I.

It has become common, in recent legal and political philosophical debates, to argue that the criminal process—from criminalization through to prosecution, trial, and punishment—has a communicative or dialogical dimension.Footnote ² On this view, the criminal process is a distinctive form of state coercion, in large part because of its expressive features: legal punishment is seen not simply as a tax on certain sorts of behavior, but rather as an expression of moral disapproval, and the criminal law as the public declaration of those acts that will occasion such disapproval. This paper explores the implications for our thinking about political authority of accepting this communicative approach. More specifically, it looks at the implications of accepting the following three propositions:

While these propositions have some intuitive force, not everyone accepts them. Although this paper will not establish that it would be wrong to reject these propositions, we will have something to say in their favor below. To begin with, however, some brief expository comments are in order.

As stated, MALA IN SE only gives particular examples of the prelegal wrongs that should be crimes. The list could and probably should be extended, but whether or how far it should be is a matter for a more developed communicative theory to settle. The point at present is just to find some crimes—perhaps central mala in se—regarding which it is intuitive that they should be part of a criminal code if one exists, and for which MORALISM and MORAL CENSURE are plausible. Certainly, it is hard to imagine a penal code or a criminal justice system that does not have laws against prelegal wrongs such as murder, rape, assault, and so on. However, MALA IN SE does not claim that all crimes are or should be pre-existing moral wrongs, or that it is a necessary condition of being a crime that an act should also be a moral wrong. It only claims that if there are to be such things as crimes then some central moral wrongs should be among them. This is important. The implications drawn here about the authority of criminal law emerge by virtue of the inclusion of some moral wrongs in that criminal law—as long, that is, as MORALISM and MORAL CENSURE are also met. To see the force of these implications, however, it is not necessary to think that all crimes should be thought of on the model of MALA IN SE.

Now, there might be various grounds for thinking that MALA IN SE is true. MORALISM offers one such ground. It claims that there being weighty, normally decisive moral reasons against performing such actions provides a central reason for criminalizing them. Some views reject this kind of legal moralism—we will hear more about this later. For now, it is simply worth noting that MORALISM is moderate, in that it does not take moral wrongness as a sufficient reason for criminalization. For instance, MORALISM is compatible with the view that only prelegal wrongs that are in some way “of legitimate public concern” should be criminalized. But MORALISM holds that any further necessary justifying conditions of criminalization do not undermine the claim that the moral wrongs mentioned in MALA IN SE are crimes, when they are, because they are serious wrongs. Being a prelegal wrong of legitimate public concern, for instance, could be a sufficient reason for criminalization in respect of at least some crimes.

MORAL CENSURE adds to this that punishment for such crimes should have an expressive dimension, where the object of disapproval is not simply that the law has been broken, but that a prelegal wrong has been done. It is important to note, however, that MORAL CENSURE does not claim that this expressive dimension should be taken to be the general justifying aim of punishment. It could be, but MORAL CENSURE can be accepted by many who think that it is not. The significance of this expressive dimension for the current discussion rather concerns the proper role of considerations of prelegal wrongness in the criminal process. According to MORAL CENSURE, such considerations should not only motivate the state to criminalize certain actions and make their perpetrators liable to punishment; the state should also publicly support or stand up for those considerations through the criminal process, in particular through the expressive dimension of punishment. The convicted offender is to be treated as one who has violated weighty and decisive moral reasons in acting as they did. This is a view held by many theorists who accept some kind of hybrid theory in which moral communication plays a role in criminal justice, even if they reject the view that it plays a central role.

To illustrate the three propositions, consider an example to which we will return: the English Hunting Act 2004, Section 1 of which states that “[a] person commits an offence if he hunts a wild mammal with a dog, unless his hunting is exempt.” It might be objected that foxhunting is not a central example of an act falling under MALA IN SE—we return to the question of whether it is disputable later—but assume for the present that there are decisive moral reasons against it such that it is an act of cruelty and a serious moral wrong. According to MORALISM the balance of those prelegal reasons—or otherwise put, the fact that foxhunting is cruel—can, if any further necessary justifying conditions are met, appropriately be the state's central reason for criminalizing it. And MORAL CENSURE says that, in convicting and punishing those who commit an offense according to the Hunting Act Section 1, the state should express disapproval of the offender for engaging in a cruel activity.

II.

We can now begin to ask what the implications are of accepting these three propositions for our thinking about political authority. The question we will be focusing on is whether the criminalization of serious prelegal wrongs should be considered an exercise of authority, and if it should, how that exercise of authority should be conceived. To be more precise, our concern lies in saying what normative change, if any, comes about for those subject to that penal code when a new law dealing with a serious prelegal wrong comes into force. We can capture this by asking what our attitude should be to a further proposition:

ALTERATION is plausible because it is plausible that some essential part of the function of criminal law is to have a normative effect on subjects. But what could the effect consist in? We assume that subjects already have weighty and decisive moral reasons to avoid those actions covered in MALA IN SE. And we assume that criminalizing the wrong meets any further necessary justificatory conditions. This being so, do subjects have any more reason—and, in particular, reason deriving from the authority of the state—to avoid those actions as a result of the law coming into force? If they do, how do those new reasons stand in relation to the prelegal reasons they already had? If our three propositions are correct, these questions arise for all systems of criminal law, since those propositions hold that, even if not all crimes involve serious prelegal wrongs, all such systems should criminalize at least some such wrongs because they are wrongs.

It might seem that there is an obvious answer to the question of whether ALTERATION is true (or would be true if the criminal justice system functioned as it ideally ought to). Prior to the law coming into force, those who engaged in foxhunting were, by hypothesis, committing a serious moral wrong and thus had decisive reasons not to do so. However, once the law comes into force, they are liable to be put through the unpleasantness of a police investigation, prosecution, perhaps a trial, and, if found guilty, punishment (in the case of the Hunting Act 2004, a fine). Given the unpleasantness of these consequences and the probability of their being apprehended, the would-be foxhunter has a strong additional prudential reason that they did not have prior to the law coming into force. In a justifiable criminal justice system, ALTERATION is true, it might be said, but what makes it true is just the creation of this prudential reason in addition to the pre-existing moral reason.

However, straightforward though this answer may appear to be, it cannot be the whole story—at least not if we think that what makes ALTERATION true is a normative change distinctive of the exercise of authority (though it may turn out, of course, that we are wrong to think this). This is because those prudential reasons are not reasons of authority, and the creation of prudential reasons is not an exercise of authority. Prudential reasons, in this context, are reasons arising from the threat of coercive power rather than right. This is not to say that no questions of right in some sense of that term arise when the only normative change is the creation of new prudential reasons. After all, the state may be justified in threatening potential foxhunters with the exercise of coercive power. And it may be that it is also justified in claiming a monopoly on the use of such coercion. However, if the justified triggering of new prudential reasons through the threat of coercion were the only normative change introduced for subjects when a new criminal law is introduced then this would not amount to an exercise of state authority. For a justified threat of state coercion is not yet the exercise of a normative right or power to alter subjects’ rights and responsibilities in relation to the bearer of that right or power. The exercise of authority involves right rather than coercion in the sense that it requires the possession of a particular normative power ranging over normative rights and responsibilities, whereby the bearer of authority can alter, by issuing a directive, what it is that those subject to the power have a duty to the bearer of the authority to do, or what their rights and liabilities in relation to the bearer of authority are.Footnote ⁴ It is in this Hohfeldian sense of “right” that we find what is distinctive about authority in contrast to coercion.

Considered as a distinctive normative power, practical authority is widely assumed to consist in a right to rule that correlates with a duty of obedience. On this understanding, practical authority is a power through the exercise of which (for instance, by issuing a relevant directive) its bearer can create new binding practical reasons for those subject to that power.Footnote ⁵ That is, an authority, if legitimate, has a right to tell subjects what to do, such that subjects have a duty, owed to the bearer of authority, to do as they are told. This claim about a right to be obeyed is what is at issue in the debate over whether the state needs to claim authority or whether the justified, unique right to coerce (and hence to create new prudential reasons for subjects) is sufficient.Footnote ⁶ If authority consists in a right to be obeyed then the exercise of authority involved in introducing criminal laws would involve the state providing subjects with a new duty not to murder, rape, assault, etc. This would be a duty that is grounded, not directly in the decisive reasons they already have not to do these things, but rather in the fact that they have been told not to. If, by contrast, the only normative change brought about by a new criminal law were the creation of new prudential reasons then we should conclude that criminalization is not an exercise of authority.

III.

Those who accept the three propositions mentioned at the start, however, think that the criminal process should treat the pre-existing reasons as decisive, rather than standing on the authority of the state. In condemning for murder and rape, on this view, the state should not say, “You did wrong because you disobeyed us”—rather, it should point to the prelegal wrongfulness of such actions. From the perspective of the communicative approach, therefore, the view that authority consists in a right to be obeyed goes wrong in a number of ways when applied to the cases falling under MALA IN SE. First, treating the decisive wrong-making feature of the crime as a violation of obligations of authority gets the focus wrong. If authority is the central issue then the victim is the state, whereas according to MORALISM, it is the wrong to the actual, proximate victim that is the central reason for criminalizing the action and making its perpetrators liable to censure and sanction. Second, the exercise of the right to be obeyed is often taken to have the effect of replacing the pre-existing reasons with new reasons, the force of which derives from the source of the directive rather than its merits.Footnote ⁷ But one who accepts the three propositions will regard such replacement as wholly inappropriate in the case of those actions covered by MALA IN SE. In regard to those actions, they think, criminalization and punishment should rather be in the business of affirming the pre-existing reasons. Third, even if the exercise of authority were taken to add reasons of authority to the existing moral reasons, rather than replacing one with the other, those who accept the three propositions will regard this as pointless—or worse. Pointless because the force of the pre-existing reasons dwarfs the force of this duty to obey, and so no significant normative change is brought about by their introduction. And potentially worse than pointless because adding a further consideration that counts in favor of not doing these things may suggest—wrongly—that there is not already sufficient reason not to do them. Thus if there is no more to authority than the right to be obeyed then claiming authority for the criminal law would be unnecessary, because citizens do not need any stronger reason not to perform these actions, or even distorting, because in demanding that citizens act for reasons of obedience the state would be displacing or interfering with the genuine, pre-existing reasons not to commit these actions.Footnote ⁸

This might suggest that accepting the three propositions outlined at the start commits us to the view that there is no exercise of authority involved in regard to crimes covered by MALA IN SE. It might suggest that what makes ALTERATION true of such crimes, when it is true, is rather that a new criminal law coming into force involves the creation of new prudential reasons, and nothing more. Nevertheless, I will now claim, this suggestion for justifying ALTERATION misses out on an important feature of the communicative view, a feature that commits it to a distinctive conception of authority.

IV.

The line of thought so far can be summed up as follows: we want to know what makes ALTERATION true in a manner compatible with the communicative view outlined at the start; we assume that this is to be explained either by the creation of new reasons of authority or the creation of new prudential reasons; however (it appears) it is not consistent with the communicative view to think that what makes ALTERATION true is the creation of new reasons of authority; therefore (it appears) if ALTERATION is true it is to be explained by the creation of new prudential reasons. This line of thought would suggest, then, that, in order to affirm ALTERATION, the communicative view is committed (a) to criminalization creating new prudential reasons through the threat of punishment, and (b) to regarding the creation of such reasons as the distinctive normative change that comes about through a change in criminal law. However, as we will now see, those who are committed to the three propositions detailed at the outset have good reason to reject either (b) or both (a) and (b).

Some influential communicative theorists such as R. A. Duff would reject (a) out of hand—and hence, by implication, (b). Duff has argued that, if the reasons for compliance that the criminal law offers to citizens are either prudential reasons or reasons to obey then the moral claims made by the law are grotesquely distorted. According to MORAL CENSURE, the reasons that the state gives in censuring the offender for those crimes covered by MALA IN SE are the pre-existing moral reasons (reasons existing independently of criminalization) that the offender violated in committing the crime. These are the reasons, we might say, with which the state demands compliance. By contrast, where the state takes itself to be creating new prudential reasons then it becomes unclear whether it is the moral or the prudential reasons with which the state demands compliance:

If, in criminalizing certain actions, the state takes itself to be creating new prudential reasons for subjects, then, Duff argues, it would be offering those prudential reasons to subjects as their grounds for compliance rather than the pre-existing moral reasons that motivated it to criminalize such conduct in the first place. Thus if one takes Duff's position one will have to explain ALTERATION otherwise than by appeal to the creation of prudential reasons—or reject it altogether.

Now many communicative theorists reject Duff's view and think that, for instance, a “mixed” censure theory such as that of Andrew von Hirsch's is acceptable.Footnote ¹⁰ They do not accept that, in intentionally creating prudential reasons, the state thereby undermines its commitment to the wrongness of the criminalized actions. However, even hybrid theorists should reject (b): that is, they should agree that the creation of prudential reasons is not all that makes ALTERATION true. The reason for this is that any version of the communicative theory is committed to the view that an important normative change, other than the creation of prudential reasons, comes about with the introduction of a new criminal law.

To see this, consider that once a new criminal law is in place, subjects’ liability to state censure has been altered. Once hunting wild mammals with dogs is criminalized, for instance, subjects have a new liability to be censured by the state for engaging in that activity. Otherwise put, once it is criminalized, the state has a right that it did not previously have to censure those who engage in hunting. Those committed to MORAL CENSURE (and the other two propositions) are committed to the existence of this normative change as long as they are also committed to the view that there can be no justified punishment without prior law (“nullem crimen, nulla poena sine lege”). If they are committed to the nulla poena maxim, but think that state censure can be justified once legislation is in place, then they must regard legislation as having changed subjects’ liability to state censure. Indeed, since punishment should only be part of a wider criminal process, the communicative view is probably also committed to thinking that, once a criminal law is in force, the state has rights that it would not previously have had to: (a) investigate those who are reasonably suspected of such actions; (b) require those who have a “case to answer” to appear before a tribunal at which their case will be heard and a judgment regarding their culpability will be made; as well as to (c) express proportionate censure of their offense if they are found guilty. If this is the case then communicative theorists who accept nulla poena are committed to the idea that legislation coming into force alters subjects’ liability in any instance in which the state is justified in pursuing an individual through the criminal process.

Is this new liability to censure a change in one's reason, rights, and duties, as ALTERATION requires? Yes, since the communicative view is presumably committed to the claim that criminalization brings it about that, for example, subjects have no right to refuse to be tried if there is reasonable suspicion that they failed to comply with such reasons, and no right to refuse being subject to public censure, should they be found at trial to have failed to comply with the law's demands, with no reasonable form of denial, justification, or exculpation.

Nevertheless, if we now return to our main question of what bearing all this has on the question of authority, it might be said that it has no bearing. Practical authority, it might be said, is a normative power to create new reasons for subjects, whereas the sense in which we have said communicative theorists are committed to the truth of ALTERATION rather concerns changing their liability for compliance with a designated set of pre-existing reasons. What we are proposing that criminal legislation does, in relation to those acts covered by MALA IN SE, is to select among putative pre-existing reasons for action and to determine those reasons the violation of which will make the violator liable to public accountability and censure. This would leave the reasons that apply to subjects unchanged and therefore, the objection concludes, does not involve authority.

By contrast, I now want to argue that this alteration of liability is precisely what the authority of the criminal law must consist in if it is to be compatible with accepting the three propositions laid out at the start. The introduction of legislation can only change subjects’ liability to state censure in the ways we have just looked at if the state has a prior right so to do. By “right” here, we mean what Hohfeld called a power—an ability to change the rights and duties of those subject to one's power. The possession of this power to which subjects are liable, and through which the state can alter their liability to investigation, tribunal, and censure, is a form of practical authority. My suggestion is, therefore, that authority is indeed required for the criminalization of the wrongs covered by MALA IN SE, but that our discussion so far should lead us to rethink the nature of practical authority. Practical authority can be not only a right to command or direct through the creation of new binding reasons, but a right to oversee compliance with a designated set of pre-existing reasons. Practical authority in this latter sense is therefore what I call the authority of moral oversight.

This can be so, for instance, not only in the case of the state, but of a professional body such as the Bar Association or the General Medical Council. The authority such professional bodies have over those who practice the associated professions is not simply that they may order lawyers or doctors to comply with their demands—though within limits they can do that too. More fundamentally it is that they have the power to select among pre-existing reasons that apply to lawyers or doctors by virtue of their role, thus to designate what the professional standards or obligations of these roles are, and then to oversee compliance with those standards and obligations. The same, my suggestion is, goes for the state. Rather than creating new binding reasons for subjects, the power invoked by MALA IN SE instead leaves the pre-existing moral reasons as they are and designates those moral reasons the violation of which will make one liable to public accountability and censure. The criminal law in relation to actions covered by MALA IN SE claims authority, not to control or guide behavior by adding reasons, but rather to oversee its moral quality.

Although the authority of moral oversight is different from the right to be obeyed, we should note that it must be a necessary part of that authority that the state also has the right to issue authoritative directives. For instance, it must have the right to order someone to attend a trial, or to submit to a sentence, etc. However, the crucial thing to note is that it has these rights to command as essential parts of its power to determine which acts will call for censure. It is because it has the power to legislate on acts that are to merit public condemnation that it has adjudicative powers to determine whether those acts have been performed, placing subjects under an enforceable obligation to answer to it for their actions, and executive powers to issue public condemnation—again by placing subjects under an enforceable obligation to serve a particular sentence—when it finds that those acts have been performed. The state may also have rights to deploy coercion in enforcing its authority. However, as we can see with the examples of the Bar Association and the General Medical Council, rights of coercion do not necessarily follow possession of such authority. The state's rights in criminal justice are therefore not simply rights to do what is necessary, within the scope allowed by individual rights, for public security. Rather the shape of its rights to coerce and command is determined, in large part, by its authority to determine the proper objects of public censure.

V.

As we have seen, Duff argues that the criminal law should not be seen as authoritatively prohibiting certain conduct. He argues that it should rather be seen as authoritatively “declaring … conduct to constitute a public wrong properly condemned by the community, for which the agent is answerable to the community through the criminal process.”Footnote ¹¹ In criminalizing certain actions, he thinks, the state should not claim to create new wrongs by making it the case that citizens have new rights to obey; rather, criminalization should be seen as an authoritative “declaration” that makes no difference to the reasons citizens have to refrain from the criminalized conduct. Nevertheless, Duff's view is committed to ALTERATION being true, since he thinks that once the state thus “identifies [these actions] as public wrongs” it thereby “imposes on citizens a duty to answer for [any] alleged commission of such wrongs through the criminal process.”Footnote ¹² This passage, coming in a section entitled “The Authority of Criminal Law,” raises, although it does not answer, the question of how the state could have such authority.

This will turn out to be a pressing question. However, if we are to be able to answer it, we need to pose it precisely. Duff's formulations are highly suggestive, but they leave it unclear what exercise of authority is involved in the criminalization of acts covered by MALA IN SE. It is compatible with his formulation that criminalization, for instance, might just involve an exercise of the right to be obeyed that creates obligations of a conditional form. That is, rather than the authoritative directive “I hereby place you under an obligation not do action X, Y or Z,” which we earlier criticized as unnecessary or distorting, the authority of criminal law could simply be reformulated as “I hereby make it the case that, if you do action X, Y or Z, then you will be under an obligation to answer for your actions through the criminal process (i.e., you will be under an obligation to appear in front of the tribunal that will decide on your guilt; and if found guilty at that tribunal you will be under an obligation to serve a sentence that expresses censure of your action).”

While this conditional formulation assimilates criminal legislation to a familiar model—a conditional in which the antecedent specifies an action and the consequent a legal or normative consequence that will come about if that action is performed—we can now see that it is in fact inadequate to understand the powers that—if we accept the propositions outlined at the start—the state claims to be exercising over its citizens in the act of criminalization. The state, we have seen, only has a right to impose such conditional obligations on subjects because it possesses the power to determine which actions merit public accountability and condemnation and to make subjects correspondingly liable. Without that power to determine what shall be condemned, the power to issue conditional commands would lack any grounding. To see the power to issue such commands as the heart of the state's claims to authority would be to fail to capture what is distinctive about the criminalization of wrongs covered by MALA IN SE. Understanding the implications of the communicative view therefore requires us to recognize that the normative power in which authority consists can involve something other than the right to issue authoritative directives and to be obeyed. More fundamentally and more interestingly, in the case of criminal law it is the power to designate a set of moral reasons and oversee compliance with them.

We should remember, however, that the three propositions set out at the start are accepted by a wider range of theorists than those who would regard themselves as fully paid-up proponents of the communicative theory of punishment. The three propositions are accepted by those who are committed to moral censure playing some role in criminal justice, even if they deny that moral communication is the fundamental justificatory purpose of criminal justice. If I am right, therefore, that accepting the authority of moral oversight is a straightforward implication of accepting those three propositions, then any theorist who thinks that moral communication should have a role in state criminal justice is committed to the state claiming the rights constitutive of the authority of moral oversight.

VI.

Having laid out the main argument of the paper, I now turn to objections. In the sections after this one, we will consider some concerns about whether a state's claims to possess the authority of moral oversight could be vindicated. First of all, however, I would like to consider an objection to the effect that I have failed to show what I have claimed to show. I have claimed that accepting the three propositions outlined at the start commits us to a particular interpretation of ALTERATION. However, it might be argued that I reach this conclusion only through an equivocation: the conclusion, that is, does not really follow from my premises. I claim that, if we accept the nulla poena maxim, the three propositions commit us to regarding the state as having new rights to censure in virtue of having legislated. The state has no right to censure prior to legislation, but it does once legislation is passed. However, it might be said that in making that claim intuitive, I give examples in which the state does much more than censure—it apprehends, summons, detains, coerces, punishes. These latter are all interventions to which the nulla poena maxim applies, and that, for good rule of law reasons (such as fair warning), the state would have no right to carry out without prior legislation. Having made it intuitive that legislation must be taken to change liability to such intervention, the criticism goes, I then use that premise to motivate my conclusion that legislation must also be taken to alter liability to censure. However, the criticism insists, this conclusion does not follow. For it is not plausible that the state only has a right to censure on the condition that it has so legislated. Just as any private individual does, the state has the right to censure whenever its censure is justified. The reason there needs to be prior legislation for the criminal process to be justified has to do with the fact that the criminal process is a coercive intervention into subjects’ freedom, and not, as I have claimed, with its communicative character.

One way to quibble with this criticism would be to query the claim that private individuals have a right to blame whenever blame is justified. Recent debates in the ethics of blame have suggested that ordinary interpersonal blame requires not only justification but standing (otherwise put, that one is within one's rights to blame).Footnote ¹³ We can see that such standing is required, it is argued, because it can be lost in three ways: in the case of hypocrisy; in the case where it is “none of your business”; and in the case where one has oneself culpably contributed to the person committing the wrong for which one is blaming them.Footnote ¹⁴ Nevertheless, a response along these lines does not approach the main issue raised by the criticism. The criticism targets the assumption I made in stating my argument, that state censure of crime is different from interpersonal blame in requiring prior legislation, and hence an exercise of authority, for its justified use. It is not clear that pointing to the standing purportedly necessary for interpersonal blame will help to answer that criticism.

In responding to the criticism, I will therefore instead draw attention to the fact that state censure of crime refers specifically to the state's right to determine what merits such censure. To illustrate this, we need to clarify what we mean by state censure of crime. The phenomenon we have in mind is not simply the issuing of a declaration to the effect that some action was wrong. States can do such things,Footnote ¹⁵ but this is not the kind of censure characteristic of criminal justice. If we are talking about the authority claims implicit in state censure of crime then we are talking about censure that comes toward the end of the criminal process. In the criminal process, liability to state censure is determined at the adjudicatory institution of the tribunal (or trial). The tribunal, however, determines liability by applying standards that are given by prior legislation. If what legislation does (in relation to those acts covered by MALA IN SE) is to designate moral reasons the violation of which is to attract state censure then the role of adjudication should be to decide whether those moral reasons were violated. But whether those are the reasons violation of which should attract state censure is not at issue at the tribunal. The role of the tribunal is only to have the offender answer to the state for compliance with the reasons designated as decisive by prior legislation.Footnote ¹⁶ Thus the liability to censure that emerges from the tribunal is liability to state censure for the violation of those reasons that the state has previously designated as decisive. Thus state censure delivered by the criminal process is quite unlike the censure issued by a private individual in that it refers specifically to the state's right to determine what merits such censure. It might be argued that state censure of crime should be reformed to make it more like the censure issued by a private individual. However, as we will see in more detail, this would be to undermine its claim to be censure that is authoritatively issued on behalf of the public as a whole.

Writers such as Duff and John Gardner have plausibly argued that the criminal trial can be understood as a kind of debate or conversation in which the defendant's reasons for action are subjected to critical attention.Footnote ¹⁷ It might seem that this analogy with interpersonal dialogue suggests that the state claims no special rights in the trial. However, despite the important parallels with moral argument and moral criticism, there is, as we can see, an important sense in which the trial cannot be a real conversation. In a real moral dialogue, it is assumed that the final say lies with the morally relevant considerations bearing on the situation. One only gets the upper hand in a real dialogue by being able to present a better grasp of those guiding considerations. Although there may be reasons of politeness or prudence or time-saving to discontinue the argument before both parties agree, and in doing so it might be necessary to give an appearance of concession, it is not that one person has a right to have their view win the day. But in the trial the state does claim to have some such right, and, on the basis of it having that right, subjects are under an obligation to answer to it for their compliance with the standards it has set for them. The guilty defendant has a duty to accept condemnation as something the state has the right to issue: the state has the right to end the conversation, settle the matter, and have its view carry the day. This is not merely the correlate of a prudential reason the defendant has to stop arguing, or of the state's monopoly of power. Neither, as we have seen, is it just the state falling back on its right to be obeyed. Rather the institution of state censure of crime requires the state to claim a right to determine, for the defendant, which moral considerations merit such public condemnation, a right that goes beyond the entitlement, possessed by all private individuals, forcefully to put forward one's view in an argument. The state claims the ultimate say about what conduct is to be treated as morally acceptable for the purposes of the trial and public condemnation. State censure of crime is therefore unlike interpersonal blame in that it is conditioned by the relation of authority: it is only by virtue of having the right to legislate and therefore to fix subjects’ liability to censure that the state could be justified in censuring in this way.

VII.

With this initial objection dealt with, we have completed our argument about the implications of accepting the three propositions outlined at the start for our thinking about political authority: specifically, that the criminal justice system that these propositions recommend could only be justified if the state possesses the authority of moral oversight. We can now ask where this leaves the idea that criminal justice involves moral communication and whether what we have argued shows it to be open to serious objection. This is to an important extent an unexplored question. Theorists who have shown us how to think about the authority of law have done so on the assumption that the authority claimed by the criminal law is either the right to be obeyed or the right to use coercion. They have explained how it might be possible for a state to gain the relevant rights—perhaps as a result of fulfilling a socially necessary function, or making it more likely that subjects will comply with the reasons that apply to them, or through some kind of transfer of authority from the individuals making up the state. But the question of how the state might acquire rights of moral oversight has not been asked.

If it is true, then, that the communicative view is committed to the authority of moral oversight, should we regard that as a strength or a weakness? Should the recognition that state punishment involves authority—public authority—lead us to abandon the idea that punishment should censure prelegal wrongness? After all, a number of alternatives are available that integrate their theories of punishment with a theory of state authority, for instance: the view that we have looked at already, that punishment is nothing more than justified coercion; the view that punishment is required by the principle of fair play;Footnote ¹⁸ or the Kantian/Hegelian view that sharply distinguishes legal from moral wrongs and sees the need to uphold public authority rather than censure moral wrongdoing as the ground for state punishment.Footnote ¹⁹ Can the communicative picture be integrated into a plausible and attractive theory of state legitimacy? While space prevents me from dealing fully with this question here, or from engaging fully with these alternative views, I want to start to address this concern by looking at two potential objections, both of which suggest that the authority of moral oversight is worryingly authoritarian.

The first objection starts by picking up on the analogy I drew above between the state and a professional association. Organizations such as the General Medical Council or the Bar Association, as I have portrayed them, exist, in part at least, in order to uphold professional standards. It may be plausible that, insofar as these associations have rights of authoritative censure over their members, their censure is concerned not just with whether practitioners obeyed directives authoritatively issued by the association, but whether they met the pre-existing role-norms that govern the conduct of the “good” or at least “minimally decent” doctor or lawyer. However, the objection goes, these are the professional standards of a particular, voluntarily adopted professional role. Through the criminal process, however, the state—at least on the communicative reading of it—claims a general kind of moral oversight over our dealings with one another and the values that are to inform those dealings that is not tied to a particular role but rather ranges widely over the moral reasons that might apply to us.

Furthermore, coming under a state's authority of moral oversight is not normally a voluntary matter. Thus, in considering my analogy, the objection continues, it is worth reminding ourselves of certain facts that are sometimes thought to make state authority particularly hard to justify. First, that individual persons do not normally have a choice but to live under the jurisdiction of some state or other. Second, that to come under the jurisdiction of a particular state it is normally sufficient simply to reside within the territory claimed by the state, and hence need not be a matter of voluntary acceptance. Third, that exiting one state and joining another is a highly onerous undertaking. Fourth, that the state normally claims the right to enforce its law—particularly its criminal law—against temporary residents as well as full citizens. The fact that the moral communication view takes no account of these issues, the objection concludes, suggests that it assumes that the authority of the state applies independently of consent and has to do with improving and overseeing our moral conduct in general. This betrays a particular, and particularly unacceptable, view of the relation between the authoritative state and its subjects—and this is reinforced by the claims to rights of moral oversight that underpin its criminal law. On this view, the state has care for the souls of its subjects and its role is as a moral authority to oversee and correct their conduct in order that they may get closer to salvation. Thus the model in the background of the communicative view is that of the state as Church.Footnote ²⁰

In response to this criticism, one who accepts the communicative picture can agree that being liable to a state's authority of moral oversight will not normally be a voluntary matter. The question that then arises is whether this fact is an argument against the authority of moral oversight. A common strategy in arguing against voluntarism about political obligation is to point out the need for certain public goods, the production of which ought not to be beholden to the voluntary agreement of each citizen, and that state action is necessary to bring about.Footnote ²¹ It looks as though this strategy might provide a line of defense for the authority of moral oversight as well. For there is another important way in which state censure of crime differs from interpersonal blame, in addition to the features we looked at in the last section: unlike interpersonal blame, state censure is not simply issued by the state as a private individual, or by officials in their capacity as private individuals, but is rather meant to express condemnation on behalf of the political community as a whole.Footnote ²² In condemning on behalf of the public, the state need not be claiming that members of the public do (or even, perhaps, that they ought to) share a certain view of the action or its perpetrator; or that they do or ought to have certain emotions in regards to the crime; or that they do or ought to have any particular disposition toward the offender, such as shunning or shaming or punishing him. It can therefore be misleading for proponents of the expressive theory of punishment to say that punishment expresses “the community's” disapproval, or blame, or indignation, etc. For punishment to have an expressive function its judgments presumably require some level of broad public support, but it is not necessary to see them as giving vent to a widely shared sense of revulsion toward the offender in the sense that Devlin wanted.Footnote ²³ Rather, the thought is that the condemnation is issued from a perspective that belongs to the public rather than just any particular private individual. An attractive way to interpret this thought that does not take us down the Devlin route is to take our lead from Joseph Raz's view about the authority of law:

As Raz suggests with his Lockean example of the authority of the arbitrator,Footnote ²⁵ we might think that such a public view emerges when some institution has the authority to adjudicate between competing views on a question that are, or might be, advanced by private individuals and has the right to determine which one will “stand” for the group as a whole, inform individuals’ dealings with one another as members of that group, and hence “settle” the question of how individuals are to relate to one another. This could be applied to the authority of moral oversight in the following way. Although there is disagreement about moral matters among citizens, state censure of crime expresses the official view taken by the body of citizens as a whole, as reached by the procedures that are authoritative to adjudicate among such disagreements. The state therefore takes itself to have a status that no private citizen has because it has the right to act as authoritative arbitrator of citizens’ moral disagreements and as the representative of the body of citizens as a whole.

The state has this power to arbitrate and make decisions on behalf of the body of citizens as a whole, it could be argued, because the authority claimed by the criminal law is part and parcel of something that needs to be done and that only the state can do, namely, standing up for a set of basic standards regarding how citizens should act so that no one can violate those standards with impunity. To back this up, it is worth briefly considering what might be lost if we were to do away with the authority of moral oversight. Possession of the authority of moral oversight allows the state to settle on a particular moral view of the situation and thereby lifts that view out of the hubbub of competing and conflicting views put forward by private individuals. In doing so it designates that view as public, shared, and collectively binding. Giving an agency this “final say” can of course be a privilege that is abused. But it can also powerfully express the non-negotiability of morality. For the fact that one moral conception can in certain important practical situations no longer be disputed means that not all moral claims are always open to question; hence, they are not always open to dissolution through sophistry, to ingenious undermining, to the clamor of the press and social media, or to the insistence of the one who can speak the loudest or who has the most impact on our attention. For a time at least, some moral claims cannot be ignored, and some moral conception has a protected status. In putting an end to discussion, and conferring finality on matters, the state can therefore speak to an important intuition that morality should not simply be a matter of who can speak the loudest. The state will not simply express an opinion that is avowedly on a par with that of any other private individual; rather, its view is distinct because it will make people answer to it. Thus in an important sense no one will have impunity in the face of that view. The authority of moral oversight therefore allows the state to become in some manner a moral community, defining, overseeing, and enforcing an evolving vision of the basic duties citizens have by virtue of their shared enterprise.

This last point is important. One as yet unaddressed strand in the objection concerned the contrast between role-morality and general morality. The concern was that the analogy I drew between the state and a professional association was misleading, because professional associations deal only with the obligations of a role, whereas the state claims oversight in regard to morality as a whole. However, we should now make it clear that the communicative picture of criminal justice being examined here need not be committed to the claim that the state claims authority of moral oversight in respect to citizens’ moral conduct in general. It is quite compatible with this communicative picture to see the criminal law as closely analogous to the model of the professional association, in that it oversees subjects’ compliance not with morality as such, but with the values that ought to inform their actions in their role as citizens. What the criminal law oversees, on this view, would be whether citizens live up to the demands that can reasonably be made of them as participants in a particular, shared, intrinsically valuable enterprise. This leaves a realm of private action in which citizens are not to be thought of as acting “in” their role and where the demands of that role do not apply to them. The boundaries between public and private here may be hard to draw cleanly, but this is not a special problem for the communicative picture.

It might be suggested that there is yet an important disanalogy between the state and a professional association. Professional associations governing a particular public service, such as medicine or law, oversee the values of a fairly well-defined collective enterprise, meaning that those associations do not have wide discretion in setting out the demands to which they hold their members; the shared enterprise governed by the state, by contrast, is less well-defined and more open to discretionary abuse. Now in response to this question it is worth first making the point that the communicative view need not, of course, be committed to the idea that the authority of moral oversight is absolute. If the authority of moral oversight exists, then, as with other forms of practical authority, its exercise is constrained by other, non-negotiable values and requirements. But beyond this, while it may be true that there is no blueprint for how to organize a polity, and while there are political choices to be made between different forms of political organization, there are clearly some important public goods that any polity needs to provide its members. Furthermore, there are well-documented problems with the state simply imposing a set of values that its members do not share, which are relevant to, and again not distinctive of, questions of the scope and limits of the authority of moral oversight.

Is it an important disanalogy between the state and a professional association that the state claims a monopoly on the use of coercion—and that criminal justice is a clear example of its coercive power? Of course, and the state's right to monopolize coercive power and use it against its members needs a justification. Nothing we have said here undermines the urgency of that question. But this is not the question we have been discussing here. Our question is whether there are distinctive claims to authority that are made by a state whose criminal justice apparatus is an institution of moral censure. To ask this question, and to attempt to answer it, in part, by means of an analogy with the authority of professional associations, is not somehow to pretend that criminal justice is not coercive. It is rather an attempt to spell out the conditions under which the state could claim such authority. If criminal justice is coercive then the fact that the state meets those conditions discussed in this paper will not by itself be sufficient to justify coercive state intervention. It will also need to meet whatever further conditions are necessary to justify coercion. But that this is the case does not show that the present investigation—geared toward showing us the rights that the state must claim by virtue of the fact that its coercive criminal justice apparatus involves censure—is not an important contribution to the general question of the legitimacy of the state.

VIII.

Our response to this first objection might seem only to open up a second. We have argued that accepting the communicative picture underpinning the three propositions outlined at the start means accepting that the state must have rights of moral oversight. But such rights, the objection begins, involve the right to some exercise of discretion. The state must have discretion, within limits, to decide which moral reasons to designate as those that will merit public censure. Note, however, that the kind of right we are now attributing to the state is an entitlement to determine practical matters that will be valid at least to some degree independently of whether the state reaches a decision that is justified, or accurate, or correct. In other words, even if the state decides that some action be designated as meriting public censure when it does not actually merit such censure (either because it does not merit censure, or because it does not merit public censure), or if it decides that some action not be designated as meriting public censure when in fact it does, it will still have the right to do so. The rights involved in the authority of moral oversight are therefore the kind of rights of which Raz says: “It is an essential element of rights to action that they entitle one to do that which one should not.”Footnote ²⁶ Otherwise put, the state's right to criminalize and condemn, as an aspect of its authority of moral oversight, would be in part content-independent.

However, the objection contends, this seems to be a problematic conclusion to draw about an institution of censure. That is, it might seem for various reasons as though there is no content-independent right to censure. To start with, it might seem that there is a difficulty simply in capturing the semantics of state censure if its force is partly content-independent. Part of the appeal of the communicative picture, as we have seen in this paper, is that it allows the content of state censure to refer to the prelegal wrongness of the action rather than the fact that it is a violation of the claims of authority. However, now it might look as though this appeal is illusory. For while any sincere expression of censure presumably contains within it a claim to be justified, censure issued by private individuals depends on the justification of its content. When it comes to state censure, on the other hand, it seems that the state has the right to make it, and the recipient a duty to accept it, even in the situation when its content lacks justification. Thus, at best, state censure will refer to the prelegal wrongness of the action only in those cases in which the state is right about the acts that merit public censure. At worst, however, it will turn out that what the state expresses in all cases does not refer so directly to the prelegal wrongness after all. What the state expresses in censure would be something more like: “Based on procedures that are authoritative in determining the proper objects of public censure, it has been decided that what you have done merits public censure.”

This objection leads to a further concern. If the state were to act as though it censured prelegal wrongness then it would ignore the fact that there must be some cases in which it gets things wrong, and where, even if it does retain the right to express the judgment that it does, this right can be grounded only in its authority rather than the content of the judgment. Particularly problematic, then, would be if the state demands—in every case in which subjects are, by the correct following of the appropriate procedures, judged to merit public censure—that those subjects should show remorse, and be subject to treatment making them likely to reform. If there is a gap between what the state determines to merit condemnation and what really does merit condemnation then requiring subjects to display remorse in every case, and perhaps disadvantaging them or subjecting them to further sanctions (or less respectful treatment) if they do not, would be highly problematic. If the state were to do this it would effectively be asserting a content-independent right to determine which content those subject to the authority must take as morally correct, or perhaps even a content-independent power to determine which content really is morally correct independent of the directive. But neither of these implications is acceptable. If the latter, such powers would be impossible; if the former, they would be rights over the conscience of subjects that could not properly be claimed by a modern liberal state.

Consider briefly the parallel with the Roman Catholic Church's teaching with regard to vexed questions of doctrine. The Church authorities take themselves to be able to make declarations on doctrine (regarding, say, contraception, abortion, or the ordination of women priests) that are binding on the faithful. And it seems as though it is in part their status as authorities, e.g., as the Pontiff, that gives them the right to settle arguments on these topics, and that they effect such a settlement by giving an authoritative ruling on the topic. Of course, the Church authorities will also take themselves to be well-informed in matters of doctrine—no doubt their authority would not be what it is were they not to be regarded as generally good guides. But their authority transcends their ability to get it right in individual cases and seems rather to depend on their status. Anyone who saw the Pope's authority as merely consisting in his being a good guide to correctness in doctrinal matters would be misunderstanding his role. Rather he is able, by considered decision and by so ruling, to determine for the faithful that which they are to give “religious assent of the mind”—in other words, belief.

The authority claimed by the Roman Catholic Church has a number of interesting features that will be worth noting in order to compare them to the features of the authority claimed by criminal law. First of all, it claims a content-independent power to determine not simply how the faithful are to act, but what they are to believe. It takes itself to have the power to place the faithful under a binding requirement, not just with respect to their external conformity with doctrine, but also that they should internalize, profess, and live such doctrine as true. It claims the right to ask that the faithful answer to its representatives for their attitudes and actions with regard to authoritative doctrine, in confession but also, if necessary, through bodies such as the Congregation of the Doctrine of the Faith (previously the Congregation for Universal Inquisition), which is charged with protecting the Church from heresy and whose duty it is to “promote and safeguard the doctrine on the faith and moral throughout the Catholic world.”Footnote ²⁷ Nevertheless, as with the authority of moral oversight, it might be asked how it is possible to square the content-independent force of the ruling with the demand for assent to its content. It seems that one must either weaken the bindingness of the ruling, weaken the demand for assent of the mind, claim that the ruling alters the content and makes it true, or claim that the ruling is infallible.

Given that the rulings of the state with respect to the proper objects of public censure cannot either be infallible nor make their verdicts true, the only remaining options are the first two. However, to weaken the bindingness of the ruling would be to forgo the benefits of having an authoritative institution of public censure. If, as I have claimed, there is something to be said for such an institution, the way forward seems to be to explore the possibility of weakening the demand for assent of the mind. In other words, we need to explore the possibility that the state claims the right to condemn, but does not thereby claim that subjects must regard its condemnation as justified on its merits. The state might hold, for instance, that its best understanding of the permissibility of foxhunting is that it is impermissible and meriting of public condemnation, and also hold that, by virtue of its position, it has a content-independent right to settle the argument definitively regarding which actions subjects can be permissibly be condemned for. But it can recognize that some matters are justifiably more controversial and difficult to discern than others, for itself as well as for its citizens. For its right to censure to make sense, it has to hold that the procedure by which the decision is made has led it to a judgment that best reflects the available evidence, not one that simply reflects a compromise. But it does not have to claim that it could not be wrong.

One implication of this, as I noted above, is that the state should be tentative about giving evidence of remorse a role in criminal justice. States may not be able to compel subjects to believe that their condemnation is justified, but they do have the power to penalize—directly or indirectly—subjects’ failure to show signs of the emotional reactions that would be appropriate to such belief (for instance, by giving longer sentences to those who show no remorse). If states were to have a right to penalize a lack of remorse, it would be a content-dependent right; that is, it would be a right the state had only where its judgments of when remorse was merited were accurate. There can be no content-independent right to require signs of remorse, or penalize their lack.

For that reason, those who favor the communicative picture outlined at the start should reject R. A. Duff's claim that punishment should be conceived of as a communicative enterprise aiming at the offender's repentance, and more broadly those views that see punishment as justified only insofar as it tends toward the rehabilitation or education of those who commit crimes.Footnote ²⁸ The state has the right to criminalize, and to punish, and it has the right to censure those who commit crimes as wrongdoers. But if the arguments of this paper are correct, it has those rights to some degree independently of the accuracy of its judgments, and for that reason the justification of its right to censure must be found elsewhere than in the likelihood that punishment will stimulate moral reform. More plausible is the claim that expressive punishment should rather be understood “ritualistically,” in the sense that the inner state of those undertaking it is irrelevant to whether it is completed effectively, and hence that punishment should be such that an offender be able to undergo it perfectly well if they are unrepentant.Footnote ²⁹ The right of the state to condemn can be content-independent, but it cannot be a right to dictate citizens’ attitudes toward content. Forms of punishment that are ritualistic and compliance with which does not demand signs of apparently authentic remorse provide one way in which the expressive theory of punishment can be reconciled with the (partially) content-independent authority of criminal law.