Strict Responsibility, Moral and Criminal

1 Introduction

Here is a curious fact of linguistic usage. Let us suppose that someone has damaged someone else’s property. If he caused that damage through non-culpable accident, inadvertence, or mistake, he should not be liable either to moral blame, or to legal conviction for criminal damage: the fact that he neither intended nor expected to cause such damage, and could not reasonably have been expected to realize that his action would or might cause it, gets him off both moral and criminal hooks. Moral philosophers would say, as would ordinary users of English, that he has an excuse for damaging your property, since non-culpable accident, inadvertence, and mistake are paradigm excuses.¹ That is why Aristotle is often taken to have offered the first sketch of a theory of excuses: the akousion, usually although misleadingly translated as “involuntary,” is defined in terms of physical constraint and non-culpable ignorance of fact.² Some legal theorists talk in similar terms about excuses in the criminal law.³ Other legal theorists, however, especially those who take the distinction between offenses and defenses seriously, do not use “excuse” in this way.⁴ Criminal damage is so defined that its commission requires intention, knowledge, or recklessness as to the fact that one’s conduct will damage another person’s property.⁵ When mistake, accident or inadvertence exculpate the agent who caused the damage, they therefore do so by negating the mens rea of the offense, and do not count as excuses; an excuse, as such theorists use the term, is a defense that comes into play only when the commission of the offense has been proved, to show that the agent should nonetheless not be convicted and condemned. A person would have what such theorists count as an excuse for damaging the other person’s property if he acted, for instance, under a type of duress that was not severe enough to justify his action, but was such that a reasonable person might have given in to it; or if he mistakenly believed that he could save himself from serious harm only by acting as he did.⁶ Such theorists therefore recognize a much narrower category of excuses in criminal law than others do in criminal law or in our moral lives.

Philosophers who recognize the broader category of excuses often also say that excuses negate responsibility. In excusing ourselves “we admit that it was bad but don’t accept full, or even any, responsibility.”⁷ Theorists who recognize only the narrower category of excuses are more likely to say that excuses presuppose, rather than negate, responsibility: in excusing ourselves, we accept or admit responsibility, but seek to avert blame or condemnation.⁸

It might be tempting to dismiss such linguistic variation as lacking substantial interest. As far as responsibility is concerned, “responsible” and its cognates lack any single, determinate meaning: we might say with equal truth both that excuses negate responsibility, and that they presuppose it, in a subtly different sense. As for excuses, we might be tempted to see this as just one of those cases in which the law adapts ordinary extra-legal terms to its own purposes, in the process giving them technical meanings that differ from their extra-legal meanings. We can also explain why this might happen to “excuse” by reference to the structure of criminal trials in adversarial systems that give the presumption of innocence a central role. In such systems the prosecutor must prove that the defendant committed the offense, and on orthodox understandings of the presumption of innocence this requires the prosecutor to prove both the conduct element or actus reus and the fault element or mens rea: to prove, for instance, not just that the defendant’s conduct caused damage to another person’s property, but that the defendant caused that damage intentionally, knowingly or recklessly.⁹ Unless and until the prosecutor proves that much, the defendant is not formally required to offer any evidence or argument, although in practice he will need to rebut or respond to otherwise persuasive evidence that the prosecutor offers if he is to avoid conviction. Proof of the commission of the offense is not yet proof of guilt: the defendant can still avoid conviction by offering what the law counts as a defense, a justification or excuse for his commission of the offense. In relation to defenses, however, the defendant bears the initial probative burden of adducing evidence to support the defense: only if such evidence is offered must the prosecution take on the burden of rebutting it, and disproving the defense.¹⁰ Accident, inadvertence, and mistake of fact therefore play, procedurally, different roles in the criminal trial from the role played by duress: accident, inadvertence, and mistake of fact must be ruled out by the prosecutor’s evidence, while the defendant must offer evidence of duress before the prosecutor acquires any probative burden in relation to it. That is why, it might be suggested, lawyers and legal theorists give “excuse” a narrower meaning, so that it covers only excuses that it is procedurally up to the defendant to offer; it is a matter of terminological convenience and clarity.

Such procedural aspects of the criminal trial are indeed important, and will figure in what follows. However, we should not be so quick to dismiss these linguistic variations between legal and extra-legal usage, since they point to some substantially interesting dimensions of responsibility in both moral and legal contexts. Nor should we be too quick to suppose that lawyers or legal theorists give “excuse” a different sense from that which it bears in ordinary extra-legal moral discourse, since while its extension does indeed differ between legal and extra-legal contexts, its intension does not: the difference in extension is to be explained by a difference not in the meaning of the term, but in the scope of responsibility in the two contexts. Section 3 will include an explanation of this point about the meaning of “excuse” and its importance; but as a prelude to that explanation, Section 2 will include a clarification of the conception of responsibility on which we will rely. Thereafter, Section 4 will include a clarification of the way in which in ordinary moral contexts responsibility can be said to be strict, whereas in criminal contexts in which the presumption of innocence has its orthodox meaning and significance, responsibility is non-strict. Section 5 will include a discussion of some of the ways in which criminal responsibility can be made strict. Criminal lawyers and theorists discuss strict criminal liability at length, but they have not yet paid enough attention to the equally significant phenomenon of strict criminal responsibility.¹¹

2 Responsibility as Relational

We can usefully draw a partially, but only partially, stipulative distinction between liability and responsibility. Our concern here is with liability to moral criticism or to criminal conviction, and so with responsibility for actions, outcomes, events, or states of affairs that are in some way untoward. In these contexts, responsibility is necessary but not sufficient for liability. It is necessary, because a person cannot be held liable for something for which she is not held responsible; she can justly be held liable only if she is justly held responsible. If John admits that Mary was not responsible for the damage that his car suffered, he cannot coherently hold her liable to blame for it; if it would be unjust to hold her responsible for that damage, it would also be unjust to hold her liable to blame for it. Responsibility is not, however, sufficient for liability, since Mary could admit that she was responsible for the damage to John’s car, but seek to avert blame by offering a defense, an excuse or a justification. So too, in criminal law, a conviction holds the defendant criminally liable for the commission of an offense, and depends on proof that she is criminally responsible for its commission; but proof of such responsibility, for instance proof that she intentionally damaged another person’s property, does not suffice to establish liability, for instance for an offense of criminal damage. The defendant can admit the commission of the offense, which is on the account offered here to admit responsibility for the offense, but she can avert criminal liability by offering a legally recognized defense.

We can understand this distinction by understanding responsibility as a matter of being answerable.¹² To say that a person is responsible for some action or result is to say that he is answerable for the action or result, which is to say that he can properly be called to answer or to account for it. That is not yet to say that he is liable to moral blame, or to a criminal conviction, for the action or result, since he can still avoid liability by offering an exculpatory defense, but it is to say that the onus lies on him to offer a defense: he is called to answer, either by accepting liability or by offering a defense, and he can be justly held liable if he cannot offer an exculpatory defense. This structure is revealed most clearly and formally in the legal process of the criminal trial. If a defendant is charged with perjury, the first question for the court is whether she committed that offense. Did she make a statement that she knew to be false, or did not believe to be true, when sworn as a witness in a judicial proceeding?¹³ It is up to the prosecution to prove this. Unless and until it is proved, the defendant has nothing to answer for in court. If it is proved, she then has an offense of perjury to answer for, and can properly be convicted as being guilty of that offense unless she offers an appropriate defense; but she can avoid conviction, by offering evidence of such a defense, for instance that she acted under duress.¹⁴ A similar logical structure can be discerned in our informal, extra-legal moral dealings. If Nancy accuses Mark of injuring her in some way, the logically prior question is whether he injured her. Was he responsible for, must he now answer for, her injury? If he is thus responsible, he might have no honest option but to admit his wrongdoing, accept her justified blame, and look for some way of making apologetic reparation; but he might instead be able to ward off blame by offering a justification or excuse for what he admittedly did.

If we understand responsibility as answerability, we must also understand it as multiply relational. Not only is a person responsible for something; he is responsible to some person or body, and as falling under some relevant, normatively laden description, typically one that places him within a normative practice. These relational dimensions of responsibility are important in part because they mark both the extent of and the limits to our responsibilities in the different areas of our lives. Thus a doctor, for instance, has particular responsibilities that others do not have, and that she would not have were she not a doctor. They include responsibilities, most obviously, in relation to the health of her patients, insofar as they put themselves in her hands, and the treatment she provides for them; they also include responsibilities under her professional code of practice, and might include a responsibility to provide medical help in an emergency even to someone who is not her patient. However, we should note two kinds of limit on her responsibilities as a doctor. First, she is responsible for her medical activities to her patients, and to her professional colleagues. They can properly call her to answer for her conduct, and can properly criticize her for misconduct if she cannot provide an answer, an explanation of her actions, that shows them to be medically appropriate. But she is not thus answerable to other people in general. If a non-medical acquaintance challenges her about some treatment she has prescribed for a patient, she can properly reply, as she could not properly reply to the patient or to a colleague: “That is not your business; I do not answer to you for my medical activities.” Second, while she is responsible for her medical activities to her patients and her colleagues, she is not answerable to them for other aspects of her life and conduct, for her religious beliefs and practices, for instance, unless they impinge on her medical activities, or for her performance as a member of a choir. There are people to whom she is answerable for such activities, perhaps her priest, other members of the choir; but those activities do not fall under her medical responsibilities.

If we are to specify a person’s responsibilities, we cannot just ask what he is responsible for. We must ask what he is responsible for under or as fitting a particular normative description; and we must ask to whom he is responsible, which is to ask who has the standing to call him to answer or to account. What makes a person responsible for another’s health is not just that she can make a difference to it, but that she is his doctor. What allows the doctor to deny responsibility for the crimes she knows her patient will commit thanks to his newly regained health, and to deny that she is responsible for aiding and abetting those crimes, is that her responsibilities as a doctor are limited to the patient’s health and what bears on it; what he then does is not her business, even if he can do it only because she has restored his health.¹⁵ What makes it appropriate for others to call the doctor to account for her treatment of this patient is that they are, for instance, his partner, or a fellow doctor, or a member of a relevant committee of the British Medical Association: she is indeed answerable to them for her treatment of him. What allows her to refuse to answer to Jones for her treatment of this patient is that he has no such standing to call her to account for it: it is not his business; she is not responsible to him.

This is not to say that the scope of our various responsibilities is always uncontroversial or clear: there is plenty of room for disagreement about both just what a doctor is responsible for, and to whom doctors should be answerable for their medical activities; the same is true of the responsibilities that attach to many other normative roles, such as parent, teacher, lawyer, employer, friend, artist, or plumber. The point is only that what a person is responsible for in a particular situation depends on what he is responsible as, on what normative description he is to be seen as satisfying in that situation, and to whom he is thus responsible. Until we have determined the “as what” and the “to whom,” we cannot determine the “for what.” There will often be disagreements about to whom and for what we are responsible. Typically, they reflect disagreements about the normative roles that generate the relevant responsibilities, about what it is to be, for instance, a doctor, teacher, or parent.

There are thus at least three different ways in which a person might deny responsibility for some prima facie untoward event, for the breaking of a window, for instance. She might, first, deny that she had anything causally to do with the event: she played no causal role in its occurrence, and had no power to prevent it. She might, secondly, deny that it was her responsibility or business: she saw the youth about to throw a large stone at the window, and could have stopped him; but she denies that she had any responsibility to protect the window or to control the youth, which is to deny that there is a failure to prevent the harm for which she must answer. She might, thirdly, deny that it is anyone else’s business: it was her own window, and she does not have to answer to anyone else for breaking it.

Our moral responsibilities and what we can call our criminal responsibilities, the responsibilities that we have under the criminal law, are similarly relational. Many of our moral responsibilities are tied to particular roles that we fill, and to particular practices within which we discharge those roles; but many are general, including most clearly our responsibilities not to attack others or their interests, not to cause harm to them, and to at least some modest degree to assist them if they are in desperate need and we can easily help. There is plenty of room for controversy both about the precise contents of such responsibilities, and about the normative descriptions to which they attach: do we have such responsibilities as, for instance, rational agents, as Kantians would say, or as human beings, as philosophers who take our humanity and our fellowship with other human beings to be crucial would say, or as inhabitants of this planet, as some ecological theorists might say, or as agents with the capacity to cause or to prevent suffering, as some utilitarians would say?¹⁶ There is also room for controversy about who has the standing to call us to answer for our alleged moral failings. Must we be ready to answer to any other moral agent, or any other human being, on the grounds that morality is everyone’s business, or can we argue that while certain kinds of serious moral wrong or failing are no doubt everyone’s business, others are properly the business only of those more intimately or closely involved in the matter? We cannot pursue these controversies here.

As for our criminal responsibilities, a full account of what we are criminally responsible for, as what, and to whom, depends on an account of the proper role of criminal law, and thus also on an account of the proper role and character of the state and of its relationship to its members. We must ask about the proper aims of the criminal law as a practice, and to whom it is addressed, by whom, and in what terms. To answer those questions, we need to appeal to political theory. One kind of answer is suggested by classical legal positivism. Criminal law, like all law, consists in the sanction-backed orders of a sovereign: the law is addressed by the sovereign to the sovereign’s subjects, in the peremptory voice of a commander; we are bound by the law as the sovereign’s subjects, and are responsible to the sovereign for our obedience or disobedience to her commands. A better answer, for contemporary liberal democracies, is that the criminal law binds us as citizens; that it is addressed by us to ourselves as a common law that is our law; and that we are therefore answerable to each other, to our fellow citizens, through the criminal courts. A further part of the answer, appealing now to some version of legal moralism, is that the criminal law is properly concerned with wrongs that count as public wrongs in the sense that they concern all members of the polity in virtue of their shared citizenship: the substantive criminal law identifies and defines such wrongs as wrongs for which we should be called to answer by our fellow citizens, while the procedural criminal law makes provision for the institutional processes, in particular the criminal trial, through which those who commit or are alleged to have committed such wrongs are called to answer.¹⁷ Given this account of responsibility and its relation to liability, we can see more clearly how excuses fit into the picture.

3 Excusing

We might do better to begin with the idea of excusing rather than that of excuse, to emphasize that we are dealing with practices in which responsibility and liability are ascribed, denied, or accepted.¹⁸ We can understand what it is to be responsible by understanding what it is to be held responsible, by others or by ourselves, and what it is to accept or deny responsibility. So too, we can understand how excuses function by understanding what it is to offer an excuse, on our own or another person’s behalf, and to accept or reject excuses. Similar points apply to the idea of justification: to understand the role that justification plays in our moral or legal lives, we must understand what it is to justify our own or another person’s actions.

To offer an excuse is to admit, at least by implication, that there is something that needs excusing. The procedural structure of the criminal trial makes this point clear. The trial calls a defendant to answer to a criminal charge: a charge that he committed a public wrong. It thus addresses him from the start as a responsible agent, who can answer for his actions, and he is expected to make a formal answer, by a plea of “Not Guilty” or “Guilty,” to the charge; that is why it is crucial that the defendant be “fit to plead.”¹⁹ If he pleads “Not Guilty,” the burden falls on the prosecution to prove that he committed the offense charged, but if the prosecution discharges that burden, the onus shifts onto the defendant to introduce evidence to support a defense if he hopes to avoid conviction. He must, that is, offer an excuse or a justification: he must seek to excuse or justify the conduct that the prosecution has proved.

A defense in a criminal trial has the form of “Yes, I did, but …”: the defendant admits, perhaps is forced to admit by the proof led by the prosecution, that he committed the offense, which is to admit criminal responsibility for that offense; but, he argues, he should not be held liable for his commission, because the defense that he offers serves to exculpate him. We can discern a similar logical structure, without the formal apparatus of charge and proof, in our extra-legal moral dealings with each other. Someone accuses her neighbor of damaging her property, and the neighbor might respond by denying responsibility: he might deny that he was the one who caused the damage, or insist that he is not answerable to anyone else for it. Alternatively, he might admit that he caused the damage, but offer an excuse or a justification; again, the form of his response is “Yes, I did, but …”: an admission followed by a defense. He admits responsibility, but averts liability by offering a defense that blocks what might otherwise be a permissible transition from responsibility to liability. She might reject his defense, and might be justified in doing so; but just as it is up to him to offer a defense, if he admits or it is shown that he caused the damage, it is up to her, and to any others who call him to account, to attend to any defense that he offers.

To understand excuses and justifications, either moral or legal, we must thus understand both what the “Yes” admits, and what can follow the excusatory or justificatory “but” to offer a defense. A key difference between moral and criminal responsibility concerns what the “Yes” admits; this also affects what can count as an excuse, what can follow the “but” when it is offered in excusatory mode. Before we attend to that issue, however, we should note two preliminary points: these will help to clarify both the logical structure of excuses, as they are portrayed here, and the way in which “excuse” is given a partly stipulative meaning.

First, let us suppose that Martha’s was broken and that the immediate cause of the breakage was that David’s foot came into violent contact with the window. When Martha accuses David of breaking the window carelessly, if not willfully, he explains that he was walking peacefully along the street when a group of men seized him and threw him bodily at the window; given their number and strength, there was nothing he could do to prevent his feet hitting the window.²⁰ This is not to offer an excuse in the sense in which “excuse” is being used here, because the “Yes, I did” that precedes an excuse admits agency, whereas here only David’s body was causally involved in the breaking of the window. It would be at best misleading, and at worst outright false, for him to say “Yes, I broke the window, but I was thrown at it.” If he was helplessly thrown at it, breaking it was not something that he did as an agent. This point is reflected in J.L. Austin’s remark that offering an excuse might involve claiming that “it isn’t fair to say baldly he did A; it may have been partly accidental, or an unintentional slip.”²¹ An excuse might qualify or modify the claim that “he did A,” but will not deny it altogether. It is also reflected in the criminal law, in the common claim that the actus reus of a criminal offense must include a voluntary act, so that in this example David has not committed the actus reus of criminal damage, and no question of mens rea, of whether he caused the damage intentionally or recklessly, even arises.²² The same is true when what a person is accused of is an omission rather than an action. If Martha accuses Robin of failing to prevent her window from being broken by a passing vandal, Robin might deny responsibility, by claiming that it was not his business, or that he lacked the capacity to intervene: he was tied up or paralyzed and could not move. He might, instead, offer an excuse: perhaps that he was too frightened to step in. That would be to admit what denials of responsibility deny: that there was a failure to prevent harm for which he must now answer.

Responsibility therefore requires at least minimal control. If a person had no control over an event or result, no capacity to make a difference to whether the event or result occurred or not, he cannot be held responsible for it or be called to answer for its occurrence.²³ If he lacked such control, there is therefore nothing for which he need now offer an excuse: he should of course regret the harm that was caused by a process that involved his body, or the occurrence of the harm that he could not prevent; but he has nothing for which he needs to excuse himself as an agent, unless it turns out that he can be held responsible for his very lack of control.

The second preliminary point concerns psychological conditions that do not negate our responsibility for particular actions so much as our very status as responsible agents.²⁴ It is a defense to show that a person’s commission of an offense was attributable to a mental disorder that satisfies the legal criteria for the insanity defense, and it seems natural to portray this defense as an excuse; it is certainly not a justification. But this excuse does not admit responsibility for the offense; instead, it shows that the person should not be held responsible for the offense. This is true, but that is why some theorists distinguish excuses from exemptions.²⁵ For a person to offer an excuse is to admit responsibility, and to answer for his actions by explaining how he came to act as he did, in a way that is meant to show why it would be unjust to condemn him; central to such an explanation will be an account of his reasons for acting as he did, in order to show that while he in fact acted as he had good reason not to act, or failed to act as he had good reason to act, he did not act in willful or culpable disregard of those reasons. By contrast, to offer an insanity defense is to say that he cannot be expected to answer for his actions, because they were not the actions of a rational agent. A person who is spared moral blame or criminal conviction on grounds of insanity is exempt, or excluded, from responsibility, from the practices of being called to answer, and of answering, for his actions.

It might be objected that to deny that the involuntariness of our bodily movements, or the insanity that explained our actions, constitute excuses is to stipulate a narrower meaning for “excuse” than it carries in our ordinary discourse. That is certainly true, but the stipulation is neither arbitrary nor unwarranted. It enables us to identify a distinctive category of liability-averting claims that admit rational agency, but deny liability: “Yes, I did it, and I did it as a rational agent who can explain his actions in terms of his reasons for action; but I should not be blamed or condemned, because ….” A central task for a theory of excuses is to explain what can properly complete that “because” clause. One important difference in the ways that that clause can be completed in extra-legal moral contexts and under the criminal law is explained in the next section.

4 Strict Moral Responsibility

In criminal law, liability is strict when it does not require mens rea, whether intention, knowledge, recklessness, or even negligence, as to some aspect of the actus reus of the offense. Thus in English law, criminal liability for possession of a scheduled drug used to be strict as to the fact that what the defendant had was a scheduled drug. The prosecution must prove that he knew that he had the relevant item in his possession, but need not prove that he knew or suspected that it was or contained a scheduled drug, or even that he had reason to suspect that it was.²⁶ Similarly, criminal liability for the “rape of a child under thirteen” is strict as to the child’s age. The prosecution must prove intentional sexual penetration of a child who was in fact under thirteen, but need not prove that the defendant knew, or suspected, or had reason to suspect, that to be so.²⁷ We can also talk of strict responsibility in criminal law. Responsibility is strict when the prosecution is not initially required to prove mens rea as to every aspect of the actus reus of the offense, but lack of mens rea constitutes a defense as to which the defendant bears the burden, if not of proof, then at least of adducing evidence that would suffice, if not rebutted, to create a reasonable doubt. Thus possession of a controlled drug is now, in English law, an offense of strict responsibility, not of strict liability. The prosecution need initially prove only that the defendant possessed what was in fact a controlled drug; but he has a defense if he can prove “that he neither believed nor suspected nor had reason to suspect that the substance or product in question was a controlled drug.”²⁸ Liability is not now strictly strict, since a proven lack of mens rea secures an acquittal. But responsibility is strict: it is up to the defendant to answer for his possession of what was actually a controlled drug; if he is to avoid conviction he must exculpate himself by offering lack of mens rea as a defense.

Strict criminal responsibility will be discussed in Sect. 5. Here, we will be concerned with strict moral responsibility. It is often claimed, or assumed, that moral liability cannot be strict, since it cannot be just to blame a person for what she does through non-culpable inadvertence, accident, or mistake. It might be appropriate for a person who non-culpably causes harm to feel a distinctive kind of “agent-regret” that marks her own involvement as an agent, but that is not to say either that it would be appropriate for her to feel remorse, or for others to condemn her or blame her.²⁹ What is less often noticed, however, is that moral responsibility is typically strict. In our moral dealings with each other, we must answer for the harms that we cause even if we cause them without the moral analogue of mens rea, through non-culpable accident, inadvertence, or mistake. We can avoid moral liability by offering an excuse; but the onus is on us to offer such an exculpatory answer for our actions.

Let us suppose that Brian has damaged Hilda’s property: he knocked over a vase in opening a door behind which it stood, trod on her glasses, which had fallen on the floor, or burned her copy of Time magazine on his bonfire. In each case, he is free from fault: he had no way of knowing that the vase had recently been put behind the door, and so knocked it inadvertently; a young child rushed past pushed him, and he trod on the glasses accidentally; he thought, and had good reason to believe, that what he was burning was his copy of Time, and he burned Hilda’s copy by mistake. Once he explains these facts to Hilda, she should realize that it would be wrong to blame him for the damage. Blame would be justified only if he had caused the damage intentionally, recklessly, or negligently, which his explanations show not to be the case. Had he been put on trial on a charge of criminal damage, it would have been for the prosecution to prove not merely that he caused the damage, but that he caused it intentionally or recklessly. Until that was proved, he would have nothing to answer for in criminal court, and would need no defense. Our moral dealings, however, are different: Hilda would reasonably expect him to answer for the damage he has caused, by explaining how he came to cause it. He can answer for his action, in a way that should avert blame: he can explain how he caused the damage through non-culpable inadvertence, accident or mistake. But unlike the criminal context, in the moral context, he does need to answer: he is rightly held responsible for the harm that he caused.

To drive this point home, we may note three things. First, Brian should apologize for causing the harm. He should express not just the regret that a concerned observer might express that Hilda’s property has been damaged, a regret that might be the same whether the damage was caused by human agency or by natural causes, but an agent’s apologetic regret for the harm he has done; and such apologetic regret admits responsibility. Hilda would be rightly annoyed if he simply denied responsibility for the damage, and expressed only a spectator’s regret, since that would fail to acknowledge the fact that he caused the damage.

Second, a natural way to fill out his apology would be to explain how he came to cause the harm, which would, in this case, involve explaining how it was not a matter of negligence. If Hilda says, in challengingly accusatory tones, “Look what you did!”, a proper response would be not to deny that he broke it, or claim that he does not have to answer to her for breaking it, but to accept responsibility by offering the exculpatory answer that he has: that he was taking all due care, which is to say that it would have been unreasonable to expect him to take the kind of care that would have avoided harm, and that he thus caused the harm through non-culpable misfortune. But such an explanation admits responsibility. It marks an acceptance that he should answer to the Hilda, as the property owner, for what he has done; it seeks not to deny responsibility, but to block the transition from responsibility to liability.

Third, what makes the apology necessary, and underpins the ascription and acceptance of responsibility, is that he did what he in fact had good reason not to do: the fact that his action would cause such damage constituted a good reason not to act thus. He did not realize that he had that reason, just as a person who mistakenly believes a glass of gasoline to be a glass of gin does not realize that she has reason not to drink from it; but that he has reason to do or not to do something does not depend on his knowledge of the facts that constitute or generate that reason.³⁰ If he realizes, as he picks up the magazine to throw it on the fire, that it is Hilda’s, he does not acquire a new reason for action, a reason not to put the magazine on the fire that he lacked before; he becomes aware of the reason that already existed. Responsibility is tied to reasons: a person is responsible, and must answer, for acting as he had reason not to act, for not responding appropriately to the reasons that bore on his action. Brian is therefore responsible for burning Hilda’s magazine, although he can offer an explanation of why he did not respond to that reason which saves him from being blamed.

The claim that moral responsibility is typically strict may well still seem bizarre to theorists who believe that responsibility for a result requires satisfaction both of the control condition, that the result was within the agent’s control, and of the epistemic condition that he knew or suspected or had reason to suspect that the result might ensue as a consequence of his action.³¹ It will seem bizarre, however, only if we do not draw the distinction between responsibility and liability. Inasmuch as it would be bizarre to argue that a person can properly be blamed or condemned when the epistemic condition is not satisfied, it is a condition of liability. Inasmuch as it is not bizarre to argue that a person can properly be called to answer for bringing a result about when the epistemic condition is not satisfied, it is not necessarily a condition of responsibility. It is a condition of responsibility within some practices, notably within the criminal law, but it need not be.

Two further points should be noted about strict moral responsibility before we turn to the criminal law. First, just what is it that we are strictly responsible for in moral contexts? John Gardner argues that what we are responsible for is wrongdoing: “the ordinary or basic kind of wrongdoing,” which includes “hurting people” and so also presumably damaging their property, is strict wrongdoing, inasmuch as it can be identified as wrongdoing without reference to any fault, or any analogue of mens rea, on the wrongdoer’s part.³² That is why we need to be able to offer a justification or excuse if we are to avoid blame. There are indeed cases in which it is appropriate to say that what needs justifying or excusing, and can be justified or excused, is wrongdoing. If a person lies to someone he does wrong to that person but might still be able to justify or excuse himself. But while it is true that when Brian damages Hilda’s property through non-culpable accident he acts as he in fact has reason not to act, and owes Hilda an apologetic explanation, it seems odd to say that he has done wrong: we should say, instead, that the explanation shows that he did no wrong and that we should see the damage as an unfortunate accident. Wrongdoing is in the air, in that if he caused the damage intentionally or carelessly he did wrong; but his exculpatory answer serves to ward off the suspicion of wrongdoing that might otherwise reasonably be created by his action. We are, properly, morally interested in and concerned about not merely the wrongs that we and others commit, but also the harms that we and others cause. That is why we are held morally responsible for such harms even if we do no wrong in causing them.

Second, however, this makes the connection between moral responsibility and liability to moral blame less tight than is the connection between criminal responsibility and criminal liability. If the prosecution proves that the defendant committed the offense, which is to prove criminal responsibility for that offense, the defendant is held criminally liable unless he offers a defense that suffices at least to create reasonable doubt: proof of responsibility thus creates a presumption of liability that it is up to the defendant to rebut. The fact that a person damaged someone else’s property, however, does not create any such warranted presumption of moral liability. More precisely, it might create a presumption of moral liability to pay for the replacement or repair of the damaged property, but not a presumption of liability to moral blame or criticism, the moral analogues of criminal conviction. The property owner might, depending on the precise context, suspect that the person caused the damage through carelessness, if not willfully. That suspicion might be reinforced if he refuses to offer any exculpatory explanation: but there might well be room for entirely reasonable doubt about the matter, and the circumstances might sometimes make it far more likely that the damage was caused through non-culpable accident or inadvertence. If he refuses to answer, and thus refuses to accept responsibility, he is liable to criticism for that refusal, unless he can explain or justify it. What we can call his primary retrospective responsibility for having caused the damage generates a secondary prospective responsibility to answer for it, and he can be called to account for failing to discharge that responsibility. But just because moral responsibility is so strict, responsibility does not by itself suffice to create a presumption of liability.

5 Criminal Responsibility, Non-Strict and Strict

Insofar as the presumption of innocence, as classically understood, holds good in the criminal law, criminal responsibility is not strict: the prosecution must prove both actus reus and mens rea before the defendant has anything for which he must formally answer. That is why what counts in extra-legal moral contexts as an excuse, non-culpable inadvertence, accident, or mistake, does not count in criminal law as an excuse. The point is not that the meaning of “excuse” varies between the two contexts: in both contexts an excuse is a plea that admits responsibility, and seeks to ward off liability by offering an exculpatory but not justificatory explanation of the agent’s conduct. The scope of responsibility varies between the two contexts, however: a person is morally, but not criminally, responsible for harm that he caused through non-culpable accident, inadvertence, or mistake.

It is worth asking why criminal responsibility should differ from moral responsibility in this way. Why should it too not be strict? If the prosecution proves the actus reus, which will normally involve proving that the defendant’s conduct caused a harm or evil of a kind that concerns the criminal law, we may wonder why the defendant should not then be liable to conviction unless she can offer a suitably exculpatory account of how she came to cause it, for instance an account that denies mens rea. It might be unreasonable to expect her to prove lack of mens rea; but it does not seem unreasonable to require that she at least adduce plausible evidence of its absence. We must answer morally to our friends for harm that we actually cause. Why should we not also have to answer to our fellow citizens, under the criminal law, for criminal harms we cause, as long as the law makes adequate provision for exculpatory answers that will avert liability?

Obvious answers to this question are that it is more important to avoid the conviction of the innocent than to ensure the conviction of the guilty, and that such a shift in the probative burden would impose unreasonable costs on defendants. Another, slightly deeper, answer is that if we see the criminal law as a practice that focuses on public wrongdoing, and on calling to public account people who are guilty of such wrongdoing, we should only have to answer in a criminal court for proved conduct that constitutes at least a presumptive public wrong: conduct that is either indisputably wrongful, although possibly justifiable or excusable, or at least such as to create a reasonable presumption of wrongfulness.³³ The mere causation of harm, while it might create a suspicion of wrongdoing, is not enough to create a reasonable presumption of wrongdoing: that is why a person should not be required to answer, in the criminal court, for the harm that he merely causes, on pain of conviction and punishment if he does not offer a suitably exculpatory answer.

However, the criminal law sometimes makes responsibility strict. Sometimes defendants must answer for the commission of an actus reus, without proof of mens rea. There are two ways in which this can be done.³⁴

First, instead of requiring the prosecution to prove mens rea in relation to all aspects of the actus reus, the criminal law can define lack of mens rea as a defense, in relation to which the defendant bears at least the evidential burden of adducing evidence of lack of mens rea sufficient to create a reasonable doubt if not rebutted by the prosecution, if not the persuasive burden of proving lack of mens rea. Thus the offense of selling “food which fails to comply with food safety requirements” is defined strictly: anyone who sells such food is “guilty of an offence.”³⁵ The prosecution does not need to prove knowledge, recklessness, or negligence as to the food’s failure to comply. If that was all there was to the treatment of this offense in the Food Safety Act, it would create strict criminal liability: but it does not do so, since a later section provides that “it shall … be a defence for the person charged to prove that he took all reasonable precautions and exercised all due diligence to avoid the commission of the offence”.³⁶ The Food Safety Act does make responsibility strict, however: the defendant must answer for committing the actus reus of the offense, for selling food that was in fact unfit, on pain of conviction and punishment if he cannot offer an exculpatory answer that negates mens rea.

Second, the law can define the offense in a way that requires proof of both actus reus and mens rea, but create a legal presumption that allows the court to presume mens rea, to treat it as being proved, given proof of the actus reus, thus laying on the defendant the burden of adducing evidence to rebut the presumption. If it is proved that Jones sexually penetrated Smith when Smith was, as Jones knew, unconscious or asleep, the court trying Jones on a charge of rape is required to presume that Smith did not consent and that Jones did not reasonably believe that Smith consented “unless sufficient evidence is adduced to raise an issue” whether Smith in fact consented, or whether Jones reasonably believed that Smith consented.³⁷ This provision makes criminal responsibility strict: what the defendant has to answer for, on pain of conviction and punishment if he cannot offer an exculpatory answer, is not sexual activity with a non-consenting person whom he did not reasonably believe consented to it, but sexual activity with someone who was unconscious. Criminal liability is not strict, since he can avoid it by offering an exculpatory explanation; but criminal responsibility is strict, since the onus lies on him to offer the explanation.

Sometimes the presumption is implicit rather than explicit. If an employee is injured by the machinery he is operating, his employer can be charged with failing “to ensure, so far as is reasonably practicable” the health and safety of her employees.³⁸ She can avoid conviction by proving that she had done all that was “reasonably practicable” to ensure their safety, but will be convicted if she cannot offer such an exculpatory answer.³⁹ The criminal law thus holds employers strictly responsible for such injuries to their employees, but not strictly liable.

Theorists who take the presumption of innocence, as classically understood, to be a central and inviolable principle of the criminal law, the so-called golden thread that characterizes a civilized system of criminal justice, will regard all such instances of strict criminal responsibility as unwarranted.⁴⁰ Many no doubt are. They serve not the ends of justice, but the convenience of the prosecution, whose probative burden they lighten, and the populist aims of what are said to be the war on crime or the war on terror. The provisions of a section of the Terrorism Act 2000 illustrate this point all too well.

The formal title of the offense is “Possession for Terrorist Purposes,” and s. 57(1) defines it as follows: “A person commits an offence if he possesses an article in circumstances which give rise to a reasonable suspicion that his possession is for a purpose connected with the commission, preparation or instigation of an act of terrorism.”⁴¹ Taken by itself, this subsection appears either to criminalize conduct that is not necessarily wrongful, or to create an unjustifiable legal presumption. If, on the one hand, the offense is simply possession under circumstances that give rise to a reasonable suspicion, it is hard to see that as a wrong that could merit public condemnation. If, on the other hand, the offense is as its title indicates, possession for terrorist purposes, then the subsection appears to require the court to presume that the defendant’s possession was for terrorist purposes, given proof merely that it was such as to create a reasonable suspicion. But such an inference from suspicion that the possession is for terrorist purposes to a verdict that the possession is for terrorist purposes is unjustifiable; however reasonable that suspicion might be, it could well still not be the case that possession is for such purposes.

Matters are not actually as bad as that, since another subsection enables the defendant to rebut the implicit presumption of terrorist purposes. “It is a defence for a person charged with an offence under this section to prove that his possession of the article was not for a purpose connected with … terrorism.”⁴² Furthermore, while it might seem grossly onerous to require the defendant to “prove” that his intentions were innocent, or at least unrelated to terrorism, a later section provides that if he can “adduce evidence which is sufficient to raise an issue with respect to the matter the court or jury shall assume that the defence is satisfied unless the prosecution proves beyond reasonable doubt that it is not.”⁴³ The Terrorism Act thus imposes not strict liability but strict responsibility. The defendant is held strictly responsible for his possession of articles that give rise to a reasonable suspicion: he must answer for that possession in the criminal court, and is liable to conviction and punishment as having possessed the articles for terrorist purposes unless he can offer evidence in support of an exculpatory explanation of his possession.

This provision for a defense of lack of guilty intent mitigates the injustice of the Terrorism Act, but does not remove it. One obvious concern is that it might still allow for the conviction of substantively innocent defendants: people whose intentions are unrelated to terrorism might not be able to adduce evidence of their innocence that suffices to create a reasonable doubt, and would then be liable to conviction. That concern might be met by a reasonably generous interpretation of “sufficient to raise an issue.” For instance, if the mere offering of an innocent explanation was taken to suffice, thus requiring the prosecution to prove a guilty purpose, the substantively innocent might not face conviction. However, there is a deeper objection: that the Terrorism Act requires us to answer publicly, on pain of conviction and punishment if we refuse or fail to offer an exculpatory answer, for conduct that might not even be presumptively wrongful. Much might depend on the interpretation of “reasonable suspicion.” But since on any plausible reading suspicion can fall well short of belief, the point of principle remains: while reasonable suspicion of criminal intent or involvement can warrant further investigation and questioning, it should not suffice to ground a criminal charge. We should have to answer, criminally, only for proved conduct that is at least presumptively wrongful, since only then does the criminal law, concerned as it is with public wrongs, have any proper interest in our conduct; but not even a reasonable suspicion of wrongful intent can suffice to ground a presumption of wrongfulness.

It will no doubt be argued that even if the letter of the Terrorism Act is too broad, we can rely on the police and prosecuting authorities to apply it with a sensible eye to its spirit, and to prosecute only when there really is a substantive case to answer: only when it would be reasonable to presume terrorist purposes in the absence of an exculpatory explanation. However, we should not be so ready to allow such discretionary power to officials. There is a clear danger that the police will apply the Terrorism Act in discriminatory ways, treating race or religion as circumstances that can render suspicions reasonable; there is the danger that the prospect of being charged under the Terrorism Act will be used as a way of pressuring individuals who might have useful information. Furthermore, apart from these dangers of abuse, there is the principled demand that the criminal law should be transparent: that its offense definitions should specify the kinds of conduct that are properly to be condemned as public wrongs, rather than specifying some much broader category that includes conduct that might be wholly innocent.

That is not to say, however, that strict criminal responsibility can never be justified. There are two kinds of case in which it can in principle be justified. The first kind of case is exemplified by the shopkeeper who sells unfit food, and the factory owner whose machinery injures an employee. Each is engaged in an activity that creates particular risks of serious harm, an activity that, while perhaps socially beneficial, is optional. We can therefore reasonably impose on them a stringent legal duty of care, to ensure that their operations are as safe as is reasonably practicable. We can also impose, as part of that duty of care, a duty to assure themselves and others that they are taking such care, by putting in place appropriate safety procedures, and making sure that they can demonstrate that such procedures are in place, and are functioning: an important part of taking care can be to assure ourselves and others that we are doing so.⁴⁴ The law can also, with the assistance of experts, specify at least in partial outline the kinds of safety provision that will be appropriate and adequate. Given such a duty of care, we can then reasonably demand that the agent answer publicly not merely for harms that she is proved to have caused recklessly or negligently, but for any harm that arises from her operations. Given the duty of assurance, and the duty that it generates to put in place verifiable safety procedures, we can also reasonably lay on the agent the probative burden of at least adducing plausible evidence that she had taken all due care: if she has taken due care, and put in place the appropriate procedures, she will normally be able very easily to show that she has. We can properly say through the law that she owes it to her fellow citizens to assure them that she took all reasonable care: if she cannot do so, by adducing evidence of the safety procedures that she operated, that proves that she was not taking due care; if she refuses to do so, she cannot complain if she is convicted. Proof that the harm arose from her activity thus justifies a legal presumption that she was not taking due care: the court is entitled to convict her unless she adduces suitable evidence that she did take due care. Normally, suitable evidence will indeed amount to proof, at least on the balance of probabilities, that she had taken due care: she just needs to produce the records that she kept of the operation of her safety procedures, records that would have been kept as part of those safety procedures. Sometimes, however, a wholly non-culpable defendant might not be able to provide such proof: perhaps a fire in the shop or the factory has destroyed the records. That is why, if we are to respect the presumption of innocence, the probative burden that is laid on the defendant should be evidential rather than persuasive: not to prove that she had taken due care, but to adduce evidence that suffices to create a reasonable doubt.

The other kind of case in which criminal responsibility could, in principle, be justifiably strict is exemplified by the agent who sexually penetrates a sleeping or unconscious person. It might not be reasonable to take the fact of unconsciousness by itself to create a presumption of lack of consent, unless the person concerned also states, after the event, that the penetration was non-consensual, though it is hard to imagine a case being prosecuted in which that was not so; but we can focus here on the presumption that the defendant did not act on a reasonable belief that the victim consented. What is it at stake here is not so much the risk of harm as the risk of wrong. The victim’s unconsciousness puts the agent on notice that what he intends to do might well constitute the serious criminal wrong of rape. We can therefore, quite reasonably, place on the agent the responsibility not just to refrain from what he knows to be rape, but to ensure that he has the victim’s consent, and to be able to assure both the victim and others that, since the context raised a real question about consent, he had made a reasonable effort to avoid committing that wrong. The conduct proved against the defendant, the sexual penetration of someone whom he knew to be unconscious, and who now says that it was non-consensual, raises a serious issue about its wrongfulness. Indeed, we can say that it constitutes a presumptive wrong: we can legitimately presume both that it amounted to the wrong of rape and that the defendant culpably committed that wrong. The onus then properly shifts onto the defendant, to offer evidence either that the penetration was consensual, that the victim is lying, or that he did at least act in the reasonable belief that the victim consented. There are actually grounds to go further than this, and so formulate the law that proof of non-consensual sexual penetration, whether the victim was conscious or unconscious, creates a presumption that the defendant is guilty of rape, as having acted without a reasonable belief that the victim consented. That would be to make criminal responsibility for rape strict, but this suggestion cannot be pursued here.

Across most of the criminal law, both liability and responsibility are properly non-strict. The presumption of innocence is rebutted, and transformed into a presumption of guilt that it is up to the defendant to rebut by offering an exculpatory explanation of his conduct, only by proof of both actus reus and mens rea. Sometimes, however, given the particular dangers of harm or of moral wrongdoing involved in his activity, we can reasonably make responsibility strict and can reasonably require the defendant to adduce evidence of his lack of culpability or mens rea if he is to avoid conviction.

By attending to the distinction between liability and responsibility, and by understanding responsibility as a matter of answerability, we can throw new light on the structures of criminal liability, on the way in which criminal responsibility normally differs from moral responsibility, and on the somewhat neglected phenomena of strict responsibility in ordinary moral contexts and in the criminal law. Much work remains to be done, within criminal law theory, to work out whether and when strict responsibility is justifiable, as well as to work out an adequate account of excuses; the aim here has been to show the direction that such work should take, by clarifying the logical structure that needs to be explained and fleshed out.