Bruce N. Waller, Against Moral Responsibility, 2011, MIT Press/Bradford Book
An original mistake, well made, can be a fine contribution to a research topic. Recalling Wolfgang Pauli’s famous putdown of a fellow physicist’s work as “not even wrong,” we can appreciate the fact that a crisp, clear argument can illuminate the field by uncovering a tempting but heretofore unexamined falsehood and forthrightly asserting it. I cannot think of a better example than Waller’s book, from which I have learned more than from the last dozen books and articles on free will that I have read, a bounty of valuable insights all marshaled on behalf of a thesis that has never before been properly defended, and is in the end, in my opinion, indefensible—but for reasons that are instructive. Waller has opened my eyes about my own project and other competing projects in the field.
Waller is engagingly candid about the uphill fight he is waging. Just about everybody takes the whole point of worrying about free will to be its role in securing moral responsibility: free will is worth wanting because we want to be morally responsible agents. Not so, he says. There are indeed varieties of free will worth wanting, but the benefits they secure for us do not reach so far as moral responsibility. In a particularly illuminating touch, he notes that his view should not seem so outlandish when we recognize that it is the heretofore unembraced combination of the two most widely held views: free will is compatible with determinism/ naturalism—but moral responsibility isn’t (p. 45). Hume, Moore, Ayer, Frankfurt, and I are right about the first, and Valla, Kant, Campbell, van Inwagen, O’Connor, Ginet, Kane and Pereboom are right about the second! Almost everybody wants to secure a justification for retributive blame and punishment, just deserts, and the first gang makes the mistake of thinking that the naturalistic varieties of free will we champion are up to the job, and the second gang sees correctly that this won’t work (Kant’s “wretched subterfuge”) and makes the mistake of opting for a miracle (Strawson’s “desperate and panicky metaphysics”). If we can just get our heads around the idea that moral responsibility is not worth wanting, is in fact an oppressive blight on our conceptual scheme, we can declare victory!
Waller’s exposition and defense of this startling idea is admirable; he is almost always fair and respectful to his targets, noting points of agreement, and articulating their arguments constructively. There are a bounty of useful novelties in his discussion, some of them discussed below. But, I think, he loses sight of a few possibilities that pull the rug out from under his intrepid project. This leads him to misinterpret some of the things I have said—but in ways that I didn’t guard against because I myself did not see what I had to deliver to make my case. Exposing these misinterpretations is my aim here.
Waller opts at the outset for a strong definition of moral responsibility, noting that there are other definitions, but this will be how he will understand what he calls the “core concept” (p. 7). He starts with a passage from Randolph Clarke:
If any agent is truly responsible . . . that fact provides us with a specific type of justification for responding in various ways to that agent, with reactive attitudes of certain sorts, with praise or blame, with finite rewards or punishments. To be a morally responsible human agent is to be truly deserving of these sorts of responses, and deserving in a way that no agent is who is not responsible. This type of desert has a specific scope and force—one that distinguishes the justification for holding someone responsible from, say, the fairness of a grade given for a performance or any justification provided by consequences. (p. 3)
The last sentence of this otherwise quite commodious definition ties moral responsibility firmly to a retributivist conception of desert from the outset: any appeal to the good consequences that can be achieved by adopting a system of moral responsibility disqualifies it, by definition. We may note that this is the way the second gang—the libertarians and other incompatibilists—tends to see moral responsibility, and it is the way unreformed, unsophisticated tradition sees it. Is this getting off on the wrong foot? I thought so, at first, but I came to see that it does permit Waller to hold the line, contrasting this benchmark notion with the attempts by us compatibilists to secure a suitable substitute, and that this perspective yields some valuable insights. And let me say right away that I agree with Waller’s main conclusion in one important sense: ˆ kind of absolutistic moral responsibility—insisting as it does on what I have called guilt-in-the-eyes-of God—is incompatible with naturalism and has got to go. Good riddance. (I don’t just think it is incompatible with naturalism; I think it is incoherent in its own terms, for reasons that I did not fully appreciate before reading Waller’s book.)
But Waller does pay a steep price for this choice: it obliges him to sort all rivals to his view by imposing a false trichotomy: if compatibilists attempt to answer the “external” question of whether there really is moral responsibility (“real” desert-based responsibility), any result is doomed to falsehood; if instead the compatibilists tackle “internal” questions, from within a presupposed system of moral responsibility, they beg the external question or change the subject: they may, at best, offer valuable refinements of the application of a substitute concept, not the concept of moral responsibility, but a consequentialist impostor. Given his starting point, he more or less has to misconstrue a proposal that I have made (along with Stephen White) about what “could have done otherwise” means:
Don’t try to use metaphysics to ground ethics . . . ; put it the other way around: use ethics to ground what we should mean by our ‘metaphysical’ criterion.” (Dennett, 2003, p. 297)
Waller sees this as question-begging. His version of my position is this: “Start from the assumption of moral responsibility—the assumption that people justly deserve reward and punishment—and use that to fix what we mean by “could have done otherwise” and by our concept of free will” (p47). Like Peter van Inwagen, he thinks, I assume the reality of moral responsibility but then, recoiling from the conclusion that van Inwagen draws—a frankly mysterious, non-naturalist indeterminism—I must fiddle with what I declare to be the meaning of “could have done otherwise” until I can get it to fit my naturalistic bent, yielding a dismal result: “an anemic account of moral responsibility joined to a shallow account of free will.” (p. 48) But that is not quite what I say, or in any event what I was trying to say; I meant that you have to start with the recognition that the concept of moral responsibility doesn’t just drop into our conceptual scheme from the sky; it is integral to a large system of social and political institutions, so understanding how it works there, in the (naturalistic) land of ethical theory, economics, political theory, psychology and other social sciences, could prove a better anchoring for the concept than any that could be mustered in physics and the neurosciences (a.k.a. metaphysics). Another way of putting the claim: get clear about the phenomena in the manifest image before going on a wild goose chase in the scientific image. From Waller’s perspective, another passage of mine also gets a harsh reading:
We ought to admit, up front, that one of our strongest unspoken motivations for upholding something close to the traditional concept of free will is our desire to see the world’s villains “get what they deserve.” And surely they do deserve our condemnation, our criticism, and—when we have a sound system of laws in place—punishment. A world without punishment is not a world any of us would want to live in. (Dennett, 2008, p. 258)
According to Waller, Dennett “rests his case for moral responsibility on our deep retributive desires” (p. 9). Not at all. I agree with him that our retributive desires have an ancient and amoral source in our evolutionary past but I don’t “rest my case” on them; I argue (or should have argued, should have made clearer) that we have devised a way to harness them—tame them, direct them down justifiable channels—in order to secure something very valuable: a secure and civil society in which people are held responsible for their promises and the other deeds they do “of their own free will.” And of course I also agree with him and many others that our American system of justice and punishment is obscenely mis-instituted for the most part, but I want to reform it, not abandon it, and that does mean securing a role for just deserts, not the role described in Waller’s chosen concept, but an ultimately consequentialist role. And don’t say that that’s a contradiction in terms, that ‘just deserts’ means—by definition—a retributive theory of punishment. A consequentialist defense of just deserts is the possibility to which Waller is systematically blind, given his starting point. Let me preview that possibility with a simple example:
Suppose you sign a contract which includes a clause about what penalty you will pay if you fail to deliver. And then when you fail, suppose you refuse to pay the penalty. What happens? What should happen? If other attempts at getting you to pay the stipulated penalty—laws suits and the like—fail, you eventually become eligible to be taken, by force if necessary, against your will, and put on trial. You become liable for punishment. Not rehabilitation, not treatment, but honest to goodness infliction-of-suffering-type punishment. Not beating, but incarceration and/or punitive fines extracted by whatever means necessary from your estate. Why? Because in entering into a contract you tacitly agreed to be part of the system that has this further escalation clause behind it. Were you “coerced” into joining this system, making this deal? (Waller, p. 109-110) Not at all. It’s a wonderful bargain. I for one don’t want to live in a world without contracts—not because I’m a pathetic dupe of capitalism (for instance) but because contracts are just the carefully articulated expression of the underlying concept of a promise made in good faith, and I don’t want to live in a world without promising. It is the very glue of civilization. I expect that Waller would agree with me on this point. He, too (I imagine—he never mentions it), wants to maintain promising and contracts and the institutional understandings that make them possible, but if so, I guess he doesn’t view the penalties and justification of taking-by-force-if-necessary of these institutions as punishment (real, retributive punishment), since after all, they have a consequentialist grounding, as just articulated. To make the contrast clear, that concept is my “core concept” of moral responsibility, an ultimately consequentialist, not retributive concept.
Waller gives us example after example in which he contrasts two people, one of whom had a salubrious upbringing while the other was deprived of most benefits. Is it fair, he keeps asking, to hold both of them responsible? Life isn’t fair. Some folks get a pretty raw deal through no fault of their own, and others thrive thanks to a great head start. Let’s face it: physical beauty gives a huge undeserved advantage to those who have it, and it is just one of a variety of unequal endowments that challenge any account of moral responsibility. None of this is fair. The state of nature isn’t fair. That’s why we have the institutions of civil society, to even the playing field as best we can, by minimizing the amplification of advantage and disadvantage that otherwise would probably occur.
Here’s another way of looking at it: think about sports. Do any games have fair rules? Is it even possible for a game to be fair, given the differences in ability among the participants, etc., etc.? And when a player commits a foul, is the penalty deserved? When? When he could have done otherwise? In what sense of that phrase? If a player gets ousted from a game, can it be true that he deserved that fate? Games provide a nice “toy problem,” as they say in AI, a simplified, clarified version of gnarly reality in which to explore the interaction of basic concepts. Games have crisp counterparts, at least, of punishment (penalties, including banishment) and reward, dishonesty (cheating) and just deserts. Might games provide a model for real moral responsibility?
Waller often resorts to sports examples, provoked perhaps by my example of the marathon with the uneven (but fair) start, which he raises (p. 117) but dismisses by just saying that it is absurd to suppose the luck averages out. It is not absurd; more on that below. He doesn’t consider the possibility that we compatibilists are saying that the “moral responsibility system” he deplores is like a game in several important but typically overlooked ways. First, it is, like a game, an artifact, a socially constructed system of rules that we have devised. More particularly, not finding “well defined boundaries” in nature (as Waller observes, p70), our game draws some sharp boundaries in order to create an environment in which desirable activities can take place. (Like goal lines and out-of-bounds lines and time limits, without which we couldn’t play football, soccer, basketball and other games.) Waller should address the question of whether his arguments also establish that fair games are impossible. After all, players of all natural abilities have different strengths and weaknesses. We compatibilist reformers claim that it is a fact, but a political fact, an artifactual fact (like a home run), that, say, Jan is morally responsible and Fran is not. Waller’s view leaves no room for such an important distinction as moral responsibility to be anything less than found in nature. His view is like saying “We should eliminate the rule in baseball that distinguishes unfairly between a 350 foot home run and a 348 foot fly ball. We should look at each struck ball very closely and stop this arbitrary distinction between hits and fly balls, dealing with each feat in all its natural, real properties.”
The home run rule is fair, I submit, and it is fair because it has been negotiated by the likely participants. If you don’t like the home run rule, don’t play baseball; play some other game. It is this dynamic of rule-reconsideration, which continues to this day, that comes close to guaranteeing that the rules are fair. (My oft-used example, one of the few favorite points of mine that Waller doesn’t discuss, is the rule that says you can’t drive till you’re sixteen years old in most jurisdictions. It’s arbitrary, artifactual, and it deliberately ignores minor differences about the abilities and maturity of individual applicants for a driver’s license. It’s fair.)
You don’t have to play the moral responsibility game; you can be a hermit on an otherwise deserted island, fishing and foraging. But if you want to enjoy the benefits of living in a civilized society, you have to play the game. Waller sees this as coercive, and addresses it in yet another sports example: “If I agree to play a game, it does not follow that the game is fair. You agree to a tennis match, but you are required to wear oversize combat boots that weight 10 pounds each, while your opponent plays in perfectly fitted, well-designed tennis shoes. The match will be unfair, your agreement to play notwithstanding.” (p. 237) That misses the beauty of the rule-reconsideration dynamic I just described: he’s right that your agreeing to play the lopsided game of tennis doesn’t mean it is fair, but the fact that almost everybody agrees to play a game means it is probably as close to a fair game as you could devise. If Waller thinks it is impossible, given the way the world is, to devise a game with fair rules that people would reasonably want to play, he should say so, and defend the claim. Now that would be an uphill philosophical battle! If, alternatively, he thinks a fair game is possible, then he can use that game, whatever it is, as his model for an alternative to the moral responsibility game.
It is open to Waller, that is, to propose a better game—and that actually is what he is trying to do—but here the most glaring deficiency in his project reveals itself: he doesn’t give the specifics. He urges us, in effect: “Let’s all play a game without hard boundaries or ironclad rules, in which we devote whatever time and effort is required to improving the playing ability of each player, and never oust anybody or penalize anybody for, oh, defective play!” Well, show us, please, how this game would go.
Waller eventually gets to this task, in a few short pages (290-304). On page 290 we get an interesting reason for eliminating moral responsibility: as pioneering reforms in industry have demonstrated, blame and shame fosters blame-shifting, and hiding one’s mistakes (in air traffic control, in medicine, . . .). Follow the example of industry and air traffic control, he says, and move to a no-blame system, to foster honest divulging of errors, which should then be studied as flaws in the system, not sins of the error-maker. The results are impressive, but they don’t generalize as smoothly as Waller apparently thinks. What is to prevent participants from abusing a no-blame system? He notes (p. 292) that Jeanne Steiner sees this problem and recommends a mixed approach, but Waller’s response to this sane advice is weaker than weak: it “seems more plausible” to extend the no-blame system further, whereas a mixed approach “seems likely to limit the effectiveness of the systems model.” He gives no details in support of these hopeful hunches (p. 292).
Not good enough. He should examine the implications further. He does make a pass at it, in his examination of a doctor in a no-blame system who makes mistakes because of a “flawed character,” but what he portrays as the “worst case scenario –in the licensing of physicians who are so greedy that they rush patient care in order to maximize their own earnings” (293) is far from being the worst case. A much worse case would be doctors who do what they jolly well please, knowing that they will not be punished or held accountable, no matter what they do. The no-blame system only works against a background of blame and the omnipresent threat of punishment, the suspenders that hold up the pants. The air traffic controllers will lose their jobs, their reputations, and maybe their freedom if they screw up badly enough. This all goes without saying in the real world, which is why Waller doesn’t notice it. Under normal conditions the threat of punishment works so well that violating the rules is all but unthinkable to almost everybody involved.
The no-blame system has to have a way of removing flawed characters from roles where they would do a lot of harm. Don’t blame them, Waller would presumably say (he doesn’t even go into this detail); just deny them the right to further employment in this profession, through a competency hearing probably. And what if the doctor refuses to cooperate with the competency hearing? What if the doctor deliberately hides the evidence of his incompetency or manufactures bogus evidence? The blame system is always there, making such noncompliance almost unthinkable.
When Waller does face this challenge of dealing with deliberate bad behavior (p. 293 ff) -- “But can one give a positive outline of a criminal justice system that rejects moral responsibility and just deserts?” – what follows is first four pages of criticism of the oh-so-unfair existing punitive system (everybody agrees about this) where what we are waiting for is the answer to his own question: can he outline a workable system that rejects moral responsibility? The answer comes at last (p. 297), and it is underwhelming to say the least: restorative justice programs (such as those being pioneered in Canada and New Zealand in native communities).
“First, it is important that the person committing a crime. . . . acknowledge the wrongdoing. . .” And if not, what? “Second, the community emphasizes restoring the wrongdoer to the community. . .” And if this doesn’t work, what? Presumably the wrongdoer is in custody during this period. If he tries to escape, he will be restrained, forcibly. That is already a kind of punishment, holding a person against his will. Or if it isn’t, if it is, as Waller says at one point, restraint but not punishment, then he is not really trying to make the case advertised. May we restrain anybody we deem to have a character flaw until such time as we deem the flaw ameliorated? Many of us think that would be worse than a system of punishment. If, to avoid that cliff, Waller agrees that only those appropriately arrested and convicted may be restrained, is he not introducing concepts of guilt and desert through the back door? Not his core concepts, but, say, mine. Might it be that all his arguments are only against the rigid, libertarian guilty-before-the-eyes-of-God concept of moral responsibility, and he can’t see that he himself is appealing to the very substitutes (the wretched subterfuges?) that he disparages among the compatibilists? He has so little to say about the actual mechanics of implementing a restorative justice program that he doesn’t notice that the suspenders of guilt and punishment must still be firmly in place, holding up the pants. He notes that restorative justice programs are still in their infancy, but that doesn’t excuse him from looking more closely at their implementation. If no-blame systems and restorative justice programs don’t “scale up” to cover whole societies because they are dependent on the very social institutions societies provide, that would shoot his alternative game out of the water.
Waller’s “last challenge” (p. 300) is what to do with dangerous criminals. He applauds Gilligan’s restraint-only institutions, with education and other such programs offered (voluntarily?) but doesn’t notice that this is still punishment, and still needs the justification of some kind of desert. We don’t send people there unless they are found guilty. He insists that this wouldn’t be “brutal Clockwork Orange” therapy, but this distracts us from the more important issue: is any therapy the answer? Suppose we have pills which accomplish what brutal Clockwork Orange therapy purports to accomplish. Not brutal, quick and easy. And if the person won’t take the pill, then what? Can we forcibly give them the pill? Is that brutal? Is it fair?
He asks what his alternative system would do with Bernie Madoff (p. 303) and then he doesn’t answer his own question. He says it is hard. He piles on (I describe the anti-thinking tool, piling on, in my forthcoming book on thinking tools).
That’s not an easy question. But in a system that rejects moral responsibility, we will not give him a lengthy prison sentence and suppose that the problem is solved without examining carefully how to put in place safeguards that will catch such wrongs before they come massive and without even trying to understand how someone with such talents and opportunities would follow a path that he obviously knew would end in personal disaster (p. 303).
Let’s see. Is he saying it’s OK to give Madoff a long prison sentence if we do examine, and try, etc.? Or will we just do the latter and let Madoff go home to his millions? Do we at least take away his ill-gotten wealth (forcibly, if necessary)? We’re not going to let him go on living the extravagant life, right? Maybe we won’t whip him or even put him behind bars, but we are going to forcibly remove his belongings from him, forcing him to make do with very modest means, and forcing him to abjure all careers in finance, aren’t we? Aren’t we? And is that fair? What right do we have to do this? If somebody’s unavoidable mistake led to similar financial loss, we wouldn’t do that, would we? It’s because we deem Madoff guilty that we consider that we have the right to rescind his rights (under the rules of the moral responsibility game) and do all these things to him that he doesn’t want us to do, and which we couldn’t justifiably do if he weren’t guilty. That’s punishment. Not retributive punishment, but punishment and blame, all the same. (And note that it is like being banished from a soccer game, given a red card. That’s a fair rule, isn’t it?)
What does Waller see as wrong with this vision of the moral responsibility system? He understands my proposal very well: there is an artificial and somewhat arbitrary plateau or threshold, and we don’t waste effort examining differences between those who make the grade (lucky them—they are not morally responsible for making the grade, but having made it, they are morally responsible). His main objection is that this involves confusing “take-charge responsibility” with “moral responsibility.”
In short, there is a take-charge responsibility plateau, and those who reach that general level of competence—though they have significant differences in talents and capacities—can take, exercise, and benefit from take-charge responsibility. But a plateau of moral responsibility is a different matter altogether, and establishing the benefits of the former offers no evidence for the legitimacy of that latter (p. 233).
But now Waller faces an odd impasse. It is trivial, given the way he has defined the terms, that take-charge responsibility is different from moral responsibility. I agree with that. I am not confusing the former with the latter: I am proposing the former—let’s call it take-charge responsibility for the role of morally competent citizen in a free society—as a replacement for the latter. And Waller has just granted that I have given a good, defensible account of take-charge responsibility. The only thing it lacks, it seems, is that it is different from the indefensible concept of moral responsibility that he is trying to abolish. Something has gone wrong here.
Waller sees me as making people a coercive offer, an offer they can’t refuse because if they refuse they are “ostracized from the human community” (p. 234). No, once again, it is like playing baseball: design a better game if you can, and refuse to play baseball if you must, but then of course you must forego the joys and benefits of playing baseball. Waller’s own alternative of restorative justice systems is a good reaction to the offer I’m proposing, and he’s welcome to flesh out the details in a way that will attract the people away from the moral responsibility game, but the advertisements to date are not alluring.
Part of what is—or should be, I think—unattractive about Waller’s alternative is its reliance on “positive reinforcement” which would be instituted “in order to maximize desirable behavior” (p. 147). It is a bit ominous that B. F. Skinner, the pioneer of this idea, in Walden Two (1948) and Beyond Freedom and Dignity (1971), gets only one mention, a quotation of his ringing declaration of the value of moving beyond freedom and dignity from About Behaviorism (1974), with none of the details examined. Skinner, notoriously, got to insist that there would be no punishment in his utopia by the simple expedient of distinguishing punishment from “negative reinforcement,” which, along with positive reinforcement, would be the method of choice. What’s the difference between punishment and negative reinforcement? Punishment is administered after the action, to discourage its repetition; negative reinforcement is administered before the desirable action, to encourage it. A society that abjured punishment might nevertheless deprive its members of food and freedom until they did what the authorities wanted. If Waller does not mean to endorse Skinner’s vision, it behooves him to make it clear what the differences are between his proposal and Skinner’s.
Until Waller actually comes up with an alternative system of justice that could work, we’re stuck with the moral responsibility game, the best game in town, and Waller concedes that it does work to a considerable degree. It just isn’t fair. He dismisses as absurd my claim that it is fair because luck averages out in the long run, and offers realistic examples that exhibit good head starts amplifying fortunate outcomes in the long run and bad beginnings spiraling down into worse and worse conclusions. There are of course an abundance of such examples to point to, but they don’t in fact make the point he thinks they do. To make the logical point, suppose that in a population of 300 million, there were 3 million examples of manifestly amplified good fortune and bad fortune of the sort he details—but only 3 million! The luck averaged out in ninety-nine percent of the population: there were as many deprived childhoods that led to happy outcomes as not, and equally many privileged childhoods that led to deplorable ends. If anything like that were true, my claim would not be absurd at all. But could anything like that be true? Of course it could, because society is actively engaged in all manner of compensatory feedback mechanisms designed (imperfectly) to make it true. For instance, by adhering firmly to “the blame system,” society puts parents on notice: if you spoil your beloved children (in either way) they will suffer as adults because we will hold them morally responsible for their misdeeds—no “get out of jail free cards” for them! Harnessing parental love (a gift from evolution), this artifact encourages the investment of considerable time and energy in moral education by parents and others who have the interests of the children at heart. If a child is growing up alarmingly wild and irresponsible, parents and teachers offer compensatory programs of one sort or another. It doesn’t always work, of course, but we keep monitoring our successes and failures, devising new, hopefully better, schemes. Compare our disciplinary practices today with those of centuries past to see the progress we’ve made. To take some relatively trivial further examples, the Boy Scouts have probably redirected millions of boys away from life trajectories ending in crime, and even television soap operas replace vulnerable naivete with street smarts in impressionable girls and young women. “Crime doesn’t pay!” is the implied moral of sensationalistic coverage of famous trials (the saga of O.J comes to mind) and is also the subtext of all the “police procedurals” that exploit our ignoble taste for representations of violence. Civilization is far from perfect (so of course it isn’t perfectly fair) but it is still a remarkable improvement over the state of nature. See Paul Seabright’s excellent book, The Company of Strangers (2nd edition, 2010, for which I wrote the Foreword) for an eye-opening vision of the security we all enjoy, fragile as it is, thanks to the trust that is routinely and mutually presupposed in most of our interactions with our fellow human beings. The spine that holds that trust in place , only intermittently visible when it is called upon for enforcement, is our mutual understanding that adults have the moral competence to make promises, both tacit and explicit, and promises are to be kept—or else.
Waller offers a forthright summary of his position on p. 102:
. . .compatibilists are half right: humans can indeed flourish in the thoroughly natural environment that shaped us, but moral responsibility cannot. And human flourishing will be enhanced once humans give up the idea that we can transplant moral responsibility from its native habitat of miracles and mysteries and make it survive in the natural nonmiraculous world.
What mainly interests me about this passage is that it ignores the social world in which humans live. Of course we can flourish in the natural environment that shaped us (and chimps and reptiles. . .) but we have also adapted to the social world, which is still natural (not at all miraculous), but contains institutions and possibilities that don’t exist in the presocial world: promising, trusting, apologizing, buying and selling, agreeing and disagreeing, and many more. Waller thinks human flourishing would be enhanced by abandoning moral responsibility. I think human beings would be disabled. That’s what I have always thought, but now, thanks to Waller, I can see better, deeper reasons for my conviction.
- DCD, October, 2012