I. Introduction
According to an intuitively appealing and widely accepted view, we should endorse the Innocence Constraint (IC). This holds that using the criminal law in response to innocent (morally permissible) actions or against innocent (non-culpable) agents is wrong.Footnote 1 The scope of this claim is contested, but it seems that it is meant to apply always,Footnote 2 or at least almost always.Footnote 3 The almost qualification appears to be limited to extreme cases in which using the criminal law against innocents is permissible where it is the only way to prevent a “catastrophic moral horror”.Footnote 4
Despite its influence, I will argue that the IC is false. While there is a (sometimes strong) pro tanto reason against imposing undeserved harm through the criminal law, this reason is not (almost) always decisive; it does not defeat every other reason, or combination of reasons, in every possible case where there is no risk of a “catastrophic moral horror”. Rather, it may be defeated by clashing reasons in more common cases, for example, when using the criminal law against innocents is the only way to prevent serious harm to other innocent persons and the legal sanction is mild, such as a small fine.
I consider four related arguments against the IC. First, a debunking argument that suggests that the intuitive appeal of the IC may be due to a conflation of different ideas. Second, a counterexamples argument, which points out that there are many cases in which using the criminal law against innocent actions or agents is justified. Third, a theoretical argument, according to which the force of the reasons for and against using the criminal law against innocents is a matter of degree, and it is therefore implausible to hold that this is never justified. Fourth, an analogy argument, which holds that it is also implausible to maintain that harming innocents is often justified in other contexts but (almost) never in the context of the criminal law.Footnote 5
The focus of these arguments is illegal actions that are morally permissible. This type of action is different from two more familiar types of action whose moral status is evaluated vis a vis their legal status. The first is “mala in se” (MS) actions, which are wrong independently of the law (and thus prior to the law). An example is harming an innocent person for a trivial reason. The second is “mala prohibita” (MP) actions, which are not morally wrong independently of the law (and prior to the law), but are morally wrong due to the law (and thus after the law). An example is not paying tax. This action is wrong, when it is wrong, only due to the law,Footnote 6 because the duty to pay a tax requires the existence of a tax system established by the law. This common contrast between MS actions and MP actionsFootnote 7 overlooks the first type of actions that are not morally wrong at all – even taking the positive law into account. I will refer to this neglected and (thus far) nameless category as “non-mala” (NM) actions.
(I define the categories of MS, MP and NM in terms of (the moral status of) actions: MS actions are wrong independently of the law, MP actions are wrong due to the law, and NM actions are not wrong even given the law. Other authors typically define these categories in terms of the definitions of criminal offences. My approach is more informative because criminal prohibitions typically cover all the relevant types of actions: MS actions, MP actions, and NM actions.Footnote 8 Therefore, defining these categories in terms of offences is coherent only if offences are classified as MS to the extent to which they apply to actions that are wrong independently of the law; as MP to the extent to which they apply to actions that are wrong due to the law; and as NM to the extent to which they apply to actions that are not wrong even given the law.)
II. The Significance of Non-mala Actions
A familiar concern regarding the IC is that it may not be compatible with using the criminal law in response to MP actions, although doing so is sometimes justified. For example, since not paying taxes is not wrong until a law that establishes the tax system is enacted, such a law appears to constitute a counterexample to the IC. Yet the more important challenge to the IC concerns NM actions that appear to be a proper target of the criminal law. This is the case for two related reasons. First, while MP actions are not morally wrong independently of, and before, legal regulation, they are wrong due to the law, once it is enacted. In contrast, NM actions are not wrong even due to (and after) the law. Thus, people who perform NM actions are innocent in a more robust way than those who perform MP actions.
Second, in response to some counterexamples arguments against the IC,Footnote 9 it has been suggested that these arguments threaten a version of the IC whose focus is (also) the enactment of criminal prohibitions (“formal criminalisation”),Footnote 10 but not a narrower version of the IC that applies only when criminal prohibitions are enforced in specific cases (“substantive criminalisation”).Footnote 11 The narrower version is compatible with enforcing the criminal law against MP actions if this is done after the enactment of offences that prohibit these actions, as is often the case, because MP actions are wrong at this time. However, even this narrower version of the IC is incompatible with enforcing the criminal law against NM actions, since such actions are not wrong even given (and after) the law. Thus, even if MP actions are not counterexamples to the narrow version of the IC, NM actions are. Consequently, the most important distinction in this respect is not between MS and MP, but rather between MS and MP, on the one hand, and NM, on the other hand. It is therefore unfortunate that relevant discussions do not highlight the significance of NM actions but rather use the term “mala prohibita” in a way that refers to every action that is not MS, without distinguishing between MP actions and NM actions.
My arguments against the IC, which focus on NM actions, apply accordingly to all versions of the IC in terms of the actions that they cover. This includes both a wide version that applies to all of the actions that are performed in the framework of the criminal law, including legislation, investigation, prosecution, conviction, and punishment, and narrower versions that apply only to some of these actions.
The fact that the category of NM actions is typically not discussed explicitly may be due to doubts concerning its existence or prevalence. However, I argue that there are cases of NM and that such cases are common even when the criminal law is used in a way that is morally justified. This conclusion is entailed by the following claims that, I believe, are all plausible and widely accepted. First, criminal offences are often overinclusive in the sense that they apply also to actions that are not wrong (even if their ultimate target is only wrongful actions). This includes offences whose rationale is a principle (a prohibition on “unjustified risks” for example) but that nevertheless take the form of a rule (one that proscribes exceeding a specific speed limit, for instance). Second, such offences are sometimes justified, since there are often conclusive reasons against distinguishing or otherwise separating the wrongful and the permissible actions. Third, we should consider specific actions rather than types of actions in this regard and accordingly reject the objection that the relevant actions are wrong since they are instances of types of actions that are wrong (perhaps because most of the specific actions that they include are wrong). The focus should be on specific actions since an action may not be wrong even if it is, in some sense, an instance of a type that is wrong. Thus, considering types of actions in this context conflates actions whose moral status is distinct: actions that are wrong and actions that are not wrong. In this respect, every way of classifying actions into types is arbitrary. Finally, even if there are sometimes reasons to obey the law, these reasons do not exist in (almost) every case, and, when they do exist, they are not (almost) always decisive. This is the case also regarding legal rules that are morally justified: while the reasons in favour of obeying the law may be sensitive to the distinction between justified and unjustified law, this is not the only relevant factor. Accordingly, there are sometimes no reasons, let alone decisive ones, to obey laws that are morally justified. Indeed, this may be the case precisely when a law is justified despite the fact that it is overinclusive in the sense that it prohibits actions that are not wrong. Therefore, the objection that, due to such reasons, there are no cases of NM but only cases of MP, fails.
The inquiry regarding the justification of the IC and of criminal offences that apply to NM actions has important theoretical and practical implications. For example, the IC plays a role in influential arguments, most notably the familiar argument that consequentialism is misguided since it entails what the IC denies, namely, that it is sometimes permissible to punish innocents.Footnote 12 A common consequentialist reply to this argument contests the premise that consequentialism entails this conclusion. However, this reply is doubtful. It is difficult to deny that plausible and common forms of consequentialism entail that using the criminal law against innocents may be justified.Footnote 13 Therefore, it seems that the more promising response must contest the other premise, namely, the IC.Footnote 14 The practical importance of considering whether the IC is plausible is due to the fact that a vast (and expanding) part of contemporary criminal law appears to be overinclusive in the sense that it applies to innocent actions and agents.Footnote 15 Many offences that belong to this part of the criminal law may be unjustified,Footnote 16 but, I argue, this is not merely because they are incompatible with the IC.
In what follows, I consider the arguments against the IC. I begin with several clarifications regarding the question under consideration that suggest that the intuitive appeal of the IC is misleading (Section III). I then consider the counterexamples argument (Section IV), the theoretical argument (Section V), and the analogy argument (Section VI). Finally, I consider the implications of the conclusion that these arguments are sound (Section VII).
III. The Innocence Constraint and Its Intuitive Appeal
I begin with a few additional clarifications regarding the IC and the arguments against it. These clarifications are important, partly because they suggest a debunking argument against the IC: that its intuitive appeal is due, at least in part, to a conflation of different ideas.
The first point that should be emphasised is that my focus is on the IC as a foundational moral standard, which reflects an intrinsic concern, rather than as a pragmatic standard that takes into account also instrumental considerations. A law that prohibits the use of criminal law against innocent actions or agents is a pragmatic standard.Footnote 17 Since the content of pragmatic standards, and accordingly the relation between such standards and foundational ones, depends on contingent facts, it is plausible to assume that there are some cases in which they diverge. Specifically, a foundational constraint that is not (even almost) absolute may entail a pragmatic constraint that is absolute. For example, if legal officials tend to use the criminal law too much, especially against innocents,Footnote 18 a pragmatic constraint (in the form of a law) that instructs them never to use the criminal law against innocents, may have the best results overall if it reduces the use of the criminal law against innocents to the appropriate level (or closer to that level than any alternative instruction). Similarly, if using the criminal law against innocents would have negative effects in terms of the degree to which people trust the legal system, a pragmatic prohibition on using the criminal law against innocents may be justified for this reason.Footnote 19 My arguments are not directed against such pragmatic rules. This is noteworthy since some of the intuitive appeal of the IC may be due to its conflation with a pragmatic standard with similar content.
Another important distinction is between standards that refer specifically to the law, such as the IC, and standards that are more general, for example, a constraint against undeserved harm that applies also – but not only – to harm that is imposed by legal means. One objection to the IC is that foundational moral standards do not refer to the law or a specific legal field such as the criminal law, as the IC does. This is because foundational moral standards do not include what is merely a means as a constitutive element and the law is merely a means.Footnote 20 However, the arguments that I emphasise in this paper apply also against more general constraints that do not refer specifically to the law, such as constraints against actions that involve undeserved harm or punishment that is not necessarily legal punishment.
The third distinction is between two senses of innocence: one that focuses on the moral status of actions and another that focuses on the moral status of persons. The IC may reflect a wrongfulness constraint, which holds that it is (almost) always wrong to use the criminal law in response to innocent actions (actions that are not wrong),Footnote 21 or a blame constraint, which holds that it is (almost) always wrong to employ the criminal law against innocent persons (who are not blameworthy).Footnote 22 These constraints can come apart. A person who performed a wrongful action may not be blameworthy, for example, if the action was done under coercion or due to an exculpatory mistake.Footnote 23 Conversely, a person may be blameworthy even if she did not perform a wrongful action, for instance, in light of her intentions or because of an inculpatory mistake.Footnote 24 It is not always clear which of these constraints is invoked by authors, although the wrongfulness constraint appears to be more common regarding the enactment of criminal offences, while the blame constraint is more salient with respect to the enforcement of the criminal law. Elsewhere, I argue that the moral status of actions is not important in itself at all for the law (and more generally to the response to these actions). The ultimate concern of the law should rather be the moral status of states of affairs (including consequences of actions) and of persons.Footnote 25 Thus, I believe that while there is a pro tanto reason against using the criminal law against people who are not blameworthy,Footnote 26 there is no pro tanto (foundational) reason against using the criminal law in response to actions that are not wrong. However, my arguments in this paper apply to both the wrongfulness constraint and the blame constraint. They demonstrate that there is no reason that is (almost) always decisive against using the criminal law against innocent actions or persons. Therefore, I often refer to the IC in its general form and assume, for simplicity, that the relevant agents are responsible for their actions such that when these actions are wrong, the agents are blameworthy, and when these actions are not wrong, the agents are not blameworthy.
The fourth important distinction is between two different senses of innocence in another respect: factual innocence and normative innocence. A person is innocent in the factual sense when she did not perform an action that is attributed to her (including a wrongful action). A person is innocent in the normative sense when an action that she did perform is not wrong.Footnote 27 Factual and normative innocence are both cases of moral innocence. What is important with respect to moral innocence is whether a person is responsible for an action that is morally wrong, and not whether she is responsible for any action or for an action that is attributed to her (by a legal official, for example). A proper declaration of innocence is “I did nothing wrong” rather than “I did nothing at all” or “I didn't do what you say that I did”. Since the IC is concerned with moral innocence, it should apply to both factual and normative innocence. My arguments against the IC are also general in this respect.
There are, however, two differences between factual and normative innocence. These differences do not affect the conclusion that both are instances of moral innocence, but they are important in other ways and, specifically, they may explain away at least some of the IC's intuitive appeal. The first difference is that a person who is factually innocent is not only morally innocent but also “legally innocent” – that is, she did not violate the law. In contrast, a normatively innocent agent is morally innocent but not necessarily legally innocent. This difference may be morally significant when there are reasons not to use the law against people who did not break the law. However, these reasons are different from the (desert-based) reason against punishing people who are morally innocent.Footnote 28 Using the criminal law in the case of factual innocence often involves a moral cost in terms of both types of reasons, whereas doing so in a case of normative innocence involves a cost only in terms of the reasons relating to moral innocence. Therefore, using the criminal law in the case of factual innocence is typically worse overall.
This difference may explain away the intuitive appeal of the IC. The IC may seem intuitively appealing, first, because of a conflation of moral innocence and “legal innocence”. Yet the answer to the question of whether a person violated the law is not, in itself, relevant to the question of moral innocence. It may be objected that this difference is relevant because it is wrong to violate the law. However, violating the law is not always wrong (overall), and the arguments that I consider focus on cases in which it is not.
The second difference between factual and normative innocence concerns the typical consequences of using the criminal law. Using the criminal law in paradigmatic cases of factual innocence involves severe punishment (and condemnation), for example when a person serves a long prison sentence or is executed for a crime that she did not commit (either by mistake or as a result of being framed).Footnote 29 In contrast, using the criminal law in the most common examples of normative innocence (against NM and MP actions) involves punishment (and condemnation) that is much less severe, for example a small fine.
This difference may also explain away some of the intuitive appeal of the IC. The IC may be intuitively appealing due to the common focus on the paradigmatic cases of factual innocence in which punishment (and condemnation) is severe. In this respect, the usual context of factual innocence is distracting, since the strong considerations against severe punishment (and condemnation), especially in the case of innocents, are not always present. The context of normative innocence is more illuminating in this respect since it highlights the fact that these considerations are not always very strong and may therefore be defeated by strong clashing considerations in favour of using the criminal law when this advances an important moral goal, such as preventing serious (undeserved) harm to other innocent persons.
These differences between factual and normative innocence suggest that the strong intuitive appeal of the IC may be due to the common focus on the paradigmatic but distracting cases of factual innocence as opposed to the more illuminating cases of normative innocence. Therefore, while the arguments against the IC apply to all types of innocence, it is easier to see their soundness when considering cases of normative innocence. Accordingly, the arguments that I consider against the IC focus on cases of normative rather than factual innocence. The suggestion that the IC is false also regarding factual innocence may seem (even more) counterintuitive than the suggestion that it should be rejected concerning normative innocence. It is therefore important to recall that this intuition may be due to features of factual innocence that do not pertain to the IC itself.
IV. The Counterexamples Argument: Non-mala Actions
The counterexamples argument holds that using the criminal law in response to actions that are not wrong, even given the law (NM actions), and against agents who are not blameworthy, is often justified. This argument thus demonstrates that there are often cases in which both the criminal law and actions that violate it are justified.Footnote 30 This is possible since employing the criminal law and violating it are different actions whose moral status may be different. Accordingly, the fact that a criminal prohibition is justified does not entail that its violation is wrong (namely, that it is MS or MP).
This argument is stronger than the analogous argument that focuses on MP actions, since MP actions are wrong once the relevant prohibition is in place. Yet this argument rests on premises that are, I believe, plausible and widely accepted. My aim here is not to provide an exhaustive defence of each of the necessary premises, but rather to point out that they are both reasonable and common. I begin by suggesting several cases that demonstrate that it is often justified to use the criminal law against NM actions and then explain why these are indeed counterexamples to the IC. The relevant actions are typically referred to as MP actions, but this may be due to a failure to make the significant distinction between MP and NM actions.
A. Examples
The content of criminal offences may not reflect the ultimate justification of these offences accurately due to pragmatic considerations relating to the cost of distinguishing or otherwise separating different types of actions or agents. For example, the ultimate justification of an offence may be actions that are wrong or harmful or agents who are blameworthy, but the offence may apply also to actions that are not wrong or harmful or to agents who are not blameworthy. This may be the case even regarding offences that are morally justified. In other words, there are offences whose rationale is a principle, but that nevertheless take the concrete form of a rule. In consequence, they are overinclusive. In what follows, I focus on offences that are directed at actions that are wrong but apply also to some permissible actions. However, analogous analysis applies also to offences that are directed at actions that are harmful or agents that are blameworthy.
Consider, for instance, a common speed limit offence. The reasoning that underlies such an offence is presumably the combination of the following assumptions. First, driving typically involves risks, which are sometimes justified and sometimes not. Second, whether or not a risk is justified depends on many variables, including the capabilities of the driver, the condition of the car and the road, the weather, the speed, and the reason for driving.Footnote 31 Third, the best way to prevent unjustified risks is often not (only) a criminal prohibition on “unjustified risks” (a principle)Footnote 32 but rather (also) one that proscribes exceeding a specific speed limit, say 60 mph, on a certain road, without referring to other pertinent variables (a rule).Footnote 33 This is because sometimes the considerations that support such a rule defeat the considerations in favour of a principle. For instance, a principle is more accurate than the rule since it applies only to wrongful actions. A rule, by contrast, is overinclusive in the sense that it captures some actions that are not wrong. But this consideration may be defeated by the advantage that rules often have over principles: they provide better guidance to drivers and to the officials who enforce the law, as it is often difficult to correctly determine what a principle requires in specific circumstances (for example, which actions involve unjustified risks). This is an advantage, for example, when better guidance prevents more cases of unjustified risks and thus more cases of wrongful (and undeserved) harms. Therefore, a speed limit prohibition in the form of a rule may be justified overall although it applies to actions that are morally permissible (NM actions).Footnote 34
Many other common criminal prohibitions reflect this type of reasoning.Footnote 35 For instance, the prohibition on driving with more than a certain level of alcohol in one's breath or bloodFootnote 36 may be justified although it applies to some actions that are not wrong because some drivers over the relevant limit can drive perfectly safely.Footnote 37 Similarly, the amount of tax that one is legally required to pay is often based only on some of the morally significant facts, for example, income, while ignoring others, such as overall wealth and needs. This may be justified sometimes, due to pragmatic considerations, and then the (justified) offence regarding non-payment of tax exceeds the parallel moral obligation and applies to some NM actions.
Indeed, criminal prohibitions may sometimes be justified even if most of the actions that they proscribe are morally permissible. For example, a specific speed limit offence may be justified even if only (say) 30 per cent. of the cases that it covers involve an unjustified risk, if having an overinclusive rule prevents (many) more accidents than a principle. Similarly, a prohibition on driving with a certain amount of alcohol in one's blood may be justified even if most of the violations of this prohibition do not involve unreasonable risk. Another example is an offence of crossing a border with more than a certain amount of money without reporting it. While the ultimate goal of such a prohibition may be to hinder wrongful activities, by detecting money that is obtained through such activities, in most of the cases to which it applies the money is not related to illicit activities. Moreover, if one did nothing wrong in obtaining or transferring the money, there may be sometimes reasons against reporting, for example, that reporting would lead to questioning and thus waste the resources (including time) of the relevant person and officials (resources that may be used better in other ways). Another example of this type is offences whose aim is to protect the environment from actions such as using inefficient cars. Sometimes, each of these actions is not harmful, and therefore not wrong, but together they cause serious harm. Therefore, if a criminal prohibition on such actions will prevent enough of them, and thus prevent serious harm, it may be justified although it applies to many actions that are not wrong.
A different example is a prohibition that is subject to an exception if a specified procedure is followed, for instance, a prohibition on euthanasia without official permission. Such a prohibition may be justified despite the fact that euthanasia is not wrong sometimes even if the required legal procedure is not followed, for example, when the pertinent facts are clear and following the procedure requires time and thus prolongs suffering. (Indeed, this is always the case if the moral status of actions does not depend on the beliefs of people and that the role of the procedure is therefore merely epistemic rather than constitutive.)
Yet another example is that of a poor person who steals money from a rich person in order to buy an expensive medicine that is necessary to save her life or prevent serious suffering. Such an action is justified if the following conditions are met: the harm to the rich person is trivial compared to the benefit to the poor person, the rich person is significantly better off than the poor person, the rich person does not deserve to be so much better off, and stealing from the rich person does not involve serious negative effects (it may be justified even if only some of these conditions are met). Still, a law that prohibits such an action (as positive law typically does) may be justified if the best way to transfer (some) resources from the rich to the poor is not such self-help, but rather taxation, because there is a substantial danger that self-help may be abused or have (other) negative effects. Yet even when these concerns justify such a legal prohibition, some specific violations of this prohibition may be morally justified, since the considerations against an explicit and general legal permission do not necessarily apply to every single violation.
This reasoning often appears to underlie positive law, which typically does not include exceptions that permit the actions in the above examples, for instance, in the framework of justificatory defences, such as “lesser evil” and “necessity”. Claims that actions such as the ones considered in the above examples are covered by these defences are usually rejected, presumably not because all of these actions are assumed to be wrong, but due to other reasons, such as concerns regarding self-help. Such concerns were often emphasised especially in English law, which, as a result, includes a very limited defence of (justificatory) “necessity”.Footnote 38
These cases constitute counterexamples to the IC – in which the criminal law justifiably applies to NM actions as well as MS and MP actions – if the following assumptions are true. The first assumption is that the criminal offences in these examples are justified. The second assumption is that these offences apply to (some) innocent actions (and agents). Finally, the third assumption is that these cases are not very rare – and thus refute even the weaker version of the IC, which holds that using the criminal law against innocents is almost always wrong. I believe that these assumptions are correct and widely accepted, or at least entailed by assumptions that are widely accepted.
The last assumption – that the relevant cases are not very rare – does not seem to be controversial. Indeed, the examples cover laws and actions that are very common. I defend the first assumption – that the relevant examples refer to criminal offences that are morally justified – more thoroughly in Section V, but this assumption seems independently plausible and widely accepted regarding at least some of the relevant examples. There may be some disputes regarding a few of them, mainly concerning the scope of the offences under consideration. For example, it may be controversial if a certain speed limit should be 60 mph or 50 mph, or if a certain speed limit is justified if its violation creates an unjustified risk in 60 per cent. of the cases or is it sufficient that its violations are unjustified in 30 per cent. of the cases. And there are of course more controversial cases of criminal offences that display a similar type of reasoning. One such example includes various preparatory offences, possession offences, and offences of risk imposition.Footnote 39 Another example is the criminal offence of sexual activity with a child that covers also a person below the age of 18 who touches a person below the age of 16 in a way that is “sexual”.Footnote 40 However, it seems undisputable that at least some of the above examples are justified. The main controversy thus seems to focus on the second premise – that these examples cover actions that are not morally wrong (even when they are illegal). Disputing this assumption appears to be the main way to try and reconcile the IC and the use of the criminal law in these examples. However, I think that this attempt fails. In what follows, I consider several objections to the claim that the actions covered by the relevant offences are not wrong.
B. Enactment and Enforcement
One objection acknowledges that the above cases are counterexamples to the IC in its wider form – the one that refers to the enactment of criminal offences – but insists that they are not counterexamples to the narrow version of the IC – which applies only when criminal prohibitions are enforced. That is, it may be thought that while it is justified to enact the relevant prohibitions, despite their overinclusive nature, it is unjustified to enforce them when they apply to innocent actions or agents.
Duff offers a detailed version of this objection. He acknowledges that there are sometimes decisive reasons to enact offences that apply to actions that are not wrong. However, he suggests that these reasons do not apply to the enforcement of these offences with respect to such actions. Accordingly, he envisions a criminal law that includes overinclusive offences that are not enforced against innocent actions. Duff suggests also that the people to whom these offences apply are not required to comply with them when they believe that the actions that are covered by these offences are not wrong. Instead, he suggests, they may violate the prohibitions in these cases and try and convince the enforcement officials that their actions are indeed not wrong and therefore that the law should not be enforced in response to these actions. Duff thus believes that there are counterexamples only to the wide version of the IC that applies to the enactment of offences but not to the one that forbids only the enforcement of these offences against innocent actions.Footnote 41
I think that this objection is unpersuasive because it is often justified not only to enact the relevant prohibitions but also to enforce them. To be sure, enacting and enforcing criminal prohibitions are different actions, and the reasons that apply to them and their overall moral status may be different. Indeed, even the reasons that apply to distinct means of enforcement may be different. Not every prohibition that it is justified to enact should be enforced in the same way with respect to every violation. This is especially evident regarding overinclusive offences that apply to innocent actions and agents. The reasons in favour of enacting such offences – for example, the cost of distinguishing wrongful and non-wrongful actions (or culpable and non-culpable agents) – do not always support, at all or with the same force, enforcing these offences in particular cases. An example is where it is clear enough to the relevant official that a particular action is not wrong. However, it does not seem reasonable to assume that the reasons that sometimes justify overinclusive offences support only the enactment of such offences and never their enforcement. It seems rather that there are sometimes reasons, and occasionally decisive reasons, not only to enact but also to enforce the relevant prohibitions even in response to actions that are not wrong (and agents who are not culpable). This may be necessary sometimes, for example, to maintain the credibility of the relevant prohibitions. This seems especially probable once we recall the range of criminal offences that are overinclusive in the relevant sense, and if the enforcement policy is not kept completely secret (Duff believes that it should not).Footnote 42
Consider, for instance, a speed limit offence. Assume that there are decisive reasons to enact such an offence despite its overinclusive nature, in light of the difficulty of separating the cases in which driving at a certain speed creates a reasonable risk and the cases in which it creates an unreasonable risk. It seems that this difficulty applies, at least to some degree, also when people are considering whether the violation of such a prohibition is justified, and when officials consider when its enforcement is appropriate. When such a prohibition is the best means of saving innocent lives and enforcing it in response to actions that are not wrong is the only means of maintaining its credibility, then enforcement appears to be justified, at least if the punishment is not very severe. I see no reason to assume that these conditions are (almost) never met. According to Duff's account, such a speed limit offence should be regarded by drivers merely as a recommendation and should be enforced only when the violation is wrong. The hypothesis that such a combination would be (almost) always sufficiently effective seems overly optimistic. Indeed, it is hard to see what could support such optimism apart from the need to preserve the IC at all costs.
Duff's view is especially ambitious in that he claims that there are not even pro tanto reasons to enforce the law in response to actions that are not wrong – and not only that such reasons are not decisive.Footnote 43 This claim is especially puzzling. Consider the following argument:
1. Some states of affairs are (overall) better than others, for example, when a virtuous person flourishes compared to when she suffers (all else being equal).
2. There is a pro tanto reason to act in a way that brings about a state of affairs that is (overall) better rather than worse; for example, to prevent the suffering of a virtuous person.
3. It is sometimes possible to bring about a better state of affairs, for example, to prevent the suffering of a virtuous person, by enforcing a criminal offence with respect to an action that is not wrong (for instance, by enforcing a speed limit offence in a way that, overall, prevents accidents).
4. Therefore, there is sometimes a pro tanto reason to enforce criminal offences with respect to actions that are not wrong.Footnote 44
This argument seems conclusive. Since this reason is, I have argued, sometimes decisive, at least some of the above cases appear to be counterexamples also to the narrow version of the IC that applies only when the criminal law is enforced.
C. Types and Tokens
Another objection is that the counterexamples argument assumes that we should consider types of actions – rather than specific actions – when evaluating the moral status of actions.Footnote 45 If we refer to types of actions, the objection continues, the relevant action types are wrong. We should, however, focus on specific actions rather than on types of actions in this respect.Footnote 46 For the question is whether we should apply the criminal law to actions that are morally permissible and an action may be permissible even if it is, in some sense, an instance of an action type that is wrong.
Moreover, it is difficult to cash out the suggestion that we should consider types of actions rather than specific actions when applying the IC in a way that is not arbitrary. Consider, for example, the speed limit case. In this case, the relevant type of action is presumably not the one that reflects the rationale of the prohibition, namely, driving in a way that involves an unjustified risk of harm to others. For this is not the prohibition under consideration here (which takes the form of a rule rather than a principle). The relevant action type is also not driving at a speed that exceeds a certain limit, since this type of action is not in itself wrong. It may be objected that the latter type is wrong if most of its specific tokens are wrong, assuming that they involve an unjustified risk of harm. This objection entails that a prohibition is unjustified if most of the actions to which it applies are not wrong – and this, I have argued, is implausible. More importantly, this and any other pertinent suggestion of classifying actions into types is arbitrary because it conflates actions whose moral status is distinct: actions that are wrong (those that involve an unjustified risk) and actions that are not wrong (those that do not involve unjustified risks). Indeed, criminal prohibitions typically cover all the relevant types of actions: MS actions, MP actions, and NM actions. For instance, the speed limit prohibition covers actions that create unjustified risks (MS), actions that are wrong because the violation of the law itself has bad consequences (MP), and actions that involve justified risks and have no other bad consequences (NM).
D. The Reason to Obey the Law
A different objection is that the actions that are covered by the relevant examples are wrong since they violate a duty to obey the law, or at least laws that are morally justified. If there is such a duty, there are no NM actions as every violation of the law is wrong – either independently of the law (MS) or due to the law (MP).
However, the claim that there is (almost) always a conclusive duty to obey the law is implausible. This is the case also regarding legal rules that are morally justified: even if the reasons to obey the law are sensitive to the moral status of the law in question, such that there are stronger reasons to obey justified laws, these reasons are not (almost) always decisive, since there may be reasons against doing so, particularly regarding laws that are overinclusive. Indeed, it seems that no one who has contributed to the literature on this topic endorses the opposite claim. The consensus is rather that even if there is sometimes a pro tanto reason to obey the law, this reason does not exist in (almost) every case, and, when it does exist, it is not (almost) always decisive.Footnote 47 Specifically, those who endorse the IC must assume that this is the case, for otherwise the IC would be trivial, as it would be impossible to violate it.Footnote 48
Consider, for example, the common suggestion that violating the law is wrong since it has various (often indirect and long term) negative effects. For example, even when driving in violation of the law does not directly involve an unjustified risk, it may encourage less competent drivers to drive in a way that does involve such a risk.Footnote 49 This may well be sometimes the case, but there is no reason to assume that such effects are (almost) always present and are (almost) always decisive.Footnote 50 The same is true regarding the claim that driving over the legal speed limit is wrong since it creates an unjustified risk because other drivers drive based on the assumption that everyone obeys the law.Footnote 51 This concern is often reasonable but it is not always decisive.
Another influential argument is that fairness requires that people who benefit from the law should bear its burdens by obeying it.Footnote 52 This argument is reasonable to the extent to which it is entailed by (the balance of) general considerations of distributive and retributive justice. It seems unreasonable, however, if it deviates from such general considerations. For there is no reason to assume that the ultimate criterion for what is just (or fair) with respect to the law (or other social practices) is different from the ultimate criterion for what is just (or fair) in other contexts. The content of general considerations of distributive and retributive justice is of course controversial, but it does not seem that plausible versions of these considerations entail a decisive reason to (almost) always obey the law – including justified but overinclusive laws. This is the case even if the situation is completely just (for example, no one is unjustifiably worse off), since it is doubtful if there is always a decisive reason to obey the law when this benefits no one, namely when violating a legal rule has no negative effects (so that it does not undermine the benefits of the law). That is sometimes the case with respect to overinclusive laws.
In addition to these general arguments, several more specific arguments for the conclusion that there is a decisive reason to obey the law focus on overinclusive criminal prohibitions. According to one argument of this type, violating the law, at least justified prohibitions such as one that sets a speed limit, is wrong since “obedience to the regulations serves a valuable assurance function – it displays to other road users my commitment to respecting the rules of this practice and to driving safely”.Footnote 53 This concern may require obedience to the law in some cases. But it does not apply to every violation of the law, for example, when others do not know that one is violating the law. And even when it does apply, it may not be decisive.
Another argument is that since we are fallible, we should “defer to laws that emerge from a democratic deliberation aimed in good faith at serving the polity's common good”.Footnote 54 Yet, first, not all justified prohibitions meet these conditions. And, second, the combination of fallibility, democracy, “good faith”, and the “common good” does not entail a reason to obey even laws that meet these conditions in every case. Most of these factors (all but the fallibility concern) support only the hypothesis that the relevant prohibitions are (often) justified. However, they may be justified despite being overinclusive, as even fallible agents are not always wrong. And, even if the combination of these factors supports obedience sometimes, it is not always conclusive.
According to another argument, using the criminal law against people whose actions are not wrong does not “violate any principle of procedural justice, because we understand that fair notice of what is illegal is sufficient to justify any of the vast range of laws that are drafted with sufficient breadth that they catch some people who know they are breaking the law but have otherwise done nothing inherently and fundamentally wicked”.Footnote 55 The suggestion appears to be that “fair notice” that a certain action is illegal entails that the action is wrong. However, an action is not wrong merely because it is declared to be illegal.
An additional argument is that the IC is compatible with the use of the criminal law in the above examples if the legal regulation is comprised of two parts: non-criminal legal rules (for example, tax, property, traffic or licensing regulations), and criminal prohibitions that forbid the violation of these non-criminal rules. This combination solves the problem, it is argued, when two assumptions hold: the IC requires only that the relevant action is wrong independently of the criminal law rather than the law in general; and there is a duty to obey the non-criminal rules (at least when these rules are morally justified). When these assumptions hold, the criminal law is used only against actions that are already wrong due to the (non-criminal) law.Footnote 56 One doubt concerning this suggestion is that the difference between such a combination of non-criminal and criminal regulation and a system that includes just one criminal element does not seem to be morally significant in terms of the concerns that underlie the IC. But even if this difference is important, this solution succeeds only if there is (almost) always a decisive duty to obey the relevant laws (namely, only assuming that the relevant actions are MP rather than NM), and this assumption begs the question against the counterexamples argument, the focus of which is NM actions.
Several other objections rely on the view that the moral status of actions depends not on the morally significant facts but rather on the mental state of the agents (for example, their justified beliefs) regarding these facts. These objections all fail if the best account of the moral status of actions is not subjective in this sense.Footnote 57 However, they fail even if it is subjective, because the premise that the relevant actions are often not wrong is true even according to common subjective accounts. For the agents in these examples may justifiably believe that the pertinent facts (such as their driving ability) obtain. Consider, for example, the objection that “people should not try to decide for themselves whether what they would like to do is safe, since they cannot be trusted (and should not trust themselves)”.Footnote 58 While the danger of overestimating one's abilities is real – indeed, it is one of the reasons that support overinclusive prohibitions that apply to some actions that are morally permissible – it does not follow that it is (even almost) never justified to believe that one's action is justified.
According to a similar objection, while a person may sometimes know that his action does not involve an unjustified risk, “he does not know that he knows this, and therefore cannot justifiably claim to be sure that he is not endangering [a relevant] interest”.Footnote 59 It is unclear why it is assumed that it is impossible or rare to know that one knows the relevant facts. (The fact that it may be difficult to verify the veracity of such claims in a way that ensures that the law does not become unenforceable is irrelevant in this regard.) Regardless, this objection fails since it assumes that knowledge that one's action does not involve an unjustified risk is not sufficient in order to justify the action. However, because such knowledge (or even justified belief) is sufficient even according to standard subjective accounts, the assumption that an action is justified only if the agent knows that he knows the relevant facts (or only if he is certain that his action involves no risk) is ad-hoc.Footnote 60
E. The Nature and Scope of the Criminal Law
Finally, some claims regarding the nature and scope of the criminal law suggest two related objections to the counterexamples argument. Consider, for example, the following statements about criminal punishment: it is “conceptually” true that punishment connotes blameFootnote 61 and wrongdoingFootnote 62; censure is an element of analytical accounts of criminal punishmentFootnote 63; and punishment is a desert-based sanction.Footnote 64 Similar claims are made about criminalisation: that it constitutes a declaration that certain actions are morally wrongFootnote 65; and that its “very meaning … would be undermined if a court added that the convicted defendant had done nothing wrong and was wholly non-culpable”.Footnote 66 And conviction is said to convey a message of culpable wrongdoing.Footnote 67 Relatedly, some, including certain English judges, distinguish between “real” or “true” crimes and “quasi” crimes, such as “public welfare” and strict liability offences, based on whether the relevant offences apply to actions that are wrong or involve stigma.Footnote 68 Others regret that English law (as opposed to German Law, for example) does not formally distinguish between offences in this way.Footnote 69
Some of these assertions are thus explicitly conceptual (those about the meaning or the analytical properties of the criminal law). Other assertions may refer to common beliefs regarding the nature of the criminal law (of the officials who make or enforce the law, or of the people to whom the law applies). But it is clear that these assertions reflect the implicit assumption that the relevant beliefs are correct.
These assertions reflect the claims that offences that apply to innocent actions or agents are not (or not “really”) criminal – by definition or based on the common understanding of the nature of criminal law. These claims are the basis for two related objections to the counterexamples argument. The first follows directly from the conceptual claims: if the relevant examples do not refer to laws that are criminal, they do not refute the IC, which governs only criminal offences. The second objection is normative: the criminal law implies (given its definition or common understanding) that the actions and agents to whom it applies are wrong and blameworthy, respectively, and using it against innocent actions or agents accordingly involves a lie (and defamation) – and this is wrong.Footnote 70
These objections are unpersuasive. First, the claims that offences that apply to innocent actions or agents are not (“really”) criminal are exaggerated. Second, and more importantly, even if these claims were correct, they do not entail objections that undermine the counterexamples argument.
Consider the first, conceptual question, first. The claims regarding the nature of the criminal law reflect several correct observations. One is that the criminal law includes an important element of condemnation. Another is that there are important differences within the criminal law, including between laws that regulate NM and MP actions and those that regulate MS actions, in terms of condemnation and punishment (among other things), although these differences are typically of degree and are not always present.Footnote 71 These observations suggest an important lesson (that I emphasise as part of other arguments against the IC): that the boundaries between different legal fields, and indeed between the law and other practices and actions, are often morally arbitrary. Therefore, no single standard is appropriate regarding all criminal offences and only these offences. Accordingly, standards that apply to the criminal law, such as the IC, should focus not on the legal classification of actions but on other factors, for example, whether they involve undeserved harm or punishment (rather than on if the punishment is legal or criminal). Therefore, foundational concerns do not depend on what is classified as the criminal law. These important observations and lessons do not, however, entail that laws that apply to innocent actions or agents are not (“really”) criminal, but rather that such definitional claims regarding the nature of the criminal law are misguided.
Moreover, to the extent to which the relevant claims appeal to the common understanding of the criminal law, they appear to be inaccurate. While no one disputes that the criminal law includes an important element of condemnation, it does not follow that its common understanding is that it never applies to innocent actions or agents. It seems rather that most people acknowledge that criminal offences are sometimes overinclusive in the sense that they apply also to some innocent actions and agents. Moreover, it seems that many understand also why this is the case, namely, the reasons that sometimes justify this.Footnote 72 Therefore, the close relation between the criminal law and condemnation does not entail that laws that apply to innocent actions and agents are not criminal or involve a lie.
These conclusions are reinforced by two additional observations. First, the legal prohibitions that apply to NM and MP actions are often the same prohibitions that apply to MS actions. For example, a speed limit offence covers both MS actions that are wrong independently of the law (those that create unjustified risks of harm to others) and NM actions that are not wrong (those that do not involve such risks). Second, some of the examples discussed above refer to prohibitions that are typically considered part of the core of the criminal law, such as the prohibition on theft, which is typically identified as an offence that covers MS actions. One example of this sort is the case of the poor person who steals from the rich in order to save her life. A more contentious example is the House of Lords’ ruling in Hinks Footnote 73 that persuading a person of “limited intelligence” to give a significant amount of his money as a gift to his friend and carer constitutes theft even if the transaction is unimpeachable in terms of civil law. The questions of whether this action is wrong (and the agent is blameworthy) and if the ruling is justified – as well as the relationship between these questions – are controversial.Footnote 74 But the ruling may be correct even if the action is not wrong (for example, since conviction may deter similar actions that are wrong).
Moreover, even if the above claims regarding the nature of the criminal law were correct, they do not entail that the counterexamples argument is unsound. The first objection – that the relevant examples do not refer to laws that are criminal, and are therefore irrelevant to the IC – is misguided since my ultimate concern is normative rather than conceptual. I argue that some actions that impose sanctions on innocent agents in response to innocent actions are justified overall despite the fact that they involve a cost that is morally significant, namely, undeserved harm. In this respect, it does not matter whether these sanctions are properly classified as criminal. This is because the way in which the criminal law is understood (by laypeople, lawyers, or (other) legal officials) is irrelevant to the normative claim that laws which impose undeserved harm are nevertheless justified sometimes.Footnote 75
Furthermore, if legal rules that apply to innocent actions or agents are not criminal (in the sense that is pertinent to the IC), the IC is trivial, since it is impossible to violate it. It is therefore not surprising that prominent adherents of the IC do not rely on these claims regarding the nature and scope of the criminal law, but rather acknowledge that the IC applies in the relevant examples and try to accommodate these examples in other ways.Footnote 76
The second objection – that using the criminal law against innocent actions or agents is wrong because it involves a lie – should also be rejected. Even assuming that using the criminal law against innocent actions or agents is misleading in some sense, and even if there is a reason against doing so, this reason is not (almost) always decisive. As I argue in the next section, even when using the criminal law against innocent agents involves a cost that is clearly morally significant, namely, undeserved harm, the claim that this cost is (almost) always decisive is unconvincing.
The objections that all of the examples discussed above concern laws that are not properly classified as criminal, or are unjustified, thus fail. The only alternatives are either to reject the IC or to reject every prohibition that applies to NM actions. I believe that the more plausible option is the former.
V. The Reasons For and Against Using the Criminal Law against Innocents
The counterexamples argument is reinforced by a more theoretical argument: the force of the considerations for and against using the criminal law is a matter of degree, both in general and specifically when it is directed at innocents. Therefore, the considerations for using the criminal law against innocents may be (relatively) strong and the considerations against doing so may be (relatively) weak. In this case, the former considerations defeat the latter. Since this is true not only in exceptional circumstances, the IC should be rejected.Footnote 77
Consider first the force of the considerations in favour of using the criminal law. This is a function of the degree to which the consequences of the criminal law are good, for example, in terms of maximising well-being or preventing undeserved harm. It may be thought that some of the most straightforward goals of the criminal law can be promoted only by using the criminal law against people who are blameworthy due to actions that are wrong independently of the law (MS actions). According to this objection, using the criminal law against actions that are not wrong at all (NM), or wrong only because of the law (MP), cannot prevent wrongful harms,Footnote 78 while using the criminal law against innocent agents cannot promote retribution. However, the criminal law may promote goals that are ultimately related to wrongdoing and culpability even when it is applied to innocent actions and agents. This is because of the rationale of overinclusive criminal prohibitions: the cost of separating MS, MP and NM actions, and the impact of such a separation on goals such as preventing undeserved harms. The force of the considerations in favour of such prohibitions depends on the degree to which they promote such goals.Footnote 79 If a legal speed limit is the most efficient way to prevent serious harms to innocents due to car accidents, or if a tax system backed by the criminal law is the most efficient way to mitigate distributive injustice, or to maximise well-being to a significant degree, then there are strong reasons for these uses of the criminal law, even if they result in the punishment of the innocent.
Consider next the claim that the force of the considerations against using the criminal law is also a matter of degree, and that these considerations are not (almost) always decisive even when the criminal law is directed against actions that are not wrong and agents who are not blameworthy.
There are considerations that apply only, or more forcefully, when the criminal law is directed at innocents, most obviously the fact that this will result in undeserved suffering. But the force of these considerations depends on the degree to which the punishment (and condemnation) involved is serious. The considerations against severe punishment (and condemnation), such as long prison sentences, are much stronger than the considerations against milder forms of punishment (and condemnation), such as small fines (assuming the amount will not have a serious impact on the offender's well-being). Similarly, while criminal offences that apply to morally permissible actions may deter such actions, performing some morally permissible actions is much more important than performing others.Footnote 80
Notice that the focus of this analysis is not the enactment of criminal prohibitions but rather the effects of their enforcement, for example in the form of prosecution and punishment. Accordingly, to the extent to which the argument is sound, it applies also to the narrow version of the IC that applies only when overbroad prohibitions are enforced against innocents.
It may be objected that there is a consideration against using the criminal law against innocents that is not a matter of degree, and that this consideration is (almost) always decisive. The most natural form of such a consideration is a deontological constraint, since consequential considerations are more clearly a matter of degree and are less often considered to be (almost) absolute. One suggestion is that there is a general deontological constraint on harming (or intentionally harming) innocents, which is typically violated when the criminal law is used against innocents. However, a constraint of this kind that is not at all sensitive to the degree of harm involved, so that it applies to every harm, including minor ones, and is nevertheless (almost) absolute, is unconvincing. For example, it is (I assume) not wrong to break the finger of an innocent person when this is the only way to save the lives of several innocent persons. Since the harm involved in using a criminal law that carries relatively mild punishment (and condemnation), such as a fine for speeding, is (I assume) typically not more serious than a broken finger, it is similarly not wrong to impose this harm if this promotes an important goal such as saving lives by preventing accidents.Footnote 81
Another objection may be that there is a special deontological constraint that forbids using the criminal law against innocents, which is (almost) absolute. However, this suggestion is even less plausible than the one that appeals to a general constraint, since it fetishises the criminal law. The criminal law is not valuable in itself, but as a means of promoting goals that are independent of it. Therefore, the criminal law is not a constitutive element of a foundational moral standard (as opposed to one that is derived from a more basic standard, such as the constraint on harming the innocent).
My claim that there is no absolute constraint on using the criminal law against innocents is compatible with the common view that, other things being equal, the costs involved in punishing innocents are (in some respects) more significant than the benefits that this sometimes brings.Footnote 82 Since these benefits may nevertheless be important, for example, when overinclusive offences are the only way to prevent serious undeserved harm, it is implausible to hold that the above costs have absolute priority over these benefits. Moreover, it is not clear that there is a reason to care more about the costs as opposed to the benefits of the criminal law. Other things being equal, the cost of undeserved suffering, for example, is the same when a criminal sanction is imposed on a person who does not deserve it, and when undeserved harm that can be prevented by the criminal law is not prevented.
While my argument is about the question of what the law should be, rather than what the law is, it is worth noting that positive law seems to acknowledge that the force of the considerations for and against using the criminal law is a matter of degree. For example, the law is sometimes sensitive to the degree to which the punishment is severe. This is especially clear in extreme examples, such as when we compare the procedural and evidentiary rules that apply when the punishment may be a long prison sentence (or death) to those that apply when the punishment is a relatively small fine. This difference is sometimes formally recognised in legislation and court judgments and sometimes informal. One example is the formal distinction in English law between “summary” and “indictable” offences regarding the question of whether a jury is required.Footnote 83 An example of an informal difference is the claim that the burden of proof is, in fact, not uniform but rather sensitive to various factors, and especially that it is applied more flexibly when the law applies to actions that are not its ultimate target, including actions that are not wrong or harmful, but are criminalised as a means of preventing other actions that are wrong or harmful.Footnote 84
There are thus cases in which the considerations in favour of using the criminal law are strong, even when it applies to innocent actions and agents, while the considerations against using the criminal law are (relatively) weak, particularly when the punishment (and condemnation) is mild, such as in the case of a small fine. NM (and MP) actions frequently involve a relatively mild sentence (and condemnation). In these cases, applying the criminal law to innocent actions and agents is justified more often than existing versions of the IC accept.Footnote 85
VI. The Analogy Argument: The Criminal Law Is Not (That) Special
The analogy argument is based on the following claims. First, harming innocents is sometimes justified in contexts other than that of the criminal law. Second, harm resulting from the criminal law is not (almost) always worse than other harms. Finally, other aspects of the criminal law are not (almost) always morally significant in a way that entails that using the criminal law against innocents is (almost) always wrong.
The first premise – that harming innocents is sometimes justified in other contexts – is compelling and widely accepted, as the broken finger example demonstrates. Moreover, many believe that (even much) more serious harm to innocents is sometimes justified, for example, that killing one person in order to prevent the death of five by diverting a trolley from the five to the one is permissible.Footnote 86 This proposition is widely accepted not only by consequentialists but also by deontologists (the latter hold that there are some ways in which harming persons is wrong even when this brings about optimal consequences, but they typically do not dispute that this is sometimes justified).Footnote 87 The law too often authorises or permits actions that harm innocents. This includes actions performed by legal officials (for example, preventive detention) as well as those by individuals who are not legal officials. The criminal law permits such actions, for instance, in the framework of justificatory defences, such as “lesser evils”, “necessity” and self-defence (which often permits also harm to innocent aggressors).Footnote 88
The second premise – that harm that is the result of the criminal law is not (almost) always worse than other harms – seems correct once we recall that criminal punishment is not always very severe.
Finally, the last premise – that other aspects of the criminal law are not decisive in terms of the justification of its use against innocents – is based on two observations. One is that the fact that a certain harm is imposed by the criminal law does not matter in itself (as opposed, for example, to the degree to which the harm is severe, and, according to some views, the way in which it comes about), since the law is merely a means. The other observation is that other effects of the criminal law – such as condemnation – are not (almost) always substantial.
Indeed, the boundary between the criminal law and other forms of legal regulation is to a significant degree arbitrary. Accordingly, the tendency to apply different legal rules in different legal domains is sometimes criticised. This criticism typically focuses on procedural law.Footnote 89 My suggestion is that a similar criticism applies to substantive law, including the IC. Both substantive and procedural law should be sensitive to factors such as the degree to which the legal sanction is harmful rather than to the formal classification of the relevant regulation as criminal or not (or indeed as legal or not). There may be sometimes instrumental reasons to preserve such classifications, even if they are arbitrary, such as the importance of certainty in some contexts. However, these reasons are irrelevant to the IC which is based on intrinsic reasons.
According to one objection, while harming innocents via defences such as “lesser evils”, “necessity” and self-defence is necessary (to prevent other harms), this is not necessary in the case of criminal punishment, but rather “merely instrumental in achieving various consequentialist goals”.Footnote 90 Yet the (significance of the) difference between being “necessary” and being “instrumental” is unclear. Since the cases in which the criminal law is justified are those in which there is no better alternative to achieve a certain goal, it seems that the criminal law is necessary in the relevant respect.
A related objection is that the law allows individuals who are not public officials to harm innocents only in “emergencies”, that is, when it is necessary to act immediately, whereas criminal punishment lacks this feature.Footnote 91 However, while the need to act immediately is often important as an indication that the action is indeed necessary, namely, that there are no better alternatives, it does not seem to be important in itself (also in terms of the justification of harming innocents).Footnote 92 Therefore, this difference is also irrelevant to the IC as a constraint that is based on intrinsic considerations.
Thus, once we recall that it is sometimes justified to harm innocents in various ways, even when the harm is severe (mainly in order to prevent more serious harm to other innocents), the suggestion that it is (almost) never justified to harm innocents through the criminal law, even when the harm is not severe, loses its appeal.
VII. Conclusion
The claim that it is always, or even almost always, wrong to use the criminal law against innocents is implausible. While there is a reason against inflicting undeserved suffering, there is also a reason in favour of preventing undeserved suffering, and the former reason does not defeat the latter in (almost) every case. Using the criminal law against innocents may be justified, specifically, to prevent serious harm that is undeserved at the cost of inflicting undeserved harm that is less severe.
This conclusion does not entail that using the criminal law against innocents is always justified, since doing so involves a moral cost. Rather, this conclusion highlights the importance of considering the questions of when and to what degree using the criminal law against innocents is justified. This inquiry is often avoided due to the common assumption, which the IC reflects, that this is never justified, and since it is unpleasant to recognise the fact that it is sometimes justified to use the criminal law against innocents. Acknowledging the fact that the criminal law often affects innocents, and that this is sometimes justified, should encourage a more thorough consideration of these important questions. We should recall that there are also other ways in which the criminal law imposes significant burdens on innocents who are not themselves its ultimate targets, such as the relatives of those who are punished and the people who could benefit from the resources that the criminal law consumes. The fact that these effects are not intentional may be significant in some ways, but it does not entail that the use of the criminal law against innocents is never justified.
More generally, the arguments against the IC suggest that it is not only exaggerated in its insistence that the reasons against applying the criminal law to innocent actions and agents are (almost) always decisive, but also that it is too narrow in two ways. First, it focuses on agents who are not culpable at all (or actions that are not wrong at all) and says nothing about agents who are culpable (or actions that are wrong) to some degree. This is noteworthy since the difference between a person who is not culpable at all and one that is culpable to a minor degree is arguably less significant than the difference a person who is minimally culpable and a person who is very culpable. Second, the IC is too limited in its application only to the criminal law, as opposed to other forms of legal regulation. To be sure, the IC does not imply that other forms of legal regulation that impose undeserved harm are justified. But the exclusive focus on the criminal law is misleading, as it suggests that imposing undeserved burdens that are not classified as criminal does not involve a moral cost of a similar type.
In these respects, instead of an (almost) absolute constraint that focuses only on innocents and only on the criminal law, a more plausible option is a non-absolute principle, which applies to all persons and all burdens, and requires that these burdens are proportional to the importance of the relevant goals and the degree to which the relevant persons are blameworthy (if they are).
