The Character Theory

The character theory of mens rea holds that a blameworthy state of mind is necessary for criminal liability because it allows us to infer a socially undesirable character trait from the wrongful actFootnote ⁹ or it permits an interpretation of the wrongful act as an expression of,Footnote ¹⁰ or as partly constitutive of,Footnote ¹¹ a disposition of character deserving blame. Unlawful acts by themselves permit no such inference, because from a single unlawful act one can tell nothing about the settled features of an agent's character unless he performed the act purposely, knowingly, with reckless indifference, or with thoughtless preoccupation with his own interests. So the presence of mens rea is essential to our being able to presume a connection (whether of evidence or causation) between an unlawful act and a blameworthy disposition or attitude. Obversely, exculpatory conditions such as involuntariness, mistake of fact, necessity, and duress either block a presumption that the defendant's act reflects a character flaw or (in the case of necessity and duress) displace that presumption with evidence of extraordinary pressures so as to make moral blame and censure undeserved.

Blameworthy character is not, of course, a sufficient condition of judicial punishment, since moral disapproval can be expressed in many ways. There must also be a harmful or dangerous act prohibited by law, because the end of judicial punishment, as distinct from that of informal types of social censure, is the prevention of crime. So the possibility of drawing an inference from act to character functions as a side constraint on the state's pursuit of the goals of deterrence, incapacitation, and reform, ensuring that these aims will be furthered only on the backs of people who deserve public condemnation.Footnote ¹²

Because it is far too difficult a question to tackle, let us grant the character theorist the assumption he needs that we are responsible for our characters and can thus be appropriately blamed for vicious traits, dispositions, habits, and attitudes. And let us also assume that the character theorist can demarcate with some precision the borders of the character for which we are responsible (does it exclude or include attributes formed through early childhood identification with parents; emotions, urges, etc. that we banish from our consciousness?), so that judgments of acts as being “in character” or “out of character” can be analytical and disciplined rather than subjective and result-driven. Even so, the character theory of mens rea is vulnerable to two lines of criticism: it fits too loosely the criminal law of which it is supposed to be an interpretation; and it provides an inadequate account of criminal desert.

Let us first consider the question of fit. Others have pointed out how precarious the act (or externalization) requirement of criminal liability becomes under a character theory of culpability, and so I do not belabor that point.Footnote ¹³ Instead I focus here on how loosely the character theory of mens rea and of exculpatory conditions fits the actual contours of the mens rea requirement in law.

Because it wants to link judicial punishment not to the culpable agency of an abstract person but to the blameworthiness of a concrete and full moral character, the character theory requires distinctions among mental attitudes much finer than the criminal law typically recognizes. Criminal law generally attaches no practical significance to the distinction between committing an unlawful act for the purpose of doing so, committing it knowingly but regretfully as a means to (or by-product of) achieving some further end, consciously imposing an unreasonable risk and being indifferent to whether the risk materializes, and consciously imposing an unreasonable risk but hoping it will not materialize. For purposes of punishability, these states of mind are treated as moral equivalents; when accompanying an unlawful act, any will suffice to move the wrongdoer across the threshold from civil to criminal liability. Yet they are far from equivalent from the standpoint of blameworthy character. Someone who, without justification or excuse, acts for the purpose of killing another human being has exhibited a more depraved character than someone who kills in anguish as a certain side effect of achieving a different purpose; someone who is indifferent to whether an unreasonable risk will materialize deserves more blame than someone who hopes it will not. A character theory must distinguish between these nuances of fault as a matter of law binding on the judge; it cannot leave them to sentencing discretion because, subtle as they are, the gradations are, for the character theory, intrinsic to penal justice, not a matter of indifference to it. Accordingly, mens rea in law is far too blunt an indicator of blameworthy character to be plausibly considered as ordered to that function.

To the same point is the observation that the character theory cannot account for the law's general practice of treating mens rea as separate from motive and motive as irrelevant to culpability. The reason someone commits an unlawful act is surely relevant to whether an inference can be drawn from the unlawful act to the actor's settled character as well as to what sort of inference can be drawn. Thus the person who, from benevolent motives, force-feeds his hunger-striking friend has revealed a character far less deserving of blame than has someone who, to amuse himself, forces food down the throat of a stranger sitting beside him at a restaurant. The criminal law treats both of these cases as simple assaults and leaves to sentencing discretion the moral differences between them. But the character theory is committed, if not to exculpating the benevolent force-feeder, at least to treating his crime as belonging to a different category (assault from benevolent motives, perhaps) from that of one who assaults from vicious motives.

If mens rea in law is too undifferentiated to be a reliable indicator of character, it is also in some cases too narrowly focused on agency to permit the blaming of the blameworthy. I have in mind here the defences of mental disorder and intoxication. Where mental disorder is not genetic or biologically caused, it comprises a set of character vices (megalomania, paranoia, narcissism, etc.) for which the agent is as much or as little responsible as he is for any other formation of character. Thus, far from a condition blocking an inference from unlawful act to character, mental disorder is itself the connection. The same may be said for voluntary drunkenness. Drinking to the point where one loses control of motor functions or even where one lacks the capacity to intend objects as means to further ends is, one may say, the sign of a man who disrespects his humanity, who is willing to “throw himself away,” in Kant's phrase. From the standpoint of the character theory, therefore, the law's practice of permitting mental disorder to exculpate when it negates criminal intent and of allowing intoxication to exculpate from crimes of complex intent must seem exactly the opposite of the way these conditions ought to be regarded. Accordingly, the character theory would seem to be committed to eliminating or drastically reducing the scope of the defence of mental disorder and to admitting all evidence of intoxication as inculpatory.Footnote ¹⁴

Not only does the character theory turn some legally exculpatory conditions into inculpatory ones; it also renders unintelligible the limitation of exculpatory conditions to a finite set of defences signifying that the defendant's act does not deserve punishment. If the only significance of the legally recognized excuses is that they preclude an inference from conduct to character, why not consider character evidence extrinsic to the act that might do so as well or that might at least reveal the act in a truer light? Indeed, since it is manifestly unfair to judge a person's settled character by an isolated act, the defendant ought, on the character theory, to be permitted to adduce character evidence not simply to support a denial of commission or in discretionary mitigation of sentence but as exculpatory in itself. Thus, if it could be shown that the criminal act was a momentary aberration from the defendant's enduring moral character, or if it could be shown that the act, while not disownable by the defendant (because no acts are), nonetheless belongs to a generally good man with all-too-human frailties rather than a bad man, he should be entitled to an acquittal or statutory mitigation. Here the fit between the character theory of mens rea and the criminal law could not be much looser. Not only could the character theory convict before a blameworthy (and dangerous) disposition issued in an act; it could also acquit on the basis of a residual character defence someone who intentionally and without justification or excuse committed an unlawful act.Footnote ¹⁵

The character theory is also intrinsically deficient as a theory of criminal desert. On this theory, mens rea functions as a connector between a dangerous act and a character deserving of blame and moral censure. But if the blameworthy disposition revealed by an unlawful act committed with mens rea does not specifically deserve judicial punishment as distinct from blame generally, how can we say that a mens rea requirement acts as a desert constraint on the state's pursuit of crime prevention through punishment? Perhaps a character vice deserves blame in the abstract, and judicial punishment is one way of expressing blame (i.e., of censuring), so that by deserving blame, a character also deserves any of the specific manifestations it might take. But judicial punishment is not just a particular instance of censuring alongside professional discipline, verbal reprimands, and social snubbing. Because it comprises actions that are ordinarily themselves wrongful (indeed criminally wrongful), judicial punishment forms an altogether different category of responses to wrongdoing needing a distinctive justification—one directed toward legitimating the violence it entails. Thus something more than deservingness of blame and censure is required to deserve it.

The same conclusion follows from attending to the difference between political and moral obligation. What is generally taken to distinguish them is precisely the authority of an official to whom a political obligation is owed to coerce obedience to norms that are otherwise binding only in conscience so that he need not rely on angry remonstrations alone. The moral obligation to obey the norm is not yet a political obligation not to resist the might of the sovereign enforcer of the norm. Rather, the political obligation requires a distinctive justification by appeal to some common interest apart from the good of virtue, whether in self-preservation (Hobbes, Locke) or in civil independence (Rousseau, Kant). But if political obligation is categorically different from moral obligation, then the authority to punish that is correlative to political obligation is categorically different from the moral standing to blame, not simply a case of it. So to explain the mens rea requirement as being needed to ensure a character's deservingness of blame, disapproval, or censure is still to leave judicial punishment undeserved. But then the pursuit of crime prevention through that means looks like arbitrary violence indistinguishable from the wrongdoer's.

The Choice Theory

The choice theory of the culpable-mind requirement is, like the character theory, an account of moral blameworthiness applied to legal punishment. The theory is quite simple and intuitively appealing. It says that, assuming punishment for a type of act is justified for forward-looking reasons, punishment of this person for an act of that type is justified only if blame is justified; and blame is justified for an unlawful act only if the agent had a choice between committing it and not committing it—only if he could have chosen otherwise than he did. If the agent had no choice but to do what he did, he is not to blame for it. The culpable-mind requirement of criminal liability embodies this truth of moral theory, for it confines punishment to those who could have done otherwise. In H.L.A Hart's words:

Here Hart has run together two different pleas that might be meant by “he could not have helped it.” That statement might mean that he lacked the capacity to choose otherwise, perhaps because he was a newborn infant. Or it might mean that he lacked a fair opportunity of exercising his capacity to choose otherwise, perhaps because he was unaware of the circumstances that made his conduct unlawful or because his will was overwhelmed by pressures much stronger than most people ordinarily face. Michael Moore, another proponent of the choice theory, distinguishes these senses by describing the first as a claim that one's choice was made impossible by factors beyond one's control and the second as a claim that one's choice was made very difficult by such factors.Footnote ¹⁷ Let us first consider incapacity to choose otherwise as a condition negating mens rea.

There is, of course, an ambiguity here as well. To say someone lacked the capacity to do otherwise than he did might mean that at the time he broke the law he lacked the capacity agents have for free choice. This might be because he was asleep or unconscious, or because a gust of wind blew him into another person. Alternatively, to have lacked the capacity to do otherwise might mean that while the defendant was a conscious agent at the time of the breach, he lacked the strength of will to resist the powerful inclination to act as he did. Just as someone might say “I couldn't lift the two-hundred-pound weights,” someone might say “I couldn't resist the impulse to steal the watch.” This second sense of “I could not do otherwise” cannot be exculpatory, however, because, so understood, the statement adds nothing to “I did not do otherwise.” Since, counterfactually, I could have had the strength of will to resist the impulse (that is, resistance did not call for superhuman fortitude), the statement that I did not have the strength to resist is tautological for “I failed to resist.” So it has no force to excuse the failure.

Let us then explore the first sense of incapacity. And let us grant the choice theorist's assumption against determinists that incapacity in this sense intelligibly distinguishes the blameless from the possibly blameworthy. Assume, in other words, that a conscious, normally functioning human being has a capacity to accept or reject any impulse or motive that its body-psyche offers up as an end of action. By “capacity” is meant here not strength of will but the bare possibility of asserting a will that distinguishes a conscious human being from other forms of life. A sunflower cannot reject the force of phototropism that causes it to turn toward the sun; a bloodhound cannot reject, except for the sake of a stronger impulse, the force of an instinct that causes it to pursue a scent. But a human being can reject, just for the sake of displaying self-mastery, the force of an emotion, drive, or inclination. It can do so because it is a self-consciousness—that is, an I distanced from these emotions and drives and so capable of taking a stance toward them. This, no doubt, is a minimalist sense of capacity. Here capacity simply means that there is a rider on the horse, to use Plato's image, not that the rider possesses the strength or know-how to control the horse. So we can say that a capacity to have done otherwise in this sense belongs to any human being who is aware of itself as an I or subject. And since this self-awareness is the possibility of choice undetermined by what is not-I, it is the condition of imputing choices to a subject and hence, as the choice theorist rightly maintains, the condition of blaming a subject for his unlawful deeds.

The difficulty for the choice theorist of the culpable mind, however, is that the presence of the capacity to have chosen otherwise says almost nothing about whether the defendant's mind was culpable. That D could have chosen otherwise when he shot V, while a minimum condition for culpability, fails by a wide margin to establish a culpable mind, for in a dark forest D may have reasonably believed V was a bear. Thus the culpable mind is determined not solely by whether the defendant had a choice to do otherwise but also (and indeed, primarily, as we shall see) by what he chose. This is why, in law, the mind's culpability is a function of its cognition rather than simply of its volitional capacity, of what the defendant knew or believed and thus thought he was doing. Thus, if he chose to shoot a bear but mistakenly shot a man he thought was a bear, he is not guilty of a homicide even though he could have chosen to flee. Even if, because of mental disorder, he delusionally mistook the man for a mouse, he is innocent of criminal wrongdoing even though he could have chosen not to kill the “mouse.”

Conversely, if, because of some condition negating self-consciousness (for example, unconsciousness) or determining physical motion irrespective of volition (for example, reflex action, forces of nature), the defendant could not have done otherwise, something more basic than a culpable mind is missing. In these situations, the law says that there is no act, because voluntariness is what distinguishes an act from an event. Accordingly, on its incapacity branch, the choice theory explains no more than the voluntariness requirement for a culpable act; it does not explain the requirement of a culpable mind—the requirement that the defendant have chosen something in particular. Hence it fails to explain the defences of mistake of fact and insanity, by virtue of which lawbreakers are acquitted despite having had the capacity that conscious subjects have to choose otherwise.

Perhaps the choice theory fares better when it argues that blameworthiness (and so punishability) depends on the defendant's having had a fair opportunity to exercise his capacity for choice. This path seems more promising, because the defendant who shot the man believing it to be a bear, though he had the capacity to choose otherwise, lacked the opportunity most people would have had for exercising this capacity, either because the mistake was reasonable (the forest was dark) or because the defendant was suffering from a mental disorder that distorted his perception of the world. Moreover, the defences of necessity and duress can also be brought within the ambit of this theory, because they are put forward by people who, though they possessed the abstract capacity to have chosen otherwise, found their will overborne by pressures far above the ordinary. In these cases, the lack of strength to resist a drive or to overcome fear does have, potentially at least, exculpatory force, because the plea does more than redescribe the failure to resist; it judges the failure as meeting the standard of moral strength that we are entitled to expect from our fellow human beings. The accused lacked a fair opportunity to exercise his capacity to have done otherwise.

Nevertheless, the fair-opportunity branch of the choice theory also fails as a theory of the culpable mind. This is so because, if the capacity test establishes too little of what makes for a culpable mind, the fair-opportunity test embraces too much, sweeping in states of mind that not even the most expansive view of mens rea found in the positive law views as culpable. So far, the outer limit to which courts have stretched the fault threshold of criminality is “gross negligence,” understood as a failure to avoid a very high risk of serious harm where the defendant had the capacity to do so but regardless of whether the defendant actually perceived the risk.Footnote ¹⁸ Any further expansion would breach the border between crime and regulatory offences based on civil negligence and thus brand as criminals those whose fault was venial. Yet the fair-opportunity test of culpability would breach the border.

To see this, consider a classic civil law case of negligence, Donoghue v. Stevenson.Footnote ¹⁹ As a matter of tort law, of course, it is immaterial whether the defendant had the mental or physical capacity to conform to the standard of care that the reasonable person would have exercised in the situation; all that matters is whether or not he measured up to the standard. Thus civil negligence does not require that the defendant have been morally blameworthy for his having breached a duty of care. Suppose, however, that Stevenson did have the basic capacities needed to meet the required standard of consideration for those likely to drink his ginger beer—to contemplate the risk that some of his bottles would become adulterated and to inspect them prior to sealing them. In that case we would say that he breached a duty of care owed the plaintiff and injured her despite having had a fair opportunity to have done otherwise—to have taken the precaution that would have avoided injury to the plaintiff. This, according to Hart's fair-opportunity test, makes Stevenson blameworthy for his negligence, and so it does. Does it also make him culpable and criminally liable to punishment (assuming there are good goal-oriented reasons to punish)? Not according to the most expansive view of mens rea found in the positive law, because his having had the capacity to foresee the risk and avoid it has not changed the magnitude of the risk, which (we may assume) was not sufficiently excessive to warrant a finding of egregious negligence. Nor does his capacity indicate anything more than venial thoughtlessness. He is not indifferent to the risk, for had it been brought to his attention, he would no doubt have taken the appropriate action to remove it. Nor was the risk of adulteration so obvious that only someone indifferent to the risk could have failed to notice it. If this is nonculpable negligence when he lacks capacity, it is still nonculpable negligence when he has it.

So here we have a case where, though the defendant is morally to blame for his negligence, it would still be unjust to punish him (though not unjust to impose noncriminal penalties that need not be deserved by either culpability or blameworthiness). But this tells us only what we already knew: that the blameworthiness that gives the wronged party the standing to blame is not yet the culpability that gives the state the right to punish.

If the presence of a fair opportunity to exercise one's capacity for choice against wrongdoing does not map onto legal culpability for wrongdoing, we should suspect that the lack of a fair opportunity to avoid wrongdoing is not the legally relevant common theme of the conditions that exculpate. Of course, it is natural to think that even if those morally blameworthy for their unlawful acts are not necessarily criminally culpable for them, at least the morally blameless are necessarily nonculpable. But this is not as obvious as it seems. Anglo-American criminal law distinguishes between exculpating conditions that negate an essential element of culpable wrong (culpable act or culpable mind) or that justify what in ordinary circumstances would be a culpable wrong and excusing conditions that presuppose a culpable wrong. Duress is (in some cases) an excusing condition. Now, let us grant that T's believable threat to kill D if D does not kill V excuses D from murdering V, because D lacked a fair opportunity to exercise his capacity for choice and is therefore morally blameless (his doing otherwise would have been morally heroic, supererogatory). On the traditional common-law understanding, D's blamelessness does not exculpate him from the wrong; rather, it (at most) gives a reason for excusing D from a culpable wrong, for holding him safe from punishment despite his legal culpability. The fair-opportunity theory, in conflating moral blamelessness and legal innocence, fails to capture this nuance of the law. Moreover, if the morally blameless are not necessarily nonculpable legally, then the lack of a fair opportunity to do otherwise cannot explain why mistake of fact and insanity exculpate. But this is what we suspected all along. For even prior to any theoretical reflection, it seems intuitively plain that the reason someone who kills a man thinking he is a bear is not culpable for his wrong (liable to judicial punishment) is not that he lacked a fair opportunity to exercise differently his capacity for choice, but that his actual choice (to shoot a bear) is nonculpable.

But (it will be protested) have we not, with this separation between exculpation and excuse, driven a wedge between the defences to a criminal accusation, thereby precluding a coherent theory of defence? Whatever their failings with respect to fit with the positive law, after all, both the choice theory and the character theory have the advantage of providing a unified theory of defences. And is it not the chief virtue of a theory that it unifies into a coherent explanation the phenomena of which it purports to be an account?

Wedge yes, incoherence no. There is no reason to consider it an advantage of a theory that it explains with one idea defences as disparate as “I thought he was a bear” and “I knew he was a man, but it was either him or me.” We should rather think that such a theory oversimplifies. But neither is there any reason to think that preserving the common law's distinction between exculpation and excuse is incompatible with a coherent theory of defences; for an account of defences can be unified without being unitary. That is, it may be possible to unify the defences not through a single idea that effaces the difference between exculpation and excuse but through a parent idea that manifests itself differently in each, that shows their kinship and compatibility, and that thus shows how they fit together, different as they are, within a coherent account of the defences. But this, too, is the subject of a different essay.

The Fault Threshold

The classical retributivist legitimation of punishment runs basically as follows. If D knowingly interferes with V's capacity to act from self-chosen ends (whether by threatening one of the essential conditions of V's freedom, by physical force, or by interfering in the proprietary domain over which V's choices rule), his act signifies something more general than an infringement of V's right so to act. Since V is a human agent with a capacity for self-actuation, D's choice to interfere with this capacity signifies for the thinking Agent standing in D's shoes a claim of permission to act that is unlimited by respect for the agency of another human being. However, claims of permission binding others can be objectively validated as rights only by free others whose liberty to act is reciprocally respected. Therefore, D's choice to interfere with another agent's freedom is impliedly a choice to act against the framework of mutual respect that makes rights (as distinct from subjective claims) to act possible. Accordingly, D's choice to interfere with V's agency signifies (to the thinking Agent) a challenge to the very idea of a valid permission to act or of a permission to act that others must respect. However, if (according to the challenge) no one has a valid permission to act, then no one can validly protest (i.e., make a protest that others must heed) if another agent impedes his movements through physical restraint or confinement—no one, including D. Thus the necessary implication of D's choice to interfere with V's agency is the denial of an inviolable sphere of liberty for D. Since D implicitly denied the existence of rights (claims recognizable by a free self) to act, he cannot now turn around and assert his right against interference with his own agency. This constitutes the aporia or self-contradictory result of the criminal's implicit claim of permission to an unlimited liberty. Because this claim was mistaken, it would be impermissible for V or anyone acting from self-interest to take advantage of the license the criminal has granted to coerce him. However, a state agent could do so for the public purpose of demonstrating the aporia so as to vindicate mutual respect as the sole ground of solid rights. Judicial punishment is justified by this purpose. By availing itself of the criminal's license to coerce him, the state shows the downfall of his claim of right to an unlimited liberty and it thereby publicly vindicates mutually recognized liberty as the framework for real rights. Since doing so also vindicates the criminal's right, the thinking Agent in his shoes would accept retribution as the actualization of valid human worth. Accordingly, the necessary implication of D's willful interference with V is his own moral vulnerability to coercion for the purpose of right-vindication, and this connection between willful wrongdoing and liability to punishment is what we call criminal desert.Footnote ²¹

We can see that what launches this story is an intentional interference with another human being's capacity for free agency. Choosing to do that from which a denial of rights of agency can be imputed to the choice triggers the circuit from wrongdoing to coercibility. Why one chose so as to imply a right-denial—one's motive, whether wicked or benevolent—is irrelevant to whether the circuit goes through, as it is to culpability in the positive law. Moreover, only a choice implying a right-denial initiates the circuit. An unwitting interference does not set the logic in motion, for if D mistakenly thinks he has permission from V, then his choice is just to pursue his object (surgery, sex, acquisition, etc.) within the bounds of respect for V's agency; and from this act description no denial of rights can be imputed to his choice, hence no authorization of his own coercion either. Coercion is then external violence incompatible with the inviolability of the person.

Observe, however, that nothing in this argument determines that an intentional interference is required for a right-infringement or a boundary crossing (i.e., a tort). Where D interferes with V's agency under a mistaken belief that he has permission from V, a boundary has nonetheless been crossed, and D will not be able to raise his mistake to defend against a civil action for battery, entry, or conversion.Footnote ²² This is so because D's subjective perceptions and beliefs cannot, consistently with mutual respect for agency, be allowed to determine unilaterally the boundary between D's and V's zones of permissible activity. That boundary must be determined impartially—in accordance with a principle securing the greatest extent of equal liberty.

In the criminal action, however, the question is whether D can be physically restrained and confined for his boundary crossing, and now his subjective perceptions become relevant as determining whether he has made a choice from which an assent to coercion may be implied. If, given his perceptions and beliefs about the circumstances or likely consequences of his action, it would be incorrect to describe his choice as one to interfere with V's free agency, then nothing connects his choice to his own coercibility. Thus even to interfere unwittingly where an ordinarily circumspect person would know of the interference (i.e., to make a negligent mistake about permission) is not yet to interfere culpably either in theory or in law.Footnote ²³ Nor is an unwitting interference culpable where the interference would be glaringly obvious to the ordinarily circumspect person, though the obviousness of the interference would be strong evidence against a plea of inadvertence. Though the ordinarily cautious person would have known of the interference, D did not, and so he made no choice from which a denial of his rights can be logically inferred.

Whereas, however, a choice to do that from which a denial of rights can be imputed to the choice is essential to culpability, knowledge of the interference is not the only way to make such a choice. A denial of rights to respect for agency can also be inferred from a choice to impose an extraordinary risk of interference—that is, a risk over and above the magnitude that agents reciprocally accept as an inevitable cost of freedom of action in a certain social context. I choose to impose an extraordinary risk, however, only when I am aware of such a risk or only when I suspect that one exists and so actually foresee the likelihood of an interference. If I am unaware of the risk, then no matter how apparent it would be to a reasonable person, I have not chosen to impose it, and so I have chosen nothing implying a denial of rights (though my bare assertion of ignorance need not, of course, be taken at face value). Thus recklessness, understood as proceeding despite advertence to one's imposing an unreasonable risk of interference with others' capacity for freedom, and a willful blindness regarding such a risk are culpable states of mind both in theory and in law; but gross negligence is an outlier for the theory and (we may say) a mistake in law. Moreover, a denial of rights can be imputed to the choice to impose an unreasonable risk whether one would regret the risk's materialization or be indifferent to it or rejoice in it. Thus mental orientations toward acts and outcomes one would want to distinguish in assessing and calibrating moral blame are juridically equivalent as denials of rights authorizing one's own coercion, and they are so treated by the positive law.

Observe that while this account of the culpable mind provides an argument for subjective mens rea as the fault threshold of criminal liability, it is not crudely subjectivist in the sense decried by many advocates of an objective standard of criminal fault.Footnote ²⁴ That is, it is not concerned solely with what is empirically going on in the mind of a monadic agent, and so it does not judge him exclusively by his own knowledge, beliefs, and self-interpretations, ignoring the interpersonal meaning of his act. True, the foregoing account of mens rea focuses on what situational facts the agent knew or believed existed or on what risks he foresaw in order to determine what he chose to do. But it does not judge the agent solely by what he subjectively thought his choice meant; rather, it judges him by what is logically imputable to his choice and thus by what the thinking Agent in his shoes meant by his choice. If D knowingly operates on V against V's will, thinking that this shows love and concern for V rather than disrespect, the foregoing account will still treat him as having implicitly denied rights of agency and authorized the interference with his own; so, too, if D reports that he thought he was only ignoring V's will, not denying rights of agency as such, or that he knew V was a human being but did not know he was an agent or even what an agent is. Accordingly, the foregoing account of criminal desert generates a subjectivism that, by virtue of incorporating the public standpoint of a thinking Agent, is immune from the charges of excessive atomism or empiricism brought against other versions.Footnote ²⁵

Observe, further, that the foregoing account of the culpable mind determines that a form of subjective fault is the threshold of criminal liability to judicial coercion in the abstract. It says that a choice from which a right-denial may be implied is essential to deserving a right-infringement. This means that once the agent has made a choice implicitly authorizing a right-infringement in the abstract, the account's subjectivist requirement is spent. There is no further need that the criminal have intended or foreseen the harmful result of his transgression in order that the severity of punishment properly reflect that result. Of course, the self-imposition requirement at the threshold rules out punishing the criminal for all consequences of his wrong, no matter how unforeseeable; for it would be self-contradictory to say that punishment per se must be implicitly self-imposed but that the measure of punishment can be abandoned entirely to chance. However, the self-imposition requirement does not entail that the criminal can be punished only for those results he intended or actually foresaw; for in choosing to coerce someone in a world of cause-effect chains he knows might extend beyond his purpose, the thinking Agent in the criminal's shoes accepts responsibility (in the appropriate degree) for those outcomes of his right-denying act that are imputable in some degree to his agency; and intention is but the highest of several degrees of imputability (foresight and foreseeability are the others). Hence the relaxed fault requirement applying to results is quite consistent with the self-imposition constraint determining the stringency of the threshold. Indeed, one can go further: a requirement of intention or foresight for results would be inconsistent with the respect for free agency determining the subjectivist threshold, for it would imply that human beings have no capacity they may be expected to exercise to think beyond the immediate effect of their acts.

It follows that nothing in the foregoing account of subjective fault as the threshold of criminal culpability determines that subjective fault is also required for criminal responsibility for harmful results once the threshold for liability to any punishment has been crossed. Because the classical retributivist account of the culpable mind determines only that subjective fault for wrong simpliciter abstracted from harm is required for punishability simpliciter abstracted from quantum, that account stands for what may be called a moderate subjectivism. Such a position is invulnerable against charges of excessive revisionism that have been brought against extreme versions of subjectivism that extend to results.

Finally, the classical retributivist account of mens rea reveals subjective fault as belonging to the sphere of crimes involving interferences with rights of agency, for it is the denial of these rights implied in intentional interferences that connects wrongdoing with the coercibility of the wrongdoer. Because the account has no application to breaches of statutes aiming at the public welfare, it leaves the fault requirement of these offences to be determined by a generic autonomy constraint on the state's pursuit of welfare, which constraint is satisfied by a due diligence defence coupled with a rule against imprisonment for negligence.Footnote ²⁶ On the other hand, the retributivist justification of physical restraint applies by extension to intentional breaches of laws (against perjury, bribing officials, etc.) supporting the enforcement of rights and even to intentional breaches of public-welfare laws. This is so because the reality of rights may be said to depend on the rule of law. A knowing breach of a statute involves a choice to which a denial of law's authority may be imputed and hence also a rejection of the condition for the possibility of effective rights. Accordingly, for every public-welfare offence involving a risk of harm to others or to a public good (a qualification required by proportionality, for otherwise an intentional breach of a no-jaywalking ordinance could incur a prison sentence), it would be permissible for a liberal state to enact a separate offence for committing the offence knowingly (to which ignorance of law would thus be a defence) punishable by imprisonment.

Exculpatory Conditions

The retributivist account of the culpable mind is also an account of the conditions that exculpate because and to the extent that their presence means that the defendant did not make a culpable choice: mistake of fact, color of right, and legal insanity. These conditions exculpate, I argue, to the extent that they preclude an implication of a right denial from the defendant's choice.

Of course, if, because his bodily movements were unwilled, the defendant made no choice at all, then he could not have made a choice to which a right denial is imputable. Thus involuntariness exculpates, not because the defendant could not have done otherwise than he did (a condition relevant only to moral blame), but because it negates a basic precondition—choice—for a right-denying choice. This explanation accords better with the positive law's own understanding of involuntariness than does the choice theory's account; for instead of developing involuntariness into a general theory of exculpation, it captures the law's sense that involuntariness negates only a basic precondition of culpability.

Mistake of fact exculpates, not because the defendant lacked a fair opportunity to exercise his capacity for choice, nor because it blocks an inference from act to character, but because (and only to the extent that) on the facts as perceived by the defendant, he made no choice that for the thinking Agent in his shoes would have been a choice to deny rights of free agency. Thus if D chose to shoot a bear and instead shot a man thinking it was a bear, he chose nonculpably unless he chose to risk shooting a man. But if he chose to shoot Peter and instead shot Paul thinking he was Peter, he in any case chose to kill a human being, which for the thinking Agent in his shoes is a denial of rights, and so his mistake is nonexculpatory even without the fiction of a “transferred intent.” If D shoots V under a mistaken belief that V is about to shoot him, then, however unreasonable the mistake, D does not make a choice from which a right-denial may be inferred; for, on the contrary, the thinking Agent with D's perceptions and beliefs is asserting his right against a transgressor, and so D's choice, while not actually justified, is nevertheless nonculpable (though D may be subject to noncriminal penalties for a negligent homicide).Footnote ²⁷ But if D shoots V, who is not threatening him or anyone else but who D mistakenly believes has forfeited her right because her evil ways have brought dishonor to the family, he has made a choice from which a right-denial may be inferred; for the thinking Agent in his shoes knows that V cannot forfeit her right by giving offence to a particular set of moral beliefs (this being contradictory to the idea of having a right) and that D has chosen to act as if V's right were held at his pleasure.

In the latter case, no doubt, D has made a mistake of law rather than of fact. However, the reason his mistake fails to exculpate is not that it belongs to one category of mistake rather than another but that, even taking into account D's beliefs, he chose to do that which for the thinking Agent is a choice to deny rights. Indeed, as it applies to right-infringements, the rule that mistake or ignorance of law does not exculpate simply reflects the idea that the thinking Agent knows the self-contradictoriness of right-claims to act unlimited by respect for others' agency and that the knowledge of the thinking Agent is imputable to the defendant. That is why Bailey was culpable for shooting Truscott on the high seas even if he could not be aware of a new statute extending the king's writ to British ships in mid-ocean.Footnote ²⁸

However, this means that if, in the particular situation, the thinking Agent would not necessarily know the illegality of D's deed, then D's mistake of law can exculpate if it means that he chose to act within the bounds of his permission and so respectfully of V. For example, if D takes V's property under a mistake of positive law that the thing is his own, he has chosen to act within his permission (under “color of right”), and so his mistake of law exculpates both in theory and in law. But then, similarly, if D makes sexual contact with V mistakenly taking an equivocal communication for consent, he chooses nonculpably (assuming no recklessness or willful blindness) even if, for reasons of social policy, consent is (unknown to D) defined by statute as an unequivocal communication of “yes” rather than a privately held attitude toward the contact.Footnote ²⁹ This is the case, too, if fifteen-year-old D has sexual contact with a thirteen-year-old willing partner believing the statutory age of consent to be thirteen when it is in fact fourteen. What matters is whether, given his beliefs, D made a choice from which a right-denial may be implied, not whether his mistake is one of fact or law; hence it should make no difference whether D mistakenly thought his partner was fourteen or mistakenly thought the legal age of consent was thirteen, for in either case he chose lawful sex. While D may be subjected to noncriminal penalties for a negligent mistake, he is not liable to punishment as a criminal, for he has made no choice from which a right-denial may be inferred.

The defence of insanity as circumscribed by the rules in the M'Naghten caseFootnote ³⁰ perfectly reflects the classical retributivist view of what constitutes culpability and exculpation. First, insanity is not treated as necessarily depriving human beings of the capacity for free choice, because this capacity—for nondetermination by causal laws—is too thin to be affected by cognitive and emotional disorders short of advanced dementia and automatism. So the mentally disordered are not regarded as infants to whom judgments of guilt or innocence, responsibility or nonresponsibility, do not apply; they are not, for example, exempted from the criminal process upon a psychiatric diagnosis of clinical insanity, as infants are exempted just because of their age. Nor are they (except in some American jurisdictions) exculpated for reasons that would not be exculpatory for other conscious agents with a capacity for choice, for example, because of “irresistible impulse.” The capacity for free choice is just the capacity to resist.Footnote ³¹ Rather, they are treated as agents who can be culpable for their choices and who are exculpated only if their disorder engenders cognitive mistakes that would be exculpatory absent their disorder. So if D kills V delusionally thinking V is a snake, he has made no choice from which a right-denial can be implied, and so the mistake is exculpatory; so, too, if D kills V thinking that he (D) is a public executioner carrying out the verdict of a court. But if D kills V under the delusion that he is guaranteed protection against penal consequences by a cosmic shield or that he is a god to whom the law does not apply, his choice is culpable regardless of his delusion, for a right-denial is imputable to it.

Just as mistakes of law may exculpate when, exceptionally, the thinking Agent would not necessarily know the law (for example, pertaining to property) determining when a boundary has been crossed, so may they when the thinking Agent with the defendant's beliefs would consider that a right-infringement is tragically justified, as in the case of a public necessity. So if the defendant delusionally believes in the existence of exigent circumstances that the thinking Agent would consider justify a wrong, he has not chosen culpably even though he has mistaken a wrong action for a right one—that is, has made a mistake of law. And this is so both in theory and in law; for the M'Naghten rules exculpate not only those unable because of mental disease to understand the nature and effect of their acts but also those incapable of knowing that what they are doing is something they ought not to do.

Some regard the latter reason for not punishing the insane as relevant to something more basic than the existence of a culpable choice, since no other adults are exculpated because they could not know right from wrong; knowledge of right and wrong is rather imputed to them. An incapacity to know right from wrong seems to make a culpable choice impossible rather than simply absent, as it does for the infant. That is, it negates a precondition of imputing culpable choices to the defendant, for an agent cannot impliedly deny the validity of rights if he does not impliedly know the purported basis of rights in human agency. Now, it may be that some mental disorders (dementia, for example) infantilize the sufferer to the point that he is no longer aware of his agency. Such a person will also be unable to understand the nature and point of the trial process and so will be unable to give a hypothetical legitimating assent to any judgment rendered against him. For this reason he will be exempted from trial. But it cannot be said a priori of all mentally disordered individuals that their disease is incapacitating to the extreme of robbing them of their humanity. Moreover, if this were the type of defendant envisaged by the insanity defence, would it make sense for a judge or jury to decide the question of capacity to know right from wrong after a trial instead of a team of psychiatrists after a series of examinations? The fact that the clinically insane defendant is tried under specific criteria for legal insanity reflects the fact that, unlike the infant, his capacity for guilt or innocence is assumed, and the inquiry is into whether his choice was actually culpable or not. Sometimes his choice will be nonculpable because he labored under a delusion that he was performing an act quite different from the one he was actually performing. But sometimes it will be nonculpable because, looking at the world from the defendant's deluded perspective, the thinking Agent would conclude that in this case a wrong is justified by public necessity and so would choose a justified act. Thus no denial of right can be imputed to the defendant's choice.

To be sure, the defendant, in erroneously believing his act was justified in the real world, has made a mistake of law; but we have seen (in the “color of right” doctrine) that if a mistake of law blocks the inference from the defendant's choice to a denial of rights, it is no less exculpatory than a mistake of fact. Accordingly, that an insane defendant may be exculpated for an incapacity to know that his act was prohibited is not a reason to think that the defence of insanity belongs with infancy rather than with mistake.

The difficult case, however, is one where the deluded defendant thinks himself justified in a wrong by a standard transcendent of the legal order. Of this kind is the case of Hadfield, who shot at the king in order to fulfill a divine commandment to sacrifice himself to avert the world's destruction.Footnote ³² This type of case presents difficulties because, though the defendant was hallucinating when he committed a crime, the thinking Agent with his perceptions chose to subordinate the public authority of law to a private belief in a higher law, which is to deny the reality of rights. Indeed, since crime on the classical retributivist theory is just the claim to a permission to act unlimited by respect for others' agency—that is, a claim of right to do wrong, the religiously inspired accused's plea in defence looks more like an admission of culpability. Nevertheless, there may be a way of distinguishing the psychotic who knowingly commits a wrong in obedience to a divine voice he “hears” from the fanatic who subordinates public rights of agency to a private religious inspiration. If the defendant's disorder is of a kind (a matter of expert evidence) that deprives him of the ability to distinguish between a private inspiration or dictate of conscience and a public command from the sovereign of sovereigns, then he is in the same position as the man who kills convinced that he is the king's public executioner. His choice is then nonculpable, for no claim of permission to a liberty unfettered by law is imputable to it.

Subjectivism and Mistake of Law

One persistent objection against subjectivism was raised by Oliver Wendell Holmes and George Fletcher and has lately been reformulated by Arthur Ripstein.Footnote ³³ It is that a consistent subjectivism is committed to abolishing the well-settled distinction between mistakes of fact and mistakes of law and to treating even negligent mistakes of law as exculpatory. This is so, argue the critics, because the subjectivist requires for criminal liability a conscious choice to do wrong or to break the law, so that if the defendant is, even through his own moral blindness or negligence, unaware of the legal prohibition, he must be acquitted. But this position obviously undermines the authority of the legal standard and runs contrary to the positive law of all Western legal systems. Hence subjectivism must be discarded as offensive to “basic sensibilities of justice.”Footnote ³⁴

The classical subjectivist position I delineated above is immune to this criticism. For while it softens the dichotomy between mistakes of fact and mistakes of law, it does so only to the extent that the positive law itself acknowledges, in the doctrine of color of right, an exception to the rule that a mistake of law does not exculpate. Moreover, far from abolishing the rule, classical subjectivism provides theoretical support for it insofar as the rule applies to the sphere of crimes and leaves it standing in the sphere of public-welfare offences. In the sphere of crimes, the subjectivism I am proposing does not require for criminal liability that the empirical agent consciously choose to do wrong; to the contrary, it treats the thinking Agent's knowledge of the rightful limit of action in the free will and ownership of others as imputable to him. What is required, rather, is that the empirical agent choose to do that from which a self-contradictory right-claim to an unlimited liberty is logically imputable to his choice. This is enough to break the objectivist's lances; for on the subjectivist position offered here, the defendant's ignorance of his wrong will not exculpate him if the thinking Agent in his shoes would be choosing a wrong; and it will positively inculpate him if his ignorance is shown in an explicit challenge to the normativity of rights such as might be expressed in the claim, “I am above the law” or “right is the interest of the stronger.”

It is true, however, that where knowledge of the law cannot be imputed to the empirical agent as a matter of reason, a mistake of law can exculpate if it precludes an imputation of a right denial from his actual choice. So if he takes or damages another's property believing it to be his own, he is exculpated from theft or mischief; or if he defends with reasonable force what he mistakenly believes is his property against someone legally entitled to possess it, he is exculpated. But these exceptions to the rule that mistakes of law do not exculpate are those recognized by the positive law itself.Footnote ³⁵

Further, the subjectivist position I described above leaves the rule regarding mistakes of law untouched in the welfare context. This is so because the retributivist argument for subjectivism applies only to right-infringements or to what are called “true crimes”; it does not apply to welfare offences punishable by noncriminal penalties, for the logical circuit between an intentional doing of what the law forbids and coercibility is not initiated (though it is by an intentional breach of the law) where the law furthers a public goal rather than protecting a right of agency. I have argued elsewhere that negligence is the appropriate fault threshold for noncriminal penalties, so that negligent mistakes, whether of fact or of law, will not excuse (a defence of mistake of law would be available only against a separate offence of knowingly breaching a statutory offence).Footnote ³⁶

Arthur Ripstein has mounted a critique specifically directed against the classical version of subjectivism I have sought to revive. He observes that classical subjectivism maintains a distinction between mistakes of law as nonexculpatory and mistakes of fact as exculpatory by imputing to empirical agents the thinking Agent's knowledge of the wrongfulness of acts interfering with other agents' capacity to act from self-chosen ends. That knowledge comes from grasping the self-contradictoriness of a claim of right based on agency to a liberty disrespectful of agency. This idea, Ripstein argues, raises three problems. First:

This objection confuses throughout the thinking and the empirical agent. The agent whose commitment can alone be self-contradictory is the thinking Agent shorn of all commitments but the one to liberty. That Agent's commitment to liberty is self-contradictory when it lays claim to a liberty disrespectful of the liberty of other agents. The criminal who prefers his gang loyalty to respect for others' freedom is an empirical agent, and it is obviously true that no single empirical commitment of his can be self-contradictory. But in giving effect to his preference in a wrongful act, the criminal implicitly (i.e., for the thinking Agent) acts from an a priori commitment (to a right-claim based on agency to a liberty nonrespectful of agency) that is self-contradictorily destructive of rights. And this right-denial can be imputed to the criminal insofar as he is an agent.

Further, the fatal presumption that Ripstein attributes to the subjectivist is one that the subjectivist will certainly disclaim. The subjectivist's argument need not presume that “the criminal's commitment to respecting agency as such flows from the fact that he acts from reasons he recognizes,” nor does it in fact presume any such thing. Rather, his argument presumes that the thinking Agent's commitment to respecting agency as such flows from his laying claim to a liberty to act, period. From this it hardly follows that the empirical agent will always end up honoring his commitment to respect agency whatever he does; for, first, he may have no such commitment, and second, he may act for empirical ends in a way that contradicts his a priori right-claim to action. Thus the subjectivist meets the condition of success that Ripstein identifies: “in order to show that some implicit commitment has been violated, the commitment must be both implicit in every act and yet violated in criminal acts.” The meaning of the last sentence in the long passage reproduced above is obscure. The question for the subjectivist is not whether the agent was aware of the law at the time of the offence, because the thinking Agent's knowledge of the law proscribing wrongs is imputable to him.

The second problem allegedly faced by the classical subjectivist, according to Ripstein:

This argument, too, rests on a misunderstanding. The subjectivist's readiness to “excuse” inadvertence about matters of fact does not flow from the mere fact that the wrongful act was inadvertent. It flows from the proposition that a right-denial cannot be imputed to an empirical agent's choice if, on the facts as perceived by the agent, he did not choose to interfere with another agent's capacity for self-actuation. This reason for allowing mistakes of fact to exculpate has no general application to mistakes of law nor any application at all to “inadvertence about one's implicit commitments.” If the agent did choose to interfere with another's agency, then a right-denial can be imputed to his choice (because he is an agent) whether or not he was empirically aware that such an interference is wrong. Indeed, since he is an agent who (unlike the infant) can discern right from wrong, his mistaking an intentional interference for a permitted act is just the right-denial that constitutes culpability. So why would the subjectivist treat such a mistake as grounds for exculpation?

The third difficulty for classical subjectivism, Ripstein says, “goes deepest”:Footnote ³⁹

What Ripstein is saying here is that the idea of mutual recognition of agency (or reciprocity, in his language) that delimits valid permissions to act must also draw the boundary of the self in reference to which a boundary crossing is determined. This means that the self's boundaries must fall out of a determination of what is a reasonable impingement or imposition of risk as between two parties in interaction, so that a boundary is crossed not by any imposition but only by an unreasonable one. The classical subjectivist, Ripstein argues, is ambivalent about the role of reciprocity in drawing boundaries to permissible action. On the one hand, he must be committed to the view that, conceptually, valid claims to individual liberty fall out of mutual respect for liberty, for otherwise the thinking Agent could not know the wrongfulness of acts disrespectful of another's liberty; but on the other hand, he is committed to atomism in drawing the self's boundary outside of interaction with others and thus independent of a standard of reasonable impositions. This is why the classical subjectivist can speak of an unreasonable belief in consent (i.e., a belief held despite outward behavior that a reasonable person would interpret as nonconsent) as a mistake of fact about what is going on in the private mind of the victim; for him, the victim's monadic will determines when her consent to bodily contact has been given, and what she inwardly wills is a fact about which one can be mistaken against all outward evidence. The monadic perspective also explains why the classical subjectivist can treat unreasonable mistakes about consent as exculpatory: if the complainant's monadic will determines when her consent to bodily contact has been given, it must be open to the defendant to say (subject, of course, to an evidentiary burden) that he was honestly mistaken about her inner will, even if his belief was held in the face of clear manifestations of resistance or without any effort to ascertain her will; and if in fact the defendant did not choose to touch (or to risk touching) the complainant against her will, then no right-denial is imputable to him.

Ripstein's own view, however, is that whether a person consents to bodily contact is determined in the same way that consent to a contract is determined—interpersonally. So if a reasonable person would interpret the complainant's words or actions as consent, then she has consented in law, no matter what lay in her inner will—a reasonable belief in consent cannot be mistaken. Conversely, if a reasonable person would interpret her manifestations as nonconsent, then she has not consented, and there is no room for an inquiry into her private thoughts. Therefore someone who claims a mistaken belief in consent despite clear manifestations of nonconsent or despite a failure to make reasonable inquiries (i.e., someone who has made an unreasonable mistake about consent) has made a mistake about what consent is; he thinks consent is a matter of one's inner will rather than one's public manifestations. This, however, is a mistake of law. And since he is making a mistake about what consent in law is, to exculpate him for this mistake is to subvert the authority of law and to allow the defendant to dictate when the complainant consents. The subjectivist, concludes Ripstein, in allowing an unreasonable mistake about consent to exculpate, gives exculpatory force to a mistake of law.Footnote ⁴¹

Now it is partly true that the classical subjectivist is monadic when it comes to drawing the boundaries of the self. I say “partly” because the subjectivist is monadic in drawing some boundaries but not others, though this discrimination reflects (I shall argue) attention to nuance rather than ambivalence. The self's body, the subjectivist believes, is a boundary that is set prior to and outside of interaction. As a consequence, there is no latitude for legally permissible but unwanted incursions on the body based on reasonable apprehensions of consent. The subjectivist, however, is not monadic when it comes to drawing boundaries against unwanted disturbances to the enjoyment of one's property or against injuries resulting from the risks attendant on action in the world. Here boundaries are determined by standards of reasonableness.

Why the difference? One could say that personal autonomy requires that the agent determine alone when her consent to bodily contact is given rather than handing this decision over to a meeting between agents. But there is a more conceptual reason. To use one's property and to act in the world are necessarily to create noises, smells, and risks, so that for my right to liberty to coexist with yours, there must be space for reasonable mutual impingements. But my self-possession can coexist with your self-possession without allowing scope for unwanted bodily incursions based on reasonable apprehensions of consent, because self-possession requires no external action. My body is openly mine without my having to take action to occupy it publicly. But because a person's body is hers absolutely, consent to bodily contact is determined from the subjective perspective of the person whose body it is. If in her mind she was not consenting, then she was not consenting, regardless of how the reasonable person would interpret her behavior.

However, the corollary to this is that the defendant must be permitted to say that he was mistaken about what was going on in her mind, whatever her outward behavior (though the unreasonableness of the belief will be strong evidence against its sincerity) and despite his failure to make inquiries. And if this was an honest mistake without advertence or willful blindness to the risk of nonconsent, then his choice was not to touch her or to risk touching her without her consent, and so no right-denial is imputable to his choice. Of course, he may be subjected to a noncriminal penalty aimed at educating him and others to sexual communication undistorted by stereotypes; but he may not be punished for a criminal assault.

So the dispute between Ripstein and the classical subjectivist comes down to this: Who is right about the boundary of the self's body? Is that boundary drawn relationally or nonrelationally? Is there scope for legally permissible unwanted incursions on the body as there is scope for legally permissible impositions of risk in acting? Is there room for a defence of mistake of fact as to whether someone has consented to bodily contact, or are all mistaken beliefs in consent mistakes of law?

To say, as Ripstein does, that consent to bodily contact is determined relationally is to say that my control over my body is not absolute vis-à-vis other persons but is limited by their qualified liberty to use my body against my subjective will. This implication is surely at odds with the liberal premise of individual inviolability. Further, the common law denies that consent is relational even with regard to possession (not enjoyment) of an ordinary item of property. If I take your Lexus from your driveway under the reasonable but mistaken belief that you gifted it to me (you said, handing me the keys, “take the car in my driveway—it's yours,” thinking it was the Lada), I do not get to keep the car. If I enter your land uninvited, you can eject me even if I reasonably believe (because of a clerical error on the deed) I had just bought it. In other words, civil liability for an interference with proprietary sovereignty (not for harm) is strict.Footnote ⁴² Perhaps Ripstein thinks this settled law is mistaken. But if not, then his relational view of consent to bodily contact requires us to believe that tort law affords greater protection to land and chattels than it does to the body.

Further, I shall pose two cases where I believe the intuition against the relational view of consent to bodily contact and in favor of the monadic view is overpowering. Suppose D is about to force himself on V, who has recently broken off a relationship with D. V thinks to herself that if D knows he is raping her, he might very well kill her to remove the only witness to his crime. So she resolves out of fear for her life to feign reconciliation and consent, and she thinks that her self-preservation depends on her being as convincing as possible. Suppose her pretense would be convincing to any former lover and that she succeeds in masking her fear. I do not believe that anyone would comfortably say that V has consented to intercourse in this case, but that is what Ripstein's relational view of consent requires us to say.

Second, suppose (counterfactually) that in Morgan the husband's story to his friends was true, that his wife had indeed invited intercourse with his friends simulating rape and that she was a willing participant in the game throughout. The relational view of consent requires us to say that no consent is given here and that the defendants' “mistaken” belief in consent is a nonexculpatory mistake of law. By contrast, the monadic view yields the sensible result in both cases (no consent in the first, consent in the second). But if consent to bodily contact is indeed determined outside of relation with another, then an unreasonable but honest belief in consent is an exculpatory mistake of fact, not a mistake of law.Footnote ⁴³

Subjectivism and Practical Indifference

R.A. Duff has fashioned an intricate argument countering the position that criminal liability ought to require a conscious choice to do that which is proscribed as wrong or a conscious choice to risk what is so proscribed.Footnote ⁴⁴ The argument is meant to establish that inadvertent recklessness—a failure to notice and avoid an obvious and unreasonable risk one had the capacity to notice—can sometimes supply the blameworthiness needed to justify punishment. Duff's strategy is first to determine what makes an agent blameworthy in what he calls the paradigm cases of responsible agency and then to see whether penumbral cases of responsible agency, namely, advertent recklessness (acting in the face of a foreseen risk of a wrongful outcome) and inadvertent recklessness can be brought within the same reasoning. The paradigm cases are “intended agency,” where the agent aims to bring about a wrongful outcome, and “intentional agency,” where he pursues an object by means that will certainly bring about a wrongful outcome but where that outcome is not itself the aim of his action. Duff's argument proceeds as follows.

If the agent's choice of a proscribed outcome underlies criminal responsibility for outcomes in the two paradigm cases, then advertent recklessness is the only kind of recklessness we can punish, since only the advertently reckless person chooses to risk a wrongful outcome. But choice, while it underlies the paradigm of intended agency, does not necessarily define intentional agency, where the agent produces an outcome he knew would certainly come about as a result of doing what he intends. The reason for blameworthy responsibility in the case of intentional agency, Duff argues, is the attitude of indifference toward the victim's important interests that it shows. The agent is willing to sacrifice the victim's important human interests for his own particular ends. But if that is why we punish agents for unlawful outcomes produced through intentional agency, then we can also punish agents for unlawful outcomes they produce through inadvertent recklessness, because (and to the extent that) they also manifest a callous indifference to the victim's important human interests. And since we are punishing agents who manifest this attitude of indifference, we are still subjectivists about mens rea, argues Duff, albeit not subjectivists in the classical sense of requiring a conscious choice to run a prohibited risk. Choosing to run an unreasonable risk of a wrongful outcome is only one way of manifesting indifference toward whether it materializes; one may also display this attitude through inadvertent recklessness. Thus criminal liability based on inadvertent recklessness can be defended on subjectivist grounds provided that we broaden the notion of the subjective to include an attitude of practical indifference to the important human interests of others.Footnote ⁴⁵

Duff emphasizes that the attitude of practical indifference is not an empirical psychological state for which the outward act is evidence. Rather, it is a practical attitude constituted by conduct that fails to consider a human being's important interests when the danger to those interests is obvious. We should say that the attitude is reflected or manifested in the conduct rather than evidenced by it. However, not all failures to notice or avoid obvious risks constitute practical indifference. Some are not culpable at all because the agent lacked the capacity to appreciate the risk; and some are blameworthy but not criminally culpable because the agent failed to notice the risk for reasons other than indifference—for example, from an undervaluation of the victim's interests in comparison with his own. This Duff calls gross negligence.

Duff comes to practical indifference as the unifying theme of culpable agency by a curious logic. Choice, he says, characterizes intended agency, but there is an ambiguity about what constitutes culpability in intentional agency: it could be that the agent chooses to bring about the outcome he is certain will occur, or it could be that he is willing to ignore another's important human interests for the sake of his ends. Duff opts for the second of these possibilities for no other reason than that an attitude of practical indifference toward others seems highly blameworthy and condemnable. He then reinterprets all the other cases of responsible agency, including intended agency, as instances of practical indifference and then brings inadvertent recklessness into the tent.

However, there is no connection between the blameworthiness of practical indifference and judicial punishment involving an infringement of rights. Nothing tells us that practical indifference deserves judicial punishment as distinct from other kinds of social responses to blameworthy conduct. So Duff's reason for opting for practical indifference rather than choice as the key to culpable agency in the case of intentional agency is not a good one. There is, however, a good reason for opting for choice. Choice of a proscribed outcome is culpable in the case of intended agency because it is the only attitude toward the outcome to which a denial of rights is imputable, and this imputation is what makes the coercion of the agent self-willed. Intentional agency is also a culpable agency because, though the agent does not aim at the proscribed outcome, he knows it will certainly occur, and so he chooses it, thus implicitly denying the validity of rights, including his own. Advertent recklessness and willful blindness are also culpable as implicit right-denials authorizing punishment, because the agent chooses a wrongful risk. And that is where the logic stops. The inadvertently reckless agent does not choose the unreasonable risk, and so no right-denial is implicit in his choice, however obnoxiously narcissistic his character may be in being oblivious to the important interests of others.

Because Duff makes the moral blameworthiness of practical indifference the key to criminal liability, he has no way of stopping his logic from extending past practical indifference to gross negligence and even to negligence, assuming the agent has a capacity to appreciate the risk. He distinguishes practical indifference from gross negligence on the basis that the practically indifferent agent does not care about others' important human interests, whereas the grossly negligent agent, in his preoccupation with his own ends, simply undervalues those interests while overvaluing his own.Footnote ⁴⁶ Yet these attitudes are surely gradations along a continuum of blameworthiness rather than qualitatively different kinds. They differ, as Duff admits, only in degree, whereas one needs a categorical distinction between culpability and other forms of moral blameworthiness to be able to attach criminal liability to one but not the others. Ordinary negligence, where the agent has the capacity to appreciate and avoid the risk, is likewise on the continuum, and so the distinction between crime and tort or between crime and regulatory offence disappears.Footnote ⁴⁷ Indeed, it would seem that any plausible rationale (whether moral blameworthiness, deterrence, or reform) for punishing practical indifference will also justify punishing gross negligence and negligence as culpable forms of fault. What is needed, therefore, is a justification for punishing criminal negligence that will not extend to ordinary negligence. But the only such justification of which I am aware is that in criminal negligence, the actor chooses to impose (what he knows or suspects is) an unreasonable risk of interference with another's agency, hence implicitly denies rights of agency, including his own.

Aware of the icy slope on which he has trodden, Duff tries to rescue practical indifference by distinguishing it from other forms of inadvertent impositions of wrongful risk. The practically indifferent agent, he says, is unaware of a specific risk, for example, of death, whereas the negligent agent is merely unaware of a general risk to safety or health.Footnote ⁴⁸ This distinction, however, is difficult to fathom. Specific and general risk can distinguish different forms of advertence (I foresaw the danger to his welfare but I did not foresee death, whereas you foresaw death), but how can they distinguish different forms of inadvertence? To be inadvertent with respect to a specific risk is either to be inadvertent with respect to risk generally or to be advertent to risk generally but not to the specific risk. The latter possibility is ruled out because the practically indifferent agent who is inadvertently reckless is not advertent to general risk either, for otherwise he would be indistinguishable from the advertently reckless agent; and the former possibility does not distinguish the practically indifferent agent from the negligent one. So for Duff's distinction (between practical indifference and ordinary negligence) to make sense, it would have to be possible for the practically indifferent agent to be at once inadvertent to risk generally and specifically inadvertent to a particular risk. But this is not possible (try not to think about artichokes).

Jeremy Horder has sought to repair this flaw in Duff's reasoning by showing how practical indifference of a certain kind can be brought within the classical subjectivist criterion of criminal culpability.Footnote ⁴⁹ Horder distinguishes between weak and strong practical indifference. The weakly indifferent agent does not deny that another agent's agency interests should be taken into account in one's practical reasoning as to what course of action to take. He simply tends in his haste to overvalue his personal interests and to be thoughtless regarding the other's agency interests. If the risk to the other's agency interests were brought to his attention, his practical reasoning would change. By contrast, the strongly indifferent person denies the weight of the other's agency interests altogether; even if the risk to those interests were brought to his attention, he would still prefer his own particular ends. Accordingly, Horder argues, in denying weight to the other's agency interests, the strongly indifferent person implicitly denies the validity of rights and so invites punishment in the sense demanded by classical subjectivism even though his recklessness is inadvertent.Footnote ⁵⁰

Horder's effort is valiant, but unfortunately it falls short. This is so because strong indifference is now ascertained by means of a counterfactual situation. The agent's act itself tells us nothing about whether the agent is strongly or weakly indifferent to the victim's agency interests. Of course, if an agent is advertently reckless, we can say that he did in fact choose an unreasonable risk of a wrongful outcome and was thus strongly indifferent toward another's agency interests. But if he does not advert to the risk, we can determine strong indifference only by imagining a hypothetical scenario; if he had known of the high risk to the victim's important human interests, would he have chosen the risk?

However, this is now a question about the defendant's underlying character, not about his culpability for this act in particular. In this act, he did not choose the risk. And if punishment is to be directed against a character who is strongly indifferent toward others' important interests, why require a wrongful act at all? Suppose that in acting, the agent did not bother to consider the likely impact of his action on the agency interests of others, but that his action was not unreasonably risky. Yet someone was hurt. Why not ask: If his act had been unreasonably risky and if he had adverted to the risk, would he have chosen the risk? If so, he is strongly indifferent and criminally culpable, on Horder's theory, for a nonnegligent injury. The actus reus requirement of culpability drops out because the agent's act cannot be interpreted as a sign of strong as opposed to weak indifference, and so it becomes a contingent piece of evidence that can be buttressed or countered by evidence of what the agent would have done in the counterfactual situation. Duff avoids this problem by emphasizing that the inadvertently reckless act is not evidence of a practically indifferent attitude lying behind the act but practical indifference itself. But this is precisely what one cannot say about strong indifference in particular. The only act that manifests strong indifference is an advertently reckless one.