choice.
In what follows, I dive into the specifics of what I think Harris is arguing and present a common rejoinder. Familiarity with some of the philosophical jargon and background information related to problems of freedom will help the reader navigate the analysis below.
Is the Notion Free Will Incoherent?
Before turning to Harris’ core argument, I think there is a bit of confusion about what, exactly, Harris is arguing. For the most part, he seems to be arguing that the very concept of a free act is incoherent. In fact, he states as much at the beginning of the book, “Free will is actually more than an illusion (or less), in that it cannot be made conceptually coherent.” (p. 16). This means that the terms used to define the notion of free will involves a contradiction. Yet, throughout the book, he uses the phrase “free will” as if it is coherent even if false. For example, In the introduction, he states, “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. As we are about to see, however, both of these assumptions are false.” (Ibid.) He even tells us in the next section what it specifically would free will would look like if someone were to have it (p. 24).
This is not a trivial issue. If he wants to argue that this free will is coherent but no one ever exercises it, then an analysis would involve an making an empirical case about whether the things he says must obtain do in fact describe human action. This gives his interlocutor a set of necessary and sufficient conditions to start from and any analysis would focus on the empirical data. However, if his claim is that it’s impossible to articulate a concept of free will that doesn’t involve a contradiction, then any empirical data would be secondary. It’s primarily a philosophical matter and those who disagree with Harris would have to find a coherent notion first (and one that is sufficiently different than Harris’ own case for how human action operates) and then look at any empirical data to see if the necessary conditions are ever met.
On my read, I do think Harris is making the philosophical argument that its not possible to make sense of the idea of free will. The empirical data is given to support the philosophical case but I think the interpretation of those data ends up resting on the assumption that his philosophical arguments are sound (and so can’t be used as their support). I’ll describe why I think this is, below. But let’s look first at what a free act might involve.
Turn the Light On If You Can
Let’s suppose I’m about to climb into bed and realize I forgot my book on the coffee table in the living room. I walk into the dimly lit room to grab the book. The moon is providing enough light that I probably could find the book and head back to bed. But I consider turning on the light to avoid tripping over something. The origin of the idea, “turn on the light” is not all that mysterious. Certainly there is a lot going on in my brain of which I am entirely unaware. But the genesis of the thought is not unanalyzable even by me—it does not appear from whole cloth. It is reasonable to think that it is generated by a host of processes and background beliefs, some of which I have access to. The idea came about because its reasonable (turning on the light will help me avoid tripping), because of past experience (in the past, I’ve tripped while walking through a dark room, light shows objects in a room I may not be able to see by moonlight), because of the influence of others (mother told me never to walk through a darkened room) and the like.
So I have the idea. Let’s say further that I have the idea at a specific time (T1) and the idea is about an action I wish to perform at another time in the future (T3) by a decision I make at an intermediate time (T2). At T1, the decision has not been made so there are two possible states of affairs before my conscious gaze that I believe I have control over: turning on the light or leaving the light off. Suppose I decide to turn on the light. In normal circumstances, I believe I made the choice freely at T2 in that I could have brought about a different state of affairs at T3 had I decided differently. Harris explores what brought about the choice to turn on the light rather than leave it off and asks if we can give any sense to the idea that it was a free action.
Harris’ option is that there is some event-causal story, the only coherent story, that we can tell that explains my action. There is a chain of causes consisting of brain events and other influences that can be isolated (in an ideal science) that describe why my body moved in the way it did to flip the switch and turn on the light. The Libertarian will claim that no such event-causal story applies if the action were truly free. The decision and subsequent action to turn on the light originated entirely from me as a “first cause.” That is, at T1 no set of event causes dictates what will happen at T3. At T2, I, as an agent, make a decision to turn on the light that initiates an event-causal chain at T3. There is a break in the causal chain at T1, I choose as a agent-caused act at T2, which creates a new event-causal chain at T3. It is the second step in this process that Harris says is false (or incoherent depending on how your read him).
Problems with causeless acts
There are obvious difficulties with this notion of an agent-caused act that can bring about new causal series.
- How is the brain involved in the act of choosing? What would it mean to say it’s not involved?
- If an act is not event caused, why did it come about?
- How do goals and desires relate to the act of choosing?
- How do we answer the ‘why’ question (why did you choose to turn on the light rather than leave it off)?
These problems seem very difficult—and they are. So why bother with them? Why not admit defeat and acknowledge that there are no such things like free acts of will? Harris lists some reason why the topic is worth exploring that I can summarize briefly. First, it appears to us that we act in a responsible way. When I turn on the light, it seems to me that I could have refrained and this “feeling” is very powerful. At the very least, this needs to be addressed. Further, it seems to me that I was not controlled in any way in making my decision. The decision seems to me to be up to me in the sense relevant here. Upon reflection, I may have difficulty describing just how, but the feeling is real and drives many intuitions about moral responsibility and human action.
Third, the implications of not being free are substantial, as Harris notes. I don’t think it’s important to decide whether things would be better—Harris thinks they would be—or worse if we conclude we’re not free. Either way, it would be a paradigm shift that would have ramifications for just about every aspect of how we think about our world. So these questions are important and Harris attempts to address some of them.
Harris sets the foundation for his position by attempting to analyze why it is so many believe they act freely in the first place. I think he does this partly because many defenses of free will start with a phenomenological claim: it seems to us that we do, in fact, act freely and are responsible for those acts sometimes. If Harris can dispense with this experience or at least tell us why the experiences aren’t what we take them to be, he can cast doubt an essential Libertarian starting point which helps his case. In the section titled, “Choices, Efforts, and Intentions,” he discusses the "feeling" of freedom and states that this feeling, while not inconsequential, is not relevant for the actual exercise of a free choice. The root of his claim is this: the experience we have of making a free choice is superficial and collapses under analysis. The feeling of freely choosing a over b appears to be caused by us but when we actually look at what is going on, we see that no sense can be given to the idea of a first cause. All action, including those which we deem freely caused by us are actually caused by a series of events most of which we have no epistemic access to or control over.
On first glance, when I have the thought “turn on the light,” it may appear to me that I initiated the thought. I “decide” to turn on the light and it is that act of deciding that “creates” the thought and kicks off a causal chain to move my body in a way that flips the switch. But on closer analysis, Harris argues, it becomes clear that I have no access to the reasons why that thought appears before my mind and not the thought “leave the light off.” I do not bring the choice-thought about. Rather, at T1, I had both ideas before my mind and then at T2, the thought “turn on the light” appears and I don’t know why and so the choice cannot be said to be made freely. In all cases where I appear to choose, I am actually the recipient of the thought not the cause of it. In support of this idea, he writes, “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.” (p. 24) We don’t have this so we’re not free.
But there is a problem with this claim. Harris seems to be toying with the counterfactual: he gives us what appears to be the necessary conditions for freedom. In other words, Harris seems to be giving us a coherent definition of freedom and then attempts to show such acts never occur. So is the idea incoherent or isn’t it? Well, there is more to the above quote. He finishes it with, "But there is a paradox there that vitiates the very notion of freedom—for what would influence the influences? More influences? None of these adventitious mental states are the real you. You are not controlling the storm, and you are not lost it in. You are the storm." (p. 24)
It is this last qualifier—the paradox—that gives credence to his claim at the beginning of the book that the very notion of free will is incoherent. Any story we could give about free will must be an event-causal story and any event-causal story makes the notion of a prime mover impossible. Despite his presentation of what a counterfactual situation involving free acts would look like, it seems clear he thinks such a counterfactual situation is not coherent.
Two Background Arguments
Kim: The Exclusion Argument
Before turning to a possible rejoinder, I want to consider two arguments that may shed some light on the problem Harris is dealing with. The first is by noted philosopher of mind Jaegwon Kim and it has been called “the exclusion argument.” Put simply, Kim argues that the only way to avoid the idea that there are actually non-physical mental states for which there are no corresponding brain states (any variety of dualism) is to argue there are no mental states. That is, there are only brain states and what we call mental states should be eliminated. The difficulty with this view is that many mental events don’t appear to us to be like brain states. If I choose to turn on the light, that decision appears to me to be an “idea” that I can articulate in a language or “image” in my brain and “consider” or “deliberate over” and then act upon. None of these descriptive terms seem applicable to what is going on in the brain at the time I’m doing those things. Mental events appear to me to be distinct from brain events and I access these events through my subjective, conscious life.
People that find arguments like Harris’ compelling typically find this idea not only not difficult, but generally unintelligible. “It’s just an appearance,” they say, “and at root its all just brain events.” In fact, this is what Harris says. He writes, “The endurance of this notion [of free will] is attributable to the fact that most of us feel that we freely author our own thoughts and actions . . . . Thus the idea of free will emerges from felt experience.” And in the conclusion, “It is not that free will is simply an illusion—our experience is not merely delivering a distorted view of reality. Rather, we are mistaken about our experience. . . . The illusion of free will itself is an illusion.” At this point, there appears to be an impasse: if if one doesn’t believe subjective experiences like performing free acts that have different properties from brain states, what evidence would convince otherwise? (I’m reminded of an exchange between John Searle and Daniel Dennett. Searle was attempting to argue that there is something like first-person conscious feelings and experiences. Dennett opts for an “objective” approach to studying the mind, something that Searle thinks amounts to rejecting first-person experience. Exasperated, Searle writes, “Now what am I to do, as a reviewer, in the face of what appears to be an obvious and self-refuting falsehood? Should I pinch the author to remind him that he is conscious?” (New York Review of Books, Volume 42, Number 20 · December 21, 1995)).
So either all mental states have a corresponding brain state and in that case, mental states appear to be superfluous and causally inefficacious. For the brain state could initiate subsequent physical action without the mental states (if both are causes then we have a case of genuine overdetermination says Kim and this is something we want to avoid). Or we deny that all bodily action is caused by a brain state and is actually caused by a mental state (physical or non-physical)—something Harris would flatly reject. A third option is to deny that we have mental states which runs counter to how the world appears to us to be. None of the options appears to be attractive for someone with Harris’ metaphysical commitments but who also wants to maintain some semblance of a common sense view.
Strawson: Freedom and Belief
Even if we accept Kim’s argument and end up concluding there are mental states that cause bodily action apart from brain states, Libertarians aren’t necessarily out of the woods. For even if there are mental states that are distinct from brain states, the idea of a free will may be in no less trouble. A powerful argument by Galen Strawson attempts to show that regardless of one’s metaphysics, free will, even on a Libertarian’s own terms, can never yield ultimate responsibility (interestingly, I don’t think he is arguing that the concept itself is incoherent—he just argues that a truly free act for which a person is ultimately responsible can never come about). Strawson argues that individuals cannot ultimately be morally responsible because one’s moral decisions are an outcome of the way one is at the time the decision is made. Strawson is only concerned with the state of one’s mental life at the moment of a moral decision—it doesn’t matter whether determinism is true or indeterminism is true, whether one is a dualist or a materialist. It doesn’t matter because the argument claims that a moral decision is the outcome of the mental state of the person (the development of much of that mental state being out of the persons direct or indirect control) and this is true regardless of one’s metaphysics. He considers a case where one passes a person on the street holding an Oxfam tin asking for money.
Suppose you face the decision to put money in the Oxfam tin. The following may be true:
- You want to help the poor person
- You believe you can afford it
- You want to feel better about yourself
But suppose the follow also are true
- You want cake really badly
- You're tired of people begging for money and don't want to support it
- You promised your spouse you'd bring home the cake
Let’s suppose you choose not to put money in the tin. One could argue that the above six antecedent beliefs and desires are in moral parity for you: they neither determine that you’ll act (put money in the tin) or refrain from acting (withhold money from the tin). Now let’s also suppose that you possess something like an active power over which you have control and that you’re rational—conditions generally considered as necessary for a free act on Libertarianism. Let’s also assume that it makes sense for to say you can choose to act or refrain from acting in this case. This means that your action is not caused by other events but that you choose to act or refrain from acting in order to meet some not-yet-realized goal. Since the antecedent conditions do not determine your actions, why did you make the choice you did?
If we ask why you refrained, there either is an answer to that question or there isn’t (this isn’t merely an epistemological claim—I’m saying there is or is not a metaphysical antecedent that explains the decision). If there isn’t, then the choice appears to be random and violates the third of the three necessary conditions for a free act. If there is a reason, let’s say you acted because you put more value in the second set of beliefs and desires more than the first, then that value judgment would be included in the first of Strawson’s premises in what he calls the Basic Argument: You do what you do—in the circumstances in which you find yourself—because of the way you then are. In this situation, we could capture this by asking, “Why did you value the second set over the first such that it influence your decision?” If your decision was truly yours, it came out of the state you were in at the time of the decision the development of which may be partly or entirely out of your control. How, then, could you be responsible for the choice to value the second rather than the first set and thus for the choice to refrain from putting money in the tin? Free, responsible action seems ultimately illusory. There is a nice summary of this argument here. Strawson lays out his full case in his recent book, Freedom and Belief.
Turtles All the Way Down
With this background, we can get a bit clearer on the root of Harris’ claims. Putting Harris in conversation with Kim, I think he would claim that all bodily action is an effect of a some process in the brain. That is, human action is a pure case of physical-to-physical causation. Mental states like free choices do appear to be something like epiphenomena for Harris. He writes, “Decisions, intentions, efforts, goals, willpower, etc., are causal states of the brain, leading to specific behaviors, and behaviors lead to outcomes in the world. Human choice, therefore, is as important as fanciers of free will believe… But the next choice you make will come out of the darkness of prior causes that you, the conscious witness of your experience, did not bring into being." (35) Harris doesn’t appear to want to make a distinction between what we might label a mental event and a brain event. But this runs counter to Leibniz’s well established law of identity. So in order to make sense of Harris’ view, I think the most charitable reading is that a mental event is an appearance, an illusion produced by the brain and correlated with a brain state but themselves are causally ineffective. The brain does all the work in moving the body.
On the causal side of the story, Harris appears to agree with Strawson. Harris argues that the explanation for an action can be given entirely in terms of one’s bodily state and environment just prior to the action occurring and that state is explanatorily sufficient for determining the action. If we could have full access to that prior state and enough data to enable us to create causal laws about the outcome of states like that, we could, as third party observers, predict your action with close-to-perfect accuracy (allowing for standard deviations in statistical models and the like).
On this view, we have to give up the idea that mental events have any causal power over action. But as Strawson shows, even if that’s not the case, free will is no better off. Harris’ presentation is weaker in this regard because he shunts the appearance of a free act to mystery. He aligns to Strawson’s argument in that actions are the result of prior causes and at least claims that one’s metaphysical commitments are secondary (he says, “It is important to recognize that the case I am building against free will does not depend on philosophical materialism . . . even if the mind were made of soul-stuff, nothing about my argument would change. The unconscious operations of a soul would grant you no more freedom than the unconscious physiology of your brain does.”). But he assigns any apparent free act to a product of mystery: the brain (or soul) on any conception of free will would be an unconscious change about which we are not aware and so could not be free.
I think the cases he gives to support this are of a weaker variety. He could have used stronger cases. But I believe he uses a particular class of examples to support his claim that choice-ideas are outside of the agent’s direct control. For example, he talks about needing to buy a present for his daughter and not knowing what to get her and a case where your life goes off track and you need to figure out how to get things right. These are cases where the options one needs to consider may be unknown and ideas, “simply arise unauthored.” (p. 35) But the cases most interesting for the Libertarian are those where the options one is considering are clear (choosing between chocolate and vanilla ice cream) or where the decision is to act or refrain from acting (turn the light on or leave it off). I suspect that Harris would say using these examples make no difference to his case. And it might not particularly if he does align with Strawson’s causal story.
But it does shed some doubt on his claim that the decisions that are paradigmatic of a free act actually do appear out of darkness and are entirely mysterious to us. I may not have any awareness of where the idea to get a tropical fish for my daughter comes from but I do seem to have some awareness of the factors involved in deciding to turn the light on or leave it off. Ultimately the cause of that thought may be a product of my prior beliefs and desires, but it seems there is at least a possibility that I could analyze this and however crudely, trace the idea’s pedigree—something that doesn’t seem possible on Harris’ model.
So Harris differs from Strawson in that Harris’ argument grounds the inability of a person to act freely in an apparent inability to know all the factors that ground intention. His apparent counterfactual to the efficient causal story entails that we “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.” Because we are not so aware, we can’t be free.
And it is just this appeal to mystery that leaves Harris’ argument open to attack. If the decision to turn the light on “comes out of the darkness” and is ultimately unanalyzable, couldn’t this just be what the exercise of a free choice by something like a first-casual agent is like? I may not know everything involved in how or why I make a choice. But the experience of making that choice seems to me to be under my control. The fact that I don’t know all those factors doesn’t in any way diminish the experience. In fact, this is precisely what Libertarians claim. From the perspective of third-person analyses (i.e. a scientific analysis), a free choice would be mysterious. A free choice is not reducible to an efficient-causal story and the genesis of the thought to turn on the light appears to come out of the darkness just because the agent herself created the thought—she made the decision. If it does “pop into the mind” it doesn’t do so without the agent being involved. It “pops” because the agent exercises a causal power to bring it about. And there is an enormous amount of experiential data (the feelings Harris references) to support this idea. That feeling should contribute to the overall empirical data we should consider, not be dismissed by hand waving.
What It Means To Be An Agent
In fact, this is essentially what Libertarians say. For most of his career*, philosopher Roderick Chisholm was a committed Libertarian and gave some of the more clearer descriptions of a free act and libertarian agency. In van Inwagen’s and Zimmerman’s Metaphysics: The Big Questions in an article titled “Human Freedom and the Self,” Chisholm describes the Libertarian scenario as one where an event is not caused by another event but by an agent. This, he says, is not incoherent but simply ascribes the “cause” of that event to something that has the power to bring events about. “We must not say that every event involved in the act is caused by some other event; and we must not say that the act is something that is not caused at all. The possibility that remains, therefore, is this: We should say that at least one of the events that are involved in the act is caused, not by any other event, but by something else instead. And this something else can be the agent—the man.” (p.359)
Why would we come up with such a thing if our experience of observing causes always seems to involve event causation? In fact, says Chisholm, our concept of event causation (which Hume showed cannot be derived from observing the world alone), probably comes from our experience of being causal agents. He writes, “How, then, do we derive [our idea of event causation]? The most plausible suggestion, it seems to me, is that of Reid, once again: namely that ‘the conception of an efficient cause may very probably be derived from the experience we have had . . . of our own power to produce certain effects.’ if we did not understand the concept of immanent [agent] causation, we would not understand that of transeunt [event] causation.” (p. 362)
The point here is that Harris is claiming that the thought to act on a rather than b appears in the mind without the agent having any kind of access to where the thought came from and this means that the agent is not acting freely. But if you ask the agent where the thought came from, the answer will be something like, “I just decided.” The phenomenology includes the mysterious but very present sense that somehow the person themself was the source of the decision in spite of (perhaps because of) the inability of that agent to give an exhaustive event-causal story. It is the very fact that such a story cannot be given that grounds the idea that the agent herself brought it about. It seems to me that only if one assumes that the story is event-causal must one demand that the agent must have access to that story in order to be free. If one abandons that assumption, then decisions (thoughts that involve choosing) along with the phenomenology that one feels as if one chose is adequate to ground at least a prima facie belief that one is responsible.
Of course the counterfactual to this is the scenario that Harris says is not possible—at least not now: an exhaustive material account that is very highly predictive of action. If such an account could be given, then it could be reasonably claimed that the phenomenology is a phantom and there is no mystery where thoughts involving choices come from. But such a story does not exist and such a story does not seem forthcoming.
Harris’ short book is intriguing and accessible. It’s a good primer on key ideas in physicalist theories of freedom and the will, written for a broad audience. Many of the ideas are challenging and he does a good job of articulating what is as stake. I think the way he frames his core argument is a bit too breezy, though he could tighten it up fairly easily by adopting some of Strawson’s framework. Given Harris’ wide following, the book will be important and hopefully start more conversations on this important subject.
* At the end of his career, a single article appeared in a collection on free will where Chisholm appears to have backed away from his commitment an agent-caused act as being sui generis. He subclasses agent caused acts under event causality since the agent’s actions is an event of some sort. He does not appear to abandon agent causality but he does appear to modify his view that agency is something unique in the world. The rather obscure but important article appears in Timothy O’Connor’s excellent collection Agents, Causes, and Events: Essays on Indeterminism and Free Will
Kim’s exclusion argument can be found here under the section titled, “THE SUPERVENIENCE/EXCLUSION ARGUMENT”