Towards a Modest Legal Moralism

After distinguishing different species of Legal Moralism (positive vs. negative; modest vs. ambitious) I outline and defend a modest, positive Legal Moralism, according to which we have good reason to criminalize some type of conduct if (and only if) it constitutes a public wrong. Some of the central elements of the argument will be: the need to remember that the criminal law is a political, not a (merely) moral practice, and therefore that in asking what kinds of conduct we have good reason to criminalize, we must begin not with the entire realm of wrongdoing, but with conduct falling within the public realm of our civic life; the need to look at the different processes of criminalization (of which legislation is only one), and to ask what kinds of consideration can properly figure in those processes; the need to attend to the relationship, and the essential differences, between criminal law and other modes of legal regulation.


Negative or Positive Legal Moralism?
First, we should distinguish negative from positive versions of Legal Moralism. Negative Legal Moralists hold that wrongdoing (the wrongfulness of the conduct to be criminalized) is a necessary condition of criminalization, but does not give us any positive reason to criminalize: we may not criminalize conduct unless it is wrongful; but our positive reasons for criminalizing it lie elsewhere, for instance in the fact that it causes or threatens to cause harm to others. By contrast, a positive Legal Moralist holds that the wrongfulness of a type of conduct gives us positive reason to criminalize it: not necessarily a conclusive reason, since we might well find stronger countervailing reasons arguing against criminalization; but a good reason to consider criminalizing it. This distinction parallels that between 'negative' and 'positive' retributivism in penal theory. For positive retributivists the central justifying aim of a penal system is to impose deserved punishment: that wrongdoers deserve punishment gives us a good, if not a conclusive, reason to institute a system of criminal law that will impose such punishments on them. Negative retributivists, by contrast, hold that the positive aims of a penal system must lie in something other than retributive desert-for instance in the achievement of some consequential benefits. Retributive desert figures as a side-constraint on our pursuit of those aims, a constraint expressed in the principle that we must not (knowingly) punish the innocent: guilt provides no positive reason in favour of punishment, but innocence constitutes a conclusive reason against it. 1 One form of negative Legal Moralism can be found in a Feinbergian version of the Harm Principle. The positive aim of a system of criminal law is to prevent harm, by criminalizing and so reducing the incidence of conduct that causes or might cause harm; but the demands of justice set a side-constraint on our pursuit of that preventive goal, forbidding us to criminalize conduct that is not morally wrongful. 2 A straightforward version of positive Legal Moralism is found in Moore's claim that criminal law's function 'is to attain retributive justice' by punishing 'all and only those who are morally culpable in the doing of some morally wrongful action' (Moore 1997, 33-35): on such a view, the only good reason for criminalizing a type of conduct is that it is morally wrongful, and therefore deserves the punishment that criminal law provides.
As thus described, these two versions of Legal Moralism are logically independent, since neither entails the other. Negative Legal Moralism does not entail positive Legal Moralism: to say that the absence of wrongfulness is a conclusive reason against criminalization is not to say that the presence of wrongfulness is any kind of reason for criminalization. Nor indeed does positive Legal Moralism entail negative Legal Moralism: for one could coherently hold that the wrongfulness of a given type of conduct is always a reason in favour of criminalizing it, whilst also holding that we can have other reasons to criminalize non-wrongful conduct. 3 However, like others who espouse positive Legal Moralism, I accept the constraint that defines negative Legal Moralism: we should criminalize conduct only if it is morally wrongful in some suitable way. Although my main focus in this paper will be on positive Legal Moralism, which is the more controversial and more often misunderstood species of Legal Moralism, I should therefore offer a brief explanation and defence of negative Legal Moralism.
If the moral wrongfulness of a given type of conduct is to be a necessary condition of its legitimate criminalization, as negative Legal Moralism holds, presumably that wrongfulness must be independent of the conduct's criminalization: conduct should be criminalized only if it is already wrongful. It is important to note, however, that the conduct need not be wrongful independently of the law as a whole. If that kind of independent wrongfulness was required, we could still justify criminalizing so-called mala in se, which consist in conduct that is pre-legally wrongful; but we could not justify any of the wide range of socalled mala prohibita that are to be found in any modern system of criminal law: for mala prohibita are, precisely, offences consisting in conduct that is not (or not determinately) pre-legally wrongful; conduct which is wrongful only because it is legally prohibited. Indeed, the conduct that constitutes a malum prohibitum might not even be possible in advance of the relevant legal regulation: it is not possible to fail to pay my taxes, or to fail to display a tax disc in my car, if there is no tax system with regulations defining who must pay what, and how, or a system of regulations for road vehicles that creates tax discs and the requirement that they be displayed. What marks a malum prohibitum is that it consists in conduct that is not wrongful prior to or independently of a legal regulation that prohibits it. Negative Legal Moralism can justify, in principle, the criminalization of some mala prohibita, so long as what it specifies as a necessary condition of justified criminalization is not that the conduct be pre-legally wrongful, but that it be pre-criminally wrongfulwrongful prior to and independently of the law that defines it as a criminal offence: for it can justify the criminalization, as mala prohibita, of some breaches of legal regulations if it can be argued that such breaches are wrongful; and such breaches will be wrongful if the legal regulation in question is well designed to serve some aspect of the common good, and if the burden that obedience to it involves is one that citizens can properly be expected to accept as a matter of their civic duty. We have, surely, good reason to create a legal system of taxation, in order to ensure that the money required for the state to discharge its tasks is raised efficiently and fairly; such a system will involve regulations about, inter alia, how citizens' tax liabilities are to be determined and discharged. Those regulations will not themselves (yet) be a matter of criminal law; they belong to the realm of tax law, not that of criminal law. But the breach of such legitimate regulations will be wrongful, as breach of one's civic duty; and that wrongfulness suffices to meet the wrongfulness requirement that negative Legal Moralism specifies. 4 Negative Legal Moralism as thus defined seems plausible. What is distinctive about criminal law is that it inflicts not just penalties, but punishments-impositions that convey a message of censure or condemnation 5 ; the convictions that precede punishment are not mere neutral findings of fact, that this defendant breached this legal rule, but normative judgments that this defendant committed a culpable wrong. The criminal law portrays 4 See further Duff (2007: chs. 4.4, 7.3), contrast Husak (2005) (Husak argues that negative Legal Moralism cannot justify mala prohibita). The account of mala prohibita at which I gesture here clearly commits me to insisting, contra Bentham and those who have followed him, that the distinction between mala in se and mala prohibita can be drawn in a tolerably clear way (contrast Bentham 1776Bentham /1977, on this 'acute distinction…. which being so shrewd, and sounding so pretty, and being in Latin, has no sort of occasion to have any meaning to it'): but I need not argue that it can always be drawn sharply. 5 See e.g. Feinberg (1970), von Hirsch (1993 crimes as wrongs; if it is to be truthful, it must therefore define conduct as criminal only if that conduct is, pre-criminally, wrongful. 6 Victor Tadros offers a counter-example to negative Legal Moralism as I have defined it here. Suppose that we are concerned about the harm that is caused by the misuse of knives, and come to believe that an efficient way to reduce such harm would be to bring it about that people do not generally carry knives in public. We might also realize that mere exhortations not to carry knives will be ineffective: too many of those who might misuse knives will not be dissuaded; others, who might have been willing to forswear knives, will then see reason to continue to carry them for reasons of self-protection-and will indeed, let us suppose, do no moral wrong in continuing to carry them for that reason. So, for the sake of more effective dissuasion, we might then pass a legal regulation, prohibiting the carrying of knives in public (allowing, no doubt, certain exceptions for those who have good reason to carry a knife). But we will also realize that a mere regulation, even if backed up by some kind of administrative sanction for those found carrying knives, will still be inadequately effective; too many people will still carry knives; and it will then be morally permissible for others to carry knives as a self-defensive precaution. The only way to secure a substantial enough reduction in knife-carrying, we come to see, is to criminalize it: for only the distinctive condemnation provided by a criminal conviction, and the distinctively condemnatory punishments that follow, will be effective in dissuading enough people from carrying knives. Now, Tadros argues, the crucial point is this. Once knifecarrying is criminalized, and once people in general are therefore not carrying knives, individuals lose the self-defensive moral permission to carry one: the threat of a knife attack from someone else is sufficiently reduced, and does not now warrant such a defensive measure. We can therefore say that, once knife-carrying is criminalized, it becomes morally wrongful. But it was not morally wrongful prior to its criminalization: for before it was criminalized, before its prevalence was thus reduced, it was morally permissible to carry a knife as a necessary measure of self-defence. So this, Tadros argues, is a case in which we might be justified in criminalizing a type of conduct as an efficient way of reducing harm, but in which the conduct to be criminalized is not wrongful prior to its criminalization. 7 I am not persuaded. Let us accept for the sake of argument that it is not generally morally wrong, in advance of any legal prohibition, to carry knives in public. Of course, it is wrong to carry a knife with the intention of using it to attack persons or their property, or to carry it in a way that creates a clear and serious risk of harm; but merely carrying, let us suppose, is not a wrong, even when it is not a self-defensive measure. However, once we collectively decide to prohibit carrying knives, for the sake of public safety, and pass a regulation to that effect, it is then wrong to carry knives, 8 in the sense that we-those on whom the regulation is binding -now ought not to carry knives (the collective plural 'we' is important here). The necessary condition for criminalization specified by negative Legal Moralism, that the conduct that is to be criminalized be pre-criminally wrongful, is then 6 There are actually two requirements of truthfulness here. One is a requirement of honesty: that legislatures define conduct as criminal only if they believe it to be relevantly wrongful. The other is a requirement of truth: that they define conduct as criminal only if it really is wrongful (compare Tadros 2007: 197-200). 7 Tadros (2012: 169-172), see also Tadros (2011: 323-325). 8 Why is it now morally wrong to carry a knife? Either because the prohibition serves this aspect of the common good in a way that does not impose unreasonable burdens, and we should obey it for that reason; or because, even if it is a 'dumb' regulation, it is not illiberal (see Markel 2012), and can claim our obedience as a matter of our civic duty to respect the democratic process that produced it. I cannot discuss the latter kind of malum prohibitum offence, or the reasons we can have for obeying misguided laws, here.
satisfied. Of course we must also provide for a range of defences that should enable defendants to avert criminal liability for committing this offence, including self-defence: one who carries a knife because he reasonably believes this to be a necessary self-defensive measure will therefore be able to offer this as a defence if he is prosecuted. But, first, someone who carries a knife for self-defensive reasons at a time when the regulation is in force but is not effectively enforced is still committing what is at the least a presumptive wrong, which requires justification, 9 and which can therefore properly be defined as a criminal offence, so long as the law also accepts such a justification. Second, if we attend not merely to the law's definition of the offence, but to the ultimate judgment that it authorizes, it does not render liable to conviction and punishment any type of conduct that was not wrongful prior to its criminalization: for what makes a defendant liable to conviction and punishment is not just that he carried a knife in public, but that he did so even though this was not a necessary or reasonable defensive measure. That-carrying a knife in public when this is not a necessary or reasonable defensive measure-is the type of conduct that we have now criminalized: even if that conduct was not wrongful prior to or independently of the law altogether, as I supposed above, it become wrongful once the regulation prohibiting carrying knives is enacted; it is thus pre-criminally wrong, which is all that negative Legal Moralism requires.
The enactment of a criminal law defining knife-carrying as an offense does play a causal role in the historical process given which a particular token of the type of conduct 'carrying a knife in public' becomes a token of the type of conduct 'carrying a knife in public although it is not a necessary or reasonable defensive measure', and thus becomes criminal: for, we are supposing, the enactment of that law brings about a change in people's behaviour, a reduction in how many people carry knives; it thus brings about a relevant change in the circumstances under which this particular token of knife-carrying is done-a change given which carrying a knife is no longer a necessary or reasonable defensive measure. But what we have decided to criminalize is not a particular act of knife-carrying, but the type of conduct 'carrying a knife in public although it is not a necessary or reasonable defensive measure'; that type of conduct is wrongful before such carrying is criminalized, as an unwarranted breach of the legitimate legal regulation prohibiting knifecarrying, even if particular tokens of public knife-carrying are wrongful only after that type of conduct has been criminalized. What matters for a legal moralist is not whether the wrongness of a particular criminal act depends, historically or contingently, on the enactment of a criminal law, but whether the wrongness of the type of conduct to be criminalized is logically prior to and independent of the law that criminalizes it; and that condition is satisfied in this case. 10 Negative Legal Moralism is, I believe, right, but it is also inadequately ambitious; it tells us when we may not criminalize a type of conduct (when that type of conduct is not wrong), but not when we should criminalize; it specifies a conclusive reason against criminalization, but does not tell us what could count as a good reason for criminalization. Although it is quite natural for theorists, faced by the apparently irresistible tendency of governments (at least in Britain and the United States) to reach for criminal law as a first rather than as a last resort, to look for negative principles that might constrain the tide of criminalization, we must also ask about the positive principles that should guide decisions 9 On presumptive wrongs and defenses, see Duff (2007: ch. 9). 10 A similar point applies to the example of the spy who ought to refrain from leaking a piece of information if, but only if, everyone else who has the information refrains from leaking it (Tadros 2012: 168).
Crim Law and Philos (2014) 8:217-235 221 about the scope of the criminal law; and positive Legal Moralism offers just such a principle.

Ambitious or Modest Legal Moralism?
However, a second distinction must now be drawn, between ambitious and modest versions of positive Legal Moralism. Moore's is an ambitious version, since he holds that every kind of moral wrongdoing is in principle worthy of criminalization, although other considerations, both principled and pragmatic, militate against criminalizing many kinds of wrongdoing, so that in the end the criminal law's scope might not be very different from that favoured by liberals who reject positive Legal Moralism. 11 A modest Legal Moralism, by contrast, holds that only certain kinds of moral wrongdoing are even in principle worthy of criminalization; for many kinds of wrongdoing, the conduct's wrongness gives us no reason at all to criminalize it. A central task for a modest Legal Moralist is of course then to explain which kinds of wrongs are in principle criminalizable, and why. The Legal Moralism that I favour is a modest one: the criminal law is, I believe, properly concerned not (even in principle) with every kind of moral wrongdoing, but only with wrongs that should count as 'public' rather than 'private'-which at once raises the question of how we are to distinguish public from private wrongs (see Duff 1998, 2010). But it is a positive Legal Moralism, since I believe that the distinctive function of criminal law, the function that marks it off from other modes of legal regulation, is its focus on wrongdoing. It aims, in its substantive mode, to define the range of public wrongs that the polity is to mark and condemn as wrongs; and, in its procedural mode, to provide for those accused of committing such wrongs to be called to public account for them through the criminal trial. In a liberal republic of the kind in which we should aspire to live, the distinctive role of the criminal law is to define those wrongs for which we will be publicly called to answer by our fellow citizens, and to provide the formal, institutional mechanisms through which we can thus be called (see Duff 2007).
The difference between such a modest Legal Moralism, and more ambitious Legal Moralism of the kind that Moore espouses, is not just one of scope: it is not that whereas Moore takes the whole realm of moral wrongdoing as his starting point, I would begin by carving out within that wide realm the particular kinds of wrongdoing that, because they properly count as 'public', are in principle of interest to the criminal law. The difference is also one in starting points. Positive Legal Moralism is often discussed, by both advocates and critics, as if it requires theorists and legislators to take wrongdoing as the starting point: we ask ourselves, as legislators or normative theorists, what we should criminalize; positive Legal Moralism's first answer is 'wrongdoing'; we must then survey the (extensive and highly variegated) realm of wrongdoing, to determine which of its species and sub-species we should at least in principle criminalize. 12 Legal Moralism can then seem to advocate a kind of moral witch hunt: we must collectively seek out wrongdoing (of the appropriate kind) in order to make sure that it is criminalized and punished. That might be how some Legal Moralists have seen the task of criminal law: to ensure that moral wrongdoers suffer the retributive pains that they deserve, or that the 'grosser forms of vice' are persecuted, in order to gratify the justified ''feeling of hatred…which the contemplation of such conduct excites in healthily constituted minds'' (Stephen 1874(Stephen /1967. But it is not, I will argue, the best way to make plausible sense of Legal Moralism, partly because it implies an over-simplified view of the criminal law as an institution that stands apart from all the other institutions and modes of legal regulation that make up a system of law. It is indeed important to gain a clear idea of the distinctive character of criminal law, as a particular mode of law. It is not simply one of the mechanisms available to governments for the control of behaviour, to be selected if and when it is likely to be a more efficient means to governmental ends. Given its distinctive meaning, the first question that we must ask about its scope is when it is intrinsically, rather than instrumentally, appropriate to use it; and what makes it intrinsically appropriate does have to do with the wrongfulness of the conduct to which it is to be applied. However, it is also important to remember that the criminal law is not simply the moral law given institutional form. It is a part of the political structure of the state-a part of our political or civic lives rather than of our more personal moral lives; its proper aims and role must be understood in that context, in terms of the contribution it makes, or should make, to that dimension of our lives. 13 This is not to say that it must be understood in political rather than in moral terms, as if the political and the moral are separate realms. It is still to be understood as an institution that is concerned with moral wrongdoing, and that aims to communicate in morally appropriate ways about such wrongdoing; but its concern is with wrongdoing that figures in the public realm of our political or civic life-not with moral wrongdoing as such (whatever that might mean).
We could put the point this way. To say that criminal law is concerned with 'public' wrongs is not to say that theorists or legislators who are thinking about criminalization should begin with the general category of wrongs, and then try to identify within that category the subcategory of wrongs that are 'public'. Rather, we begin with the idea of the public-the res publica, the realm of our civic or political life; as we think about how to organise and regulate that realm, we will find a role for a system of criminal law as an appropriate way in which we can mark and respond to wrongs committed within it; but such wrongs come into consideration as already being public wrongs. We must indeed, any polity must, work out a distinction (or set of distinctions) between the public and the private realms-between those aspects of citizens' lives which belong to the civic enterprise of living together as a polity, which are therefore of legitimate interest to their fellow citizens in virtue simply of their citizenship; and those aspects that fall instead within the various non-political, non-civic, spheres of the citizens' lives. That distinction will then generate a distinction between 'public' and 'private' wrongs-between the wrongs that could in principle fall within the reach of the criminal law, and wrongs that are in principle not the law's business. But this distinction between public and private wrongs itself flows from a more fundamental distinction between the public and the private realms, since a public wrong is simply a wrong committed within the public realm; any adequate normative account of the proper scope of the criminal law must therefore begin, not with wrongs as such, nor even with public wrongs, but with the public realm. 14 13 Compare Thorburn (2008Thorburn ( , 2011, on criminal law's 'public' character: even if one does not accept his particular account of its role, he is right to focus on its place within the political structure of the state. Also relevant here is the German idea of criminal law as ultima ratio or a last resort, and the related principle (or slogan) that it has an essentially 'fragmentary' and 'subsidiary' character: for a very helpful discussion, see Jareborg (2005). 14 As might be evident, I think that the republican tradition provides the best grounding for an account of the civic enterprise: see e.g. Dagger (1997), Pettit (1999). But my argument here is that any account of the proper scope of the criminal law must be grounded in a political theory of some kind, whether or not its spirit is republican. I will try to explain this view, and render it plausible in the rest of this paper. To that end, it will be useful to think about criminalization not simply as an act, for instance of legislation, but as a process, or rather a series of processes.

Processes of Criminalization
The processes of criminalization are those processes through which conduct comes to be or to be treated as criminal. My concern here is not with how those processes actually operate (that would be a complicated, messy, often depressing story, based on historical, sociological, psychological and political evidence), but with how they should operate: with what considerations should guide people in deliberating over whether to criminalize or not; and in particular, how-at what stages, in what ways-considerations about the moral wrongfulness of potentially criminalizable types of conduct should figure in such deliberations. More precisely still, my concern is with the logic rather than with the chronology of criminalization processes: not with the ways in which or times at which such considerations are actually adverted to in deliberations about criminalization, but with the proper logical structure of the deliberations that could justify criminalization.
The most salient criminalization process is legislation: a type of conduct is criminalized by the enactment of a statute defining it as criminal; the process of criminalization is the process through which the statute came to be enacted. 15 That process includes the deliberations of policy makers, civil servants and legislators as they think about whether to introduce a new statute, what it should cover and how it should be worded; it can also include various earlier processes of deliberation and argument by, for instance, law reform bodies, pressure groups and campaigners, and so on. For simplicity's sake I will focus here on governmental or official deliberations, by policy makers, law commissions and the like. We must also note, however, that criminalization involves other processes than these.
The enactment of a new criminal statute, defining as criminal conduct that was not until then criminal, constitutes criminalization 'in the books' 16 : the statute is now in the books; the criminal code, if there is one, includes this new offence; those who engage in that conduct are formally liable to be prosecuted, convicted and punished. Whether they are, and how many of them are, actually prosecuted, convicted or punished, however, depends on the decisions of other actors related to the criminal justice system, who help to determine what is treated as criminal 'on the streets'. The police must decide which crimes to investigate, and what resources to devote to their investigation; which cases to refer to the prosecutor, which to dispose of in other (often informal) ways. Prosecutors must decide which cases to pursue as criminal cases, and which to drop, or to divert from the criminal process. A key question here is how far prosecutors should observe a 'legality principle' requiring them to prosecute whenever there is sufficient evidence to do so 17 ; or how far should they also attend to 'the public interest' in deciding whether or not to prosecute? 18 Both police and prosecutors will make many such decisions on a more or less ad hoc, case by case basis; but they might also adopt policies identifying kinds of cases in which, or factors given which, they will or will not generally take a case through as a criminal case. When a policy is adopted of not prosecuting in certain specified types of cases, it could be argued that in such cases the conduct is now effectively decriminalized: it is, as a matter of official policy, no longer treated as criminal; if a prosecutor pursued a case, the court might dismiss the prosecution as an abuse of process. 19 Judges, especially in the appellate courts, also play an obvious role in criminalization processes: their interpretations of the laws that they must apply help determine what kinds of conduct are counted as criminal. 20 If a legal system includes lay participants in the criminal process as arbiters of guilt or innocence, for instance as jurors or as lay magistrates, they too can play a significant role, given their effective power to nullify the law by acquitting defendants who are clearly, in law, guilty as charged. 21 Nor should we ignore the role played in the processes of criminalization by a variety of non-state actors, particularly those who find themselves with the power to treat some type of conduct to which they have been subjected, or which they have witnessed, or indeed which they have themselves committed, as criminal by informing the police, or not to do so. Supermarkets, for instance, must decide whether to report shoplifters to the police for prosecution, or deal with shoplifting in other ways; businesses must decide whether to treat minor pilfering by employees as criminal theft, by involving the police, or as an 'internal' matter (or to ignore it). Individuals often have a similar power: if I suffer some harm at another's culpable hands, I might respond by calling the police and reporting it as a crime; or I might decide to ignore it; or I might pursue the matter by other means (by a civil suit for damages; by an attempt at informal mediation….): it is often in effect, if not in law, up to me to make it a criminal, or a non-criminal matter by the way I choose to respond. In such cases, as in cases of jury nullification, we are of course dealing not with a change in legal rules or doctrines, but with whether or not individual actions are treated as criminal; but they are still relevant as processes through which conduct is, or is not, effectively criminalized. If we are going to work towards an adequate account of criminalization-if not of what conduct should be criminalized, at least of the kinds of consideration that should determine decisions about what to criminalize-we need to attend to these kinds of criminalization decisions as well.
These are some of the processes through which conduct can come to be, and can come to be treated as, criminal, or as non-criminal. In each of these cases, we must ask what kinds of factors should figure in the agents' deliberations, and how they should figure, as reasons for or against criminalizing the conduct or treating it as criminal. I will have time in what follows to discuss only the processes that result in formal legislation-in the 18 See Rogers (2006), Ashworth and Redmayne (2010: 204-206), Crown Prosecution Service (2010a: 10-15). 19 For a good example see Crown Prosecution Service (2010b), specifying the factors that would guide decisions about whether to prosecute those who helped others to travel to the Dignitas clinic. 20 For a very straightforward example from England, see R [1992] 1 AC 599, holding intra-marital rape to be criminal, although it had been clear until that case that, in the law's eyes, a man could not rape his wife. A striking feature was the lack of protest about such judicial, and retrospective, expansion of the criminal law. 21 This is not the place to debate the proper role (if any) of jury nullification in a democratic criminal law (see, e.g., Hreno 2007;Rubenstein 2006). Also worth considering are cases in which English lay magistrates acquitted peace demonstrators who caused symbolic damage to American fighter planes in protest of the Iraq war: the demonstrators pleaded the necessity of stopping an illegal war-a kind of plea that had been firmly and formally rejected by the House of Lords (Chandler v Director of Public Prosecutions [1964] AC 763). enactment of new criminal statutes; but we can hope that this will throw some light on the other processes noted here.

Criminalization and Wrongdoing
Processes of criminalization do not arise from nowhere: they begin in response to a perceived problem. We see some state of affairs U, actual or prospective, as problematic; and we come to consider criminalization as a way of dealing with or responding to U. 22 So to ask about the legitimate processes of criminalization is to ask what kinds of U, or what lines of deliberation about how to deal with or respond to U, could properly lead us to see criminalization as an, at least in principle, appropriate option.
To illustrate this, consider a simple version of ambitious, positive Legal Moralism. 23 On this view, criminalization enters the frame if and only if U consists in moral wrongdoing of some kind, and we think of criminalization because such wrongdoing deserves punishment, which it is the criminal law's function to provide. I will argue that we should indeed think of criminalization only when and because moral wrongdoing appears: but first, U need not itself involve any moral wrongdoing to serve as the starting point of a deliberative process that leads to criminalization; second, not all kinds of moral wrongdoing makes criminalization even in principle appropriate; and third, the proper reason for criminalizing conduct is not (or not simply or merely) to ensure that it is punished as it deserves. These points should become clear in what follows. Consider next, by way of sharp contrast, a simple instrumentalist view, according to which the question we ask is simply: what will be the most cost-effective way of dealing with U? On such a view, U is problematic because it impairs or threatens a good (for instance, for utilitarians, happiness); we have reason to use the criminal law as part of our method of dealing with U if and only if we can expect that to be (part of) an efficient way of minimising the evil that U causes, or of protecting or restoring the good that it threatens. We will, of course, need to determine what distinctive contribution the criminal law can make: what beneficial effects it can produce at what cost, as compared to other available mechanisms. But the criminal law is one technique among others: we cannot ask whether it is intrinsically (in)appropriate in dealing with U, since for the instrumentalist cost-effectiveness in producing valued outcomes is all that matters. It would be absurd to deny that instrumentalist considerations are relevant to criminalization: attention to the likely effects of criminalizing (or of not criminalizing), and of this or that particular mode of criminalization, is crucial to final decisions about whether, and how, to criminalize. But such considerations are not the whole story, and should not figure in the initial stages of deliberation: our first questions must rather be about whether criminal law is intrinsically apt as a part of our response to U; whether such a response is suitable or appropriate to the character of the problem that it is supposed to address. What makes criminal law thus intrinsically apt, I will argue, is the wrongfulness of the conduct to be criminalized.
Whilst those must be our first questions about criminalization as a specific mode of legal regulation, they will often not be the first questions on the road that ultimately leads towards criminalization. Our first question in practical deliberation might simply be some version of 'What are we to do about U?'; and U, as a problem about which we must do something, will often not appear in the guise of wrongdoing to which criminalization is a possible response. We might indeed begin simply with the thought of U as something untoward-undesirable, harmful or dangerous. But if such a thought is to initiate practical deliberation, to move us to action, it must become the thought of U as a matter about which something could, and should, be done. That thought might initially be expressed in a passive, impersonal voice-'Something should be done about U'; but that is not yet apt to initiate practical deliberation. It implies that there is an unspecified 'they' who should do something about U, but also that it is not our business; we speak as detached, sympathetic, observers rather than responsible agents. 24 For deliberation to begin, 'Something should be done' must turn into 'We should do something', and that transition is not merely grammatical: it involves an acceptance (or a claim) of responsibility, that U is our business; and such responsibility ascriptions, to ourselves or to others, can be problematic. That transition can reflect a proper realisation that what we had previously observed from a detached stance is our business-that, for instance, the suffering of those afflicted by starvation elsewhere in the world is our business, and that we should do something about it. 25 But it can also reflect a more arguable claim that U is our business-a claim that others, including those on whom U most directly impinges, might deny.
Suppose for instance that I deplore my students' choice of newspapers: instead of reading quality broadsheets, they corrupt their minds by reading tabloids whose journalistic standards and aims are, to put it mildly, low. If I suggest that 'we', the university, should do something about this, for instance by offering guidance on which papers are worth reading, or allowing only certain papers to be sold on campus, I might be accused by both my students and my colleagues of interfering in something that is not my, or our collective, academic business; to sustain my claim that 'we should do something about it' I must argue not merely that it would be a good thing if students read better quality newspapers, but that improving their habits in this respect is part of what the university should be doing in educating them. I do not say that this could not be argued; but my point is that the transition from 'Something should be done about U' to 'We should do something about U' is mediated by the claim that U is our business-a claim which might be controversial.
To determine whether a university should seek to guide its students' newspaper choices, we must articulate a conception of what universities are for: what belongs to, what falls within, the academic enterprise of educating students and pursuing research? Similarly, when the 'we' who are to 'do something' is a polity, or a government acting on a polity's behalf, the claim that U is our business must be grounded in some conception of the res publica-of just what the polity's enterprise, the civic enterprise of living together as citizens, includes; and some familiar battles between liberals and their opponents (and among those who call themselves liberals) about the proper scope of the criminal law reflect competing conceptions of the res publica. I will not engage with those debates here: but we must remember their relevance, and that of the debates in political theory from which they flow, to questions about criminalization, and to other questions about what (if 24 That is one reason why the notorious remark by the soon to abdicate Edward VIII that 'something must be done' for the unemployed miners he met in Wales was crass: it distanced him from any responsibility for helping them-hardly an appropriate attitude for someone presenting himself as their king. 25 I leave aside here the question of whether that realisation is grounded simply in the fact that we are able, collectively, to do something without incurring unreasonable costs, or also partly in the fact that we share in the causal responsibility for their present suffering. anything) we should do collectively, as a polity, about various kinds of U that we find problematic or disturbing.
Which kinds of U, about which we think we should do something, might lead us to think of criminalization as part of that something? The simple answer to this question is that there is no simple answer. There is no single kind of starting point from which we are properly led towards criminalization, no single route to that end; there are, as I will illustrate shortly, a number of different starting points and different possible routes. However, I will suggest, any legitimate route towards criminalization must pass through three gates 26 : the conduct to be criminalized must be wrongful; it must require a collective response; and we must have good reason to make its wrongfulness salient in that collective response.
One kind of route towards criminalization begins with simple kinds of harm: with damage, or the threat of damage, to interests that it is, on any plausible conception of the res publica, the business of the polity to protect. We come to realise, for instance, how harmful smoking is, to smokers and to others around them, and how addictive nicotine is; or how dangerously driving competence is impaired by drinking alcohol; or how dangerous certain manufacturing processes are to those engaged in them; or how HIV can be transmitted by unprotected sexual intercourse. 'We should do something about it'-but what? Given confirmation that the risks are real, and sufficiently serious, our first strategy might plausibly be one of information and encouragement: the government, or its health and safety agencies, mount campaigns to make clear the dangers involved, and how they can be reduced-by refraining from the dangerous activity altogether (by giving up smoking), or by taking suitable precautions when engaging in it. Such warnings and advice are likely, however, to be ineffective: too many people will be unmoved by them; the risks, and the harms that ensue when those risks are actualised, will continue at a level that we regard as unacceptable. We might then think about doing more, in particular of using the law-but what more, and to what end?
It is here that advocates of Mill's, as distinct from Feinberg's,27 Harm Principle should first be concerned, before there is any question of criminalizing anything. So long as what we do is limited to 'remonstrating', 'reasoning', 'persuading', or 'entreating' people not to engage in such risky activities, or to take precautions when doing so, the Harm Principle is silent; but as soon as we exercise any kind of 'compulsion and control, whether the means used be physical force in the form of legal penalties, or the moral coercion of public opinion', we must ensure that we do so only 'to prevent harm to others' (Mill 1859: ch. 1, para. 9). I will not discuss the plausibility of this restrictive doctrine here, since my concern is with the question of when and why, once we consider some kind of legal regulation, we should then think of involving the criminal law.
There are plenty of ways, other than criminalization, in which we could mobilise the law in our efforts to control such dangers. We could make the dangerous activity harder or more costly to engage in, for instance by taxation; we could require those engaging in it to obtain a licence, to secure which they need to demonstrate their competence; we could make provision for those who are harmed or endangered to bring civil cases for damages or compensation; or we could (which will bring us closer towards criminalization) create a regime of regulations specifying the various kinds of precaution that those engaging in the activity should take; we could even, if we could see no tolerably safe way of regulating the activity, and also see its value as insufficient to make it reasonable to take, and accept, the risks it creates, prohibit the activity altogether. But we are not yet talking of criminalization. However, once we think of regulating, or prohibiting, the activity, we must ask how we should deal with breaches of the regulations, or with those who still engage in the activity. Now we might wonder about using the criminal law-about making such breaches criminal. But what could give us good reason to think in such terms?
There are, I suggest, three conditions that must be satisfied if we are to have good reason, in principle, to criminalize. 28 First, the criminalizable conduct must be morally wrong 29 : for the criminal law condemns and censures, and such condemnation and censure must, as a matter of justice, be directed at wrongful conduct and its culpable agent. We can note three ways in which this condition may be satisfied. Some kinds of risk-creation are so obviously pre-legally wrongful, in subjecting others to manifestly unreasonable risks of harm, as to be eligible for criminalization in advance of any particular regime of legal regulation; we have good (albeit not necessarily conclusive) reason to criminalize them when they actually cause harm, 30 and when the harm does not in fact ensue. 31 In other cases, the risk created might not be so obvious, or so obviously serious, or so obviously one that it is not reasonable to expect others to accept. In such cases, it would be unjust to make the conduct criminal, thus making those engaging in it liable to public censure and punishment, without giving them fair warning that it would attract such a response: that is one reason why, in cases such as smoking, or drinking and driving, or HIV-transmission, if criminalization is to be justifiable at all it must follow a campaign of public education which seeks to make clear that and why the conduct in question is dangerous and therefore wrong. Finally, in yet other cases, we might think that the only efficient way to reduce the risks to a tolerable level is to introduce regulations prohibiting a wide category of conduct that includes both actually dangerous conduct and conduct that might not itself be dangerous: this is, for instance, why our driving regulations include not only a requirement to drive safely, but also a requirement to obey the speed limits-although we know that exceeding the speed limit is not always dangerous. In this kind of case, the wrongfulness of conduct that violates the regulation lies not necessarily in any actual danger that it creates (for it might not be dangerous), but in the violation of a justified legal rule that serves public safety. 32 Second, the wrongful conduct must require, or make appropriate, a collective response: it should not be left to the individuals directly affected by it to deal with it, for instance through a civil suit for damages or compensation for any harm they have suffered. This is one central difference between a criminal law and a civil law method of dealing with wrongful harms. A criminal law response is a collective response by the whole political community, which takes over, or some would say 'steals', the wrong from its direct, individual victims (when there are any). 33 By contrast, a civil law response remains in the 28 I do not have space here to compare the approach taken here with that taken in Husak (2007), which looks for constraints on criminalization, and has little to say about the 'substantial state interests' that could generate positive reasons for criminalization; or with that taken in Simester and von Hirsch (2011), which specifies both wrongfulness and harm as criteria of legitimate criminalization. 29 It must indeed be a public wrong, which concerns us all as citizens: but since we have got this far only because the matter in hand is a public matter, wrongs related to it will also be public wrongs. 30 As with offences of recklessly causing either physical harm to the person or damage to property: see e.g. Offences Against the Person Act 1861, s. 20; Criminal Damage Act 1971, s. 1. 31 As with the Model Penal Code offences of reckless endangerment in § § 211.2, 220.1. 32 On this kind of malum prohibitum offence, see Husak (2007: 73-76, 103-119), Simester andvon Hirsch (2011: 24-29), Duff (2007: 89-93, 166-174). 33 See Marshall and Duff (1998), on 'stealing', see Christie (1977). hands, under the control, of the victim -the person who claims to have been wrongfully harmed or endangered. We can collectively support victims in bringing civil suits, but we leave it to them to decide whether to pursue the issue or not. A criminal case is our collective case; whether and how it is pursued is a public, not a private matter. We could have various reasons for treating the wrong in this way as a public matter that merits a collective response. Perhaps there is no individual victim to bring a civil case: the risk affects the public generally, rather than identifiable individuals. Or perhaps we think it important to deal systematically with this kind of conduct, rather than leaving it to the vagaries of individual victims' choices whether each instance attracts a legal response. Or we might think it is unreasonably burdensome to expect the individual victims to pursue the case-that we owe it to them to take on the burden as a collective responsibility. Or our reasons might relate to the third condition, to be discussed next: that we should make the conduct's wrongfulness salient in our response, as something that needs to be collectively marked and censured. 34 The satisfaction of the first two conditions does not yet bring us to criminalization. We now have a reason to criminalize: we have conduct that constitutes a public wrong (wrongful conduct directly related to a matter that is our collective business), which is therefore in principle apt for criminalization. But we must now ask whether our collective response should make this feature, the wrongfulness, salient. That is what criminal law does: it defines conduct as a public wrong, and provides for a response, the criminal process of trial and punishment, which focuses on the wrong as something for which the perpetrator is to be called to public account, on pain of formal censure and punishment if he cannot provide an exculpatory explanation of his actions. Such a response is appropriate to public wrongs: it takes the wrong seriously, and thus does justice to the victim (if there is one) as someone who has been not just harmed, but wronged; it takes the perpetrator seriously as a responsible citizen who can be held to account by his fellows; and it displays our commitment to the values that the wrong violated. But it cannot be an unqualifiedly absolute demand that we respond in this way to all public wrongs, whatever the cost of doing so: the polity has other values, other goals, which may sometimes conflict with, and may sometimes outweigh, the reasons we have for responding to public wrongs as wrongs whose perpetrators must be called to public account, censured, and punished. 35 We should also, to give just a few examples, be concerned with preventing future harms and wrongs; with repairing harm that has already been caused; with preserving or restoring civic relationships. A proper criminal process of calling to account can serve such ends, 36 but it can also conflict with them: we will sometimes do better to focus on repairing harm, or dissuading risky conduct, or resolving conflicts, rather than calling wrongdoers to account. This is another reason why the Legal Moralism advocated here is a modest one: it does not demand even that all public wrongs must be criminalized; it holds only that we have reason to criminalize them, whilst recognising that we might have weightier countervailing reasons either for doing nothing formally, or for preferring legal mechanisms other than that of criminalization. This is not to fall back into an instrumentalist approach that would favour criminalization if and only if it is likely to be the most efficient among those available to us to produce independently specifiable social goods: criminalization is an intrinsically appropriate response to public wrongs in that it takes the wrongs, and their victims and perpetrators, seriously. Such wrongs, especially if they are egregious, make a categorical, non-instrumental demand on our attention-a demand that criminalization can satisfy (see Jareborg 2005: 534). But categorical demands are not always absolute demands, and they must compete with other kinds of demand for our attention and our resources.
I have focused so far on a route towards criminalization which lies close to that implied by a Feinbergian Harm Principle: we have reason to criminalize conduct when it wrongfully harms or threatens an interest that the polity should protect. But the structure of practical reasoning that is exemplified in this context is not limited to contexts in which the U that motivates deliberation about what we should do consists in Feinbergian harm: other starting points, and other routes to criminalization, are possible, some more direct (because wrongfulness appears earlier) than the route that starts with harm. I have space here only to note some of these possibilities; a central task for any would-be theory of criminalization will be to explore others.
The route discussed above began with harms that are detachable not merely from wrongful conduct, but from human conduct altogether-harms that could be caused by human action, but also by natural processes 37 ; we begin on the road to criminalization when we focus on the ways in which they can be caused by human conduct that is wrongful either because it is dangerous or because it violates a regulation that serves public safety. Sometimes, however, we begin with a wrongful harm: with a harm that cannot be understood independently of the wrongful human action that produces it. This is a feature of such central mala in se crimes as murder, rape, and burglary, and other crimes that involve attack on protected interests 38 : they typically involve serious or potentially serious harms, but the character of those harms is determined in crucial part by the wrongful character of the attacks that produce them. In these cases the initial reason for criminalization comes into view straight away, as soon as we recognise the harmful wrong as a public wrong. 39 It is also usually a stronger reason: because the harm that concerns us is partly constituted by the wrongfulness of the conduct, an adequate response to that harm will need also to take account of, to make salient, the wrong.
In other kinds of cases, we begin not so much with a concern to prevent harm (at least to any concrete interests), as with a concern to maintain the institutions and practices on which the functioning of the civic enterprise and the common good depend. Those institutions might of course have to do with harm-prevention: this is true, for instance, of the whole apparatus of health and safety regulation, of many driving regulations, not to mention defence, health care and social welfare provisions. But we will also recognise a need for regulations that serve not so much directly to prevent the kinds of harm with which the institutions are concerned, as to ensure the functioning of the institutions themselves: regulations concerning various kinds of license provide one obvious kind of example here; tax regulations provide another. We have to ask again how we should respond to, or dissuade, breaches of such regulations; and we will have reason to criminalize such breaches if, and only if, they constitute public wrongs that merit or require a 37 Compare Feinberg's account of a 'harmed condition ' (1984: 31-36). 38 See further Duff (2007: 126-130, 147-158). 39 Even here we must begin, therefore, not with wrongfulness as such, but with a conception of the public realm. But since the public realm is specified partly in terms of the values which structure the civic enterprise, wrongs that violate those values cannot but be seen ab initio as public wrongs.
formal, collective response that condemns the wrong and calls the wrongdoer to public account.
In yet other cases, harm plays an even less central role in our starting point: what strikes us is the wrongfulness of a certain kind of conduct; although it might also be a source of harm (partly in virtue of people's reactions to it), our thought that 'we must do something' is motivated by the wrong rather than by the possible consequential harm. I take this to be true, in different ways, of 'extreme' pornography that graphically depicts the humiliation or torture of members of one sex or group as a source of sexual gratification for readers or viewers; and of certain kinds of racist or otherwise discriminatory insult, especially if they are directed against members of an already disadvantaged or vulnerable group, which violently deny their fellow membership of the polity. What makes it plausible to see these as public wrongs (even when, in the case of pornography, they are perpetrated in the 'privacy' of the person's home) is that they are serious violations of the respect that we owe each other, and thus denials (at least implicitly) of the moral status of those who are their objects. 40 In other cases we may find a starting point in conduct that might not be straightforwardly harmful, but that seriously violates the dignity of those subjected to it (or taking part in it), even if they freely consent to it. This is true, I think, of 'dwarf-throwing', and provides the only morally plausible (but still inadequate) basis for the notorious decision in the sadomasochist case of Brown. 41 In other, typically less serious cases, what is at stake is what we would naturally count neither as harm nor as a dignity-violating wrong, but as one kind of 'offence': the kind that involves inconvenience or annoyance, or damage to civic amenities, which is covered by regulations concerning noise, littering, minor vandalism of public spaces, and the like. 42 We can note finally that, whilst the Harm Principle is traditionally interpreted as allowing criminalization only on the basis of harm to other people, we need not accept that limitation. We can see reason to criminalize cruelty to non-human animals, without having to argue that it is likely to generate harm to human beings, by recognising such cruelty as a wrong that is our proper business. 43

Concluding Comments
I have briefly noted several examples of possible routes to criminalization that do not begin with Feinbergian harm to illustrate one of the central points that should be recognised in debates about criminalization. We should not assume (as too many theorists tend to assume) that we can create a rational and properly limited system of criminal law only if we can articulate a single master principle, or set of principles, that provides substantive general criteria by which we can identify the kinds of conduct that are in principle criminalizable; the search for such a principle or set of principles is doomed to failure (see Duff 2007: ch. 6). Deliberation about criminalization must be structured by principles that can guide and constrain it; but those principles might be procedural rather than substantive, as the principles sketched in this paper are. If we are to criminalize a type of conduct, we must show that it falls within the public realm, the civic enterprise, and that it is therefore of proper interest to all citizens in virtue of their participation in that enterprise; that it constitutes a public wrong within that realm; that it is a wrong that requires the particular kind of response that the criminal law provides-one that condemns the conduct and calls its perpetrator to public account for it. The question of which kinds of conduct constitute public wrongs is then a matter for public political deliberation, which might focus on whether the conduct in question is a public matter, or on whether and how it is wrong: given the plurality of values by which a liberal polity will define itself, we should expect to find a similar plurality among the starting points for the processes of criminalization.
One further point is worth noting. I have taken for granted the main structural features of the criminal law as we know it-as it functions in, for instance, Britain and the United States (which is already to skate over a number of significant differences). In particular, I have so far taken it for granted that to criminalize a kind of conduct is to locate it within a distinctive kind of process that combines several potentially separable features: a process whose central defining purpose is to call people to answer to charges of public wrongdoing, and to answer for that wrongdoing it is proved against them; a process which culminates, if the alleged wrongdoing is proved and is not exculpated, in a condemnatory conviction and punishment; a process that is controlled not by the direct victim(s) of the wrongdoing, but by a public official, who may proceed against the wishes of the victim (or decline to proceed despite the victim's wishes). We can contrast this paradigm of a criminal process with a paradigm of a civil process, in which the central aim is to determine an appropriate allocation of the costs of some harm that has been caused; in which wrongdoing figures only as a condition bearing on the allocation of such costs; in which the outcome is not condemnation and punishment, but an award of damages which are proportioned not to fault or wrongdoing but to the costs of repairing the harm; and which is controlled by the party claiming to have suffered the (wrongfully caused) harm, who can decide whether and how far to pursue the case, and whether to insist on the payment of any damages awarded. 44 However, if we are to think creatively about the enterprise of criminalization, we must also think about ways in which those paradigms can be de-or re-constructed. Are there, for instance, ways in which we can keep a proper focus on wrongdoing and the need to call its perpetrators to account, but without adding punishment as the typical outcome? Are there, for another instance, ways in which we can (and should) give victims of crime more control over the criminal process, for instance by giving them a formal right to insist on prosecution, or to veto prosecution? 45 Advocates of various kinds of 'restorative justice' offer one kind of answer to such questions (although they are often unhappy with the focus on wrongdoing that is, I think, an essential feature of anything that is to count as a system of criminal law); but we should also be ready to explore other possibilities. 44 I do not suggest that either of these paradigms is neatly exemplified in our existing systems, which permit of no such clear distinction between the criminal and the civil; but as ideal models they can still serve a purpose. 45 The German provisions for victims to have a formal role in initiating, requesting, or blocking prosecutions are of interest here: see Bohlander (2012: xxx-x).
It will be evident that what I have offered here is neither a theory of criminalization (would that I had one, although I have indicated my doubts about whether one is to be had), nor even a sketch of such a theory. My aim has rather been, more modestly, to sketch the kind of approach that I think we should take if our aim is to work towards a tenable normative account of the conditions under which, the deliberative procedures through which, the ways in which, the ends for which, a contemporary liberal polity should criminalize conduct; and to suggest, in particular, the shape that a plausible (and plausibly modest) version of Legal Moralism can take.