Law and Philosophy (2010) 29:483–504 DOI 10.1007/s10982-010-9075-9
Springer 2010
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
RESPONSE TO CRITICS (Accepted 19 March 2010)
We would like to begin our response to these comments on and criticisms of our book by expressing our deep gratitude to a number of people: to those to whom we will respond, who have found our book worthy of their attention and their criticisms; to those who did not write for this issue but who attended the conference in Chicago at which these papers were presented and who commented orally on the book; to Heidi Hurd and Michael Moore, who organized that conference and solicited these papers; and to Law & Philosophy, for agreeing to devote an entire issue to a discussion of our book.1 The rewards of a scholarly life are many, but surely the greatest is to have oneÕs contributions taken seriously by others in oneÕs field. Although our aspiration in writing the book was to have had the last word on the topics we took up, we realized that such an aspiration is almost never achieved. What one more realistically hopes is that oneÕs contribution be a constructive one, one that advances the discussion even if it does not end it. We hope that at least that aspiration has been realized. I. ALEC WALEN
Alec Walen makes four criticisms of our book. He argues that we do not justify treating mistakes of how much concern is sufficient concern as nonexculpatory when we do treat mistakes of fact as exculpatory. He argues that we do not justify treating
1
Kim Ferzan was not an Associate Editor at the time at which Law and Philosophy agreed to devote an issue to the book.
484
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
negligence as nonculpable. He argues that our repudiation of resultÕs lack of ability to affect retributive desert overlooks the many ways luck affects how we fare in life. And in a footnote, he quarrels with our position on incomplete attempts and other inchoate crimes. With respect to the first criticism, we hold firm to our conviction that mistakes of fact are exculpatory. If someone does not realize that a given act is risky to an extent that would evidence insufficient concern for othersÕ interests, his lack of awareness of the actÕs riskiness negates the inference from the act to his having insufficient concern. And if his act does not display insufficient concern, then he is not culpable and thus not deserving of punishment. Well, what if someone mistakenly believes that imposing high risks on others just is sufficient concern? Our answer: someone who treats imposing huge risks on others for trivial reasons is displaying insufficient concern, even if he does not so regard it. If WalenÕs criticism cuts no deeper than this, then he has failed to draw blood. There is a deeper issue, however, which does not turn on whether an actor thinks an act shows Ôsufficient concernÕ for othersÕ interests. That issue is whether we can ultimately be deemed morally responsible for the weights that various reasons appear to us to carry. If reasons and their weights are like factual beliefs and just come to us unbidden, beyond our voluntary control, then can we ever be morally responsible for our acts? This is perhaps the most difficult issue in the free will-determinism debate – not whether our choices are ÔfreeÕ but whether we have the appropriate degree of control over the reasons for which we act – and it is an issue we elide in the book. We assume that someone who acts in a way that displays insufficient concern just is morally responsible for his act. That does not depend on whether he thinks his concern is sufficient or insufficient. But if he has made a factual mistake and does not believe he is imposing an untoward risk, then he is not displaying insufficient concern – again, regardless of whether he believes he is or is not. Moreover, it is not so easy to draw a parallel between mistakes with respect to the weight of reasons and mistaken factual
RESPONSE TO CRITICS
485
beliefs. Even those who would punish for negligence would distinguish between an intentional killer who believes he has the right to kill others and the person who mistakenly (and unreasonably, as these theorists see it) believes that a deadly poison is actually sugar. The former killer receives substantially more blame and punishment. This asymmetry suggests that no one believes that mistakes regarding the weight of reasons are morally on par with mistakes regarding facts. All theorists need a good answer for this asymmetry and thus, this puzzle cannot itself undermine our argument against negligence. WalenÕs second criticism, that we do not justify treating negligence as nonculpable, is something of an ipse dixit in that he does not engage the arguments we advance in Chapter Three. His remarks suggest – but only suggest – that he holds this position: culpability for negligence comes from the character defects that led the person to be unaware of the risk. We deal with that position at some length in Chapter Three, and Walen nowhere engages our arguments against this position. Because we cannot be sure that is the position Walen holds, we must just demur to this criticism and refer readers to Chapter Three. WalenÕs third criticism, that in rejecting result luck we neglect the many ways luck affects how we fare in life, is misdirected. Of course luck affects how we fare, and in a multitude of ways. However, one need not be a luck egalitarian to deny that luck affects the blame or punishment we deserve. Our position is that result luck does not affect culpability, and culpability is the determinant of retributive desert. Walen gives some consequentialist reasons for singling out harm-causers among those deserving of punishment. However, he does nothing to show that those who imposed what they believed were the same risks and did so for the same reasons, but whose acts did not result in harm, are less deserving of punishment. WalenÕs final criticism is found in a lengthy footnote. We argued in Chapter Six that one has not acted culpably until one takes what he believes is the final step in unleashing a risk of harm. Therefore, one who intends, say, to kill with a gun has not acted culpably with respect to killing until he has pulled the
486
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
trigger, for prior to that point he will believe the risk of death has not left his control. Walen believes this position is subject to a reductio. It forces us to conclude that someone, intending to kill, who is chasing his victim through a crowd, but who has not yet fired at him, is culpable only for creating fear or the possibility of an accidental death or injury, so that he is close to the moral equivalent of a practical joker recklessly brandishing a fake but realistic gun. Walen says, ÔThe latter may be guilty of a serious crime; that sort of practical joke could get someone – most likely the actor, but possibly an innocent bystander – killed if the intended victim acts in self-defense or a third-party acts in other-defense. Nevertheless, the crime of creating such risks should be much less [serious] than the crime of trying to murder anotherÕ. We do hold the position Walen finds ludicrous. ItÕs a bullet we bite. However, if one rejects our position on incomplete attempts (and other inchoate crimes), we demonstrate that one is led to the position that the act of intending harm is itself a culpable act. And that position leads to the many puzzles and difficulties we describe in Chapter Six. So there is a different bullet to bite if one finds ours too hard on the teeth. We lay out that challenge in Chapter Six, but Walen does not take it up. We should add, in connection with this discussion of inchoate crimes, that perhaps punishing inchoate crimes represents a subtle but real transition from retribution based on culpability to preemption and preventive detention based on dangerousness. Preemptive, preventive restrictions on liberty are a topic we do not take up in our book, as our focus is on a desert-based criminal law, not on preemptive measures aimed at preventing harm. Perhaps thatÕs a topic for a future book. It is surely a topic worthy of more treatment than it has thus far received. II. JIM STAIHAR
Jim StaiharÕs main criticisms are directed at our thesis that someone is not culpable for acting when he believes facts exist that would justify his act but is not motivated by those facts but by facts that do not justify the act. Prior to making those
RESPONSE TO CRITICS
487
criticisms, however, Staihar quarrels with our rejection of a substantial risk requirement for culpability in favor of an unjustified risk requirement. Our main thought experiment to support our position involved Deborah, who loves to play involuntary Russian Roulette on passers-by just for the thrill she gets from doing it. Our intuitions, shared by everyone we have heard from, is that Deborah is culpable for her act, no matter how many empty chambers her gun contains and therefore even if the risk of death she imposes is far less than the risks imposed by such ordinary and nonculpable activities as driving (carefully) to the store, playing golf, and so on. We concluded that the magnitude of the risk was doing no independent work in the culpability calculus, and that all the work was being done by the reasons for which the risks of various magnitudes were being imposed. Staihar does not directly gainsay our argument. Rather, he asks how a risk can be insubstantial and yet be a gross deviation from a reasonable standard of care. We see no difficulty here. Deborah may have a thousand-chambered gun and only one live round. The one in a thousand risk of death she imposes for no reason other than a thrill seems to us to be a gross deviation from the concern for othersÕ interests that it is reasonable to expect. If she were driving an ambulance with a very sick patient to a hospital, however, then imposing an even higher risk of death on bystanders than she imposes playing Russian Roulette would not be a gross deviation. We fail to see the problem. Turning to StaiharÕs main points, we think Staihar has misunderstood our arguments. He admits that the ambulance driver who likes to speed, or the executioner who likes to kill, are justified in speeding and killing despite the fact that they have misanthropic motivations. He claims that this is so because they appropriately restrict their speeding and killing to situations where the facts justify the acts. But he argues that this does not apply to the person who redirects the trolley from the five workers to the one, but only does so because he wants to kill the one. We argued that he, like the ambulance driver and the executioner, does not act culpably in redirecting the
488
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
trolley if he is aware of the facts that justify it (saving a net four lives), even if he is not motivated by those facts. That misanthropic person may not be admirable and may be dangerous, but his act does not itself display insufficient concern. He has a bad character but has not acted badly. Indeed, because he has no obligation to redirect the trolley, and because we want him to do so, we may well try to motivate him by pointing out how much he hates the one worker. Staihar misunderstands our argument on this point. It is not based on the opacity of the actorÕs motivations. Nor is it based on the effects of threatening the actor with blame or punishment. For the actor may not merely want not to be censured. The actor may want not to be culpable. So if he says to us – ÔIÕll redirect the trolley, but only because I hate the one worker, and only if I will not be culpable for doing so; otherwise, I will just let the trolley hit the fiveÕ – what will be our reply? He will not be culpable for doing nothing, and he does not want to be culpable. We think he will not be. Staihar disagrees and must therefore counsel him to do nothing and let the five die. So it is a red herring to bring up feelings of resentment and their consequences. We are interested in the actorÕs culpability, and the actor may not want to be culpable, even if no one expresses or even feels resentment. StaiharÕs misconception of our argument against the materiality of motivation to culpability also infects what he says about the lesser versus least evil issue. In a case where the trolley will kill five unless switched, and it can be switched to a track with two workers or to a track with one (or no) worker, we would prefer that the actor switch the trolley to the track on which it will do the least harm. But if the actor is not motivated to do so, and will not be culpable if he does not switch the trolley but lets it kill the five, then we should be grateful that he is motivated to switch the trolley to the track with two workers. After all, he would have been justified in doing so had that been the only siding. Moreover, had he switched the trolley to that siding and only then discovered a siding to that siding, on which there was only one worker, he would have had no obligation to switch the trolley to the second siding. The same arguments we
RESPONSE TO CRITICS
489
have given for why motivation does not matter in the original trolley case apply here as well and suggest that the actor is not culpable for switching the trolley to the siding with two workers, even if he is aware of the second siding, and even if he is motivated by hatred of the two workers. III. PETER WESTEN
Peter Westen raises some very provocative points in his critique of our book. His principal objections are to our view that resulting harms are immaterial to punishment and our view that risk is an epistemic notion and represents judgments of probability. He also denies that our consequentialist account of justification is correct, even with a deontological constraint built into it. And he criticizes our view that self-defense (and other-defense) against innocent aggressors cannot be impersonally justified but only excused (or personally justified). Let us begin with the last point. Westen says that Ô[a] selfdefender who kills an innocent aggressor to avoid being killed leaves consequentialist evils in equipoise and, hence, does nothing wrongÕ. But suppose that there are two innocent aggressors, or five. Suppose five young children find loaded guns and, believing them to be toy guns, point them at the selfdefender. If the self-defender is justified in killing the five children, it cannot be because the consequences are in equipoise; for he has taken five lives to save one, his own. Westen says that although innocent aggressors are not culpable, they are doing something wrongful, namely, invading the defenderÕs sphere of rightful autonomy. Note, however, that this response begs the central question, which is whether the defenderÕs Ôsphere of rightful autonomyÕ includes freedom from a risk of death or injury when maintaining that freedom requires killing nonculpable people. The gun-brandishing children do represent a threat to others; but the latter, if entitled to use deadly force to protect themselves, represent a threat to the children. And we see no reason to privilege the defenders over the children, the insane, the mistaken, and other nonculpable aggressors when allocating risks.
490
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
Westen believes our deontologically-constrained consequentialism is a mistaken account of justification. The only counterexample he provides is one where the actor, escaping a death threat, causes bystanders a couple of concussions and the blinding of an eye. Westen correctly notes that the actor violates no deontological constraint – he is not ÔusingÕ the bystanders – and that his death is, on the consequentialist calculus, a greater evil than the harms he causes. If that is true, then, as Westen correctly concludes, we would deem the actor to be justified. Westen states, however, that the actor clearly is not justified, so our account of justification must be in error. We stand by our conclusion. We embrace WestenÕs supposed reductio. And Westen does nothing more than assert that our conclusion is mistaken. He provides no alternative account of justification. Westen believes that he has an account of objective risk, one that does not cash out risk in terms of probabilities. Westen concedes that probabilities are epistemic and perspective-relative. He claims, however, that the correct account of risk for criminal law is a counterfactual one, namely, that when harm does not occur, its risk was how ÔcloseÕ it came to occurring. (In a nearby possible world, would the harm have occurred?) We do not wish to delve into the metaphysics of counterfactuals, nor do we have a view on the truth value of claims about possible worlds and their proximity to the actual one. When the would-be killerÕs gun jams, or the intended victim has donned a bullet-proof vest, how ÔcloseÕ is the world in which the killer succeeds? We have no answer for this, and we doubt that Westen does, or that juries and judges will be able to give consistent, nonarbitrary answers. Moreover, because the epistemic account of risk for us fully determines culpability, the relevance of WestenÕs account turns entirely on the proposition that as between two defendants, each of whom estimates the same probability that his act will cause harm, the one whose act ÔobjectivelyÕ risks more harm is – what? More culpable? More deserving of punishment? Westen does not say. Because we believe that the subjective estimate of probability determines culpability, and culpability should fully determine punishment,
RESPONSE TO CRITICS
491
we cannot see how WestenÕs ÔobjectiveÕ probabilities are material to anything the criminal law should care about.2 WestenÕs lengthiest criticism is directed at our position that resulting harm should not affect punishment. Westen thinks that it should, but not because he thinks results affect desert. He agrees with us that two actors who impose what they believe is the same risk of the same harms for the same reasons are equally culpable, no matter if oneÕs act causes harm and the otherÕs does not. And if they are equally culpable, then Westen agrees they are equally deserving of punishment. Where he disagrees with us is over whether their equally deserving punishment means that the harm-causer should receive no more punishment than the one whose act causes no (or less) harm. Westen, making an argument he attributes to Plato, maintains that although the actor who causes no harm deserves as much punishment as the actor who causes harm, it does not follow that he should receive the same punishment. Indeed, says Westen, he should not. Why not? Because we (society) should feel relieved that he failed to cause harm, which should cause us in turn to feel less need to punish to the full extent deserved. And if we feel less need to mete out the full measure of retributive desert, then it would evince overweening self-righteousness to do so. Westen contends that this position is in accord with widely shared intuitions, which although not intuitions about moral desert, are moral intuitions nonetheless. We doubt this. In Chapter Five we give several examples in which there is both harm and more than one equally deserving actor, but in which we doubt that there is an intuition supporting singling out the harm-causer from the others. (Think of the Satanic Cult example, the children sword-fighting, or the children hitting 2
We also wonder whether conceptions like Ônear miss,Õ Ôdangerous,Õ or ÔscaryÕ are ultimately hostage to other objective probability accounts. That is, when we say something was Ôclose,Õ although we may not fully specify the probabilities, we are likely doing a rough and ready assessment wherein we select a reference class and then think about the probabilities. This is also true for the actor, as well as the observer – for an actor to think his act is ÔdangerousÕ is to have some notion of probabilities (though not fully specified at the Ph.D level).
492
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
baseballs into the neighborÕs yard.) Nor do we think it overweening self-righteousness to punish those who did not cause harm in those examples equally with those who did. Westen denies that the mysteries of proximate cause would need to be unlocked in order to make resulting harm a material element in punishability. Instead, Westen argues, we can just look to see if society regards the actor with more righteous indignation than it regards those who risk but fail to bring about harm. If it does, then he is a proximate cause, and causal metaphysics are immaterial. We do not see why we should abandon reason in favor of jury indignation if we are going to allocate different punishments to equally deserving actors. We should and probably do feel equally indignant towards both Pierres in our PierreMonique scenarios even though one looks like the quintessential non-proximate causer. If Westen is looking at the right consideration in directing our attention to jury indignation, we may as well delete altogether any concern with proximate causation. We are happy to do so, but only because resulting harm should not matter to punishment. IV. ERIC JOHNSON
Eric Johnson argues that our critique of objective probability is problematic. He claims we settle on subjective probability through a process of elimination, but we fail to take account of another method for determining objective probability that presents a viable alternative to a subjective account. Johnson offers us an alternative account of objective probability – determine the probability from Ôwhat the actor believes, knew, or should have known about the underlying factsÕ. Specifically, he supports the approach that focuses on what the actor knew about the facts and circumstances and determines the objective probabilities from those facts and circumstances. Although JohnsonÕs response merges our arguments about risk, it is worth separating the arguments pertaining to recklessness and negligence. Indeed, at times, Johnson appears to take our arguments out of context, thus misunderstanding our points. We donÕt, for example, adopt his Subjectivity #2
RESPONSE TO CRITICS
493
approach. Rather, in the context of focusing on what an actor could have done, we argue that, considering everything about the actor (including his beliefs), he could not have done otherwise in the sense of having any internal reason to do so. But that isnÕt our account of culpability. Our account of culpability is based on what the actor believes and his estimate of the probabilities. There is no objective calculation of probabilities required. The calculation is purely subjective. So, letÕs take our first argument for subjectivity. We argue that the risk that the actor imposes is the risk that he believes that he is imposing and not some ÔobjectiveÕ risk. That is, we reject the position that in instances of risk creation there can be an actus reus formulation of the risk that departs from the actorÕs subjective approximation of the risk. There is, in other words, no attempting to be reckless, just recklessness. So, if David believes that he is going 80 miles per hour, but his speedometer is broken and he is actually going 50 mph, we should understand the risk he is imposing based on his 80 mph belief, not the 50 mph speed. We argue that the actual risk of an accidentÕs occurrence is always 0 or 1; and although objective probabilities can be calculated based on relative frequencies, the problem is that there is no principled way – no way that bears on an appropriate actus reus formulation for the criminal law – to select the reference class within which the probability would be calculated. Johnson tells us that there is an objective probability – a probability based on what the actor believed or knew about the underlying facts – and thus we ÔoverreachÕ when we infer that there is Ôno gap between the actorÕs subjective estimate of the risk and the ÔtrueÕ or ÔobjectiveÕ riskÕ. But, to finish the sentence that Johnson quotes, we say there is no gap because the objective risk is either illusory (because the risk is really zero or one) or arbitrary (because there is no principled way to select among references classes), and that same complaint applies to JohnsonÕs view. If we look at what David ÔknowsÕ, then he doesnÕt know the speed he is traveling. He has a false belief. How are we to glean the probability of harm based on DavidÕs knowledge? Moreover, even if we could calculate the proba-
494
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
bility, how would this be a principled measure of the actual risk that David presented? The risk was either 0 or 1. Some alternative probability culled from the true beliefs that David has hardly represents a principled way to determine the ÔobjectiveÕ risk that David presented. This approach hardly seems sufficiently objective to serve as an appropriate formulation of an actus reus. Not only does JohnsonÕs notion of Ôobjective riskÕ seem unprincipled, but when it is used for the culpability determination of recklessness, it appears conceptually confused. Johnson claims that the definitions of recklessness and negligence within the Model Penal Code require an analysis, not of what the actor believes, but what the actor knows. Let us return to David. Now, Johnson would have us define not only the risk presented based on what the actor knows but also DavidÕs culpability. But under JohnsonÕs approach David would not even be attempting to be reckless. If what David believes – no matter how misguidedly – does not determine his culpability, then there is no conceptual room for us to say that David was reckless for believing that he was imposing this risk because his mistaken beliefs would not be part of the calculation. JohnsonÕs approach renders impossible attempts conceptually impossible. JohnsonÕs framework is also not normatively desirable for determining recklessness. Consider a variation on Malone. Assume that the defendant does not know that two bullets were put in the gun before he received it. He believes that he had already loaded the gun with one bullet (but he is incorrect about this), and then he puts another bullet in. How many bullets does Malone know are in the gun? One. He has a justified true belief as to one bullet. He believes there are two bullets. And, there are actually three bullets. In such a case, it seems clear to us that MaloneÕs culpability is a product of the risk he thought he was creating with two bullets in the gun. It is the decision to pull the trigger given the belief that there are two bullets in the gun that reveals MaloneÕs degree of indifference to the lives of others. Finally, let us look at negligence. If there is merit to JohnsonÕs proposal, it would be here. JohnsonÕs claim is that an
RESPONSE TO CRITICS
495
actor who does not foresee a risk is acting unreasonably if given what he knew, he should have foreseen the risk. We do not find this to be a viable option. First, Johnson quotes Holmes who notes that Ôrecklessness in the moral senseÕ would be purely subjective, but this view is Ôtoo subjective for the criminal lawÕ. But then we have to ask what Johnson thinks justifies punishment. If one wants to punish only those who deserve it and only as much as they deserve, then the moral sense is the criminal lawÕs sense. If Johnson seeks to use the criminal law for some other purpose, well, we got off that boat in Chapter One of our book. Second, JohnsonÕs objective alternative lacks any moral bite. As George Sher has recently argued, Ôit is hard to see how the truth of S, construed simply as the claim that agents have such evidence, can render them responsible for acting wrongly. For how can any mere abstract relation among a set of facts or propositions and the further proposition that a given act is wrong or foolish bring the person who performs the act into a close enough relation to its wrongness or foolishness to justify us in blaming, punishing, or making him pay the price for it?Õ3 V. SUSAN BANDES
Bandes attacks our views about negligence. We claim that the negligent actor cannot be blamed because at the time of the negligent act, the actor is impervious to guidance by any conception of the riskiness of his conduct other than his own estimation of that risk. Bandes claims that we assume a sharp distinction between conscious states (which are correctable) and subconscious states (which are not). She argues this is descriptively inaccurate because consciousness is a spectrum. Moreover, she claims that the Ôawareness of controlÕ that we have is Ôthe capacity that permits an actor to choose whether to risk or cause harm, and this type of awareness may take place unaccompanied by cognitive awarenessÕ. Bandes claims that human beings operate primarily through subconscious states, and our consciousness has little to do with it. Moreover, we can ÔeducateÕ people to make their sub3
Sher George, Who Knew? Responsibility Without Awareness (Oxford New York: Oxford University Press, 2009), pp. 75–76.
496
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
conscious operations accessible. ÔPolice can be trainedÕ. ÔDoctors…can improve their clinical intuitionÕ. ÔPeople can be trained to improve their empathetic capacityÕ. From our capacity to be trained, Bandes jumps to blameworthiness. She also questions our conclusion that Sam and Ruth could not recall the child in the bathtub, asking if we are inferring that they Ôdid not act differentlyÕ because Ôthey could not?Õ She claims that their behavior was within their control. Clearly, this pinpoints the matter for debate, which is the locus of responsibility. Should we take a broad view of an individual, including both his conscious and subconscious aspects? Or should criminal law take the narrow view? We claim the latter. Criminal law expresses blame and condemnation, and that should require that the agent consciously reason and exercise his will with respect to the harm that eventuates. We also donÕt see how far the training point carries. So Sam and Ruth can be trained to be better parents. To not forget. To not be selfish. To be aware of these issues. But even if the state should be involved in such training and educating (which, to us, smacks of overweening state perfectionism), it is not the case that at the crucial moment and in the absence of that training that Sam and Ruth can remember what they forgot. Even if Joe can be trained in karate to prevent a deadly attack, thus alleviating his need to use a gun, such a possibility does not mean that Joe can at a given moment, without such training, deliver the kick. And then, as the karate example reveals, we need to determine what duty an individual had to be trained thusly, whether he was aware of the availability training, and therefore whether he was culpable for not getting the training. (Bandes later voices concern about what to do with Sam and RuthÕs second, third, or fourth child, but her argument, that the preconscious and conscious work together and provide feedback loops does provide the answer here, which is to question the descriptive accuracy that Sam and Ruth would fail to be aware of the risk of harm the second or third time around.) Moreover, we believe that Bandes makes too much of what we assume about Sam and Ruth and their inability to remember the child in the bath and too little of her own claimÕs
RESPONSE TO CRITICS
497
implications. As she notes, in our discussions of opaque recklessness, we do discuss how individuals operate at both the preconscious and conscious levels. What we deny is that conduct that is not conscious at any level is sufficient to ground blame and punishment. On the other side of the coin, we question whether Bandes is truly committed to abandoning the conscious/subconscious distinction. Will she punish for dreams that can be indirectly influenced? Will she abandon the distinction between recklessness and negligence and punish all negligent actors at the same level as reckless ones? After all, if she denies that there is anything to a conscious/subconscious divide, then she must deny that reckless actors – those who consciously choose – are more culpable than those who do not. But if she is willing to draw a descriptive line between those who are conscious of risks and those who are not, then she can no longer question our descriptive assumption, and she must turn to and answer the normative question: Are actors blameworthy for acts over which they only have indirect and atemporal control? Bandes also notes that proving mental states always relies on external evidence, but she takes things one step too far in our view when she argues that Ôthere is no good reason to believe that the conscious mind provides a reliable account of what one adverted to or perceived before engaging in risky behaviorÕ as Ô[o]ur conscious reasons are often post-hoc explanations rather than reflections of an actual processÕ. We find these claims hard to take seriously. To think that we do not have an understanding, as we write this, of what we are doing and why we are doing it, is to give up the central role that practical reasoning plays in responsibility. If all our consciousness amounts to is an after-the-fact explanation of something our brains and bodies did, then we might as well stop now. The criminal law does not assign blame then; it merely controls creatures. Although there are times when we are self-deceived, more frequently we are not. And the distinction between when we do something ÔaccidentallyÕ and when we do it Ôon purposeÕ is a central component of our understanding of ourselves as responsible
498
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
creatures capable of determining how we will interact with the world. VI. ANDREW LEIPOLD
Andrew Leipold criticizes our abandonment of negligence, worries about our theoryÕs practical application, and questions whether grand theories are desirable. To begin with his critique of negligence, Leipold argues against both what he deems our ÔstrongÕ argument – that the failure to attend to risks is not blameworthy in the criminal law sense – and what he deems our ÔweakÕ argument – that even if the negligent actor is criminally blameworthy, there is no principled place to draw the line between those who are morally culpable and those who are not. Let us take the strong argument first. Leipold thinks that we are doing away with an entrenched basis of liability Ôbecause our willingness to impose serious sanctions on a negligent actor is deeply rootedÕ. We demur. Sure, one may get fired from a job for falling below some standard of performance – there isnÕt a lot of work for clumsy ballerinas and blind pilots (though the absent-minded may find employment within the academy). The issue, though, is punishment. And here, the law is rather skeptical of negligence. Negligence is not sufficient for most crimes. And, indeed, when harm does not occur, reckless endangerment is often only a misdemeanor, and there is no crime of negligent endangerment. In fact, one begins to wonder whether most folks care about negligent action at all when harm does not occur, thus leading to the conclusion that negligence liability is most likely not about culpability but about harm-causing. Indeed, even those theorists who wish to claim we are responsible for our negligent acts cannot establish that we are culpable and blameworthy for our negligent acts. For different reasons, Joseph Raz and George Sher have both argued that we are responsible not just for the acts that we choose, but also for those failures to choose to exercise greater care because those failures are, to grossly oversimplify their positions, still a part
RESPONSE TO CRITICS
499
of us.4 So, when the baby dies because her mother forgot to drop her at the daycare (because the father usually does so), the mother takes responsibility for the harm caused by her forgetfulness. But even if she wishes to claim responsibility for the harm, this is not to say that she is blameworthy for forgetting. It is certainly true that one feels obligated to account for the harm, but it is not the case that this is sufficient to make the individual culpable and blameworthy. Moreover, when we blame the individual, it is often that we blame the underlying trait of character (forgetfulness, clumsiness, etc.) and not the individualÕs exercise of will to harm others. It is only in the cases in which the actor chooses to harm others, and the potential harm factors into the actorÕs deliberations, that his action is sufficiently his own in a way that warrants blame and punishment for his actions instead of just criticism of the kind of person he is.5 Leipold argues that the law does not require an injunction to notice everything, to not forget anything, etc., but rather, an injunction to notice, remember, etc. those items that a reasonable person would notice. His claim, then, is that the moral injunction is not as broad as we claim, and we are not constantly violating its command. The problem with this argument, however, is that, at the time of acting, no actor can recall what he should recall, pay attention to what he should pay attention to, without recalling, noticing, etc. That is, one has to have the power to remember or recall before one can cull out those specific items that it would be ÔunreasonableÕ not to. Leipold believes that our Sam and Ruth hypothetical undermines our position, but the move that he makes ulti4
For Sher, the Ôbasic idea is that a responsible agent is best identified not only with his subjectivity or rationality but also with their causes.Õ Id. at 121; For Raz, we are shaped by our Ôcompetence in using our capacities for rational agency.Õ ÔActions due to malfunction of our capacities of rational agency result from failure to perform acts of which we are masters. In acknowledging our responsibility for these unintentional acts and omissions we affirm our mastery of these abilities, deny that we are disabled in the relevant regards.Õ Joseph Raz, ÔResponsibility & the Negligence StandardÕ. Oxford Legal Studies Research Paper No. 29/2009; Columbia Public Law Research 09-207, available at http://ssrn.com/abstract=1436022. 5 Cf. Sher, supra note 3, at 74.
500
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
mately supports our approach. He claims that we should not focus on the moment when the baby drowns, but the moment when Sam and Ruth leave the baby in the bathtub to begin with. Leipold argues that Sam and Ruth knew (or should have known) that this posed an unjustifiable risk, and he goes onto give other examples where the actor makes a prior culpable choice. Leipold notes that in these instances, the prior act is culpable whether the actor recognized the risk or not. But the rhetorical force of each hypothetical derives from the awareness that one should not leave a child in the bathtub because one knows the risk of later forgetting. But our hypothetical posits the opposite: ÔSam and Ruth both go downstairs to greet the guests, both realizing that the child would be in grave danger if they failed to return and turn off the water, but both correctly believing that at the rate the tub is filling, they will have plenty of time to return to the child after they have welcomed the guestsÕ. (p. 77). There is no prior culpable choice. Moreover, as we argue against Bandes, the criminal law is not about incentivizing conduct but punishing culpable actors. If we want people to guard against errors, then perhaps we should employ government resources to warn them. We should not punish those who did not notice or remember in order to promote othersÕ noticing and remembering. Leipold relies on the jury to determine when failures to notice or remember are sufficiently culpable for criminal liability and when they are not. But there is a difference between faith in juries and the complete failure of criminal law theorists to articulate the standard in the first instance. Why should we trust that juries can do what we ourselves cannot? Leipold also raises some implementation questions. At what levels do we pitch punishment? How do we take into account criminal history? Is our crime counting view capable of being implemented? Will our focus on the actorÕs reasons for imposing risks encroach on liberty? He then concludes by arguing that grand theory is, even if theoretically elegant, not practically possible – and the game is not worth the candle. Clearly, we cannot answer every practical implementation question here, given that we could not even begin to delve into
RESPONSE TO CRITICS
501
the depths of these concerns in our book. But we believe these questions actually show the value of searching for a Grand Theory rather than the fruitlessness of such an endeavor. If we start with a pure theory of criminal law, we can then ask how criminal law should interact with issues such as providing notice, constraining the state, allowing police intervention, using preventive detention, and the like. Once we know what it means to say that an individual deserves punishment, we can make progress on the practical problems of implementing that theory. We need not take the current practices of criminal law as givens. As just one example, consider our complaint against incomplete attempts – that they punish an actor for what she will do, and not what she has done. We argue (persuasively, we hope) that if we intervene, we are intervening to prevent harm and not to punish for a culpable act. Now, as a society, we may want to intervene, but we are going to need to search for other principles to do so; we cannot simply bastardize the criminal law and claim to be punishing the deserving. A grand theory may not be easy to implement, but it is where we should start. Moreover, if our theoretical arguments stand and the question is their implementation, then we hope that our book has still provided a useful prism through which to evaluate current doctrines. When one looks at incomplete attempts (or, even more troublesome, possession of burglarÕs tools), one will know if desert exists, and if not, what values appear to be trumping it. If we cannot implement an appropriate counting of criminal acts because it is just too hard given our epistemic resources, then at least we know the ideal – the answer – that we are striving for. VII. RON ALLEN
The gist of Ron AllenÕs critique is that we are disconnected from the real world. First, he claims that we have oversimplified the complexities of decision theory and the obligations of government. How could we not have, given that we devoted only a paragraph to it? Allen raises some fair points, but none of them would cause us to rethink our theory of culpability.
502
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
First, with respect to the stateÕs obligation not to punish the innocent, Allen notes that the calculation is not just a matter of burden of proof. We have to look at how many people are guilty, how many go to trial, and what the error rate is. True enough. The content of the criminal law is only one part of the crime and punishment calculus. Our view accommodates this. As we say in Chapter One, we subscribe to a moderate retributivist view wherein giving the guilty what they deserve is an intrinsic good but one that has to be weighed against other goods (and evils). The government has some complicated calculations to do.6 Punishing the guilty is only one goal. Moreover, if the government could devote resources to poverty or health care that would prevent the need to spend any money on the criminal law, that would be fine with us. Then, even though we have a theory of what it means to give people the punishment they deserve, the world simply wouldnÕt need it. AllenÕs critique, however, denies that we can simply cabin off the Ôreal worldÕ as we analyze culpability. And true, we have to take some things as givens (such as the existence of the criminal law). Moreover, in our book, we do occasionally make concessions to practical considerations; for instance, we argue that although the unit of culpable action is the volition, the actual product of that volition – the willed bodily movement – is more practicable for criminal law. Allen worries about the expansiveness of our conception of culpability. If any act may be producing an unjustifiable risk if we know the actorÕs reasons are bad, then any act that produces any risk gives the police probable cause to investigate it. Allen claims that even if standards are used in current law, they have the sharp edges of being distinct crimes and requiring dead bodies. That constrains the cops. Under our view, he claims, anything goes. However, despite being able to stop anyone, our view, says Allen, cannot result in any convictions. In a world where there is a privilege against self-incrimination, how do we 6
At one point, Allen extends our argument that results donÕt matter beyond the realm of culpability and into the realm of government responsibility. But our claim that results do not matter for purposes of culpability and blameworthiness does not entail the absurdity that results never matter.
RESPONSE TO CRITICS
503
get the sort of fine-grained evidence that we need? And how would the commentaries that we suggest might be read to juries square with the rules of evidence? But notice that once again these are questions about how to trade off the intrinsic good of giving people what they deserve against other goods and evils. Before we settle for a second-best criminal law that has been contorted by the needs of constraining government or working with juries, should we not know first what the criminal law would look like if it were shaped with the sole aim of giving people what they deserve? Should we not start from the ideal and then settle for less? ShouldnÕt we know that the criminal law we have is shaped by these less than ideal considerations and evaluate whether the trade-offs are worth it? Consider one case in which the criminal law has constantly struggled with the purity of theory and the limitations we wish to place on government – impossible attempts. If Alex shoots at a tree stump believing that he is killing a person, Alex is, to us, very culpable and deserving of punishment. If one is concerned with punishing Alex, it is because one worries about how this looks objectively: A man shoots a tree and goes to jail. Should we not worry that the government might overreach? That any sort of innocuous behavior will be punished? But, on the other hand, if we do have significant evidence that Alex thought that tree was a person (his diary, what he told his friends, what he said as he fired the gun, how he had detailed plans of how to dispose of the body), then we might worry that putting within the law of attempts some objective danger requirement will fail to punish some actors who are deserving and who we can prove are deserving of punishment. Although our preference would be to figure out some new rules for criminal procedure (why should we take the current rules as a given?), we can certainly see how the criminal law might be tinkered with to add these protections. But we still maintain that we need to know what the person deserves and what a pure theory would look like before making real world concessions. Chapter Eight and its discussion of these issues is just the beginning of this discussion, not the final word.
504
LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN
LARRY ALEXANDER
School of Law, University of San Diego, 306 Legal Research Center, 5998 Alcala´ Park, San Diego, CA 92110-2492, USA E-mail:
[email protected] KIMBERLY KESSLER FERZAN
Rutgers School of Law – Camden, 217 North Fifth Street, Camden, NJ 08102, USA E-mail:
[email protected]