Th ere is also a second signifi cant problem with punishing negligence, and that is that there is no principled and rationally defensible way to defi ne the “reasonable person in the actor’s situation” (the RPAS). Th ere is no moral diff erence between punishing for inadvertent negligence and punishing on the basis of strict liability, and the lack of a moral diff er- ence evidences itself in the inability to draw a distinction between strict liability and negligence on any basis other than arbitrary stipulation.
32 See United States v. Carroll Towing Co., 159 F.2d. 169 (2d Cir. 1947).
Th ere are two clear boundary lines for the RPAS. First, the RPAS could be a person apprised of all the facts about the world that bear on a correct moral decision. At the other possible conceptual boundary, the RPAS could be someone with all the beliefs that the actor actually held.
Put somewhat diff erently, where action falls below the standard of reck- lessness – the conscious disregarding of an unjustifi able risk – the action will appear reasonable to the actor and thus to the RPAS if the RPAS has exactly the same beliefs as the actor.
Th e two possible boundaries that provide the frame for character- izing the RPAS present us with this dilemma. If the RPAS knows all the facts, the RPAS always chooses the action that averts the harm (in the absence of justifi cation, of course). But if this is the standard of the RPAS, then every case of strict liability will be a case of negligence as defi ned by the RPAS standard. It will never be reasonable not to know.
(Notice, as well, that because risk is epistemic, the omniscient actor deals only in certainties: for her, the “risk” of a particular harm entailed by any act is either one or zero. On this construct, then, not only is there liability for every avoidable and regrettable – unjustifi able – harm, but there is also no negligence where harm does not occur.)
On the other hand, if the RPAS knows only what the actor knows, there is never any negligence either, only recklessness. Th e RPAS will always act as the actor acted where the actor is not conscious of the level of risk, and will act diff erently only where the actor is conscious of the level of risk, that is, is reckless.
At either conceptual boundary, therefore, RPAS collapses negligence into either strict liability or recklessness. Th e question, then, is where between those boundaries the RPAS is to be located.
Th e answer is that any location between these two boundaries will be morally arbitrary. Between the boundaries, any RPAS will be a construct that will include some beliefs of the actual actor together with beliefs that the constructor inserts. Which beliefs are inserted other than the ones the actor actually had will determine whether the RPAS would act as the actor acted. But there is no standard that tells us which of the beliefs of the actual actor should be left intact and which should be replaced by other (correct) beliefs. Th e RPAS standard, cut loose from the alternative moorings of the actor’s actual beliefs or of the world as it really was at the time the actor acted, is completely adrift
in a sea of alternative constructions, none of which is more compelling than others.
Some commentators at this point assert the possibility that the RPAS is like the actual actor in all material aspects but claim that the RPAS
“would have” adverted to and properly assessed the risks because the actual actor “could have” adverted to and properly assessed them.33 But there is an equivocation here in the reference to what the actor “could have” adverted to and assessed in the actor’s situation. If we take the actor at the time of the “negligent” choice, with what he is conscious of and adverting to, his background beliefs, and so forth, then it is sim- ply false that the actor “could have” chosen diff erently in any sense that has normative bite. For although it may be true that the actor “could have” chosen diff erently in a sense relevant to the freewill-determinism issue, it is false that in that situation the actor had any internal reason to choose diff erently from the way he chose.
To have such a reason, an actor will have to advert to that to which he is not adverting. But one has no control at such moments over what one is adverting to or is conscious of: try thinking of what you are not thinking of, but should be! Th e “could have adverted to the risk” posi- tion is directly at odds with the voluntary act principle as a refl ection of the value of restricting punishment to choices over which the actor has fair control.
To put the point another way, an actor may fail to form a belief (or a correct belief) if he (1) lacks the requisite background beliefs, (2) lacks the intellectual ability, or (3) lacks the motivation to form the belief. 34 With respect to (1), we do not see how it can be fair to say that an actor
“could have” believed a if only he had believed b without articulating how it was within the actor’s control to believe b. One would then have to show why it was, at the time belief b was available to him, a culpable decision not to form that belief.
In addition, as to (2), we do not see how we can blame an actor who either intrinsically (because of limited intelligence) or extrinsically (because of momentary distraction) fails to be able to form belief a at
33 Hart, supra note 2, at 148; Brenda M. Baker, “Mens Rea, Negligence, and Criminal Law Reform,” 6 L. & Phil. 53, 83–85 (1987); Fletcher, supra note 2, at 401, 417.
34 See Garvey, supra note 19, at 351–352.
the appropriate moment. An agent lacks the requisite control over her ability to form the correct belief at the appropriate time.
Finally, with regard to (3), there may be two reasons why an agent is not motivated to form a belief. First, an agent may simply not see any reason to gather additional information. Suppose someone more epistemically favored than the actor – an epistemically favored observer (EFO), who we shall assume is the equivalent of a RPAS – would esti- mate the risk that the actor is imposing on others to be higher than the risk the actor actually estimates. Th e EFO we shall assume would be culpable for acting as the actor acts. Th e problem for the actor, how- ever, is that for him to become an EFO would require the expenditure of resources – time, energy, and perhaps material resources, plus the cost of forgoing the act in question at the time in question – and given his appraisal of the situation, that expenditure of resources appears unrea- sonable. Aft er all, the actor does not know what risk the EFO estimates.
Th e risk of harm could be higher than the actor estimates, but it could as well be lower, or the same. (If the EFO is God, then the risk will be higher or lower; it will be one or zero.) Nothing that the actor is aware of indicates to the actor that he should spend the resources to become an EFO.35
Second, she may be a bad person. However, we fail to see how someone who has no internal reason to form a belief is culpable for not doing so when, for whatever characterological reason, she does not rec- ognize that such a belief should be formed. And if the negligence pro- ponent wants to hold us responsible for our characters, he will have to off er some evidence that we have control over them.
In this last respect, consider the recent attempt by Peter Westen to construct the RPAS.36 Westen argues that we can construct the RPAS with all of the actual defendant’s beliefs, desires, education, intelligence, and so forth, but the jury must then ask the question whether an actor who truly had suffi cient concern for others would have come to form
35 See John Conlisk, “Why Bounded Rationality,” 34 J. Econ. Literature 669 (1996); Philippe Mongin, “Does Optimization Imply Rationality,” 124 Synthese 73 (2000). See also Larry Alexander, “Foreword: Coleman and Corrective Justice,” 15 Harv. J. L. & Pub. Pol’y. 621 (1992).
36 Peter Westen, “Individualizing the Reasonable Person in Criminal Law,” 2 Crim. L. &
Phil. 137 (2008).
the belief that the defendant did not form. So, Westen argues, the proper question in Williams is whether otherwise identical parents with suffi - cient concern would have recognized the life-threatening nature of the risk that the defendants did not recognize.
Th is inquiry takes us out of the individuation frying pan and back to the character fi re. Here, we have the same concerns that we voiced with Tadros’s view – can we fairly hold people responsible for faulty characters of which they may or may not be aware, and which they may or may not be able to change? On the other hand, and consistent with our views about genetic recklessness, we are somewhat sympathetic to the claim that when a defendant becomes aware of a potentially dangerous trait, he may then have a duty to correct it. A man who discovers that he has a short temper that has caused him to lose control and hit his child does have a duty to take some affi rmative steps to control his temper. He will be culpable for failing to do so (irrespective of whether he again loses control of his tem- per) if the risk of harm to others outweighs the burden of trying to get his temper under control (discounted by the likelihood of success). 37
A culpability-based criminal law will not include liability for negli- gence. Culpability entails control, and the negligent actor does not have this requisite control. Any control a negligent actor has over her char- acter can only partially and indirectly aff ect whether the actor will fail to advert to a risk at the requisite time. Moreover, because none of us is perfect, even those with relatively “good” characters will constantly fail to advert to risks that they should have chosen to avoid had they adverted to them.
Moreover, even if we could control these failings, there is simply no nonarbitrary way to determine the standard against which we should be judged. Th e reasonable person is neither the actual actor nor the omni- scient god, but some construct that lies in between. Because there is no principled way to determine the composition of this construct, punish- ment for negligence is morally arbitrary.
37 We also wonder how far this duty extends. For instance, if one discovers that because of his racist beliefs, he is more likely to misinterpret an act by a black man as an attack, does he have a duty to attempt to rid himself of these beliefs? Or must a liberal society tolerate this sort of potentially dangerous character? For the claim that a liberal society must tol- erate the chosen retention of illiberal character traits, see Garvey, supra note 27.
86
I V
Defeaters of Culpability
As we have established, primarily in Chapter 2 and by exclusion of inadvertent negligence in Chapter 3, an actor’s culpability is the product of the risk(s) to others’ protected interest(s)1 that he believes he has unleashed2
1 As we point out in Chapters 2 and 8, sometimes the law will deem certain conduct to be an illegitimate risking of others’ interests even when not every token of that conduct creates signifi cant risks of that type. In such cases, the actor may believe that he is taking what would be an otherwise unjustifi able risk of engaging in that conduct – or may have the purpose to so engage in it – but also believe that he is taking only a negligible risk of the pri- mary interest(s) that the prohibition of that conduct is supposed to safeguard. He may, for example, be risking exceeding the speed limit or intending to exceed it in circumstances where he perceives little or no risk to others’ persons or property. How to assess his culpa- bility in such cases is a topic we take up in Chapter 8.
2 And as we elaborate in Chapter 5, because risk is an epistemic, not an ontic notion, and because the relevant risk for culpability is the risk the actor believes exists, there are no objective components to culpability. Whether we are interested in results – harms to oth- ers’ interests – or in result-threatening conduct (such as speeding), what we care about for culpability-desert-punishment are solely the risks that the actor perceived himself to be unleashing beyond his control through an act of his will. If he perceives himself to be fi ring
by a temporally fi xed3 act of his will and the reason(s) he had for so willing. We believe that the criminal law should intervene at the point at which the imbalance of risks and reasons represents a gross deviation from the standard of care that a law-abiding citizen would observe in the actor’s situation.
Th is formulation – encompassing both a comparison of risks to rea- sons and a comparison of the actor to the law-abiding person – contains within it much of the current doctrines of justifi cation and excuse.4 Th at is, whenever the actor’s reasons are suffi cient to justify the risk, the actor is justifi ed. Even if the actor’s actions were not justifi ed, if she has lived up to all that we can fairly expect of her, then she is excused. Despite the fact that much of what currently constitute defenses are subsumed within our culpability calculation in Chapter 2, we off er our thoughts here on those special reasons that justify risks oft en thought to be “justifi cations,” and we off er our view on how excusing conditions should be construed.
We begin with agent-neutral justifi cations. First, we discuss the lesser-evils defense, the paradigmatic consequentialist justifi cation. We also explore deontological side constraints on the consequentialist calculus.
Second, we turn to defense against culpable aggressors and address myriad issues that bear on when an individual may act defensively.
a bullet at someone, it does not matter that he is actually a brain-in-a-vat and that his will- ing to move his fi nger was unsuccessful, much less that his fi nger moved but the trigger was stuck, the trigger moved but the gun was unloaded, the gun was loaded but the bul- let missed, and so on. Similarly, if the actor believes he is depressing the accelerator in a manner that will result in the car’s speeding instantaneously and thus with no further action or omission on his part, it does not matter that the car is actually in neutral, or that the actor is unbeknownst to him in a car simulator rather than a car, or that his foot is paralyzed and will not depress in response to his will, or that he is a brain-in-a-vat. Indeed, in some cases it may not matter that the speed that he is trying to achieve is not prohibited at this location if he believes that it is and is unaware of the facts that make it safe to go that speed at that location.
3 Another topic that we address in Chapters 6 and 7 concerns risks that are inappropriately created at T1 but that the actor can reduce through subsequent action at T2. When the actor drives very fast, he may be aware that he is risking others’ persons and property unjustifi - ably as long as he continues to drive at that speed. But he may then slow down. Or when the actor lights a fuse to burn down another’s home, he may realize there is a small risk that even if he has a change of heart in the next moment, he will for some reason fail to snuff it out. So he has created an unjustifi able though small risk, though one that becomes larger as time goes by and his ability to eff ectuate any change of heart declines.
4 Cf. George P. Fletcher, Th e Grammar of the Criminal Law: American, Comparative, and International § 8.4.2 (2007) (Model Penal Code defi nitions of recklessness and negligence embody within them questions of justifi cation and excuse).
Next, we turn to excuses. We begin with duress and discuss how this “excuse may actually encompass a personal justifi cation. Th at is, an actor may have an agent-relative justifi cation – a moral permission – when she defends herself or her loved ones. Aft er discussing duress, we turn to instances of mistake. Although the determination of whether an actor is culpable is based upon her epistemic perspective – so that mis- takes regarding either harms or reasons are ordinarily exculpatory – we also consider whether the actor is entitled to an ignorance-of-the-law excuse when she is unaware of the existence of a proxy crime.
Finally, we address the sorts of rationality impairments that may exempt an individual from liability; excuse his conduct on a given occa- sion; or, at the very least, mitigate his culpability. Among the candidates for excuse are infancy, insanity, and altered states of consciousness.
Sometimes, of course, these facts defeat culpability without regard to demonstrating defective reasoning, as when they explain why the actor did not perceive substantial risks from his conduct or believed that cer- tain justifying facts were present. Because ignorance regarding risks and reasons straightforwardly negates culpability, and because infancy, insanity, and altered consciousness explain ignorance, there is nothing special about using the latter to negate culpability.
Sometimes, however, the actor is aware of the risks and does not per- ceive any facts that would justify imposing such risks but nonetheless is deemed not culpable for acting in the face of such risks and the absence of justifying reasons. Th e explanations for deeming such actors noncul- pable are varied and controversial. Lack of suffi cient maturity or sanity to grasp fully the import of otherwise culpable acts and lack of the abil- ity to access fully all of one’s reasons for not acting are oft -cited though not fully explained. Are the young, the insane, the hypnotized, the som- nambulant, or those acting in the grip of “automatism” diff erent in kind or only in degree from the impulsive, the habitual, the daydreamer, and so on? We cannot answer these questions fully here. What we can do is identify them and show where they fi t into our schema.