Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Politics. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 03 October 2022

Law and Political Philosophyfree

Law and Political Philosophyfree

  • Jens Damgaard ThaysenJens Damgaard ThaysenDepartment of Learning and Philosophy, Aalborg University


Modern states pursue most of their (domestic) ends by creating law and acting in accordance with the law they create. Moreover, many believe states ought to pursue most of their ends this way. If a state ought to do something, then chances are it ought to do it by creating, abolishing, changing, upholding, or enforcing some law. Therefore, almost any kind of political philosophy with bearing on what states should do has bearing on what law should be like. Justifying the legal proscription of some conduct involves more than just showing that citizens ought to refrain from that conduct. Legally restricting conduct is an exercise of coercion and must be justified as such. Criminal prohibitions in particular require special justification, as they are not only coercive but also commit the state to deliberately inflict the harm and stigma of punishment on some of its own citizens. Nevertheless, if the state must coerce its citizens, it ought generally to do so through a law that conforms to the rule of law. Law conforms to the rule of law if it is capable of guiding the citizens as they act and plan for the future. This the law can do only if it is open, clear, prospective, and stable, such that citizens can know what it demands now and predict with reasonable certainty what it will demand in the future. Conformity to the rule of law promotes freedom and is required to respect human dignity. Much of the debate about the justification and scope of legal coercion revolves around several principles that advance claims about what considerations are relevant to the justifiability of law. These principles all have the following structure: The fact that a legal restriction of a certain kind is related in a certain way to a certain type of conduct has a certain impact on whether that restriction is justifiable. Common principles include (a) legal moralism, according to which it is always a good reason to criminalize conduct that the conduct is wrongful; (b) the wrongness constraint, according to which criminalizing morally permissible conduct is never justified; (c) liberalism, according to which it is always a good reason to criminalize conduct that the conduct is either harmful or seriously offensive to others, and criminalizing conduct that is neither harmful nor offensive is never justified; (d) the public wrong principle, according to which it is always a good reason to criminalize conduct that the conduct is a public wrong, and criminalizing conduct is never justified unless the conduct is a public wrong; (e) the sovereignty principle, according to which the only legitimate restrictions on conduct are those that secure independence. Which, if any, of these principles one should accept is the subject of an extensive and sophisticated academic debate.


  • Political Philosophy


Modern states pursue most of their (domestic) ends by creating law and acting in accordance with the law they create. Moreover, many believe states ought to pursue most of their ends this way. If a state ought to do something, then chances are it ought to do it by creating, abolishing, changing, upholding, or enforcing some law. The law fundamentally shapes society. Laws confer rights and powers, impose legal obligations, shape the distribution of society’s wealth, and create the legal framework required for major social institutions to function. Without laws defining the powers of political office, the frequency of elections, and criteria for being eligible to vote, sustaining a large-scale democracy would be impossible. Without laws pertaining to contracts and transfers of property, a modern market economy could not exist. A just law is necessary for a just society, though it may be insufficient. Indeed, one of the most important ways in which a society can be unjust is by having unjust laws. Therefore, almost any kind of political philosophy with bearing on what states should do has bearing on what the law should be like.

This article proceeds as follows. It first examines what a law is and why laws must be justified. Next, the nature and value of the rule of law are discussed before turning to the question of how laws can be justified by examining some general features of normative principles about the justification of law. The remaining sections discuss specific principles about the justification of law asserting, respectively, that wrongness, harm, public wrongness, and interference with sovereignty are of special relevance to the justification of law. The focus of this article is thus on the nature, justification, and scope of coercive domestic law. As this area of legal philosophy has direct consequences for what states should do, its relation to political philosophy is particularly clear, immediate, and strong.

What Is Law, and Why Must It Be Justified?

A law is a rule, but not all rules are laws. In order to know what law is, one needs to know what distinguishes legal rules from other rules, such as those of morality. H. L. A. Hart’s famous The Concept of the Law (2012) sees the central feature of the legal system as the combination of primary and secondary rules. Primary rules confer duties to engage in or refrain from specific conduct (e.g., “do not kill”). Secondary rules confer powers to manipulate or control primary rules, such as the power to create, modify, abolish, and determine compliance with primary rules. They also specify the procedure one must follow to exercise these powers successfully. Rules of recognition specify the criteria by which to determine whether a suggested rule has the force of law. For instance, it may be that a bill must be approved by a majority of the legislature and made public in some appropriate place to have the force of law. Rules of change specify who has the power to create, modify, and abolish primary rules. The legislature (typically) has the power to create, abolish, and modify laws by following a certain procedure (typically involving the formal submission of a proposal followed by voting during a session of parliament). However, private individuals also possess some limited power to shape their own legal obligations, such as the power to enter into a legally binding contract or create a will by following a certain procedure (e.g., having witnesses present when creating the will). Rules of adjudication specify who has the power to settle disputes about the existence and compliance with primary rules. Rules of sanction specify who has the power to sanction non-compliers (Hart, 2012, chap. 5). Morality also contains primary rules, but unlike legal rules, moral rules are not joined with secondary rules of recognition, change, and adjudication.

The Justification of Law

Before examining why law is in need of justification, it is useful to distinguish between three types of law that differ with respect to what is required to justify them: criminal offenses, civil offenses, and power-conferring legal rules. Criminal and civil offenses both proscribe certain conduct. They differ in that criminal offenses make violators liable to state punishment (Husak, 2008; Simester & Von Hirsch, 2011), whereas civil offenses, such as those of the law of torts, do not. Instead, civil offenders are liable to civil suits by private parties. Power-conferring legal rules grant citizens who meet certain criteria the power to create or alter their legal obligations by following a certain procedure. Power-conferring law does not create offenses, and failing to follow power-conferring law does not result in sanctions. Rather, those who try to exercise some power without meeting the criteria or following the procedure established by law will simply fail to exercise the power they were trying to exercise (Hart, 2012). A same-sex couple does not break the law by performing the actions by which heterosexual couples marry in a jurisdiction that does not recognize same-sex marriage; rather, they simply fail to marry (in the eyes of the law). This article now turns to the question of why legislation is generally in need of some justification other than (or in addition to) the democratic legitimacy of the legislative process.

Consider first the justification of criminal law. Criminalization has a number of general downsides in virtue of which any criminal law is prima facie problematic and in need of justification. These downsides create a presumption of liberty, a standing case against criminalization that must be overcome by the reasons in favor of criminalization in order to justify criminal prohibitions (Feinberg, 1987, p. 9; Moore, 2014). The first general downside of criminal proscriptions is their coerciveness. Suppose some state criminalizes queue-jumping. By doing so, this state orders citizens not to queue-jump and threatens to punish citizens who disobey this order. This is a textbook example of coercion, depriving citizens of their negative freedom queue-jump (Simester & Von Hirsch, 2011). Even though the world could do with fewer instances of the genuine, if minor, wrong of queue-jumping, such an exercise of coercion is in need of justification. Even if the negative liberty to engage in wrongful conduct has little or no intrinsic value, this liberty might be instrumentally valuable as a means to other goods, such as positive freedom, autonomy, or the possibility to refrain from wrongdoing of one’s own accord (Moore, 2014). The second general downside of criminalization is that criminal laws commit the state to deliberately harm and stigmatize some of its own citizens. A state that criminalizes conduct thereby declares that it will punish citizens who engage in this conduct without a legally valid excuse or justification for doing so (henceforth: for no good reason). Because criminalizing conduct will rarely, if ever, cause literally everyone to stop engaging in that conduct for no good reason, criminalization is almost certain to commit the state to actually punishing someone (Husak, 2008, pp. 78ff). Accordingly, one must justify the punishment of those who queue-jump for no good reason, in order to justify the criminalization of queue-jumping. As punishment is commonly understood as treatment intended to harm the punished and censure her for engaging in the conduct she is being punished for (Husak, 2008; Ryberg, 2016; Simester & Von Hirsch, 2011), this means that criminalization commits the state to deliberately harm some of its own citizens as well as stigmatize them as wrongdoers.1 This comes with a heavy burden of justification. Indeed, the lengthy prison sentences imposed on those found guilty of the most serious crimes are so immensely destructive to their interests (Feinberg, 1987, p. 4) that it is probably one of the worst things decent states do to their own citizens (Husak, 2008, p. 95). Even if genuine offenders deserve the harm of criminal punishment, criminal prohibitions are highly likely to cause some undeserved harm as well (Edwards, 2014, pp. 259–262). The incarceration of criminal offenders often causes emotional as well as economic hardship to their children and other innocent family members (Edwards, 2014, p. 260; Tadros, 2013, pp. 356–359). Moreover, even a good legal system will make some mistakes, punishing the guilty more severely than they deserve and punishing the innocent who deserve no punishment at all. Furthermore, a system of criminal justice must necessarily confer powers on wardens and officers that some of them will abuse (Edwards, 2014, p 262; Husak, 2008, pp. 203–204). Third, enforcing criminal prohibitions is costly (Husak, 2008, p. 203). The resources used to enforce a criminal law could have been used elsewhere or left in the pockets of the taxpayers. In sum, criminal prohibitions must be justified, because they are coercive and costly, and commit the state to harming its own citizens, some deliberately, some accidentally. To what extent do these considerations apply to the creation of civil offenses and power-conferring law?

Like the creation of criminal offenses, the creation of civil offenses imposes legal duties and makes offenders liable to sanctions. Accordingly, the creation of a civil offense is coercive much the same way as criminalization is: the state attaches costs to some conduct with the known (and often intended) effect that fewer will engage in that conduct. Being forced to compensate the plaintiff makes civil offenders worse off, but unlike criminal offenders, civil offenders are not harmed deliberately, but in order to compensate the plaintiff for the damage done by the defendant’s actions (Tadros, 2010). Nor are civil offenders stigmatized to the same extent as criminal offenders. Being labeled a criminal is a lot more damning than being labeled a tortfeasor (Simester & Von Hirsch, 2011, p. 5). Nevertheless, the creation of civil offenses still restricts liberty in a way that requires justification.

Power-conferring legal rules makes it possible for individuals to enter into certain legal relationships (e.g., a contract) provided they meet certain criteria and follow certain procedures (Hart, 2012, pp. 27–28). Power-conferring legal rules are coercive and must be justified as such when they confer the power to coerce. However, there is no coercion to justify when legal rules confer other powers. Instead, it seems to be the limits on who gets what powers and how these powers can be exercised that must be justified in order to justify such power-conferring laws. This touches on questions about when enabling particular options only for some citizens is compatible with treating all citizens equally. It also touches on questions about what constitutes a legitimate reason for the state to refrain from enabling a currently unavailable option—a question that might very well have a different, more permissive answer than the question of what constitutes a legitimate reason to remove a currently available option. A great number of debates within political philosophy involve power-conferring law. What, if anything, can justify restricting the legal power to marry to only heterosexual couples? Should the right to vote be dependent on a certain level of rationality and political knowledge (Brennan, 2016)? Should citizens be allowed to use their powers to enter into any legally binding contract, even those that are obviously unfair, exploitative, or otherwise extremely disadvantageous (Shiffrin, 2000)?

The Rule of Law

The section entitled “The Justification of Law” argued that justifying law, especially criminal law, can be demanding indeed. However, it is widely held that if the state must coerce its citizens to refrain from certain kinds of conduct and harm those who do not comply, then it should do so by creating a criminal offense. To the extent that the state is justified in ruling—through coercive interference—its subjects, it should rule by law. But how should one understand the rule of law and its value? Joseph Raz writes:

[The rule of law] has two aspects: (1) that people should be ruled by the law and obey it, and (2) that the law should be such that people will be able to be guided by it.

(Raz, 2009, p. 213)

It follows from the first aspect that where the law rules, no one is above the law, including those who make it. The reasoning behind the second aspect is roughly as follows. The law does not rule citizens, unless it guides their actions, but the law can guide the citizens only if they know what the law demands. A number of familiar features of the rule of law follow from this second aspect. First, the law must be prospective, open, and clear in order to guide the citizens. Obviously, citizens cannot adjust their past conduct on the basis of a retroactive law. Nor can citizens be guided by legislation unless they know what it means. Therefore the law must be open, as opposed to secret, and its meaning must be clear (Raz, 2009, p. 214). Second, the law must not change too often. If legal change is too frequent, it will become overly difficult to keep track of what the law demands. However, a stable law is also important, because it is required for the citizens’ long-term plans to be guided by law (Raz, 2009, p. 214). Third, the rules establishing who has the power to make particular legal orders (e.g., orders issued by a police officer regulating traffic) and how these powers must be exercised, must be clear, open, and general. Though the authority of various public servants to issue particular legal orders is difficult to do without, it does create an element of unpredictability that makes it harder to be guided by law. To minimize the unpredictability introduced by particular legal orders, the rules specifying who has these powers and what duties they are bound by when exercising them should also be clear, open, and general (Raz, 2009, p. 216).

A law that conforms to these three principles will be such that it can guide citizens when making and acting on plans. To this Raz adds a number of further principles, which are

designed to ensure that the legal machinery of enforcing the law should not deprive it of its ability to guide through distorted enforcement and that it shall be capable of supervising conformity to the rule of law and provide effective remedies in cases of deviation from it.

(Raz, 2009, p. 218)

These principles demand that the courts should be independent, be fair in their application of the law, have the power to review whether the law conforms to the other principles, and be accessible to the general public. Moreover, the police and other enforcement agencies should not have discretion to decide to prosecute only some crimes or some offenders (Raz, 2009, pp. 217–218).

In sum, a law that conforms to the rule of law is a law that citizens are able to figure out what requires, so they can act and plan accordingly in the short as well as the long term. This is desirable for several reasons. First, conformity to the rule of law reduces the possibility of arbitrary exercises of power by legal officials, though it does not eliminate such possibilities entirely (Raz, 2009, p. 219). Second, conformity to the rule of law makes the law stable and predictable enough to constitute a “safe basis for individual planning” (Raz, 2009, p. 220), thereby enhancing individual freedom of action. Third, only a law that conforms to the rule of law respects human dignity, by treating the citizens as autonomous beings capable of making and acting on plans for the future. An unclear or unstable law undermines human autonomy because uncertainty about what the law demands now or in the future makes it difficult for citizens to make plans for the future. An arbitrarily enforced or retroactive law tends to frustrate the plans citizens do make by causing the law to appear compatible with certain plans at one moment, only to suddenly and unpredictably thwart those plans at another (Raz, 2009, p. 222). A further reason to consider it desirable that criminal law, in particular, conforms to the rule of law is to ensure citizens are given a fair chance to avoid the harm and stigma of criminal punishment (Tadros, 2013, p. 329). This requires that citizens are realistically able to know what conduct makes them liable to punishment, which requires a clear, open, and prospective criminal law.

How Can Law Be Justified?

The question of how law can be justified is the subject of an extensive and sophisticated literature revolving around a number of principles, all of which advance claims about the relevance of certain considerations to the justifiability of legal restrictions (Duff, 2007; Feinberg, 1987; Husak, 2008; Mill, 1992; Moore, 1997; Ripstein, 2006). Before moving on to discussing particular principles, this section examines some general features of principles concerning the justification of law. Each principle makes a claim of the following type:

The fact that a legal restriction of a certain kind is related in a certain way to a certain type of conduct has a certain impact on whether that restriction is justifiable.

Principles can vary with respect to (a) what type of conduct they pick out; (b) what kind of legal restriction they concern; (c) what relation between the relevant legal restriction and the relevant they concern; and (d) how it impacts the justifiability of legal restrictions of the relevant kind that they relate in the relevant way to the conduct picked out by the principle. At the core of each principle is (a) picking out a certain type of conduct. Prominent examples of such “types of conduct” include wrongful conduct (Dworkin, 1999; Moore, 1997), harmful conduct (Feinberg, 1987; Mill, 1992), offensive conduct (Feinberg, 1988), publicly wrongful conduct (A. Duff, 2007; Husak, 2008; Marshall & Duff, 1998), and conduct that violates independence (Ripstein, 2006). The differences in the type of conduct picked out by various principles are arguably the most fundamental, such that principles are often categorized according to what type of conduct they pick out. With respect to (b), one might distinguish between those principles that restrict their concern to the justification of criminal law (A. Duff, 2007; Feinberg, 1987; Husak, 2008; Moore, 1997) and those concerned with the justifiability of (state) coercion in general (Devlin, 1963; Dworkin, 1999; Mill, 1992; Ripstein, 2006). For present purposes, the interest in the latter principles is restricted to their applicability to legal restrictions, but they also apply to extralegal coercion.

As concerns (c) one might distinguish between act-centered (A. Duff, 2007; Moore, 1997) and effect-centered principles (Feinberg, 1987; Mill, 1992; Ripstein, 2006). Act-centered principles focus on the properties of the legally restricted conduct as such (e.g., is it wrongful?). Effect-centered principles focus on the outcome of legally restricting that conduct (e.g., does it prevent harm?). In general, principles that pick out (publicly) wrongful conduct tend to be act centered, whereas other principles tend to be effect centered. To see why the distinction between act-centered and effect-centered principle matter, suppose that a criminal proscription of the morally permissible act of carrying one’s wallet in the back pocket would prevent wrongful pickpocketing, and that a criminal proscription of wrongful lying would increase the number of wrongful lies. In such a scenario an act-centered principle picking out wrongful conduct as relevant to the justifiability of criminalization would permit the criminalization of wrongful lying, but not of carrying one’s wallet in the back pocket, whereas the opposite would be true of an effect-centered version of the same principle.

With respect to (d) one might distinguish between justifications and constraints. Justifications advance claims about what constitutes a good reason to intervene. Constraints advance claims about what properties a legal restriction must possess in order to be justified. Note that there is no tension in simultaneously affirming multiple justifications (or multiple constraints), but that constraints that pick out one type of conduct will often (though not necessarily) conflict with justifications that pick out another type of conduct. As there might be a good reason to legally restrict some conduct, even though legal restriction is not justified, all things considered, one might ask why no principles advance claims about what is sufficient, all things considered, to justify legal restrictions. Such principles are omitted because they have no defenders. The number of considerations that could potentially undermine the justifiability of legal restrictions (especially those of the criminal law) is so great that it is impossible to state a sufficient justification of legal restrictions as a principle that is both tolerably precise and simple enough to be of any use (Tadros, 2016, p. 111). Constraints have bearing on the scope of the law. To defend a constraint on criminalization is to identify the normative limits of the criminal law by identifying a sphere of conduct that ought principally to remain outside the scope of the criminal law. To defend a constraint on legal restrictions is to identify the normative limits of any kind of coercive law. As most scholars working on the justification of criminalization explicitly take the view that too much conduct is criminal (Husak, 2008; Moore, 2014; Ripstein, 2006, p. 245; Tadros, 2016, p. 91), constraints on the criminal law are often motivated by and advertised as ways to address a problem, or even a crisis, of over-criminalization. Table 1 provides an overview of the general structure of different types of principles, substituting “X” for the type of conduct picked out.

Table 1. The General Structure of Principles About the Justification of Law



Justification (criminal law)

It is always a good reason to criminalize conduct that the conduct is X.

It is always a good reason to criminalize conduct that criminalization prevents X.

Constraint (criminal law)

Criminalization is never justified unless the criminalized conduct is X.

Criminalization is never justified unless it prevents X.

Justification (all legal restrictions)

It is always a good reason to legally restrict conduct that the conduct is X.

It is always a good reason to legally restrict conduct that legal restrictions prevent X.

Constraint (all legal restrictions)

A legal restriction is never justified unless the restricted conduct is X.

A legal restriction is never justified unless it prevents X.

The distinctions drawn in this section result in a large number of possible principles: 40 when one multiplies the eight types listed in Table 1 with the five candidates for the type of conduct the principles might pick out. Nevertheless, this article actually simplifies the theoretical landscape significantly in order to keep the discussion at a reasonable length, as it is possible to draw many other distinctions in addition to those drawn in this section, resulting in even more different principles (cf. Edwards, 2014). Even so, discussing all of the 40 combinations of the distinctions this article does draw is manifestly unfeasible. Accordingly, the article only discusses the most prominent combinations. The remainder of this article now moves on to a discussion of particular principles.

Legal Moralism and the Wrongness Constraint

Consider first the justificatory principle picking out wrongful conduct:

Legal moralism: It is always a good reason to criminalize conduct that the conduct is wrongful.

Legal moralism is historically associated with strongly conservative views, particularly support for criminal laws against homosexual intercourse between consenting adults (see Devlin, 1963; Hart, 1963), which are widely discredited as of the early 21st century. However, legal moralism itself is only as conservative or as liberal as the underlying moral theory. The prominent legal moralists of the early 21st century emphatically reject the moral views associated with conventional “moralism” (Arneson, 2013, p. 452; Dworkin, 1999, p. 946; Moore, 1997, p. 756, 2009, p. 32). This liberal conception of morality severs the connection between legal moralism and the most embarrassing examples of conventionally moralistic laws, such as criminal proscriptions of homosexuality. However, even the sparsest plausible account of morality will contain wrongs nobody thinks should be criminalized (Simester & Von Hirsch, 2011, p. 22): breaking promises, cheating on one’s romantic partner, and queue-jumping are all plausibly morally wrong, yet these kinds of conduct ought not to be criminalized. The legal moralist can explain this by saying that although there is indeed a good reason to criminalize any wrong, the downsides of criminalizing some minor wrongs (e.g., ordinary lies) are so significant that this reason is outweighed (Moore, 1997, pp. 661–665, 2014).

Thus, those who support a liberal criminal code need not therefore reject legal moralism. But why affirm legal moralism in the first place? One argument for legal moralism proceeds from retributivism about punishment (Moore, 1997, 2009, 2014). Retributivism is, roughly, the position that punishment is justified because it is intrinsically good that offenders suffer in proportion to their moral culpability, which is determined by the gravity of the wrongs committed and the offender’s responsibility for committing them (Moore, 1997, p. 71).2 The argument proceeds as follows: particular criminal laws should aim at doing what criminal law is generally good for. Whatever criminal law is generally good for must be something criminal law can bring about. What criminal law can bring about is the punishment of offenders. Therefore, the justifying good of punishment determines what counts as a good reason to criminalize. Retributivists think punishment is justified because it is good that wrongdoers suffer in proportion to their culpability. Punishment serves this good if, and only if, it is inflicted in response to wrongdoing. However, it is permissible to punish wrongdoing only if the wrong was criminal at the time it was committed. Therefore, there is always a good reason to criminalize wrongful conduct so that it can be permissibly punished (Moore, 2014, pp. 189–192). One can resist this argument by denying that retributivism is the correct theory of punishment (Tadros, 2016, p. 115). Another objection is that criminal offenders are not merely liable to punishment, but to state punishment (A. Duff, 2007; Husak, 2008). Thus, it is not enough to show that wrongdoers deserve punishment; one must show that they deserve punishment by the state. This may be difficult, as showing that wrongdoers deserve to suffer does little to show that taxpayers deserve to finance their suffering. This requires showing that the state has a proper interest in criminalizing all wrongdoing (Husak, 2008, pp. 134ff, 199ff). Duff argues that legal moralists cannot show this, as only a “totalitarian” state (R. A. Duff, 2016, p. 107), which is presumably therefore objectionable, that takes an interest in every aspect of its citizens’ lives could have a proper interest in proscribing all wrongdoing (A. Duff, 2007, p. 50).

A second argument in favor of legal moralism, the conceptual argument, proceeds as follows. To say that some conduct is morally wrong is to say that it ought not to be done. But if conduct ought not to be done, then there is a good reason to discourage it, and then there is also a good reason to criminalize, or at least legally restrict, wrongful conduct in order to discourage it (Dworkin, 1999, pp. 943–944). This is a powerful argument. Indeed, even paradigmatic opponents of legal moralism seemingly concede the point (Feinberg, 1990, p. 322; Mill, 1992, p. 157). It is difficult to deny that the assertion “wrongful conduct ought not to be done” is conceptually true in virtue of the meaning of “wrongful.” However, like the argument from retributivism, the conceptual argument can be criticized for failing to take into account that law is something enforced by the state. If a good reason to discourage wrongful conduct is to be a good reason for criminalization, or another legal restriction, then it must be the state that has this reason. But while it is plausible that calling conduct wrongful is always to say that someone has a good reason to discourage this conduct, the conceptual argument does not show that this someone should be the state.

The Wrongness Constraint

The wrongness constraint is a constraint that picks out wrongful conduct. One might distinguish between two versions (Tadros, 2012, p. 158):

The strong wrongness constraint: A criminal law is never justified unless the criminalized conduct is wrongful independently of criminalization.

The weak wrongness constraint: A criminal law is never justified unless the criminalized conduct is wrongful after it is criminalized.

The strong and the weak wrongness constraints both hold that criminal laws are justified only if they target wrongful conduct, but diverge with respect to when the moral status of the criminalized conduct matters. Whereas the strong wrongness constraint cares about the moral status of the criminalized conduct at the time of criminalization, the weak wrongness constraint cares about the moral status of the criminalized conduct subsequent to criminalization (typically at the time of punishment). Support for some sort of wrongness constraint is remarkably widespread (A. Duff, 2007, p. 80; Feinberg, 1965; Husak, 2008, pp. 73ff; Moore, 1997, p. 72; Ripstein, 2006; Simester & Von Hirsch, 2011, p. 22), although it is rejected by some (Tadros, 2016, pp. 98ff). The wrongness constraint is typically affirmed as an act-centered principle about the criminal law.

A common defense of the wrongness constraint proceeds as follows. Punishment is justified only when imposed in response to wrongful conduct. Because engaging in criminal conduct without excuse or justification makes one liable to punishment, criminalization is never justified unless punishment of the criminalized conduct is justified. Therefore, a criminal law is never justified unless the criminalized conduct is wrongful. The claim that punishment is justified only in response to wrongful conduct can be defended on the grounds that punishment is justified only when deserved, and deserved only if one has done something wrong (Husak, 2008, pp. 71–89; Moore, 1997). Alternatively, one can argue that punishment inevitably expresses moral condemnation of the punished action, which is only appropriate when that action is wrongful (A. Duff, 2007, p. 80; Feinberg, 1965; 1990, pp. 150–53; Husak, 2008, p. 88; Simester & Von Hirsch, 2011, p. 23). Another line of argument proceeds from an observation about specific legal defenses. Defendants who fulfill the offense elements of a crime are nevertheless acquitted if they can demonstrate that the circumstances were such that their actions were either justified, that is, were not wrongful, all things considered; or excused, that is, it would not be appropriate to consider these defendants responsible for their conduct. As justifications and excuses deny, respectively, the wrongfulness of the defendant’s actions and her responsibility for those actions, they only make sense as legal defenses on the assumption that defendants in criminal trials stand accused of being responsible for wrongdoing. Why else should it matter whether the actions of the defendant are justified or excused? (Dimock, 2014, pp. 157–159; Husak, 2008, pp. 72–73). Thus, the legal practice of the early 21st century presupposes some form of wrongness constraint.

Both arguments defend the wrongness constraint on the basis of observations about trial and punishment. These arguments thus seem to support the weak wrongness constraint, as they suggest that the problems of criminalizing permissible conduct arise only at the stage of trial and punishment (Simester & Von Hirsch, 2011, p. 27). Nevertheless, these arguments are compatible with the strong wrongness constraint, and proponents of the wrongness constraint might be interested in exploring whether the strong version can be defended solely in virtue of being the more ambitious version, especially insofar as they see constraints as tools with which to fight a dangerous trend of over-criminalization (Husak, 2008). Furthermore, the weak wrongness constraint is in danger of being trivialized by the existence of a general duty to obey the criminal law, as most merely permissible conduct would then be made wrongful by criminalization.

At this point it is useful to introduce a classic distinction between mala in se offenses, which proscribe pre-legally wrongful conduct and mala prohibita offenses, which proscribe conduct that is not pre-legally wrongful (Husak, 2008, pp. 104–105). Prima facie, the wrongness constraint appears hard to reconcile with quite uncontroversial mala prohibita prohibitions, such as the legal demand to drive on the left side of the road. What exactly is wrongful about driving on one side of the road rather than the other? However, criminalization can certainly make it wrong to drive on the right side, as doing so when everyone else is driving on the left will recklessly, and wrongfully, endanger others. This reconciles the weak wrongness constraint with the existence of justified mala prohibita. However, the strong wrongness constraint cannot be reconciled with the existence of justified mala prohibita in this way, as conduct must be pre- legally wrongful in order for criminalization to be justified according to the strong wrongness constraint. Thus, another strategy is required to reconcile the strong wrongness constraint with the existence of justified mala prohibita. One such strategy is to argue that although it is true that driving on the right side of the road only violates a moral duty not to recklessly endanger others, after the adoption of a rule that everyone drives on the left, the general duty not to recklessly endanger others exists independently of criminalization. The role of the rule that everyone should drive on the left side of the road is merely to settle the concrete implications of that duty (Moore, 1997, p. 73). Another solution is to argue that although the conduct prohibited by mala prohibita offenses cannot, by definition, be pre-legally wrongful, it can still be pre-criminally wrongful (R. A. Duff, 2014). Although driving on the right side of the road is not pre-legally wrongful, a non-criminal legal rule that vehicles should all drive on the left side of the road can make it pre-criminally wrongful. Thus, the justifiability of a criminal prohibition against driving on the right side of the road can be reconciled with the strong wrongness constraint through a two-step process of first creating a non-criminal legal rule and then criminalizing violations of that rule. Although both solutions arguable handle the notable mala prohibita offenses found in the law of traffic, it is unclear whether they are able to reconcile the strong wrongness constraint with justified mala prohibita in general. Consider the example of tax evasion. It is plausible that tax evasion is wrongful only if one can reasonably expect most other people to pay their taxes. After all, it does not make much of a difference that you pay your taxes if almost nobody else does. But it may be that the only way to ensure that a sufficient number of people pay their taxes is to criminalize tax evasion. In that case, tax evasion is not wrongful independently of criminalization, and non-criminal legal rules cannot make it pre-criminally wrongful, but it is hardly plausibly that tax evasion cannot therefore be justifiably criminalized (Tadros, 2013, pp. 323–325).

The weak wrongness constraint has also been subject to some criticisms on the grounds that it might fail to hold in cases, where the point of criminalization is unrelated to the criminalized conduct itself. Imagine, for instance, that a small country will be invaded by its powerful neighbor unless it criminalizes the possession of a particular kind of drug. This country might be justified in criminalizing the possession of this drug in order to appease the powerful and aggressive neighboring country. Because the purpose of such a criminal law has little to do with the properties of the criminalized conduct itself, possessing the drug might remain permissible after criminalization, rendering such a proscription incompatible with the weak wrongness constraint (Tadros, 2016, p. 98).


Let us now move on to the liberal position outlined by Joel Feinberg in his influential four-volume work on the moral limits of the criminal law (Feinberg, 1986, 1987, 1988, 1990). Liberals affirm the following three principles:

The narrow harm justification: It is always a good reason to criminalize conduct that criminalization prevents harm to others.

The offense justification: It is always a good reason to criminalize conduct that criminalization prevents serious offense to others.

The liberal constraint: Criminalization is never justified unless it prevents harm or offense to others.

Feinberg sees his position as a spiritual heir to John Stuart Mill’s immensely influential harm principle, according to which “the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others” (Mill, 1992, p. 12), which, though different in some ways, is “faithful to Mill’s liberal motivating spirit” (Feinberg, 1987, p. 15).

One needs to know the meaning of harm and offense in order to assess the implications of liberalism. Feinberg understands harms as wrongful setbacks to interests (Feinberg, 1987). This implies that all harms are wrongs, but not all wrongs are harms. One has an interest in “those things in which one has a stake” (Feinberg, 1987, p. 34), such that one gains or loses depending on the state of those things, and “flourishes and languishes as they flourish or languish” (Feinberg, 1987, p. 34). Although there are many other understandings of harm (Cf. Holtug, 2002), Feinberg’s account of harm has the attractive implication that merely disliked experiences do not qualify as harmful, as one lacks a general interest in getting what one wants and avoiding what one dislikes. This implication prevents practices from qualifying as harmful merely because some citizens object so strongly to these practices that they experience disliked mental states as a result of the mere knowledge that engaging in those practices is legal (Feinberg, 1987, p. 50). Needless to say, disqualifying practices, such as homosexual intercourse, from counting as harmful merely because they are the subject of widespread prejudice and dislike is essential if liberalism is to realize its motivating aims. Feinberg understands offense as “a disliked state of mind [. . .] produced wrongfully by another party” (Feinberg, 1988, p. 2). The inclusion of the offense justification caters to the intuition that there is a good reason to criminalize some conduct—racist abuse and various forms of harassment come to mind—that produces a disliked state of mind in, but does not set back the interests of, its victims (Feinberg, 1988, p. 25). Prima facie, this threatens to let moralism (in the pejorative sense) in through the back door. Prejudice and misguided moral objections will cause some to experience disliked states of mind when confronted with innocent practices that no liberal will want to criminalize (again homosexual expressions of affection seem an obvious example). This makes the requirement that disliked states of mind must be wrongfully produced to qualify as offense of vital importance, as it excludes such disliked states of mind from consideration under the offense justification. Furthermore, Feinberg stresses that only conduct that causes very serious offense may justifiably be criminalized. This provides another safeguard against an overly permissive offense justification (Feinberg, 1988, p. 26).

Having clarified the meaning of harm and offense, the case for liberalism can now be examined. The grievance argument points out that harms and offenses constitute a special class of wrongs.3 Harming or seriously offending others is not victimless conduct, which merely provides a reason for impersonal moral criticism on behalf of no one in particular. Rather, harms and offenses are wrongs to somebody who has a legitimate grievance against the person responsible for harming or offending her. Thus, complaints about harmful or seriously offensive conduct are made on behalf of somebody who can reasonably demand legal protection and remedy against it. Therefore, it is a good reason in favor of criminalization that it prevents harm or serious offense to others. By contrast, no one is wronged by free-floating evils, wrongs that cause neither harm nor offense, and complaints about such victimless wrongs cannot be made on behalf of anybody:

At most, they [free-floating evils] are matters for regret by a sensitive observer. To prevent them with the iron fist of legal coercion would be to impose suffering and injury for the sake of no one else’s good at all. For that reason the enforcement of most non-grievance morality strikes many of us as morally perverse.

(Feinberg, 1990, pp. 79–80)

Therefore, criminal prohibitions are never justified unless they prevent harm or serious offense to others. “Extreme” liberals who reject the offense justification can also endorse this argument, provided they make the case that causing serious offense to others is either harmful or fails to give the victim a grievance. A related argument in favor of liberalism points out that because criminal proscriptions cause harm, including some that is undeserved,4 criminal proscriptions are justified only if they produce benefits of a type that can compensate for this harm, namely, the prevention of harm to others (Edwards, 2014, p. 262). It is, however, unclear why the harm of criminal proscriptions cannot be compensated for by other benefits than the prevention of harm to others (e.g., the prevention of harm to self). Moreover, this line of argument seems to undermine the offense justification. Alternatively, the idea could simply be that because criminal proscriptions always make some worse off, they are justified only if they also make some better off by preventing harm or offense.

A general worry about liberalism is that the wrongfulness of harmful conduct might be doing all the work in explaining why there is a reason to criminalize harmful conduct. If so, the intuitive justifiability of criminal proscriptions against harmful conduct supports legal moralism rather than the narrow harm justification. Because the strongest intuitions about which criminal laws are justified and which are not tend to involve, respectively, harmful wrongs (e.g., murder, assault, rape) and conduct that is both harmless and permissible (e.g., homosexual intercourse), these intuitions cannot help distinguish the relevance of harmfulness from that of wrongfulness. However, one reason for suspicion is that there seems to be no reason to criminalize permissible conduct that seriously sets back the interests of others, such as killing in self-defense (Moore, 1997, p. 649). It is less clear that there is no good reason to criminalize the few examples of genuinely wrongful conduct that do not set back interests (Moore, 1997, p. 646). Indeed, there might be no harmless wrongs at all (Mill, for one, might have held this view; see Mill, 1992, pp. 61, 74–75, 78; Moore, 2014, p. 193). Moreover, it sometimes seems impossible to identify setbacks to interests independently of the wrongfulness of the actions that cause them (A. Duff, 2007, pp. 127–128). An earthquake and a burglar might damage your property in the same way, but only the burglar will set back your interest in safely enjoying your private property. The interests set back by the burglar cannot be identified without reference to the wrongfulness burglary. Therefore, the good reason to criminalize burglary must be in virtue of its wrongfulness (A. Duff, 2007, pp. 127–128).

A further objection specifically to the effect-centered version of the narrow harm justification is that criminalizing the harmless conduct of one person will sometimes prevent a second person from harming third persons. Imagine, for instance, that Albert will harm Betty if Curt violates a religious taboo. The effect-centered narrow harm justification implies that this is a good reason to criminalize the violation of this taboo. But restricting Curt’s freedom in order to prevent Albert from harming Betty is to treat Curt as a mere means to an end (Ashworth & Zedner, 2012). In general, there seem to be a number of worries about the ability of liberalism to provide plausible answers about conduct that is harmless in itself but increases the probability that someone will subsequently choose to harm others. Call this remotely harmful conduct (cf. Simester & Von Hirsch, 2011, pp. 75ff). The causal connection of such conduct with harm is sufficient to make its criminalization prevent harm. However, including all remotely harmful conduct under the narrow harm justification and the liberal constraint would render liberalism problematically permissive. Even the criminalization of gay (and heterosexual) sex will prevent some dysfunctional, abusive, and therefore harmful, romantic relationships, thus preventing remote harm (Tadros, 2016, pp. 103–104)! If all conduct with a remote causal connection to harm qualifies as harmful, liberalism will be compatible with almost anything. Anthony Duff expresses this worry when he claims that “if the Harm Principle extends to cover conduct that is not itself directly harmful or dangerous, it is hard to see how it can set tight or determinate limits on the scope of the criminal law” (A. Duff, 2007, p. 137). If, on the other hand, no conduct with a remote causal connection to harm qualifies as harmful, the liberal constraint will exclude criminal proscriptions of remotely harmful conduct such as direct incitement to violence, the sale of handguns, and distributing a how-to guide on nuclear terrorism! Most liberals would be loath to reject all of these criminal proscriptions as unjustified. Thus, liberals need some criterion for distinguishing between justified and unjustified proscriptions of remotely harmful conduct. Although some plausible suggestions exist, these risk undermining liberalism another way, namely, by reducing the question of remote harmfulness completely to a question of wrongfulness (cf. Simester & Von Hirsch, 2011, pp. 70ff). Consider now:

The wide harm justification: It is always a good reason to criminalize conduct that criminalization prevents harm.

The narrow and wide harm justifications both imply that there is a good reason to criminalize conduct in order to prevent citizens from harming each other. However, the wide harm justification also implies that there is a good reason in support of paternalist criminal laws that prevent citizens from harming themselves. Examples include laws making it mandatory to wear seat belts and proscribing the use of certain recreational drugs. A principled point in favor of the wide harm justification is that harm is, by definition, something that is ceteris paribus good to prevent, regardless of who inflicts it (Feinberg, 1986, p. 25; Simester & Von Hirsch, 2011, pp. 143–144). It is therefore the liberal who must explain why harm matters only when inflicted by others.

However, liberals can make two forceful objections to paternalism in the criminal law. First, they can argue that since citizens cannot be wronged by conduct to which they consent, and they always consent to their own conduct, citizens are not wronged by harms they inflict on themselves (Feinberg, 1986, p. 115). This is what makes harm to self, different from harm to others. Because one has a grievance only if one has been wronged, no one has a grievance when someone harms herself. The permissibility of harming oneself5 also leads legal moralists (Moore, 1997, pp. 759ff, 2009, p. 32) and proponents of the sovereignty principle (Ripstein, 2006, p. 245) to object to the wide harm justification. Indeed, the permissibility of harming oneself puts paternalist criminal laws at odds with the wrongness constraint and anyone who affirms it. Second, paternalist criminal laws are perversely counterproductive under realistic levels of compliance. Consider the example of criminalizing the possession of recreational drugs for personal use in order to prevent citizens from harming themselves through drug abuse. Such a criminal proscription leads the state to inflict the further harm of criminal punishment on those citizens whom it fails to deter from harming themselves through drug abuse. Rather than attaining its objective of preventing harm, this (and other) paternalist criminal laws end up causing harm to the very same citizens they are supposed to help (Husak, 2008, pp. 151ff; Simester & Von Hirsch, 2011, p. 157).

These objections are forceful but apply only to paternalism in the criminal law. Hence, the more interesting question concerns paternalism outside the criminal law, that is, a wide harm justification applied to legal coercion in general. The main objection to paternalism in general is based on its apparent conflict with respect for individual autonomy (Midtgaard, 2016). Because adults are capable of making their own choices, interfering with these for their own sake amounts to treating these adults as children (hence paternalism). However, paternalistic interference seems to preserve autonomy in some situations, such as when paternalist interventions sometimes prevent harm that would have undermined the agent’s capacity for autonomous choice, or prevent conduct that is misaligned with the long-term goals of the agent because of some mistake by the agent (Simester & Von Hirsch, 2011, pp. 149–150). In such cases, objections to paternalist legal interventions lose a lot of their force, leaving room for justified paternalist laws.6

The Public Wrong Principle

Consider now:

The public wrong principle: There is always a good reason to criminalize public wrongs, and criminalization is never justified unless the proscribed conduct is publicly wrong (A. Duff, 2007, 2018; R. A. Duff, 2014, 2016; Duff & Marshall, 2010; Marshall & Duff, 1998).

This principle asserts both a constraint and a justification, and is exclusively concerned with the justification of criminalization. A particular wrong is public in some polity if, and only if, that wrong threatens or violates the defining aims and values of the polity (A. Duff, 2007, pp. 141–142). The case for the public wrong principle proceeds from a relational account of responsibility. On this account, persons are responsible as participants in some practice, to other participants in that practice, for matters determined by the defining aims and values of that practice (A. Duff, 2007, p. 37). Thus, a married doctor is responsible to his wife for adultery, but not to his patients, as infidelity violates the defining aims and values of marriage, but not of medicine. Indeed, his fidelity is not the business of his patients at all. Criminal responsibility is responsibility as citizens to our fellow citizens. The practice in which the members of society participate together as citizens is “living together as and in a polity” (A. Duff, 2007, p. 50). Thus, citizens are criminally responsible only for those wrongs that violate or threaten the defining aims and values of the polity (A. Duff, 2007, pp. 37, 49–50).

The most striking feature of the public wrong principle is that it is political, in the sense of implying that facts about the state have non-trivial consequences for what conduct there is a good reason to criminalize. Neither the legal moralism of Moore (1997) nor the liberalism of Mill (1992) and Feinberg (1987, 1988, 1986, 1990) are political in this sense. Indeed, the fact that the criminal justice system is part of the state plays virtually no role in their accounts (A. Duff, 2007, pp. 126ff; R. A. Duff, 2016; R. A. Duff et al., 2014, pp. 17ff). As an apolitical principle, legal moralism struggles to explain territorial jurisdiction as well as the underlying rationale for distinguishing between criminal and civil wrongs (presumably the equally apolitical liberalism is vulnerable to the same criticism). Legal moralism struggles to explain territorial jurisdiction, because the focus on wrongdoing as such implies that all legislators have at least some reason to criminalize all wrongdoing everywhere (A. Duff, 2007, pp. 48–49). However, Sweden does not criminalize or claim jurisdiction over the murder of a Brazilian by a Chilean in Argentina, and nor should it. The public wrong principle can explain why by noting that even the grave wrong of murder is no threat to the defining aims and values of Sweden unless it involves a Swede or is committed on Swedish territory (A. Duff, 2007, pp. 48–49). Legal moralism also struggles to explain the principled reasons for distinguishing between civil and criminal wrongs, as it holds that there is a reason to criminalize all wrongs, which would vastly expand the criminal law (Husak, 2008, p. 199). The public wrong principle can nicely explain the distinction between criminal and civil wrongs as follows. Criminal wrongs are public wrongs that concern the whole polity in virtue of threatening its defining aims and values, which is why the community, rather than the victim, is listed as the complainant (by titles such as “People,” “State,” “Commonwealth”), and controls the prosecution. Civil wrongs are private, and it is thus for the victim to decide whether to seek legal redress, in which case she will be listed as the complainant (Marshall, 2014, p. 298).

Critics of the public wrongness principle have pointed out that the public wrong principle does not in itself rule out severely unjust laws, as the principle does not imply that it is never (or always) justified to criminalize any particular wrong. Any wrong qualifies as public if, and only if, it threatens the defining aims and value of the polity, but there are almost no constraints on what those aims and value can be (Tadros, 2016, p. 126). To see why this is a problem, imagine a hypothetical polity, Homeria, which defines itself by values such as fidelity, honor, and honesty, but not (female) sexual autonomy. In Homeria, adultery and slights to the honor of others would constitute public wrongs, whereas many sexual offenses would not. However, many would consider the criminalization of infidelity to be unjust and the criminalization of sexual offenses to be morally required in any polity, including Homeria. Those who think that the criminalization of some genuine wrongs would be unjust in any polity, whereas the criminalization of other wrongs would be just—indeed morally required—in any polity, will thus think the public wrongness principle constrains the criminal law both too much and too little. Duff does point out that no liberal polity can consider all wrongdoing public (A. Duff, 2007, p. 50). Furthermore, he mentions two substantial constraints on what conduct can be publicly wrong. First, public wrongs must be wrongs. Second, “murder, serious physical assault, rape, attacks on property” (A. Duff, 2007, p. 143) must constitute public wrongs in any decent polity. However, these remarks will hardly be sufficient to alleviate concerns that the public wrong principle fails to rule out repugnant legal regimes. To say that any liberal polity must see some wrongs as private precludes no particular wrong from counting as public, even in a liberal polity. Thus, it does almost nothing to rule out repugnant laws. Similarly, the list of universally public wrongs omits many wrongs that the legal moralist will think any polity is obligated to criminalize. A more promising line of reply to the criticism that the public wrong principle fails to rule out unjust laws is to concede that there is something unjust about a polity such as Homeria, but insist that the problem is not that the criminal law of Homeria fails to reflect the polity’s defining aims and values, but these aims and values themselves (A. Duff, 2018, p. 166).

The Sovereignty Principle

We now turn to the following principle:

The sovereignty principle: “The only legitimate restrictions on conduct are those that secure the mutual independence of free persons from each other” (Ripstein, 2006, p. 229).

This principle is a constraint concerned with (state) coercion in general. In order to know what the sovereignty principle means, one needs to know what it means to secure the mutual independence of free persons from one another. According to Ripstein, “You are independent if you are the one who decides what ends you will use your powers to pursue, as opposed to having someone else decide for you” (Ripstein, 2006, p. 231). Thus, securing mutual independence is to secure the equal freedom from intentional interference with one’s freedom to decide what ends one uses one’s powers to pursue. Ultimately this rests on a conception of the ability of setting and pursuing one’s own ends as one of the fundamental interests of human beings (Ripstein, 2006, p. 231). One can violate the independence of others in two ways. First, one can coerce or manipulate others to do what they would not otherwise have done. This violates their independence, because it uses them and their powers in pursuit of an end chosen by others (rather than by themselves), thus depriving them of the ability to choose not to pursue this end (Ripstein, 2006, p. 234). Second, one can destroy or damage the capacity of others to set and pursue their own ends. Standard harms such as physical injuries violate independence in this way. For instance, someone who breaks your legs limits the ends you are able to pursue by (temporarily) destroying your power to walk. This violates independence because the person doing you harm makes it a matter of her choice whether you retain your powers to set and pursue ends (Ripstein, 2006, p. 235).

The case for the sovereignty principle takes its starting point in a counterexample to the liberal constraint. Ripstein asks the reader to imagine that a stranger enters her home and takes a nap in her bed without damaging her property at all. The stranger picks the door lock without damaging it, brings her own bed linen, sleeps with a hairnet in lint-free pajamas and takes great care to clean up after herself before leaving (Ripstein, 2006, p. 218). Ripstein argues that this sort of trespassing is justifiably criminal, even though it is perfectly harmless. This is incompatible with the liberal constraint. However, the justifiability of such legislation is compatible with the sovereignty principle. Trespassers use the powers of another—in the form of their property—to pursue their own ends without consent (Ripstein, 2006, p. 240). Thus trespassing, whether harmless or not, violates mutual independence in the first of the two ways outlined by Ripstein. Ripstein furthermore argues that the sovereignty principle realizes the ambitions motivating liberalism at least as effectively. It does so because the sovereignty principle implies that protection of individual liberty itself is the only valid reason to interfere with liberty, thus demarcating an expansive sphere of individual liberty (Ripstein, 2006, p. 245). Moreover, the sovereignty principle is able to accommodate the liberal rejection of paternalism because harming others violates their independence, but harming oneself involves no subjection of someone else to one’s will (Ripstein, 2006, p. 236).

It is not clear that Ripstein is right to think that liberals cannot account for the justifiability of criminalizing harmless trespassing. Even though it is possible to imagine instances of harmless trespassing, a blanket prohibition of trespassing might still prevent harm to others. This is so because prohibiting only harmful trespasses would invite citizens to use their own fallible judgment to determine whether a given instance of trespassing is harmful. This will almost certainly result in some mistakes, thereby leading to more acts of harmful trespassing than would have occurred if trespassing was generally prohibited (Edwards, 2014, pp. 272–273). However, Victor Tadros provides an alternative example of a justified criminal law that is incompatible with the liberal constraint but compatible with the sovereignty principle, namely, criminal proscriptions of interfering with attempts at self-sacrifice (Tadros, 2011b, pp. 53–59). The scenario Tadros has in mind is something like the following. A falling brick will hit Arthur unless Betty pushes him out of the way, in which case the brick will hit her. Being hit by the brick will cripple Arthur but kill Betty. As Betty attempts to push Arthur out of the way, Carl prevents her from doing so by tackling her. Because Carl prevents more harm to Betty than Arthur would suffer, his interference with attempts at self-sacrifice is not harmful. Thus, liberals cannot justify criminalizing interference with attempts at self-sacrifice. However, Tadros thinks that interfering with self-sacrifice in this manner should indeed be illegal and that the sovereignty principle can explain why because such interference violates the independence of the would-be hero by preventing her from pursuing the end of preventing harm to others by harming herself (Tadros, 2011b, p. 62). However, Tadros also formulates an important objection to the sovereignty principles. Like any normative principle rooted in the human capacity to autonomously set and pursue ends, the sovereignty constraint has difficulty dealing with beings who lack this capacity but are nevertheless subjects of moral concern: infants, the severely mentally disabled, the comatose, and animals. All of these lack the capacity to set and pursue ends, yet surely it should be a criminal offense to harm them; however, the sovereignty principle implies that such a criminal law would be unjust (Tadros, 2011b, pp. 62–63).

Concluding Remarks

This article has discussed the nature of law, why law is in need of justification, and what is attractive about the rule of law. The major part of this article, however, has discussed the content, strengths and weaknesses of different principles concerning the justification of law, notably, legal moralism, the wrongness constraint, the wide and narrow harm justifications, the liberal constraint, the public wrong principle, and the sovereignty principle. How should one react to the fact that the principles about the justification of law were all vulnerable to various objections? One might take this to show that searching for grand principle(s) is not a useful way to think about the justification and scope of law. This is the view of Victor Tadros, who argues that any tolerably precise principle will fail, whereas any principle that is not tolerably precise will fail to illuminate (Tadros, 2016, pp. 92ff). He suggests that we should instead think about the limits of criminal law in terms of the limits of what conduct it is permissible for the state to punish (Tadros, 2016, pp. 131–133, 170–172). Alternatively, one might note that the main objections to at least some of the principles do not show that these principles are bound to fail, but merely that more work needs to be done by the defenders of those principles. This is particularly clear in the case of legal moralism and the public wrong principle. The objections to legal moralism show that legal moralists must do more to explain why the good reasons to respond wrongdoing are always reason for the state, but do not show that such an explanation cannot be provided. The main objection to the public wrong principle is that the principle was compatible with repugnant legal regimes in the absence of a specific theory about the defining aims and values of the state, but one might be able to avoid this problem by joining the public wrong principle with a particular theory of the state. After all, nothing about the public wrong principle dictates relativism about the defining aims and values of states.


  • Arneson, R. J. (2013). The enforcement of morals revisited. Criminal Law and Philosophy, 7(3), 435–454.
  • Ashworth, A., & Zedner, L. (2012). Prevention and criminalization: Justifications and limits. New Criminal Law Review: An International and Interdisciplinary Journal, 15(4), 542–571.
  • Brennan, J. (2016). Against democracy. Princeton, NJ: Princeton University Press.
  • Devlin, P. (1963). The enforcement of morals. Indianapolis, IN: Liberty Fund.
  • Dimock, S. (2014). Contractarian criminal law theory and mala prohibita offences. In R. A. Duff, L. Farmer, S. E. Marshall, M. Renzo, & V. Tadros (Eds), Criminalization: The political morality of the criminal law (pp. 151–181). Criminalization Series 4. New York, NY: Oxford University.
  • Duff, A. (2007). Answering for crime: Responsibility and liability in the criminal law: Legal Theory Today. Oxford, U.K.: Hart.
  • Duff, A. (2018). The realm of criminal law. Oxford, U.K.: Oxford University Press.
  • Duff, R. A. (2014). Towards a modest legal moralism. Criminal Law and Philosophy, 8(1), 217–235.
  • Duff, R. A. (2016). Legal moralism and public wrongs. In K. K. Ferzan & S. J. Morse (Eds.), Legal, moral, and metaphysical truths (pp. 95–109). Oxford, U.K.: Oxford University Press.
  • Duff, R. A., & Marshall, S. E. (2010). Public and private wrongs. In G. H. Gordon, J. Chalmers, F. Leverick, L. Farmer, & A. Ashworth (Eds.), Essays in criminal law in honour of Sir Gerald Gordon (pp. 70–85). Edinburgh Studies in Law (Vol. 8). Edinburgh, Scotland: Edinburgh University Press.
  • Duff, R. A., Marshall, S. E., Farmer, L., Renzo, M., & Tadros, V. (2014). “Introduction.” In R. A. Duff, S. E. Marshall, L. Farmer, M. Renzo, & V. Tadros (Eds.), Criminalization: The political morality of the criminal law. New York, NY: Oxford University Press.
  • Dworkin, G. (1999). Devlin was right: Law and the enforcement of morality. William and Mary Law Review, 40(3), 927–946.
  • Edwards, J. (2014). Harm principles. Legal Theory, 20(4), 253–285.
  • Feinberg, J. (1965). The expressive function of punishment. The Monist, 49(3), 397–423.
  • Feinberg, J. (1986). The moral limits of the criminal law, Vol. 3: Harm self. Oxford, U.K.: Oxford University Press.
  • Feinberg, J. (1987). The moral limits of the criminal law, Vol. 1: Harm to others. Oxford, U.K.: Oxford University Press.
  • Feinberg, J. (1988). The moral limits of the criminal law, Vol. 2: Offense to others. Oxford, U.K.: Oxford University Press.
  • Feinberg, J. (1990). The moral limits of the criminal law, Vol. 4: Harmless wrongdoing. Oxford, U.K.: Oxford University Press.
  • Hart, H. L. A. (1963). Law, liberty, and morality. Stanford, CA: Stanford University.
  • Hart, H. L. A. (2012). The concept of law (3rd ed). Clarendon Law Series. Oxford, U.K.: Oxford University Press.
  • Holtug, N. (2002). The harm principle. Ethical Theory and Moral Practice, 5(4), 357–389.
  • Husak, D. N. (2008). Overcriminalization: The limits of the criminal law. New York, NY: Oxford University Press.
  • Marshall, S. E., & Duff, R. A. (1998). Criminalization and sharing wrongs. Can JL and Jurisprudence, 11(1), 7–22.
  • Marshall, S. E. (2014). “It isn’t just about you”: Victims of crime, their associated duties, and public wrongs. In R. A. Duff, L. Farmer, S. E. Marshall, M. Renzo, & V. Tadros (Eds.), Criminalization: The political morality of the criminal law (pp. 291–306). Criminalization Series 4. New York, NY: Oxford University Press.
  • Midtgaard, S. F. (2016). Paternalism. Oxford Research Encyclopedias—Politics.
  • Mill, J. S. (1992). On liberty and utilitarianism. Everyman’s Library 81. New York, NY: Knopf.
  • Moore, M. S. (1997). Placing blame: A theory of the criminal law. Oxford, U.K.: Oxford University Press.
  • Moore, M. S. (2009). A tale of two theories. Criminal Justice Ethics, 28(1), 27–48.
  • Moore, M. S. (2014). Liberty’s constraints on what should be made criminal. In R. A. Duff, L. Farmer, S. E. Marshall, M. Renzo, & V. Tadros (Eds.), Criminalization: The political morality of the criminal law (pp. 182–212). Criminalization Series 4. New York, NY: Oxford University Press.
  • Postema, G. J. (2005). Politics is about grievance: Feinberg on the legal enforcement of morals. Legal Theory, 11, 293–323.
  • Raz, J. (2009). The authority of law: Essays on law and morality (2nd ed.). Oxford, U.K.: Oxford University Press.
  • Ripstein, A. (2006). Beyond the harm principle. Philosophy & Public Affairs, 34(3), 215–245.
  • Ryberg, J. (2016). Punishment and political philosophy. Oxford Research Encyclopedias—Politics.
  • Shiffrin, S. V. (2000). Paternalism, unconscionability doctrine, and accommodation. Philosophy & Public Affairs, 29(3), 205–250.
  • Simester, A. P., & Von Hirsch, A. (2011). Crimes, harms, and wrongs: On the principles of criminalisation. Oxford, U.K.: Hart.
  • Tadros, V. (2010). Criminalization and regulation. In A. R. Duff, L. Farmer, S. E. Marshall, M. Renzo, & V. Tadros (Eds.), The boundaries of the criminal law (pp. 163–190). Criminalization Series. Oxford ; New York, NY: Oxford University Press.
  • Tadros, V. (2011a). Consent to harm. Current Legal Problems, 64(1), 23–49.
  • Tadros, V. (2011b). Harm, sovereignty, and prohibition. Legal Theory, 17(1), 35–65.
  • Tadros, V. (2012). Wrongness and criminalization. In A. Marmor (Ed.), The Routledge companion to philosophy of law (pp. 157–173). New York, NY: Routledge.
  • Tadros, V. (2013). The ends of harm: The moral foundations of criminal law. Oxford Legal Philosophy. Oxford, U.K.: Oxford University Press.
  • Tadros, V. (2016). Wrongs and crimes. Criminalization Series. Oxford, U.K.: Oxford University Press.


  • 1. For further discussion of punishment, see the article dedicated to this subject (Ryberg, 2016).

  • 2. Cf. the article on punishment (Ryberg, 2016).

  • 3. For this argument, see Feinberg (1988, p. 2, 1990, pp. 79–80, 321–322), Edwards (2014, pp. 257–258), and Postema (2005, pp. 294–295).

  • 4. See the section on “The Justification of Law.”

  • 5. This could be disputed on the grounds that one violates self-regarding duties by harming oneself (Tadros, 2011a). Note further that harm to self might be wrongful in virtue of its effects on others, such as the distress caused to loved ones, even though it is not wrongful qua harm to self.

  • 6. For further discussion of paternalism, see Midtgaard (2016).