Criminalization: In and Out

In this paper I explore Antony Duff’s claim that there are categorical constraints on the scope of the criminal law that are set by its internal standards. I argue against his view that such constraints are categorical, and I suggest that his account of the nature of the criminal law is partial, and narrows the focus of our enquiry into the scope of the criminal law too much. However, I suggest that the project is an important contribution to our understanding of one central element of the criminal law.

(1) The contrast between content-focused principles that are concerned with what is criminalized and effect-focused principles that are concerned with the effects of criminalization.
(2) Principles that state uniform reasons for and against criminalization and principles that state duties to or not to criminalize. (3) Principles that state all things considered duties to criminalize, or not to criminalize, and those that state only pro tanto duties to do so, or not to do so.
Criminalization principles can also be stated with more or less restricted domains. For example, they might apply to all possible worlds, all possible states governed by liberal democracies, all circumstances that are likely to occur, and so on.
Antony Duff's magisterial work, most recently and carefully developed in The Realm of Criminal Law 1 (hereafter ROCL) illustrates another way of distinguishing principles of criminalization: it suggests that we distinguish principles that are internal to the kind of practice or enterprise that criminal law is, and those that may influence how we utilize the criminal law, but that are external to it. Let us label these just as 'internal' and 'external' principles for short.
Duff seeks to rescue some principles concerned with wrongness from sceptical challenges by showing that these principles are true when understood as principles that are internal to the criminal law. He also seeks to show that there are categorical principles of this kind. Whilst such principles are 'thin', in that they do not guide the content of the criminal law in themselves in a very precise way, without further information about the particular values of the political communities that seek to apply them, they nevertheless structure debates about criminalization.
One set of objections to Duff's view is its relativistic features. Whether a polity is warranted in criminalizing conduct depends on the actual values of that community, and not simply on the nature and gravity of the wrong. I have expressed some doubts about this idea elsewhere, and I do not pursue it further here. 2 I focus on two things. First, I argue that Duff's argument that there are categorical internal principles of criminal law fails. Any such principles are thus best understood as pro tanto. I then investigate what general merits there might be in the idea of distinguishing internal and external principles of criminalization.
The question is whether this idea can meet a challenge to the significance of pro tanto principles that I have mounted elsewhere. The challenge is this. Pro tanto content-focused principles of criminalization, of the kind that Duff endorses, have the following form: we are pro tanto required to criminalize, or not to criminalize, conduct if it has certain features. To be significant, such principles had better not just be reducible to the view that there are reasons to criminalize, or not to criminalize, conduct if it has certain features. For such principles are trivial: everyone recognises that criminalizing or failing to criminalize conduct almost always has downsides that count against the decision to criminalize, or not to. To be better than trivial, then, such principles must rely on a distinction between something being pro tanto required and there being a reason to do or not to do that thing. But no one has developed a general theory of this distinction; it is not clear how to do it; and it is not clear that doing it would advance our understanding of criminalization in a significant way. For that reason, I argued, we have good reason to abandon the project of seeking pro tanto principles of criminalization.
One response to this objection inspired by Duff's work is that there is something special about internal principles of criminal law; that internal principles are the proper subject of criminalization; and that violating such principles is wrong (either categorically, as Duff suggests, or pro tanto). I am not sure whether this response works, for reasons I set out below.

Categorical and Pro-Tanto Principles
Let us begin with Duff's defense of the idea that there are illuminating categorical internal principles of criminalization. To focus our discussion, let us start with two principles that he thinks structure our understanding of the scope and content of the criminal law: Public Wrong (Reason): we have reason to criminalize a type of conduct if, and only if, it constitutes a public wrong, and a type of conduct constitutes a public wrong if, any only if, it violates the polity's civil order.
Wrongness Constraint (Categorical): the state is categorically required not to criminalize conduct that is not publicly wrong.
I doubt that these principles are true.

What is Criminalization for Duff?
One crucial feature of any principle of criminalization concerns the nature of criminalization itself. Any principle with the form 'only criminalize c if f' or 'we always have reason to criminalize c if f' relies on an understanding of the conduct-criminalizationthat these principles are concerned with. And 'criminalization' might refer to a range of different things. As these are Duff's principles, let's work with Duff's own understanding of criminalization, which is somewhat complex. It is set out in chapter 1 of ROCL, where he suggests that to criminalize conduct is to 'define it as a crime in the substantive criminal law, and thus to render those alleged to have engaged in it liable to be prosecuted and tried, and to render those convicted of it liable to criminal punishment.' 3 This seems like a fairly conventional idea of what criminalization involves, but because Duff has a distinctive vision of prosecution, trial and punishment, his notion of criminalization is narrow and distinctive. Criminalization, for him, involves defining conduct as a crime for the purposes of prosecution, trial and punishment, according to his conception of these things. 4 Conduct that is not defined for these purposes may be regulated, but it is not criminalized, on Duff's account, even if the sanctions are harsh.
Given this view, we can unpack Duff's principles thus: Public Wrong (Reason): we have reason to portray a type of conduct as publicly wrongful and thus to render those alleged to have engaged in it liable to be prosecuted and tried, and to render those convicted of it liable to crim-inal punishment if, and only if, it constitutes a public wrong, and a type of conduct constitutes a public wrong if, any only if, it violates the polity's civil order.
And his categorical version of legal moralism, with his understanding of criminalization inserted is: Wrongness Constraint (Categorical): the state is categorically required not to portray a type of conduct as publicly wrongful and thus to render those alleged to have engaged in it liable to be prosecuted and tried for it, and to render those convicted of it liable to criminal punishment if that conduct is not publicly wrong.
These principles are not trivial, because they are categorical. But there is nevertheless something problematically trivial about them, in how far they narrow the scope of our enquiry into the criminal law. The general idea that the state should not portray conduct as f unless it is f is pretty uncontroversial, though it is controversial that the duty not to do this is categorical.
Here is why these principles lack much bite in the real world: they tell us nothing about whether the state is permitted to do other things, that do not involve criminalization as Duff understands it, including punishing people for things that are not publicly wrongful without portraying the conduct as publicly wrongful. Any practice that did this would not be a criminal practice, as Duff defines it, but it seems that he has nothing to say about practices of this kind. And as I think that much of what we ordinarily understand as the criminal law is like this, Duff's principles have little to say about whether what most of us understand as the criminal law is warranted.
So, for example, consider the Wrongness Constraint (Categorical). This only tells us not to make out that something is publicly wrong and punish people for it if it is not publicly wrong. It does not tell us not to make out that something is wrong, but not publicly wrong, and punish people for it if it is not publicly wrong. So as far as this principle is concerned, it may be permissible to declare publicly that some conduct is wrong, not declare that it is publicly wrong, and condemn people for that conduct. This would not involve criminalization, for Duff, and as far as I can see the principles in ROCL have nothing to say about this practice.
Of course, Duff may object to that practice on other grounds, for example that it involves condemning and punishing conduct that is none of the state's business. But condemning conduct that it is not one's business to condemn is not plausibly categorically wrong. At most, it is pro tanto wrong. For example, suppose that my condemning a person for something wrong that it is not my business to condemn will lead other people not to commit murder. Doing so is a justified pro tanto wrong at most. Now suppose that the whole 'non-criminal' system is concerned only with declaring things to be wrong, and not publicly wrong. Public officials just don't portray people to have committed public wrongs at all, because they don't refer to the polity's civil order, or the shared values and commitments of the society. There is then no criminal system as Duff understands it. As far as I can see his principles say nothing about that system. And even if that system involves the state in committing pro tanto wrongs, by regularly condemning conduct that it is none of its business to condemn, it may be justified all things considered.
I am not sure that our actual system, the system that we (wrongly according to Duff) call the system of the criminal law, is a system of condemning people for public wrongdoing as Duff understands it. I think that when public officials condemn other people for murder, rape, theft and so on, they often just condemn them for doing things that are serious wrongs to the victim, and don't think at all about the civil order. And when these things are criminalized, they are criminalized simply because they are seriously morally wrong.
Furthermore, many offences in the criminal system are not, I think, concerned with wrongs at all. For example, the UK parliament recently passed the Offensive Weapons Act 2019. After the recent set of knife crimes and acid attacks, the government proposed new preventive offences, including further restrictions on delivery of blades by post, possessing corrosive material in a public place and so on. There was absolutely nothing at all in the white paper that preceded the legislation about whether these new criminal restrictions concern conduct that is morally wrong.
Perhaps some officials think about the civil order, or the values of the society that they live in, when deciding what to criminalize. But many don't. To the extent that they don't our system is not a criminal system according to Duff. And his principles tell us nothing about whether our 'non-criminal' criminal system is justified or not.
We all agree that 'criminalizing' such conduct may be permissible if the good is great enough, even if the conduct is not morally wrong, let alone publicly wrong, and a theory of criminalization should, I think, have plenty to say about this issue. For this reason, I think it better to start with a more neutral and inclusive account of what criminalization involves, so that our theories of criminalization have something more direct to say about things that we currently think of as decisions about criminalization.

Categorical or Pro Tanto?
Now let us focus on Duff's actual principles-in particular on their supposed categorical nature. There is an obvious objection to this feature of them, and it is an obvious objection to the general idea that there are important categorical principles like this. Take some principle that claims that it is categorically wrong to v. Now let failing to v have some absolutely horrific consequence. Then it is not plausibly wrong to v; indeed, ving is required. Any plausible categorical principle must then have the kind of content that makes it impossible that failing to adhere to it will have horrific consequences ('don't act in a way that has horrific consequences overall' might be one such principle…). But criminalizing or failing to criminalize conduct can have horrific consequences, and this casts doubt on the validity of a very wide range of putative categorical criminalization principles.
To see the application of this idea in this context, let some horrific thing occur if a state does not criminalize some non-wrongful conduct, and the state ought to criminalize it. It thus both has reason to criminalize it, and is not required not to do so; indeed, it is required to do so. 5 We might then conclude that any content-focused principle of criminalization like the Wrongness Constraint must be pro tanto rather than categorical to be plausible.
Duff argues against this argument on the following basis. In any case where it seems that criminalization of non-wrongful conduct will have horrific consequences, the state need not respond by criminalizing the conduct, but only by pretending to do so. Thus, the state ought to avoid the horrific consequences by abandoning the practice of criminalization, and pretend to engage in it instead. 6 I am not sure how consistent this is with Duff's account of what criminalization is. As I understand it, criminalization, for Duff, involves portraying that something is publicly wrong, and it is a bit hard to understand what pretending to portray something as publicly wrong would involve. Duff's response would be better if criminalization, for him, not only involved portraying conduct to be publicly wrong, but doing so with some kind of belief or commitment to its being wrong.
More importantly, recall the general objection that any principle categorically requiring a person not to v in any circumstances will fail because there will be cases where that person is required to v to avoid horrific consequences. Now let those consequences occur unless the person actually vs rather than, for example, pretending to v. Then the person will normally be required to v, and not merely to pretend to v, for that is the only way to avoid the horrific consequences.
To see how this response applies in this context, suppose that preventing the awful thing from occurring depends on the state actually criminalizing the conduct (whatever that amounts to), and not merely pretending to do so. Then the state ought to actually criminalize the conduct and not merely pretend to do so.
Here is a case that I discuss in previous work that Duff considers, and that will help to illustrate the objection: suppose that the US will sanction a state severely unless it criminalizes possession of some harmless drug. That will result in a shortage of essential medicine, which will seriously impact the health of the population (US Threat). I claim that the state is required to criminalize possession of the harmless drug, and that the Wrongness Constraint (Categorical) is thus false. 7 Duff responds that the state is not required to criminalize the conduct, but only to pretend to do so-to create a simulacrum of the criminal law rather than an actual criminal law.
Duff's view relies on a controversial view about the properties of actual criminal laws. A law that declares that those who possess the drug will be punished, and actually does so, but where officials do not believe the conduct is wrong only counts as a simulacrum, but this seems doubtful even on Duff's own account of what criminalization involves-the state criminalizes the conduct if it portrays possession as publicly wrong on that account, and it can do that without any public official believing it to be wrong. 7 Wrongs and Crimes 98. 5 See Tadros Wrongs and Crimes ch. 6. 6 ROCL 249-50.
But even if we work with this revised view of what criminalization involves, Duff's argument won't work. To see this, suppose that the US is very good at detecting the difference between actual criminal laws and simulated ones. It will sanction the state unless the officials are sincere in criminalizing the conduct. They are then required sincerely to declare that the conduct is publicly wrong.
Perhaps Duff might argue that it is impossible for the state to actually criminalize conduct that is not wrongful. It cannot do so, because to do so requires it to sincerely condemn conduct that is not wrong, and it cannot do this where it does not believe the conduct to be wrong. If 'ought implies can', then the state is not required to actually criminalize the conduct.
But even if 'ought implies can' is true, and has this implication, and even if it is impossible to manufacture sincerity (and I doubt that it is), this argument fails. It would show only that those officials with the right moral view are unable sincerely to criminalize the conduct. But many officials may wrongly believe that the conduct criminalized is wrong. They can criminalize the conduct sincerely, and they are required to do so.
Suppose, for example, that state officials believe that possessing the drug is wrong in US Threat. They can then sincerely condemn people for acting in the relevant way. And, let us suppose, they know that if they do not do so, horrendous things will occur. When they criminalize the conduct, they do what they are required to do, their false beliefs about the wrongness of the conduct notwithstanding. So here is a clear case where the state is required to criminalize non-wrongful conduct, and it can do so. This suggests that any categorical version of the Wrongness Constraint, or just about any other content focused constraint like this, fails. Only pro tanto versions of such constraints meet the challenge of terrible consequences.

Trivial Pro Tanto Principles
In the remainder of this paper, I consider whether Duff's general picture of how we should understand criminalization can nevertheless usefully illuminate pro tanto principles of criminalization. The promise of that view is that it can help us respond to a challenge to the significance of such principles for our understanding of the content and scope of the criminal law.
This section is concerned with outlining the challenge. The next explores how Duff's theory might help us respond to it. Consider some familiar principles of criminalization, such as: Harm Constraint (HC): It is wrong to criminalize conduct that is not harmful. Wrongness Constraint (WC): It is wrong to criminalize conduct that is not morally wrong. Public Wrong Constraint (PWC): It is wrong to criminalize conduct that is not publicly wrong.
The argument in the previous section suggests that we should consider pro tanto versions of these principles, where 'pro tanto' is put in front of the first 'wrong' in each of them.
To understand these principles, then, we need to understand what makes some conduct pro tanto wrong, as well as what it means for conduct to be pro tanto wrong. The challenge is to provide a well-worked out theory of pro tanto wrongdoing that makes these principles both plausible and informative. And that challenge is especially difficult, not only because no one has a well-worked out theory of pro tanto wronging, but also because plausible accounts of what this involves make these principles trivially true.
To see the concern, note that everyone agrees that criminalization has downsides, in every case, or almost every case. If everything with a downside is pro tanto wrong, all of the principles described above are true, but they are trivially true. Consider any principle of the form: 'It is pro tanto wrong to criminalize v if f.' Now note that all instances of criminalization, whether f obtains or not, have downsides. And suppose that downsides make for pro tanto wrongdoing. Then all principles of this form turn out true, whatever f is. For example, suppose that f is very serious public wrongdoing. Criminalizing f has downsides, and if those downsides make criminalization pro tanto wrong, this principle is true: Public Wrongdoing Reversed (PWR): It is pro tanto wrong to criminalize public wrongdoing.
Some well-known views about crime and punishment are especially vulnerable to this line of argument-they make principles such as HC and PWC true, but they also make principles such as PWR true.
Consider Doug Husak's view that it is always pro tanto wrong to punish anyone, including wrongdoers. 8 Then it is pro tanto wrong to criminalize anything at all, including very serious public wrongdoing, on the basis of the following argument: (1) Criminalizing very serious public wrongdoing results in very serious public wrongdoers being punished; (2) It is pro tanto wrong to punish them; (3) It is pro tanto wrong to do something that results in a pro tanto wrong; (4) Criminalizing very serious public wrongdoing results in a pro tanto wrong (5) Therefore, criminalizing very serious public wrongdoing is pro tanto wrong.
Perhaps some might resist (3); but even if (3) is not always true, it does not seem that the falsity of (3) is relevant in this case. Now consider this objection to standard constraints, such as HC, WC, and PWC. If my argument for PWR succeeds, unsurprisingly HC, WC and PWC will turn out true as well. But that also shows that these principles are no more interesting, and actually significantly less interesting, than PWR. In fact, we are better with just the general principle that incorporates these narrower principles into it; a principle that seems warranted if Husak's view of punishment is right: Any Content Constraint (ACC): It is pro tanto wrong to criminalize anything.
If Husak is right about punishment, ACC is true.
Indeed, this principle is probably true even if Husak is wrong about its being pro tanto wrong to punish wrongders (as I think he is). It is probably true on the basis of the following argument: (1) Criminalizing anything results in some innocent people being punished for doing that thing; (2) It is pro tanto wrong to punish innocent people; (3) It is pro tanto wrong to do something that results in a pro tanto wrong; (4) Criminalizing anything results in a pro tanto wrong (5) Therefore, criminalizing anything is pro tanto wrong.
This argument is not completely uninteresting, and nor is ACC. But it hardly provides much guidance in determining the scope of the criminal law.
The concern, then, is to spell out pro tanto principles of criminalization that help us to distinguish just and unjust criminal laws. ACC doesn't help us to do that, because according to it, all criminal laws are pro tanto wrong, and the question is whether it is justified to perpetrate these wrongs all things considered.

Wrong Kind of Reasons
Duff's work suggests a way of meeting this objection. It might be met by distinguishing between internal and external principles. Both sets of principles might be pro tanto, but internal principles tell us something special-they tell us when criminalization is contrary to the internal values or spirit of the criminal law as such.
It is plausible, at least at first blush, that something like WC or PWC are grounded in the internal values or spirit of the criminal law. In contrast, even if PWR, or ACC are true, they are not grounded in the internal values or spirit of the criminal law. They are true because sometimes the enterprise of the criminal law goes wrong, and that is contrary to its spirit. Thus, we might distinguish internal and external pro tanto principles, where these principles are distinguished by their grounds.
Those sympathetic to this basic structure for the evaluation of the criminal law might disagree about a wide range of topics. First, they might disagree about what the intrinsic features of the practice are. Second, they might disagree about what conduct is suitably related to these intrinsic features. Third, they might disagree about whether criminalization is always pro tanto wrong where these intrinsic features are not present. Fourth, they might disagree whether criminalization can be pro tanto wrong even where these intrinsic features are present, and when extrinsic goals outweigh them. And fifth, they might disagree about when extrinsic reasons are sufficiently weighty to justify any pro tanto wrong involved in criminalizing unsuitable conduct. Here I focus on a way of understanding the distinction between intrinsic and extrinsic features of the practice. This structure for evaluating the criminal law has a close relationship with the more general problem in philosophy of the Wrong Kind of Reason. Understanding that problem helps us to clarify the distinction between internal and external principles. Here is the basic idea. Certain kinds of reasons seem like the right kinds of reasons for a person to perform certain actions, have certain states of mind, have certain emotional reactions, and so on. They are the right kind of reason because of the intrinsic relationship between these reasons and the actions, states of mind, and emotional reactions they relate to.
For example, the fact that there is evidence supporting p is the right kind of reason to believe that p; the fact that v is a choice worthy action is the right kind of reason to form the intention to v; the fact that f is regrettable is the right kind of reason to regret f; the fact that X is admirable is the right kind of reason to admire X; and so on. Now note that 'external' reasons might justify all of these actions, states of mind, reactions, and so on. A standard way to show this is to imagine an evil demon that will cause a catastrophe if the person does not have the relevant feature, but the right kind of reason for that feature is not present. Then there is decisive reason for that feature, but it is of the wrong kind. So for example, suppose that an evil demon will cause total world destruction if I don't believe p even where there is no evidence for p; or intend to v even where ving is in no way choice worthy; or regret e where e is in no way regrettable; or admire x even where x is in no way admirable; and so on. Then I have decisive reason to make it the case that I have the relevant feature if I can do so, but it is the wrong kind of reason. It may be difficult to get myself to have the relevant feature-it isn't easy to form beliefs without evidence, for example. But I would have a reason to do so were I able.
Exactly how to draw the contrast between the right and wrong kinds of reason has been the subject of extensive debate, and it has proved by no means easy to draw that contrast in a principled way. But it seems clear that there is an important contrast to draw. For shorthand, I will say that the right kinds of reasons are aptness reasons. The wrong kinds of reasons are those that justify making it the case that I have the relevant feature but are not aptness reasons. I thus call them 'non-aptness reasons'. For example, the fact that there is evidence that p is an aptness reason to believe p, where the fact that an evil demon will destroy the world if I don't believe p is a non-aptness reason to do so. Obviously, aptness and non-aptness reasons can coincide-where there is decisive evidence that p and the evil demon will destroy the world if I don't do so, for example.
This structure allows us to develop principles about such features of the following kind: only make it the case that you have feature, f, if certain facts, a, are true. In such principles, a are the set of facts that make f apt. Here are some examples: 'only believe propositions that are supported by sufficient evidence'; 'only form intentions to do acts which are choice worthy'; 'only regret things that are regrettable'; and so on. Obviously, 'only' in these principles is restricted to the aptness considerations that apply to the relevant feature. We know that from the fact that non-aptness considerations can provide people with decisive reasons to have the relevant feature. Such principles could thus be rewritten: there is an aptness reason to have feature f only if certain facts, a, are true.
Such aptness reasons might also generate deontic principles. For example, 'X ought not to have some feature, f, unless certain facts, a, are true.' And some might state moral duties. For example, 'it is morally wrong to have some feature, f, unless certain facts, a, are true.' As moral wrongness can be pro tanto or all things considered, there are also two variations of this last idea. And as non-aptness reasons can sometimes be decisive, the pro tanto version is the only plausible version.
Consider blame. There are aptness and non-aptness reasons for blame. Aptness reasons include the fact that a person has acted wrongly, is responsible for doing so, and so on. Non-aptness reasons include being threatened by an evil demon with total world destruction. It is also plausible that it is pro tanto morally wrong to blame people for conduct who have not acted wrongly. Thus, this principle is plausible: 'it is pro tanto morally wrong to blame a person for conduct where that person has not acted wrongly.' But not all aptness reasons are also duty-conferring. It may be irrational to believe that p without evidence that p, and thus I ought to believe p only if I have evidence that p. But there is nothing wrong, or at least nothing morally wrong, with believing p without evidence of p.

Aptness and Criminalization
It may seem that criminalization is not a good candidate for the aptness/non-aptness distinction, because criminalization is a practice. But it is at least somewhat plausible that practices have internal standards of correctness that set standards where there can be right or wrong kinds of reason to perform those practices in certain ways.
Mark Schroeder gives the example of knot-tying. 9 The internal purpose of knottying is to make ropes secure. That sets the standards of correctness for knot-tyingcorrect tying is tying in a way that makes ropes secure. There are right kinds of reason to tie a knot in a certain way-those that meet these standards of correctness. And there are wrong kinds of reason to do so-an evil demon who threatens me that if I do not tie a knot in a certain way, where that way is less sound than some alternative, arguably gives me the wrong kind of reason to tie the knot in that way. That explains why we say that the demon gave me a reason to tie the knot incorrectly.
A similar thing might be said of criminalization. We might be able to identify the central purpose of the criminal law. This sets standards of correctness for criminalization. And that determines what we ought to criminalize, just as there is a particular way in which we ought to tie knots. We can have incentives to do something else with the criminal law, but if we respond to those incentives we do not meet the relevant standards of correctness.
As we can see from the knot-tying example, it is a further step to show that failing to meet the relevant standards is pro tanto wrong, let alone that it is pro tanto wrong because it fails to meet the relevant standard. 10 In the knot-tying example, for example, tying the knot incorrectly need not be pro-tanto wrong. In fact, it need not even be irrational to fail to meet the relevant standards of correctness, and thus there need be no reason at all against doing the relevant thing. If there is no reason to make the rope secure, there is no reason at all against tying the knot incorrectly. Indeed, sometimes there are only reasons in favour of tying knots incorrectly-for example to demonstrate to others what happens when this is done, or where the knot ties a rope that others intend to use to hang an innocent person. This is in contrast with, for example, forming a belief where there is no evidence because of a threat, or forming an intention where the action intended is not at all choice-worthy because of some threat. Doing these things is in some narrow sense irrational (even if there is a wider sense in which doing them is rational). And a person who is justified in responding to the relevant incentives nevertheless fails to conform to all of the reasons that apply to her. Even if I am justified in believing p with no evidence that p, given the threat I face, there is something regrettable about my believing that p with no evidence that p.
This suggests that an account needs to be given why violating the principles of correctness that are established by the internal aims of the system of criminal law is also pro tanto wrong. It cannot be assumed at the outset that this is so, merely because there are internal standards of correctness in the criminal law. But it may seem that this challenge is also easily met. If the internal aims of the criminal law are to publicly blame those who have acted in a publicly wrongful way, the standards of correctness for criminalization are to criminalize conduct that warrants public blame: public wrongdoing. Criminalizing conduct that does not warrant public blame involves warranting publicly blaming people for public wrongdoing where they have not acted in a publicly wrongful way. And it is pro tanto wrong to do that. Later, we will see a challenge to this argument, but it at least gives us some reason to think that abiding by standards of correctness in the criminal law is morally required, at least pro tanto.
It is also worth highlighting that the relationship between the intrinsic features of criminal justice and criminalization is more complex than examples such as knottying and belief formation may suggest. The principles just outlined for those practices and states directly to the feature that aptness is concerned with. For example, because it is inapt to blame a person for wrongful conduct, and because that inaptness has pro tanto deontic force in all cases, there is a general principle that it is pro tanto morally wrong to blame a person for conduct where that person has not acted wrongly. Now consider the criminal law. If the practice of criminal law is holding wrongdoers publicly to account for public wrongdoing, and if that practice is apt only if the person has acted in a publicly wrongful way, there is a general principle that it is pro tanto wrong to exercise the practice of criminal law on non-public-wrongdoers. But this does not yet establish that it is pro tanto wrong to criminalize non-wrongful conduct. To establish that, a bridging principle is required-something like: Bridging Principle (BP): if it is pro tanto wrong to exercise the practice of criminal law on people for ving, it is pro tanto wrong to criminalize ving.
For reasons similar to those outlined earlier, BP is only plausible as a principle of pro tanto wrongness. The wrong kinds of reasons might justify criminalizing conduct where it is wrong to exercise the practice of criminal law on these people. For example, if an evil demon would destroy the world were some conduct not criminalized, but it would be pro tanto wrong to exercise the criminal law on those who acted in that way, criminalization would be justified all things considered.
Thus, BP is best understood in this way. It is only apt to criminalize conduct where it is apt to exercise the practice of criminal law on those who perform that conduct. Where it is inapt to do this, the inaptness always grounds pro tanto moral wrongness. Therefore, it is pro tanto wrong to criminalize conduct where it would be inapt to exercise the practice of criminal law on those who perform that conduct. Aptness thus enters the theory of criminalization twice-once in the exercise of the criminal law on people, and again in the relationship between criminalization and the exercise of criminal law on people.
How could this view be vindicated? I think the best way is this. Criminalizing conduct, v, is a warrant to the state to exercise the practice of the criminal law on those who v. Such a warrant is obviously subject to certain constraints-prosecutions need not occur where doing so is extremely costly, or would be overly burdensome to the victim or the wrongdoer, for example. But criminalization is at least a warrant that claims that certain facts that justify the exercise of the criminal law on those who v are met-those that make the exercise of criminal law apt. And it is apt to warrant this only if it is true. As the inaptness of doing this makes doing it pro tanto morally wrong, something like BP is true.

What is Intrinsic to the Criminal Law?
It may seem that the argument in the previous section is sufficient to show that there are internal standards of the criminal law, that make it pro tanto wrong to criminalize non-publicly-wrongful conduct (or, alternatively, just non-wrongful conduct).
Here is a standard view about the criminal law. It is an institution that governs holding people to account for certain kinds of wrongdoing in order that an appropriate response to that wrongdoing can be made both by the wrongdoer and by state institutions.
There is little doubt that this view is correct. Given my suggestion in the previous section, this may seem to vindicate a principle like the following in a particular way: Wrongness Constraint: It is wrong to criminalize non-wrongful conduct.
Here is the argument.
(1) Criminalizing v warrants the exercise of the criminal law on those who v.
(2) It is apt to warrant the exercise of the criminal law on those who v only if it is apt to exercise the criminal law on those who v.
(3) The exercise of the criminal law involves holding people to account for certain kinds of wrongdoing. (4) It is apt to hold people to account for wrongdoing only if they have acted wrongly.
(5) Therefore, it is apt to criminalize conduct only if that conduct is wrong.
(6) It is pro tanto wrong to criminalize conduct where it is inapt to do so. (7) Therefore, it is pro tanto wrong to criminalize non-wrongful conduct.
However, although this seems compelling, the argument fails. First note that the argument just outlined is invalid. It is made valid only if premise (3) is replaced with: 3*) The exercise of the criminal law necessarily involves holding people to account for certain kinds of wrongdoing.
Here is why. Anything weaker than 3* leaves it open that the exercise of the criminal law need not involve holding people to account for wrongdoing. But then, the rest of the argument fails. It may be apt to criminalize conduct that is unsuitable for the practice of holding people to account for wrongdoing, for in criminalizing that conduct, the state need not warrant holding the person to account for wrongdoing.
What Duff needs to make the argument go through, then, is an argument for the view that the exercise of the criminal law is uniquely a process of holding people to account for wrongdoing. Only with an argument for that view do we have the result that there is only one set of standards of correctness for criminalization, where criminalizing non-wrongful conduct fails to meet those standards.
An alternative view is that the criminal law is a multi-function system. On that alternative, there are various different aims of the criminal law. These different aims set different standards of correctness. Criminalization is apt as long as one set of standards of correctness for criminalization is met.
Compare claw hammers (my tool of choice for hypothetical examples). They are for banging in nails and pulling out nails. There are standards of correctness for both these things-use the blunt end for banging in nails, and the claw bit for pulling out nails. Banging in a nail with the claw bit is an incorrect use of the hammer, as is pulling out a nail with the blunt bit.
So now we should work out how to determine what facts are intrinsic to the criminal law. Is it uniquely a blaming system, or a public blaming system? Or is it a multifunction system? Of course, we could just stipulate that the criminal law, as it is to be understood for the purposes of some argument, is a public blaming system. And we could call the remainder a system of public regulation. But we have already seen that this narrows our inquiry very substantially. It leaves out offences that are created that determine a person's liability to state punishment, where there is no ambition publicly to blame people; offences that people think of as part of the criminal law.
Furthermore, when we consider what principles are internal to the criminal law, a natural thing to consider is the reasons that we have to create and sustain a system of criminal law. Just like hammers, the main way in which we identify what the system is for is to assess what is created for.
For both core, and non-core, offences, the main reasons have nothing to do with public blame. They include a wide range of social goods that the criminal law helps to protect, including protecting people against serious wrongdoing, and stabilizing expectations to enable people to plan and flourish in their work and private lives. When visiting societies where criminal law begins to break down, the main loss is not a loss of adequate accountability, but rather a loss of adequate protection and stability.
None of this is to say that accountability is unimportant. Often, it plays a crucial role in achieving the instrumental ambitions of the criminal law, and it is valuable in itself. But the idea that this is the main aim of the system, or the main aim that people ought to have in creating and sustaining it, is false. 11 For these reasons, I prefer to see the criminal law as a multifunction system-a system that creates some offences that warrant public blame, and others that do not. There is then a question which of these offences is justified. Where an offence is created that warrants public blame, the conduct had better be wrong. And, if there is a genuine idea of public wrongs that limits the scope of the law, it had better be publicly wrong. But as long as some offence is not created with a warrant for state officials publicly to blame people for wrongdoing, it is no objection to the creation of such offences that the relevant people are not public wrongdoers, or not wrongdoers at all.
There may be other objections to punishing non-wrongdoers for reasons of deterrence. I think that such objections are often, but not always, decisive. But no argument about the conditions under which it is apt to condemn people for public wrongdoing touches such offences. What Duff needs to vindicate his account, then, is an argument for the view that parts of the criminal justice system that do not aim at holding people to account for public wrongdoing are either unwarranted, or are not part of the criminal law.
I don't think that there is much prospect for an argument of this kind. As a conceptual matter, I think that Duff's account of what the criminal law involves does not capture all of the things that are appropriately called criminal law. And as a matter of critical enquiry, I see no reason to narrow our investigation into criminalization to those things that Duff calls criminal law. Furthermore, I think that there are instances of criminalization that Duff does not call criminal law, but that are appropriately called criminal law, and that are fully justified. There is a question whether these laws infringe the rights of citizens, but then we have seen that all acts of criminalization infringe the rights of citizens.

Conclusion
My conclusion, from this critical argument that I have ended with, though, is not that ROCL is a failed project. Rather, ROCL is better seen as a deep and important exploration of a central part of the practice of criminal law. That part of the criminal law may well have its own internal standards, and for all that I have said, Duff may be correct about what those standards are. That may give rise to a set of standards and principles that are similar to those Duff defends (though these can only give rise to pro tanto duties). Even if that is right, though, there are other aspects of the criminal law that Duff's project leaves untouched, and that provide an important focus of attention for the philosophy of criminal law.