1. Individual Justice

Following Aristotle (Reference Crisp2000: 1105b), we can provisionally define the individual virtue of justice as a settled disposition of character to perform just acts with (a) knowledge, (b) sound emotion, and from (c) rational choice (prohairesis). Aristotle also thinks of virtue as hitting a mean between two extremes of vice in terms of excess and deficiency, that is, performing the right action, with the appropriate emotion, at the right place, and at the right time. As some commentators have noticed, the virtue of justice fits uncomfortably within Aristotle's doctrine of the mean (Annas Reference Annas1993: 312). However, as I discuss later, there is a sense in which justice is a mean or at least, as Aristotle puts it, ‘a sort of mean’ because ‘it is related to a mean’ (Aristotle Reference Crisp2000: 1134a).

Aristotle cites two ways in which a person is said to be unjust, reflecting two senses of the positive Greek adjective dikaion: (1) as an unlawful person, and (2) as an unfair/greedy person (Aristotle Reference Crisp2000: 1129a). Aristotle's observation forms the basis of a contrast between two broad conceptions of individual justice: a lawfulness conception and a fairness conception. According to a fairness conception, the lawful person and the just person are not necessarily the same because some laws may be unfair from the standpoint of justice and therefore not acceptable to the just person. In this sense, ‘unfair’ is an unspecified synonym for ‘unjust’. Thus, the just person acts in accordance with what justice requires as a matter of fairness (however specified). Since existing laws, whether just or not, also raise considerations of fairness in terms of their equal application, the just person need not sanction lawlessness. But the lawful and the just remain distinct kinds of reasons.

In contrast, the lawfulness conception asserts an identity: the just person is the lawful person. The just person will not act in accordance with what justice requires in any sense unconnected from what is lawful, understood as following the law of an existing political community. This use of ‘law’ is broadly consonant with F.A. Hayek (Reference Hayek1973: 72–73), who distinguishes the unguided evolutionary process of law from the deliberately designed and planned process of legislation. By following the law in the Hayekian sense, the just person follows and upholds the law of an existing political tradition, which, following John Kekes (Reference Kekes1998: 38), is ‘a set of customary beliefs, practices, and actions that has endured from the past to the present and attracted the allegiance of people so that they wish to perpetuate it’. The just person, on the lawfulness conception, has a disposition to abide by these norms, whereas on the fairness conception, the just person has a disposition to act in accordance with reasons of fairness, which may or may not support such norms.

Aristotle (Reference Crisp2000: 1134b–1136a) distinguishes the just person from two other modes of justice: the just act and what makes something just. What makes something just, for Aristotle, is determined by whatever secures the happiness or flourishing (eudaimonia) of a political community's members (Aristotle Reference Reeve1998: 1282b15–1284a). And thus, the just act is an action that contributes to the happiness of a political community's members. Further, an act of injustice performed through ignorance or excessive emotion can be separated from acts that are performed from an unjust character, which, following Aristotle's account of virtue, requires choice and a settled disposition or habit. In addition to (i) the just person, (ii) the just act, and (iii) what makes something just, we should also add for our purposes (iv) the just institution, which is an institution that is in accordance with (iii) what makes something just.

How are the four modes of justice related? While sympathetic to the possibility of beginning with the just person, David Wiggins (Reference Wiggins2004: 487–91) observes that the notion of the just person seems to depend logically on a prior analysis of the just act (i.e., what acts the just person characteristically performs), which in turn depends on what makes something just (i.e., what features make those acts just). Indeed, this is precisely the point on which fairness and lawfulness conceptions of individual justice diverge. A fairness conception claims that just acts are fair, whereas a lawfulness conception claims that just acts are lawful. Further, what makes something just seems to be tightly connected to the idea of a just institution because many questions of justice invoke or assume the existence of institutions that can be either just or unjust. But what makes an institution just? Must we answer this question prior to defining the just person? Or is there something central about the just person?

2. Institutional Justice

The structural view claims that the just person is not central to institutional justice. Instead, the structural view suggests we begin by trying to understand the justice of institutions and then think of the just person as someone who is disposed to uphold and follow the requirements of institutional justice. The compositional view, meanwhile, claims the opposite: we begin by trying to understand the just person and then think of just institutions in terms of maintaining the relations between just persons. This section examines these competing views in more detail and argues for the compositional view.

The compositional view, while not explicitly identified there, has roots in ancient Greek ethical theory, particularly the ‘depoliticized’ outlook of the Stoics, who in their emphasis on the individual virtue of justice, according to Julia Annas (Reference Annas1993: 311), ‘[did] not regard the justice of institutions as a centrally important ethical matter’. Despite its ancient roots, the possibility of a compositional view of justice seems to have been overlooked by contemporary political philosophers. For instance, responding to a possible objection that his treatment of justice is a break from the philosophical tradition represented by Aristotle, John Rawls writes, ‘Aristotle's definition [of justice] clearly presupposes … an account of what properly belongs to a person and of what is due to him’ (Rawls [1971] Reference Rawls1999c: 10). Moreover, ‘such entitlements are … very often derived from social institutions and the legitimate expectations to which they give rise’ (ibid.). Following the compositional view, however, one may argue that what properly belongs to a person, while logically connected to social institutions, cannot be determined by a theory of justice for what Rawls calls ‘the basic structure of society’ (Rawls [1971] Reference Rawls1999c: 6). Instead, just institutions maintain the relations between just persons, and these institutions, which the just person has a disposition to recognize as just, define what properly belongs to a person and of what is due to him.

While the compositional view is seldom noticed, the structural view has dominated contemporary political philosophy since Rawls's ([1971] Reference Rawls1999c) A Theory of Justice. The structural view maintains that justice is primarily a property of institutions. ‘Justice’, in Rawls's famous phrase, ‘is the first virtue of social institutions’ ([1971] Reference Rawls1999c: 3). Moreover, it is thought that we can develop an idea of what the institutions of a just society would be like. Theorizing about justice then normatively guides our evaluation of existing institutions. And this normative guidance extends to the just person, who is understood to have a disposition to follow and promote the institutions recommended or supported by the most plausible (or reasonable) theory of justice. As Rawls puts it, the duty of justice for individuals ‘requires us to support and to comply with just institutions that exist and apply to us. It also constrains us to further just arrangements not yet established’ (Rawls [1971] Reference Rawls1999c: 99).

The structural view rightly focuses on institutions as a site of justice (cf. Cohen Reference Cohen1997: 3–30). This is an important contribution. However, the structural view fails to respond to the sense in which the individual virtue of justice is part of a good life as a whole, which is the central normative concept of ancient virtue ethics (Annas Reference Annas1993: 27–46). In modern philosophical language, the structural view divorces the reasons that motivate the just person as part of a good life from the reasons that support the institutions of a theory of justice. Famously, for example, Rawls requires that we leave behind ideas about the good (and so virtue) in order to enter the original position and consider justice impartially behind the veil of ignorance. This creates a bifurcation in practical reasoning because institutions are not justified in terms of the virtues characteristic of a good life that an actual person might pursue, but in terms of an ideal theory of just institutions for ideally just persons. But when we cease to reason behind the veil of ignorance, ideally just institutions may or may not answer to the concerns of actual persons and how they conceive of the good for themselves and their communities.

One reason for this bifurcation in practical reasoning is Rawls's claim that the concept of the right is prior to the concept of the good (Rawls [1971] Reference Rawls1999c: 28). This means, among other things, that individuals have broad latitude to pursue different conceptions of the good so long as these are consistent with and (ideally) supportive of what institutional justice requires as a matter of right (Rawls [1988] Reference Rawls and Freeman1999d: 449–50). In this regard, Rawls follows Henry Sidgwick ([1874] Reference Sidgwick1981: bk. 1, ch. 9) who thought of the right as an imperative notion, what we are required to do even if it does not appeal to us, whereas the good is an attractive notion, that is, what is a desirable (but optional) way to live (see also Gaus Reference Gaus2011: 7).

However, with some exceptions (Watson Reference Watson1990: 449–70), these philosophers have not noticed that the virtue of justice has both notions. Justice is a constraint because it requires or prohibits certain actions. This is the aspect of the right, justice as an imperative notion. This imperative sense of virtue supports what Rosalind Hursthouse (Reference Hursthouse1999: 36–39) calls ‘v-rules’, which are rules that follow from acting in accordance with the virtues. But justice is not simply an imperative notion. Justice is also a directive because it directs the virtuous person to be just and to establish just institutions where possible, both of which we ought to desire because they are demanded by virtue. This is the aspect of the good: justice as an attractive notion. The main defect of the structural view, then, is that it does not adequately account for the directive aspect of justice; it might tell us what institutions are required by a theory of justice, but it does not integrate that idea into a theory of the just person that we are compelled to emulate.

Rawls's structural view was evidently motivated to avoid a utilitarian version of a compositional view, according to which just institutions are defined in terms of satisfying the principle of utility for individuals. For this type of utilitarian, just institutions are composed of individuals who are just insofar as they act on the principle of utility through the medium of institutions. In this way, Rawls observed, utilitarianism extends to ‘society as a whole the principle of rational choice for one man’ (Rawls [1971] Reference Rawls1999c: 24). And Rawls rejected utilitarianism because it allowed individual interests to be sacrificed whenever doing so might contribute to the aggregate good of society. Hence, ‘[u]tilitarianism does not take seriously the distinction between persons’ (Rawls [1971] Reference Rawls1999c: 24). Can the same objection be made to compositional views generally? Does the compositional view of justice allow the sacrifice of the individual for the group? There is good reason to think the objection does not apply to a compositional view that takes virtue instead of utility as central. As Hursthouse explains:

The italicized portion of this passage is what Mark LeBar (Reference LeBar and Russell2013b: 273) calls ‘Hursthouse's Constraint’. According to LeBar, we should interpret Hursthouse's claim as setting a constraint on the authority of political institutions, such that the individual good cannot be sacrificed for the overall good precisely because the good is understood in terms of virtue. In other words, Hursthouse avoids Rawls's objection to utilitarianism because she recognizes the constraining aspect of individual justice. Individual justice constrains what can count as institutional justice because just institutions are composed of the relations between just persons.

To sum up this section, the virtue of justice has what I have called a constraining aspect and a directive aspect. I rejected the structural view because it does not adequately account for the directive aspect of justice. A theory of the virtue of justice should integrate both aspects, as the compositional view does. There is still a remaining issue, however, for we could understand individual justice either in terms of fairness or lawfulness, and this will make a difference in how we understand the claim that institutional justice is composed of the relations between just persons. The compositional view remains incomplete until we settle on a conception of individual justice that determines what it means for institutional justice to be composed of the relations between just persons. Thus, we need to examine more closely the fairness and lawfulness conceptions of justice as an individual virtue.

3. A Fairness Conception of Individual Justice

While fairness has a strong association with post-Rawls egalitarian work on distributive justice (e.g., Cohen Reference Cohen2008), it is also an important concept in the classical liberal tradition (see Tomasi Reference Tomasi2012: 96–99, 162–96). In recent work, following this tradition, LeBar (Reference LeBar and Hooft2014, Reference LeBar2013a, Reference LeBar and Russell2013b, Reference LeBar2009, Reference LeBar2005) defends a version of what I have called a fairness conception of individual justice that incorporates the Kantian idea of respect for persons in a way that is suitable to approaching issues in political philosophy from the perspective of virtue ethics. LeBar's view is significant because, unlike Rawls and many subsequent political philosophers who combine a fairness conception of individual justice with the structural view, LeBar defends the compositional view, and he does so from the perspective of virtue ethics (LeBar Reference LeBar and Hooft2014: 273–74). It is worthwhile therefore to examine LeBar's view as a competitor to ‘justice as lawfulness’.

One form LeBar's view takes is an argument for ‘a certain liberal conception of political authority—what is sometimes called ‘justificatory liberalism’—[as] something virtue ethical theories have reason to endorse’ (LeBar Reference LeBar and Russell2013b: 265, emphasis in original). Justificatory liberalism is a liberal conception of the authority of political institutions that ‘requires the justification of that authority to each individual subject to it’ (LeBar Reference LeBar and Russell2013b: 279). Drawing on Rawls (Reference Rawls1993), justificatory liberals (or sometimes ‘public reason liberals’) ground their account of political authority in a conception of persons as fundamentally free and equal insofar as they are reciprocally placed to determine their obligations and duties (Gaus Reference Gaus2011: 14–35). No person, according to this view, has any natural or preexisting obligation to obey another.

LeBar (Reference LeBar and Russell2013b: 278; Reference LeBar and Hooft2014: 267–70) argues that an understanding of this ‘justificatory relationship’ between persons is missing from virtue ethics, particularly in the ancient sources. To understand the justificatory relationship better, LeBar draws on work by T. M. Scanlon (Reference Scanlon1998) to propose that we see ourselves as standing in a relation of ‘mutual recognition’ with each other, and he identifies this relation as involving a virtue. LeBar cites the following passage from Scanlon:

Elsewhere LeBar (Reference LeBar and Hooft2014: 268–70) is clear that this relationship between persons is a way of understanding the reasons we have to respect persons, and that respect for persons is plausibly the core of an adequate conception of individual justice.

There are three Kantian elements to LeBar's conception of the virtue of individual justice. Each contributes to the idea that justice is about fair dealings with others. The first is the familiar Kantian principle of always treating persons as ends in themselves and never merely as means. Here the idea is that we cannot simply use other people as means to our own ends because they have ends of their own sufficient to demand respect (LeBar Reference LeBar2013a: 10–14). A just person recognizes persons as ends in themselves. The force of the first Kantian idea leads to a second idea that the just person, in his relations with others, ‘refrains from imposing his will on them in any way that is incapable of reciprocation’ (LeBar Reference LeBar and Hooft2014: 268–69). Third, this recognition of reciprocal respect suggests that others have a kind of moral standing to demand that we treat them in ways that can be reciprocally justified.

LeBar's version of a fairness conception of individual justice places the idea of reciprocity at the heart of just relations between persons (see also Schmidtz and Thrasher Reference Schmidtz, Thrasher, Timpe and Boyd2014: 59–74). It is worth noting, however, that unlike most fairness conceptions in the vast literature on distributive justice, LeBar's focus on reciprocity only amounts to a formal conception of individual justice because it does not yield substantive conclusions about what actions are just (or unjust) in determinate circumstances. In this respect, LeBar's view is much closer to Rawls's earlier work (e.g., Rawls [1958] Reference Rawls and Freeman1999a: 59–60), which linked justice and fairness through the idea of reciprocity (see Rawls [1971] Reference Rawls and Freeman1999b: 190). Nevertheless, unlike on a lawfulness conception, the just person acts on reasons of justice (as reciprocity or fairness), which are logically prior to what existing law requires. For LeBar, the just person is disposed to treat others as free and equal persons, deserving of respect and reciprocal justification for the imposition of duties and obligations. This is what it means to treat someone fairly in accordance with reasons of justice. And since LeBar also defends the compositional view, institutional justice is constrained by the relations between just persons who act according to a principle of reciprocity in their fair dealings with each other.

LeBar's conception of individual justice rightly focuses on what we might term our ‘accountability relations’ with others, that is, the ways in which we feel ourselves required to justify our actions to each other as equal members of a political community. And plausibly, part of treating others fairly in this way is to respect their status as moral persons with ends of their own. But LeBar's attempt to determine the scope and justification of our accountability relations is not successful. As I will argue, the issue of justifying our duties and obligations to one another is fundamentally about having a means of public justification, and this is a function that existing law uniquely serves.

LeBar accounts for our evident substantive disagreements about justice by focusing on the higher-order procedural question of how we justify our actions to each other (i.e., ‘who gets to decide?’). And his answer is that we are reciprocally placed insofar as moral obligations must be justified to each person (i.e., each decides for himself or herself). But this formal principle of reciprocity cannot substantively determine what justice requires in our accountability relations with others. This is because such relations are essentially a matter of jurisdiction, and jurisdiction cannot be determined without first accounting for existing relationships, institutions, and practices, many of which are publicly justified in terms of the law of an existing political community and few of which require reciprocal justification for the imposition of obligations and duties (e.g., the duties of family). To be sure, reciprocity might be a worthy ideal of how to act within our existing accountability relationships (friendship comes to mind), but it cannot determine or justify them a priori. As J. R. Lucas (Reference Lucas1980: 27) puts it, ‘[the law] tells us not what is to be done in a situation, but who is to decide what shall be done. And to have the say is to have a right’.

The foregoing indicates that what LeBar refers to as the ‘justificatory relation’ between persons is something already embedded in the relationships, institutions, and practices that constitute a political community and are the pretheoretical given of ethical life. Such entities determine our relationships to others in ways that make our obligations and duties transparent, even if not fully determinant. We have what Michael Hardimon (Reference Hardimon1994: 333–63) calls ‘role obligations’, for example, the obligations of family and country, or what the Romans called pietas. But what these obligations amount to concretely must be worked out in practice by individual persons in their particular circumstances. They cannot be determined by a theory of justice. This is partly because the virtues, much less a single virtue like justice, do not provide moral guidance that is reducible to a set of principles (McCloskey Reference McCloskey2007: 346–68). Rather, the virtues (including justice) consist in the proper way of acting—virtuously—within our existing relations with others (Annas Reference Annas2002: 103–23). The just person recognizes the normative status of these relations and acts well within them, thus serving as an ideal of the virtuous person toward which we aim.

4. Justice as Lawfulness

Justice as lawfulness begins from the premise just stated, namely, that the virtuous person always acts in the context of a shared social structure. According to justice as lawfulness, the just person is disposed to act in accordance with the nomoi of her political community. I use the ancient Greek term nomoi deliberately to indicate the usefully broad sense of ‘law’ in ancient Greek political thought. As Richard Kraut explains,

While the ancients did not make Hayek's distinction between the processes of law and legislation, Aristotle (Reference Reeve1998: 1287a–1287b10) does recognize the general superiority of customary law over written law. Since legislation may be a necessary part of a legal tradition based in law, I use ‘nomoi’ to encompass both as well as the broader sense indicated by Kraut.

Given this understanding of nomoi, in terms of Aristotle's four modes of justice (see section 1), justice of lawfulness consists of the following four claims. First, what makes something just refers to the nomoi. Second, a just act is an action performed in accordance (or not in conflict) with the nomoi. Third, the just person acts from a settled disposition to observe the nomoi, not out of blind obedience, but as a virtue, chosen with knowledge and sound emotion. Finally, given the compositional view, a just institution will be an institution that maintains the relations between just persons— that is, an institution in accordance with the nomoi.

The critical role of the nomoi in understanding the virtue of justice is that they function as public judgments for how persons should relate to each other. They constitute what LeBar calls the ‘justificatory relation’. The nomoi structure our relationships with and consequent obligations to others. A thoroughly lawless society (Durkheim's anomie) is one in which persons act solely based on their own judgment without regard to their relationships within the broader community. While there can never be complete unanimity, full transparency, or determinacy in the nomoi of a society, the public nature of law, either written or informal, allows people to interact beneficially since they know what to expect from each other. Law is the institutionalization of trust.

Justice as lawfulness is superior to a fairness conception of justice because it captures the sense in which existing law serves the function of public justification. We determine our accountability relations with others by referring to the offices and relationships that are defined by the institutions and practices of an existing political tradition, expressed by and in its nomoi. This is not a Rawlsian constructivist conception of public justification understood as public reason, but rather a conception of public justification that accounts for the actual role played by institutions and practices in ordinary moral life. Justice as lawfulness focuses on public practices rather than public reason.

Constructivist accounts of public justification appear partly motivated by what Rawls calls ‘the fact of reasonable pluralism’ (Rawls [1997] Reference Rawls and Freeman1999e: 573), ‘the fact that a plurality of conflicting reasonable comprehensive doctrines … is the normal result of [a democracy's] culture of free institutions’. For it is thought that we need a way to model and explain how the diversity of values that exist in modern societies can morally coexist with the coercive nature of political institutions (Larmore Reference Larmore1994: 61–79). While the fact of reasonable pluralism is important, especially with respect to competing conceptions of the virtues in a pluralistic society, I believe it carries too much weight in the discussion of public justification. Any stable society characterized by reasonable pluralism also consists of a set of shared beliefs, the nomoi. Reasonable pluralism is only possible if there exists some set of stable beliefs about how groups with fundamentally different belief systems ought to behave toward one another. Otherwise, while pluralism may temporarily flourish, in the long run, chaos is inevitable. And it is a virtue of justice as lawfulness that it recognizes the important function of these stabilizing beliefs as realized in the laws and institutions of existing political communities.

We are now able to reconsider Hursthouse's contribution that individual virtue—more specifically individual justice—is a constraint on institutional justice. How does the just person, who is disposed to follow the nomoi of her community, help us understand institutional justice? Since the just person upholds and follows the nomoi, which consist of both legal and social norms, we can start by thinking about the relationship between these two kinds of norms. The legal theorist Lawrence Solum (Reference Solum2008: 188–89) proposes three forms this relationship can take:

1. 1. Congruence: a legal norm can be congruent with a social norm (or vice versa), e.g., criminal laws against murder.

2. 2. Support: a legal norm can be supported by a social norm, e.g., social norms about which authority determines the rules of the road (i.e., a public transport authority).

3. 3. Inconsistency: a legal norm can be inconsistent with a social norm, either in terms of a) content or b) lack of recognized authority.

The congruence relation is one that raises no special problems from the standpoint of justice. When legal norms are congruent with social norms, there exists a harmony between what the (positive) law requires and what the community believes the law ought to require as a matter of justice. The same legal norm, of course, will be incongruent if it is exported to a place where the appropriate social norms do not exist (e.g., exporting democracy to parts of the Middle East). The support relation is similarly unproblematic. Every functioning society relies on social norms that sanction specific authorities to have jurisdiction over some realm of social life, such as a public transport authority. The just person is disposed to recognize these authorities, where established by the legitimately made law of a political community.

The crux of individual justice as a constraint on institutional justice resides in the inconsistency relation. Following Solum (Reference Solum2008: 188–9), inconsistencies between legal and social norms can take at least two forms: (a) conflict in content and (b) conflict in recognized authority. I would also add (c) conflict in implementation. What does the just person do when faced with such conflicts? Either she acts on the legal norms, in which case she is betraying the social norms, or she acts on the social norms, in which case she is betraying the legal norms. Neither preserves the just person as a lawful person.

First, consider a conflict in content. In cases of new or proposed legal norms, it is easy to see how social norms that are part of an established political tradition take precedence. For instance, if a civil authority attempted to pass a law of mandatory government informing on fellow citizens, this would violate a clear social norm regarding privacy in the United States. We also have norms about which institutions have recognized authority to perform certain functions. For example, there are norms sanctioning the traffic management authorities to regulate speed on public highways, but a traffic authority that mandated what color cars may drive on the highways would be grossly overstepping its mandated authority. The point, in both cases, is not that the prior social norm has been around for a long time, but rather that it has persisted as part of a political tradition that has successfully met the needs of its citizens, who look to their political tradition as a constitutive good in their lives, a good that allows them to cooperate peacefully for mutual benefit. The just person, then, has a disposition to uphold and defend such a tradition from the encroachment of political, social, and individual forces.

A third possibility is a law that, while having some plausible justification, requires action that is inconsistent with a social norm in its implementation. This seems to be the kind of case Hursthouse had in mind with laws that require persons to engage in vicious activity. If, for instance, there is some justification for a law prohibiting certain drugs, but in practice the enforcement of the law requires the authorities to treat minor drug offenders in an inhumane way (as arguably the War on Drugs does), there is some reason to think that the law as such is inconsistent with a social norm and therefore not a law acceptable to the just person, at least in the vicious form of implementation.

I close this section with what the reader may have felt as a lingering objection that sets up the final section's discussion of social change. I have explained the constraint of individual justice in terms of the relationship between legal and social norms (collectively, the nomoi), where the just person has recourse to one or the other to maintain the constraint of justice. But what should be said if the nomoi themselves are vicious or unjust? How does justice as lawfulness account for unjust nomoi? For the view seems to depend on the Aristotelian claim that the nomoi reliably track the happiness (eudaimonia) of the members of a political community. But we know this is often not the case, as illustrated by the antebellum South where both the legal and social norms marginalized and brutalized the slave population or in other cases where institutions fall short in less extraordinary but altogether unsurprising ways. As I will argue, this is why justice as lawfulness needs to be understood in conjunction with practical wisdom (phronesis), so that the nomoi can be reformed in accordance with an implicit standard of justice that the just person approaches incrementally through the political tradition in which he or she acts.

5. Unjust Laws and Social Change

Part of the task of this final section is to explain why justice as lawfulness is not essentially conservative. For the requirement that political institutions are constrained by the existing nomoi of a political community might seem unduly prejudiced toward the status quo and hostile to new proposals for beneficial (and sometimes necessary) social change. Indeed, one of the attractions of the structural view when combined with a fairness conception, which dominates contemporary political philosophy, is that it begins from the premise that existing institutions are very likely unjust (because unfair) and therefore must be made just in accordance with a theory of (institutional) justice. Undoubtedly fairness—especially political equality—has been an important value in the development and spread of liberal societies. But fairness is a multidimensional and highly contested political value, and it is unclear how to construct an ideal of justice that can serve the function of public justification in a political society characterized by substantive disagreement.

An important feature of the nomoi is that they are bounded in time and place. They are the path-dependent result of successful responses to social problems that existed in the past, problems to which their mere existence testifies. The existing norms of a political community are successful because they have gone through a process of social evolution; they have answered to an implicit standard of justice that is grounded in solving the perennial problems of social life. The nomoi of a society are justified not because they comport with an ideal theory of justice, but because they approximate ideal solutions to nonideal problems (Schmidtz Reference Schmidtz2016: 9). And since one of our nonideal problems is the imperfectability of human beings, we must continually fine-tune our institutional instruments to adapt to our own fallibility (Russell Reference Russell2012: 268).

What is more, changes in the moral and social landscape, for example, new economic conditions, will necessitate reform because old solutions may be maladapted to new problems. For example, while contingencies having to do with the division of labor made women's participation in politics a near impossibility in earlier ages of human society, economic progress and social change toward the end of the nineteenth century made the prospect seem desirable to enough people to win women's suffrage democratically. In this case, women (and men) sought to reform the existing social norms of their time that confined women to a fixed gender role in society. As John Stuart Mill argued, why should women not be given a fair chance to succeed on the same terms as men? It is tempting to view Mill's plea as an appeal to a theory of institutional justice or a theory of fairness. But arguably Mill was not trying to replace an old social norm with a new one. He was trying to reform the existing social norm, namely, that political participation is reserved for those who are capable of political deliberation and action, to include the other half of the population, who, in Mill's estimation, eminently met that criteria.

The necessity of new proposals for social change is grounded in the fact that over time the nomoi of a society inevitably ossify. Part of the reason for that ossification is natural decay. But it is also the result of human iniquity, fallibility, and our inability to keep pace with the dynamics of a complex world. It is, therefore, the responsibility of the just person not only to abide by the existing nomoi, but to reform the nomoi when necessary and possible, so that the political tradition can be revived and perpetuated. In the immortal words of Edmund Burke ([1790] Reference Burke1999: 108), ‘[a] state without the means of some change is without the means of its conservation’. Correspondingly, a state without just persons is without the means of its change.

According to justice as lawfulness, change—at least salutary change—comes about because of the just person who looks to change the nomoi, not by reference to abstract considerations of justice, but by reforming the existing nomoi in a way that improves upon the existing political tradition, sustaining it across time. This change in individual judgment, represented by the just person, then eventually becomes widespread enough to change the nomoi, and ultimately, though perhaps slowly and uncertainly, move the political community closer to an ideal of justice. For instance, refusing people opportunities and services based on an immutable characteristic (i.e., race) was once an accepted and widely practiced social norm in the United States. But over time individual judgments about the evils of this practice consolidated to make this kind of discrimination one of the worst transgressions of a social norm one can commit.

The moral judgment of the just person alerts us to ways in which the nomoi are deficient and therefore fall short of a standard of perfect justice that is imperfectly embedded in the institutions of a political tradition. But attempts to transform existing institutions to satisfy an ideal of justice can be quixotic or even dangerous if they are not based on shared judgments, which must balance contested ideals of justice with the very real risks of social change. This is one reason why just persons must argue for their reform proposals in public, and if they are to be persuasive, ground their proposals in the ideals of a shared political tradition. Like Martin Luther King's pronouncements on the evils of segregation—pronouncements which were deeply grounded in the highest ideals of the American political tradition—such proposals have the power to reform the nomoi over time—if they are heard and considered. Likewise, harmful proposals can be heard and defeated.

The nomoi must also include second-order judgments about justice, such as what kinds of social arrangements are considered just by the community. These judgments are widely shared by definition. Such judgments can create social pressure to reform legal norms that are no longer considered just, which is to say they no longer serve the function of maintaining a society that is perceived as just by its members. The process can also work in the other direction. For sometimes the social norms are defective, and new legal norms must be passed politically to force obstinate persons to change their behavior (this is one reason why law needs legislation).

This process of balancing justice with the vagaries of political and social change requires care and, above all, practical wisdom. The distinguishing feature of justice as lawfulness is that it places the emphasis on reform from an initial place of conformity, rather than on transformational or revolutionary change. On the one hand, insisting that institutions conform to an ideal of justice without paying attention to the existing nomoi undermines the very institutions that make reform possible. On the other hand, maintaining manifestly unjust institutions for the sake of short-term stability can undermine the long-term legitimacy of existing institutions and fail to respond to the dynamic needs of a complex society. While there is no perfect balance between stability and change, and there are no guarantees that the just person will be successful in achieving a recognizably just (much less liberal) society, justice as lawfulness nevertheless locates justice in the actions of just persons, rather than in the design of just institutions.

I have drawn the basic contrast between the structural and compositional views as residing in the relative logical priority given to either just institutions or just persons in the order of explanation. But as Annas (Reference Annas1993: 316) observes, Aristotle, whom I have taken as a paradigm of the compositional view, ‘has much to say about both [i.e., just institutions and individuals], and, in his usual fashion, has no drive to reduce one to the other or to force them under a specific common account’. While I have defended the compositional view over the structural view, I believe the view Annas attributes to Aristotle is right, and justice as lawfulness explains why. The just person is disposed to follow the nomoi. But the nomoi are inextricably linked to the institutions of an existing political community. And so, if existing institutions are in accordance with the nomoi of a political community, the just person qua lawful person is, in effect, someone who follows just institutions (per the structural view). However, institutions are just not because they comport with a theory of just institutions, but because they are in harmony with the nomoi of a political community (contra the structural view). Thus, while justice may or may not be ‘the first virtue of social institutions’, the just person strikes a necessary balance between justice (as lawfulness) and the other virtues of a good society (e.g., efficiency, fairness, artistic achievement, etc.).

This last idea is why I believe justice as an individual virtue can be usefully understood in terms of a mean. Aristotle famously defines virtue as a mean with respect to a vice of excess and a vice of deficiency. These are two ways of going wrong. For example, the temperate person avoids gluttony (excess) and asceticism (deficiency) with respect to bodily pleasures. The virtue of justice, it has been argued by Bernard Williams (Reference Williams and Rorty1980: 189–200) and others, allegedly does not fit this model because there seems to be no special feeling associated with justice, no corresponding way to do the just thing in a way that is intermediate between two unjust things, being either too just or not just enough. According to justice as lawfulness, however, there remains a sense in which justice is a mean between two ways of going wrong. The two ways of going wrong are: (a) excessive conformity to existing law and institutions without regard for reform and (b) deficient regard for existing law and institutions in pursuit of an ideal of justice. Hence, justice is a mean, either between an excess of stability and a deficiency of change or between an excess of change and a deficiency of stability. Justice is a matter of getting it right.