The Marionette’s Lament

A Response to Daniel Dennett
February 12, 2014

Dear Dan—

I’d like to thank you for taking the time to review Free Will at such length. Publicly engaging me on this topic is certainly preferable to grumbling in private. Your writing is admirably clear, as always, which worries me in this case, because we appear to disagree about a great many things, including the very nature of our disagreement.

I want to begin by reminding our readers—and myself—that exchanges like this aren’t necessarily pointless. Perhaps you need no encouragement on that front, but I’m afraid I do. In recent years, I have spent so much time debating scientists, philosophers, and other scholars that I’ve begun to doubt whether any smart person retains the ability to change his mind. This is one of the great scandals of intellectual life: The virtues of rational discourse are everywhere espoused, and yet witnessing someone relinquish a cherished opinion in real time is about as common as seeing a supernova explode overhead. The perpetual stalemate one encounters in public debates is annoying because it is so clearly the product of motivated reasoning, self-deception, and other failures of rationality—and yet we’ve grown to expect it on every topic, no matter how intelligent and well-intentioned the participants. I hope you and I don’t give our readers further cause for cynicism on this front.

Unfortunately, your review of my book doesn’t offer many reasons for optimism. It is a strange document—avuncular in places, but more generally sneering. I think it fair to say that one could watch an entire season of Downton Abbey on Ritalin and not detect a finer note of condescension than you manage for twenty pages running.

I am not being disingenuous when I say this museum of mistakes is valuable; I am grateful to Harris for saying, so boldly and clearly, what less outgoing scientists are thinking but keeping to themselves. I have always suspected that many who hold this hard determinist view are making these mistakes, but we mustn’t put words in people’s mouths, and now Harris has done us a great service by articulating the points explicitly, and the chorus of approval he has received from scientists goes a long way to confirming that they have been making these mistakes all along. Wolfgang Pauli’s famous dismissal of another physicist’s work as “not even wrong” reminds us of the value of crystallizing an ambient cloud of hunches into something that can be shown to be wrong. Correcting widespread misunderstanding is usually the work of many hands, and Harris has made a significant contribution.

I hope you will recognize that your beloved Rapoport’s rules have failed you here. If you have decided, according to the rule, to first mention something positive about the target of your criticism, it will not do to say that you admire him for the enormity of his errors and the recklessness with which he clings to them despite the sterling example you’ve set in your own work. Yes, you may assert, “I am not being disingenuous when I say this museum of mistakes is valuable,” but you are, in truth, being disingenuous. If that isn’t clear, permit me to spell it out just this once: You are asking the word “valuable” to pass as a token of praise, however faint. But according to you, my book is “valuable” for reasons that I should find embarrassing. If I valued it as you do, I should rue the day I wrote it (as you would, had you brought such “value” into the world). And it would be disingenuous of me not to notice how your prickliness and preening appears: You write as one protecting his academic turf. Behind and between almost every word of your essay—like some toxic background radiation—one detects an explosion of professorial vanity.

And yet many readers, along with several of our friends and colleagues, have praised us for airing our differences in so civil a fashion—the implication being that religious demagogues would have declared mutual fatwas and shed each other’s blood. Well, that is a pretty low bar, and I don’t think we should be congratulated for having cleared it. The truth is that you and I could have done a much better job—and produced something well worth reading—had we explored the topic of free will in a proper conversation. Whether we called it a “conversation” or a “debate” would have been immaterial. And, as you know, I urged you to engage me that way on multiple occasions and up to the eleventh hour. But you insisted upon writing your review. Perhaps you thought that I was hoping to spare myself a proper defenestration. Not so. I was hoping to spare our readers a feeling of boredom that surpasseth all understanding.

As I expected, our exchange will now be far less interesting or useful than a conversation/debate would have been. Trading 10,000-word essays is simply not the best way to get to the bottom of things. If I attempt to correct every faulty inference and misrepresentation in your review, the result will be deadly to read. Nor will you be able to correct my missteps, as you could have if we were exchanging 500-word volleys. I could heap misconception upon irrelevancy for pages—as you have done—and there would be no way to stop me. In the end, our readers will be left to reconcile a book-length catalogue of discrepancies.

Let me give you an example, just to illustrate how tedious it is to untie these knots. You quote me as saying:

If determinism is true, the future is set—and this includes all our future states of mind and our subsequent behavior. And to the extent that the law of cause and effect is subject to indeterminism—quantum or otherwise—we can take no credit for what happens.  There is no combination of these truths that seems compatible with the popular notion of free will.

You then announce that “the sentence about indeterminism is false”—a point you seek to prove by recourse to an old thought experiment involving a “space pirate” and a machine that amplifies quantum indeterminacy. After which, you lovingly inscribe the following epitaph onto my gravestone:

These are not new ideas. For instance I have defended them explicitly in 1978, 1984, and 2003. I wish Harris had noticed that he contradicts them here, and I’m curious to learn how he proposes to counter my arguments.

You see, dear reader, Harris hasn’t done his homework. What a pity…. But you have simply misread me, Dan—and that entire page in your review was a useless digression. I am not saying that the mere addition of indeterminism to the clockwork makes responsibility impossible. I am saying, as you have always conceded, that seeking to ground free will in indeterminism is hopeless, because truly random processes are precisely those for which we can take no responsibility. Yes, we might still express our beliefs and opinions while being gently buffeted by random events (as you show in your thought experiment), but if our beliefs and opinions were themselves randomly generated, this would offer no basis for human responsibility (much less free will). Bored yet?

You do this again and again in your review. And when you are not misreading me, you construct bad analogies—to sunsets, color vision, automobiles—none of which accomplish their intended purpose. Some are simply faulty (that is, they don’t run through); others make my point for me, demonstrating that you have missed my point (or, somehow, your own). Consider what you say about sunsets to show that free will should not be considered an illusion:

After all, most people used to believe the sun went around the earth. They were wrong, and it took some heavy lifting to convince them of this. Maybe this factoid is a reflection on how much work science and philosophy still have to do to give everyday laypeople a sound concept of free will…. When we found out that the sun does not revolve around the earth, we didn’t then insist that there is no such thing as the sun (because what the folk mean by “sun” is “that bright thing that goes around the earth”). Now that we understand what sunsets are, we don’t call them illusions. They are real phenomena that can mislead the naive.

Of course, the sun isn’t an illusion, but geocentrism is. Our native sense that the sun revolves around a stationary Earth is simply mistaken. And any “project of sympathetic reconstruction” (your compatibilism) with regard to this illusion would be just a failure to speak plainly about the facts. I have never disputed that mental phenomena such as thoughts, efforts, volition, reasoning, and so forth exist. These are the many “suns” of the mind that any scientific theory must conserve (modulo some clarifying surprises, as has happened for the concept of “memory”). But free will is like the geocentric illusion: It is the very thing that gets obliterated once we begin speaking in detail about the origins of our thoughts and actions. You’re not just begging the question here; you’re begging it with a sloppy analogy. The same holds for your reference to color vision (which I discussed in a previous essay).

And when you are not causing problems with your own analogies, you are distorting mine. For instance, you write that you were especially dismayed by the cover of my book, which depicts a puppet theater. This cover image is justified because I argue that each of us is moved by chance and necessity, just as a marionette is set dancing on its strings. But I never suggest that this is the same as being manipulated by a human puppeteer who overrides our actual beliefs and desires and obliges us to behave in ways we do not intend. You seem eager to draw this implication, however, and so you press on with an irrelevant discussion of game theory (another area in which you allege I haven’t done my homework). Again, I am left wishing we had had a conversation that would have prevented so many pedantic digressions.

In any case, I cannot bear to write a long essay that consists in my repeatedly taking your foot out of my mouth. Instead, I will do my best to drive to the core of our disagreement.

Let’s begin by noticing a few things we actually agree about: We agree that human thought and behavior are determined by prior states of the universe and its laws—and that any contributions of indeterminism are completely irrelevant to the question of free will. We also agree that our thoughts and actions in the present influence how we think and act in the future. We both acknowledge that people can change, acquire skills, and become better equipped to get what they want out of life. We know that there is a difference between a morally healthy person and a psychopath, as well as between one who is motivated and disciplined, and thus able to accomplish his aims, and one who suffers a terminal case of apathy or weakness of will. We both understand that planning and reasoning guide human behavior in innumerable ways and that an ability to follow plans and to be responsive to reasons is part of what makes us human. We agree about so many things, in fact, that at one point you brand me “a compatibilist in everything but name.” Of course, you can’t really mean this, because you go on to write as though I were oblivious to most of what human beings manage to accomplish. At some points you say that I’ve thrown the baby out with the bath; at others you merely complain that I won’t call this baby by the right name (“free will”). Which is it?

However, it seems to me that we do diverge at two points:

1. You think that compatibilists like yourself have purified the concept of free will by “deliberately using cleaned-up, demystified substitutes for the folk concepts.” I believe that you have changed the subject and are now ignoring the very phenomenon we should be talking about—the common, felt sense that I/he/she/you could have done otherwise (generally known as “libertarian” or “contra-causal” free will), with all its moral implications. The legitimacy of your attempting to make free will “presentable” by performing conceptual surgery on it is our main point of contention. Whether or not I can convince you of the speciousness of the compatibilist project, I hope we can agree in the abstract that there is a difference between thinking more clearly about a phenomenon and (wittingly or unwittingly) thinking about something else. I intend to show that you are doing the latter.

2. You believe that determinism at the microscopic level (as in the case of Austin’s missing his putt) is irrelevant to the question of human freedom and responsibility. I agree that it is irrelevant for many things we care about (it doesn’t obviate the distinction between voluntary and involuntary behavior, for instance), but it isn’t irrelevant in the way you suggest. And accepting incompatibilism has important intellectual and moral consequences that you ignore—the most important being, in my view, that it renders hatred patently irrational (while leaving love unscathed). If one is concerned about the consequences of maintaining a philosophical position, as I know you are, helping to close the door on human hatred seems far more beneficial than merely tinkering with a popular illusion.

Changing the Subject

We both know that the libertarian notion of free will makes no scientific or logical sense; you just doubt whether it is widespread among the folk—or you hope that it isn’t, or don’t much care whether it is (in truth, you are not very clear on this point). In defense of your insouciance, you cite a paper by Nahmias et al.[“1”]

E. Nahmias, S. Morris, T. Nadelhoffer & J. Turner. 2005. “Surveying Freedom: Folk Intuitions about Free Will and Moral Responsibility,” Philosophical Psychology, 18, pp. 561–584.
  It probably won’t surprise you that I wasn’t as impressed by this research as you were. Nahmias and his coauthors repeatedly worry that their experimental subjects didn’t really understand the implications of determinism—and on my reading, they had good reason to be concerned. In fact, this is one of those rare papers in which the perfunctory doubts the authors raise, simply to show that they have thought of everything, turn out to be far more compelling than their own interpretations of their data. More than anything, this research suggests that people find the idea of libertarian free will so intuitively compelling that it is very difficult to get them to think clearly about determinism. Of course, I agree that what people think and feel is an empirical question. But it seems to me that we know much more about the popular view of free will than you and Nahmias let on.

It is worth noting that the most common objection I’ve heard to my position on free will is some version of the following:

If there is no free will, why write books or try to convince anyone of anything? People will believe whatever they believe. They have no choice! Your position on free will is, therefore, self-refuting. The fact that you are trying to convince people of the truth of your argument proves that you think they have the very freedom that you deny them.

Granted, some confusion between determinism and fatalism (which you and I have both warned against) is probably operating here, but comments of this kind also suggest that people think they have control over what they believe, as if the experience of being convinced by evidence and argument were voluntary. Perhaps such people also believe that they have decided to obey the law of gravity rather than fly around at their pleasure—but I doubt it. An illusion about mental freedom seems to be very widespread. My argument is that such freedom is incompatible with any form of causation (deterministic or otherwise)—which, as you know, is not a novel view. But I also argue that it is incompatible with the actual character of our subjective experience. That is why I say that the illusion of free will is itself an illusion—which is another way of saying that if one really pays attention (and this is difficult), the illusion of free will disappears.

The popular, folk psychological sense of free will is a moment-to-moment experience, not a theory about the mind. It is primarily a first-person fact, not a third-person account of how human beings function. This distinction between first-person and third-person views was what I was trying to get at in the passage that seems to have mystified you (“I have thought long and hard about this passage, and I am still not sure I understand it…”) Everyone has a third-person picture of the human mind—some of us speak of neural circuits, some speak of souls—but the philosophical problem of free will arises from the fact that most people feel that they author their own thoughts and actions. It is very difficult to say what this feeling consists of or to untangle it from working memory, volition, motor planning, and the rest of what our minds are up to—but there can be little doubt that most people feel that they are the conscious source of their own thoughts and actions. Of course, you may wish to deny this very assertion, or believe it more parsimonious to say that we just don’t know how most people feel—and that might be a point worth discussing. But rather than deny the claim, you simply lose sight of it—shifting from first-person experiences to third-person accounts of phenomena that lie outside consciousness.

It is true, of course, that most educated people believe the whole brain is involved in making them who they are (indeed, you and I both believe this). But they experience only some of what goes on inside their brains. Contrary to what you suggest, I was not advancing a Cartesian theory of consciousness (a third-person view), or any “daft doctrine of [my] own devising.” I was simply drawing a line between what people experience and what they don’t (first-person). The moment you show that a person’s thoughts and actions were determined by events that he did not and could not see, feel, or anticipate, his third-person account of himself may remain unchanged (“Of course, I know that much of what goes on in my brain is unconscious, determined by genes, and so forth. So what?”), but his first-person sense of autonomy comes under immediate pressure—provided he is paying attention. As a matter of experience (first-person), there is a difference between being conscious of something and not being conscious of it. And if what a person is unconscious of are the antecedent causes of everything he thinks and does, this fact makes a mockery of the subjective freedom he feels he has. It is not enough at that point for him to simply declare theoretically (third-person) that these antecedent causes are “also me.”

Average Joe feels that he has free will (first-person) and doesn’t like to be told that it is an illusion. I say it is: Consider all the roots of your behavior that you cannot see or feel (first-person), cannot control (first-person), and did not summon into existence (first-person). You say: Nonsense! Average Joe contains all these causes. He is his genes and neurons too (third-person). This is where you put the rabbit in the hat.

Imagine that we live in a world where more or less everyone believes in the lost kingdom of Atlantis. You and your fellow compatibilists come along and offer comfort: Atlantis is real, you say. It is, in fact, the island of Sicily. You then go on to argue that Sicily answers to most of the claims people through the ages have made about Atlantis. Of course, not every popular notion survives this translation, because some beliefs about Atlantis are quite crazy, but those that really matter—or should matter, on your account—are easily mapped onto what is, in fact, the largest island in the Mediterranean. Your work is done, and now you insist that we spend the rest of our time and energy investigating the wonders of Sicily. 

The truth, however, is that much of what causes people to be so enamored of Atlantis—in particular, the idea that an advanced civilization disappeared underwater—can’t be squared with our understanding of Sicily or any other spot on earth. So people are confused, and I believe that their confusion has very real consequences. But you rarely acknowledge the ways in which Sicily isn’t like Atlantis, and you don’t appear interested when those differences become morally salient. This is what strikes me as wrongheaded about your approach to free will.

For instance, ordinary people want to feel philosophically justified in hating evildoers and viewing them as the ultimate cause of their evil. This moral attitude is always vulnerable to our getting more information about causation—and in situations where the underlying causes of a person’s behavior become too clear, our feelings about their responsibility begin to shift. This is why I wrote that fully understanding the brain of a normal person would be analogous to finding an exculpatory tumor in it. I am not claiming that there is no difference between a normal person and one with impaired self-control. The former will be responsive to certain incentives and punishments, and the latter won’t. (And that is all the justification we need to resort to carrots and sticks or to lock incorrigibly dangerous people away forever.) But something in our moral attitude does change when we catch sight of these antecedent causes—and it should change. We should admit that a person is unlucky to be given the genes and life experience that doom him to psychopathy. Again, that doesn’t mean we can’t lock him up. But hating him is not rational, given a complete understanding of how he came to be who he is. Natural, yes; rational, no. Feeling compassion for him could be rational, however—in the same way that we could feel compassion for the six-year-old boy who was destined to become Jeffrey Dahmer. And while you scoff at “medicalizing” human evil, a complete understanding of the brain would do just that. Punishment is an extraordinarily blunt instrument. We need it because we understand so little about the brain, and our ability to influence it is limited. However, imagine that two hundred years in the future we really know what makes a person tick; Procrustean punishments won’t make practical sense, and they won’t make moral sense either. But you seem committed to the idea that certain people might actually deserve to be punished—if not precisely for the reasons that common folk imagine, nevertheless for reasons that have little or nothing to do with the good consequences that such punishments might have, all things considered. In other words, your compatibilism seems an attempt to justify the conventional notion of blame, which my view denies. This is a difference worth focusing on.

Let’s examine Austin’s example of his missed putt:

Consider the case where I miss a very short putt and kick myself because I could have holed it. It is not that I should have holed it if I had tried: I did try, and missed. It is not that I should have holed it if conditions had been different: that might of course be so, but I am talking about conditions as they precisely were, and asserting that I could have holed it. There is the rub. Nor does ‘I can hole it this time’ mean that I shall hole it this time if I try or if anything else; for I may try and miss, and yet not be convinced that I could not have done it; indeed, further experiments may confirm my belief that I could have done it that time, although I did not. (J.L. Austin. 1961. “Ifs and Cans,” in Austin, Philosophical Papers, edited by J. Urmson and G. Warnock. Oxford, Clarendon Press.)

This is a good place to start, because you say the following in your review:

I consider Austin’s mistake to be the central core of the ongoing confusion about free will; if you look at the large and intricate philosophical literature about incompatibilism, you will see that just about everyone assumes, without argument, that it is not a mistake.

I am happy to take the bait. I see no problem with using Austin’s example to support incompatibilism. I should emphasize, however, that I am discussing only the implications of Austin’s point for an account of free will, not how it functions in his original essay (which, as you know, was an analysis of the relationship between the terms “if” and “can,” not a sustained argument against free will).[2]

Let’s make sure you and I are standing on the same green: We agree that a human being, whatever his talents, training, and aspirations, will think, intend, and behave exactly as he does given the totality of conditions in the moment. That is, whatever his ability as a golfer, Austin would miss that same putt a trillion times in a row—provided that every atom and charge in the universe was exactly as it had been the first time he missed it. You think this fact (we can call it determinism, as you do, but it includes the contributions of indeterminism as well, provided they remain the same[3]) says nothing about free will. You think the fact that Austin could make nearly identical putts in other moments—with his brain in a slightly different state, the green a touch slower, and so forth—is all that matters. I agree that it is what matters when assessing his abilities as a golfer: Here, we don’t care about the single NMDA receptor that screwed up his swing on one particular putt; we care about the statistics of his play, round after round. But to speak clearly and honestly (that is, scientifically) about the actual causes of what happens in the world in each moment, we must focus on the particular.

What are we really saying when we claim that Austin could have made that putt (the one he missed)? As you point out, we aren’t actually referencing that putt at all. We are saying that Austin has made similar putts in the past and we can count on him to do so in the future—provided that he tries, doesn’t suffer some neurological injury, and so forth. However, we are also saying that Austin would have made this putt had something not gotten in his way. He had the general ability, after all, so something went wrong.

Then why did Austin miss his putt? Because some condition necessary for his making it was absent. What if that condition was sufficient effort, of the sort that he was generally capable of making? Why didn’t he make that effort? The answer is the same: Because some condition necessary for his making it was absent. From a scientific perspective, his failure to try is just another missed putt. Austin tried precisely as hard as he did. Next time he might try harder. But this time—with the universe and his brain exactly as they were—he couldn’t have tried in any other way.

To say that Austin really should have made that putt or tried harder is just a way of admonishing him to put forth greater effort in the future. We are not offering an account of what actually happened (his failure to sink his putt or his failure to try). You and I agree that such admonishments have effects and that these effects are perfectly in harmony with the truth of determinism. There is, in fact, nothing about incompatibilism that prevents us from urging one another (and ourselves) to do things differently in the future, or from recognizing that such exhortations often work. The things we say to one another (and to ourselves) are simply part of the chain of causes that determine how we think and behave.

But can we blame Austin for missing his putt? No. Can we blame him for not trying hard enough? Again, the answer is no—unless blaming him were just a way of admonishing him to try harder in the future. For us to consider him truly responsible for missing the putt or for failing to try, we would need to know that he could have acted other than he did. Yes, there are two readings of “could”—and you find only one of them interesting. But they are both interesting, and the one you ignore is morally consequential. One reading refers to a person’s (or a car’s, in your example) general capacities. Could Austin have sunk his putt, as a general matter, in similar situations? Yes. Could my car go 80 miles per hour, though I happen to be driving at 40? Yes. The other reading is one you consider to be a red herring. Could Austin have sunk that very putt, the one he missed? No. Could he have tried harder? No. His failure on both counts was determined by the state of the universe (especially his nervous system). Of course, it isn’t irrational to treat him as someone who has the general ability to make putts of that sort, and to urge him to try harder in the future—and it would be irrational to admonish a person who lacked such ability. You are right to believe that this distinction has important moral implications: Do we demand that mosquitoes and sharks behave better than they do? No. We simply take steps to protect ourselves from them. The same futility prevails with certain people—psychopaths and others whom we might deem morally insane. It makes sense to treat people who have the general capacity to behave well but occasionally lapse differently from those who have no such capacity and on whom any admonishment would be wasted. You are right to think that these distinctions do not depend on “absolute free will.” But this doesn’t mean nothing changes once we realize that a person could never have made the putt he just missed, or tried harder than he did, or refrained from killing his neighbor with a hammer.

Holding people responsible for their past actions makes no sense apart from the effects that doing so will have on them and the rest of society in the future (e.g. deterrence, rehabilitation, keeping dangerous people off our streets). The notion of moral responsibility, therefore, is forward-looking. But it is also paradoxical. People who have the most ability (self-control, opportunity, knowledge, etc.) would seem to be the most blameworthy when they fail or misbehave. For instance, when Tiger Woods misses a three-foot putt, there is a much greater temptation to say that he really should have made it than there is in the case of an average golfer. But Woods’s failure is actually more anomalous. Something must have gone wrong if a person of his ability missed so easy a putt. And he wouldn’t stand to benefit (much) from being admonished to try harder in the future. So in some ways, holding a person responsible for his failures seems to make even less sense the more worthy of responsibility he becomes in the conventional sense.

We agree that given past states of the universe and its laws, we can only do what we in fact do, and not do otherwise. You don’t think this truth has many psychological or moral consequences. In fact, you describe the lawful propagation of certain types of events as a form of “freedom.” But consider the emergence of this freedom in any specific human being: It is fully determined by genes and the environment (add as much randomness as you like). Imagine the first moment it comes online—in, say, the brain of a four-year-old child. Consider this first, “free” increment of executive control to emerge from the clockwork. It will emerge precisely to the degree that it does, and when, according to causes that have nothing to do with this person’s freedom. And it will perpetuate its effects on future states of his nervous system in total conformity to natural laws. In each instant, Austin will make his putt or miss it; and he will try his best or not. Yes, he is “free” to do whatever it is he does based on past states of the universe. But the same could be said of a chimp or a computer—or, indeed, a line of dominoes. Perhaps such mechanical equivalences don’t bother you, but they might come as a shock to those who think that you have rescued their felt sense of autonomy from the gears of determinism.

In your review, you called my book a “political tract.” The irony is that your argument against incompatibilism seems almost entirely political. At times you write as though nothing is at stake apart from the future of the terms free will and compatibilism. More generally, however, you seem to think that the consequences of taking incompatibilism seriously will be pernicious:

If nobody is responsible, not really, then not only should the prisons be emptied, but no contract is valid, mortgages should be abolished, and we can never hold anybody to account for anything they do.  Preserving “law and order” without a concept of real responsibility is a daunting task.

These concerns, while not irrational, have nothing to do with the philosophical or scientific merits of the case. They also arise out of a failure to understand the practical consequences of my view. I am no more inclined to release dangerous criminals back onto our streets than you are.

In my book, I argue that an honest look at the causal underpinnings of human behavior, as well as at one’s own moment-to-moment experience, reveals free will to be an illusion. (I would say the same about the conventional sense of “self,” but that requires more discussion, and it is the topic of my next book.) I also claim that this fact has consequences—good ones, for the most part—and that is another reason it is worth exploring. But I have not argued for my position primarily out of concern for the consequences of accepting it. And I believe you have.

Of course, I can’t quite blame you for missing that putt, Dan. But I can admonish you to be more careful in the future.