2. The Debate Between Practice-Independence and Practice-Dependence

Before turning to the debate between practice-independence and practice-dependence, it is necessary to clarify the key notions at issue. Mainstream practice-dependent theorists tend to understand practice in broadly institutionalist terms, as an ongoing social structure of activity that organizes patterns of interaction through institutions or rules. In this sense, practices include not only formal institutional arrangements but also informal social rules (Valentini, Reference Valentini2011, p. 403). It is precisely these different understandings of the relation between social practices and principles of justice that give rise to the methodological contrast between practice-independent and practice-dependent approaches.

The practice-independent approach holds that the relation between principles of justice and social practices takes the form of what Sune Lægaard calls an “application model” (Lægaard, Reference Lægaard2019, p. 88). On this model, theorists first identify a higher-order moral principle with general normative authority and then apply it to specific contexts to derive the relevant rules of regulation. Here the rules of regulation and the moral principle stand in a means-end relation: Social practices merely provide the background within which given moral aims are to be realized, and their specific features matter mainly for determining which rules of regulation are most conducive to promoting those aims. Utilitarianism is the classic example of a practice-independent approach. Utilitarians affirm a single general moral principle: Institutional arrangements ought to maximize overall utility. Guided by this principle, whatever the form of a practice, agents should adopt those rules of regulation that are most favourable to utility maximization (Sangiovanni, Reference Sangiovanni2016, p. 15). For practice-independent theorists, then, principles of justice are in effect the set of optimal rules of regulation for particular practices. Although the concrete content of such principles may vary with the form of the practice, their normative basis is entirely independent of the structure and point of the practice.

By contrast, the practice-dependent approach is committed to a “justification model” (Lægaard, Reference Lægaard2019, p. 89). On this model, principles of justice are not moral standards external to practice, but social rules internal to it. To provide an adequate normative justification for principles of justice, we must therefore appeal to an analysis of the point of the relevant practice. The ambition of practice-dependence is to steer a middle course between pure moralism and pure interpretivism. That is, it differs from the practice-independent approach by refusing to exclude practice altogether from the justification of principles of justice, while also rejecting the view that principles of justice are simply constitutive elements of practice whose legitimacy can be secured by interpretation alone (James, Reference James2012, p. 27).

But this immediately raises a crucial question: How does practice enter the justificatory process? In what sense can a descriptive analysis of the structure, point, and organization of an institutional scheme impose normative constraints on the content of principles of justice? Why should practice affect not merely the scope of application of principles, but the very structure of reasons offered in their support? It is precisely at this point that practice-independent theorists challenge the coherence of practice-dependence. In their view, “[a]ny argument supporting a substantive moral standard for a given group of people must be justifiable by reference to at least one moral value or independently derived moral principle” (Meckled-Garcia, Reference Meckled-Garcia2013, p. 108). If that claim is right, then any justificatory constraint imposed by practice must ultimately appeal to some moral value independent of practice. And, if so, practice can affect only the application of principles, not the justificatory stage itself.

To appreciate the force of this criticism, we need to look more closely at the way practice-dependent theorists typically proceed. They often draw on Ronald Dworkin’s method of constructive interpretation, and divide the derivation of principles of justice from practice into three stages: (1) a pre-interpretive stage, in which one describes the institutional features, internal workings, and interactional structure of a given practice in order to fix its conceptual boundaries; (2) an interpretive stage, in which one articulates the values or purposes the practice aims to realize and explains, from the participants’ point of view, why they have reason to accept these aims; and (3) a post-interpretive stage, in which one derives, on the basis of the practice’s identified structure and point, the fundamental normative principles it ought to follow (Dworkin, Reference Dworkin1986, p. 47; James, Reference James2005, p. 301; Sangiovanni, Reference Sangiovanni2008, p. 148). Within this framework, the pre-interpretive stage is purely descriptive: It provides, in roughly sociological fashion, the factual basis for normative argument. The crucial shift occurs at the second stage. At that point, the theorist must supply what may be called a “moralized characterization” of the point of the practice. Thus, in characterizing the point of global trade, Aaron James does not merely describe it as a matter of factual exchange among states, but rather as a scheme of mutual cooperation through which states become interdependent in order to increase national income. Once interpreted in this moralized way, the point of the practice is taken to impose normative constraints on principles of justice. James is explicit that the practices entering the constructive procedure are not just any existing practices; they must first be understood as serving some socially valuable purpose, and that purpose must in turn be regarded as worthwhile and legitimate (James, Reference James2012, p. 44). Only then does a practice qualify as an appropriate object for the construction of principles of justice.

But once a practice must first pass through an independent normative filter before it can enter the construction of principles of justice, its role in justification becomes obscure. Is practice really constraining the reasons that support the principles, or is it merely helping to specify how some higher-order moral principle is to be applied? This difficulty is visible in James’s own account of justice in international trade. He argues that the gains from international trade ought to be distributed equally, and one of the central reasons supporting that conclusion is the equal moral standing of participating states. Yet market practices themselves are often characterized by asymmetries of power, and they do not by themselves generate a normative requirement of equal moral standing. If so, James’s view may invoke the aims, organization, and role of practice at the surface level of construction, but at the crucial justificatory point, the real normative work is still being done by a higher-order moral principle independent of practice. As Robert Jubb puts the point, “What matters is whether features of the context a given principle or set of principles are to govern play a role in justifying those principles beyond simply constraining the extent to which they can be achieved” (Jubb, Reference Jubb2016, p. 83). If the point of a practice cannot itself shape our understanding of which higher-order moral principles are desirable or appropriate, then practice has not genuinely entered into the justification of principles of justice.

Andrea Sangiovanni is sensitive to the difficulty facing constructive interpretation, and accordingly turns to a more modest form of practice-dependence. On his view, the difference between practice-independence and practice-dependence is, at bottom, a difference between two ways in which higher-order moral principles may be applied. The former treats practice merely as an instrument for realizing those principles, whereas the latter assigns practice a mediating role in making those principles more determinate. In Sangiovanni’s terms, higher-order moral principles have an open texture: They are marked by indeterminacy of content and openness of interpretation, and so cannot by themselves generate concrete requirements for particular practical contexts. What is needed, then, is the point, structure, and characteristic features of a practice as the context within which higher-order moral principles are rendered more specific, thereby yielding a set of principles of justice appropriate to that practice. He offers the European Union (EU) as an example. If we appeal only to a general principle of reciprocity, that still does not tell us which principles of justice ought to govern the EU. We must also explain the nature and structure of the EU as a practice: what kinds of institutionalized mutually beneficial interaction it facilitates, whether those interactions are among Member States or among citizens, and what the point and purpose of the EU as an institutional arrangement is. Only on the basis of such a descriptive interpretation, he argues, can we formulate a determinate set of principles of justice. Practice is thus a necessary condition for higher-order values to acquire concrete content (Sangiovanni, Reference Sangiovanni2016, pp. 14–22).

The difficulty with this position is that, even if we grant that higher-order moral principles must be mediated through practice, it remains unclear what could make one mode of mediation normatively prior to others. Real social practices are typically complex and multifaceted; they are capable of sustaining several different interpretations. The EU, for instance, can be understood as a framework of long-term reciprocal cooperation, but also as an authoritative order that shapes Member States’ status and opportunities, or again as an institutional mechanism for managing common risks and sustaining political union. Different interpretations of the same practice will support different routes of specification, and will allow different higher-order moral reasons to enter into justification. How, then, are we to determine which interpretation — and so which mediating route — ought to be given priority? If nothing can be said here, then what determines the justification of principles of justice is still not the structure and point of practice, but rather the higher-order moral principles we independently bring to the practice in the first place. As Jubb argues, the central difficulty for practice-dependence is not merely whether practice is relevant, but that different interpretations of practice can generate different normative conclusions; theorists must therefore explain why the interpretations on which they rely are the appropriate ones (Jubb, Reference Jubb2016, p. 99). It is precisely here that Sangiovanni’s mediating-deduction argument reveals its weakness.

What all this shows is that existing practice-dependent theories have not yet explained how practice can impose substantive constraints on justification. Still, this is not yet a defeat of practice-dependence. What it shows, rather, is that a different strategy is needed. In what follows, I develop a functionalist account of practice-dependence, which aims to ground it more securely by uncovering the functional connections between principles of justice and social practices.

3. The Function of Justice and Structural Need

If practice is to enter into the justification of principles of justice, we first need to explain how principles of justice are related to practice at all. Otherwise, practice-dependence is all too easily reduced to a vague slogan, sliding between application and justification. Functionalism, I suggest, offers exactly the right perspective here. In general, functionalism is the view that the nature of a thing is to be understood in terms of the function it performs. From this perspective, natural objects, artifacts, and social practices alike occupy specific functional roles: The function of the heart is to pump blood; the function of a hammer is to strike objects; the function of a parliament is to generate collective decisions. If we import this functionalist perspective into a theory of justice, we can begin to see more clearly the functional role justice plays within social practice.

But how, exactly, should the function of justice be characterized? Merely listing the effects that justice happens to produce will not do. Equal protection of individual rights, the cultivation of civic virtue, and even the maintenance of some ideological structure may all be consequences of justice under particular historical conditions. The crucial task is to distinguish, among the many effects associated with justice, those that are merely contingent by-products from those that belong to its core function. My strategy here is to understand function in terms of need. The basic thought is that, if we want to understand the function of a concept, we should engage in a kind of reverse engineering: We reconstruct the problem situation that would have led to the emergence of the concept and ask what prompted a group lacking the concept to invent it in the first place. In this way, we can identify the concept’s functional significance. As Mathieu Queloz puts it, the task is like that of “an archaeologist who digs up a mysterious relic [and] will try to reverse-engineer its point by imaginatively reconstructing the life of those who used it and hypothesizing what useful difference it might have made to that life” (Queloz, Reference Queloz2021a, p. 3).

Pursuing this line of thought, we must begin from a situation in which justice is not yet in place, and identify the practical pressures that would have made justice emerge. David Hume’s discussion of the circumstances of justice provides a natural starting point. Hume argues that justice, as an artificial virtue, arises only under specific social conditions, such as moderate scarcity of resources, limited benevolence, and approximate equality of power (Hume, Reference Hume and Selby-Bigge1978, pp. 483–486). Although Hume’s account of these conditions is subject to various controversies, his key insight is to expose the strategic structure of social life. Interaction among members of a society can be understood in terms of a game-theoretic relationship, in which agents’ choices often depend on their expectations about the choices of others. Building on Hume’s account, we can understand the circumstances of justice in a more general way and reduce them to two basic sets of conditions. The first set consists of conflict conditions: persistent, non-eliminable disagreements among individuals regarding interests, beliefs, life plans, and values. The second set consists of coordination conditions: the fact that cooperation can generate mutual benefits and improve agents’ situations, together with common knowledge among the parties that a certain set of strategies can secure cooperative outcomes (Vanderschraaf, Reference Vanderschraaf2018, p. 110). The conjunction of these two sets of conditions gives interactions under the circumstances of justice a structure of interdependence. In such a structure, agents can neither simply disregard the actions of others, since others’ choices may affect their own interests, nor can they unilaterally realize the optimal outcome, since more favourable results depend on some predictable and stable arrangement of cooperation. It is precisely in such circumstances that justice becomes not only necessary, but also possible.

The stag hunt is often regarded as a paradigmatic case of the problem of cooperation. Consider Jean-Jacques Rousseau’s stag hunt. If both hunters cooperate in hunting stag, they secure the higher payoff; but that outcome is available only if each expects the other to persist in cooperation. If, by contrast, one worries that the other may defect and chase hare instead, then hunting hare alone, though less rewarding, becomes the safer strategy, while the one who sticks with stag may end up with nothing at all (Rousseau, Reference Rousseau and Gourevitch2019, p. 167; see also James, Reference James2012, Chapter 3; Skyrms, Reference Skyrms2004, Chapter 1; Van Schoelandt, Reference Van Schoelandt2020, p. 419). Thus, in the absence of a stable mechanism of cooperation, even when all agents know that cooperation is collectively preferable, they may still abandon it because of uncertainty about what others will do, and retreat to lower but safer strategies. The result is that individually rational, self-interested choice leaves collective interaction stuck at a suboptimal equilibrium. More than that, even when agents are willing to cooperate, cooperation still requires some shared expectation about how action is to proceed, how risks are to be allocated, and how benefits are to be distributed; and these, of course, are often themselves the subject of disagreement. Precisely for this reason, neither ad hoc individual judgement nor occasional concession is enough to sustain stable cooperation. What is needed is some public rule capable of coordinating expectations and adjudicating disputes, so that interaction can move from instability toward enduring social order.

The predicament revealed by the circumstances of justice is thus an unavoidable practical problem internal to relations of interdependence and cooperation, and with it an equally unavoidable structural need for justice. By “structural need” I mean the need for social rules that arises whenever agents stand in relations of interaction marked by both conflict conditions and coordination conditions. The “structural” part matters because the need does not depend on the particular content of the resources at issue, but on the structure of the interaction itself. In different contexts, people may, of course, have different first-order needs — for land, water, livestock, data, supply chains, or carbon-emission space. But wherever conflict conditions and coordination conditions obtain together, the structure of interaction itself places pressure on agents and generates a need for justice. That is why structural need is counterfactually robust (Queloz, Reference Queloz2021a, p. 92). And the “need” part matters because what is at issue here is not a contingent desire, preference, or subjective choice, but an inescapable practical problem. As Queloz puts it, “[i]t is only when the needful concept is not available that its lack becomes a salient feature of experience, and the need becomes a felt need” (Queloz, Reference Queloz2025, p. 215). In the circumstances of justice, if no effective principle of justice is available to guide interaction, human cooperation will become seriously difficult and may even break down altogether. Meeting structural need is thus a functional condition of possibility for human cooperation.

This is why principles of justice are so closely related to practice: They occupy the functional role of responding to this structural need. By constraining self-interest, stabilizing expectations of cooperation, and adjudicating competing claims, they make otherwise unstable relations of cooperation durable and predictable, and thereby provide a normative framework for common life. Chad Van Schoelandt argues that this requires a principle of justice to possess several features: (1) it must take a general form and be capable of ranking outcomes; (2) it must be usable in actual reasoning, and therefore cannot be excessively complex; (3) it must be able reliably to coordinate expectations, providing groups characterized by both common interests and conflict with a publicly recognizable basis for settlement; and (4) it must be capable, over time, of giving rise to shared understandings and interpretations among those who rely on it (Van Schoelandt, Reference Van Schoelandt2020, pp. 420–422). Only if these conditions are met can members of a group tend to arrive at roughly similar conclusions in applying principles of justice, and only then can justice successfully perform its function.

From the functionalist perspective, then, we should distinguish between two senses of “practice.” The first is the structure of the relations of interaction that generate the problem of justice in the first place — namely, the conjunction of conflict conditions and coordination conditions. The second is the functional mechanism through which justice is realized and structural need is met; such a mechanism may take the form of formal social rules, but it may also consist in informal conventions or norms. Functionalism, as I understand it, takes principles of justice to depend on the former rather than the latter. This helps explain two things. First, the function of justice is multiply realizable: The same structural need can be met through a variety of different functional mechanisms. Democracy, monarchy, and aristocracy, for instance, may all be understood as different ways of responding to the cooperative needs that arise at the level of the state. Functional mechanisms are historical, contextual, and variable; but the underlying structural need remains. Second, principles of justice can apply even where no determinate institutional framework is yet in place. On this point, Miriam Ronzoni offers an important revision of practice-dependence. She argues that the dependence of justice on practice does not imply that principles of justice become operative only once institutional arrangements already exist. On the contrary, when existing forms of interaction progressively erode the fair background conditions of cooperation, and when those conditions cannot be sustained through individual conduct alone, justice itself may require the creation of new institutions capable of restoring and securing background justice (Ronzoni, Reference Ronzoni2009, p. 241). Functionalism, then, need not be tied to existing institutional forms. It can criticize, reform, abolish, and indeed justify the creation of new ones.

What all this shows is that the close relation between practice and justice is grounded in the pressure that problems of cooperation exert within human relations of interaction: Such problems generate a structural need that calls for some social rule capable of answering to it. Without such a rule, the cooperative relation itself becomes difficult, perhaps impossible, to sustain. Principles of justice occupy precisely this functional role. In the next section, I argue that it is because principles of justice perform this function that practice can go on to affect not merely their application, but also their mode of justification.

4. How Function Affects the Justification of Justice

Justice is not an abstract norm external to practice, but a social rule internal to relations of cooperation. Still, that by itself is not enough to vindicate practice-dependence. It leaves unanswered the most important question: How can practice enter the justification of principles of justice? Why should a functional description of the problems of cooperation do more than explain the social role of justice? Why should it also constrain the content of principles of justice? In what follows, I distinguish between two ways of understanding functionalism, and argue that only the second can help us answer these questions.

One initially tempting view is to understand functionalism as a first-order theory of the justification of principles of justice. On this understanding, structural needs themselves are the source of the normativity of principles of justice. Whether a social rule can adequately respond to the relevant structural needs would then directly determine whether it qualifies as a principle of justice. Some functionalists have indeed argued in something like this way. David Copp, for example, maintains that the justification of moral principles depends on whether they answer to social needs (Copp, Reference Copp1995, Chapter 10). And Carlo Burelli argues that the functional needs of politics can serve as a source of political normativity, providing a standard for assessing whether political institutions are adequate (Burelli, Reference Burelli2022, pp. 639–640). If that line is simply imported into a theory of justice, it may seem to follow that principles of justice are justified merely in virtue of their capacity to respond to the relevant structural needs.

But this first-order version of functionalism is, I think, untenable. First, functional success is not the same thing as moral justification. A rule may help sustain order or reduce conflict, but it does not follow from that alone that it is normatively desirable. Otherwise, any efficient but oppressive institutional arrangement could count as justified, so long as it were capable of sustaining stable cooperation (Leader Maynard, Reference Leader Maynard2025, p. 21). Second, first-order functionalism gives rise to what Queloz calls an “alienation effect” (Queloz, Reference Queloz2021b, p. 1371). Principles of justice can perform their function stably only if participants subjectively regard them as norms worth following, rather than merely as expedient devices for bringing about some external effect. If agents accept a principle of justice only because it helps secure some cooperative surplus, that acceptance is liable to remain at the level of prudential calculation. And once individuals expect that their own departures from the principle will make no significant difference to the stability of the overall order, they will have reason to free ride. This means that a motivational structure in which agents comply with principles of justice on purely functional grounds will itself tend to undermine those principles’ capacity to sustain stable cooperation.

A more plausible view is to understand functionalism as a second-order meta-normative theory. On this understanding, the question is no longer “which principles are themselves just?” but rather “what kinds of moral reasons are suitable for justifying principles of justice?” This shift is necessary because principles of justice are not just any moral principles. They are normative standards of a special kind, occupying a distinctive social role. They count as principles of justice only because they are meant to serve, within conditions of social interdependence and persistent disagreement, as terms of cooperation, standards of distribution, and bases for adjudication. For that very reason, justifying a principle of justice is not merely a matter of showing that it has moral appeal in the abstract. It is also a matter of showing why it is fit to serve as a public normative standard of that kind. If a given reason cannot explain how the relevant principle could occupy that role, then it may still be a good reason in some other context, but it is not enough to justify that principle as a principle of justice.

At this point, it is helpful to think of the reasons functionalism offers as second-order reasons. A second-order reason is not a reason directly supporting some substantive moral principle; rather, it is, as it were, a reason about reasons, specifying which first-order reasons ought to be admitted and which ought to be excluded. Joseph Raz makes the structure clear with a familiar example: Colin promises that, in decisions concerning his son’s education, he will act solely on the basis of his son’s interests and will exclude other irrelevant reasons (Raz, Reference Raz1999, p. 39). The promise itself neither supports nor opposes any specific educational option. What it does is require him to exclude self-interested considerations and to treat only his son’s well-being as a relevant reason in that context. Scott Keeling sharpens the point by explaining that “the SOR [second-order reason] causes a … certain factor to have influence that it wouldn’t otherwise have” (Keeling, Reference Keeling2024, p. 806). On the basis of second-order reasons, an agent can treat some first-order reasons as trustworthy, or as reasons of the appropriate kind for the context at hand (Keeling, Reference Keeling2024, p. 800). Functionalism, I suggest, should be understood in exactly this way: as a meta-normative theory about second-order reasons. It does not add yet another substantive reason competing with first-order moral reasons. Rather, it explains which moral reasons may appropriately enter into the justification of principles of justice, and which should be excluded.

More specifically, functionalism can play two roles in the justification of principles of justice. The first is negative. Functionalism can exclude higher-order moral principles that are ill-suited to serve as the justificatory basis of principles of justice. Ronzoni’s criticism of G. A. Cohen’s camping trip example can be read in precisely this way. Ronzoni argues that the appeal of Cohen’s principles of community and radical equality of opportunity depends on conditions such as a high degree of agreement among participants, low exit costs, and a tightly limited scope of cooperation. Social cooperation, by contrast, is long-term, not easily exited, and must accommodate a plurality of competing life plans (Ronzoni, Reference Ronzoni2012, pp. 178–181). The difference in practical structure means that the problems the two contexts pose are not the same. And once the conditions of the cooperative problem change, the principles in question may no longer be capable of responding adequately to the structural needs of social cooperation. So even if a higher-order moral principle is valid under some particular conditions, it may still lose its standing as a basis for justifying principles of justice because it is simply not fit for the social level.

The second role is positive. Functionalism can supply threshold constraints for selecting among candidate principles. True, structural needs do not by themselves yield any single determinate principle. But they can help show what minimal conditions a principle must satisfy if it is to play a justificatory role within a certain form of cooperation. Consider John Rawls’s insistence that political justification must satisfy a requirement of publicity. Some theorists take this to smuggle in a substantive commitment to the value of equal respect. But it can also be read as a functional requirement (Larmore, Reference Larmore2008, pp. 146–150). If a principle of justice is to function as a shared assurance for group cooperation, it must be capable of generating stable mutual expectations, so that each agent can reasonably believe that others, too, will act in accordance with it. If a principle of justice depends on the comprehensive doctrine of some particular group, then those outside that group will lack sufficient motivation to accept it. In that case, even if the principle has considerable normative appeal, it will be unable to do the work required in a pluralistic society. So the reasons supporting principles of justice must possess a public character: They must be recognizable as relevant reasons by agents with different backgrounds of belief. It is for this reason that Rawls excludes highly controversial comprehensive doctrines and instead seeks the higher-order moral resources for justification within the public political culture itself (Rawls, Reference Rawls1996, pp. 13–14). Functionalism cannot replace the argument for higher-order moral principles; but it can provide a comparative standard enabling us to explain why some such principles are better suited than others to enter into the justification of principles of social justice.

Once functionalism is understood in this way, it can respond to the practice-independent objection. Practice-independent theorists argue that any argument supporting or constraining substantive moral principles must ultimately appeal to independent moral values; the functional role of practice, being a merely descriptive fact, cannot do that justificatory work (Meckled-Garcia, Reference Meckled-Garcia2013, p. 108). From that perspective, it may seem that the second-order reasons invoked by functionalism must themselves rest on some deeper moral value, since otherwise it would be unclear how they could affect justification at all. But this criticism presupposes that moral value is the only dimension relevant to justification, and, in doing so, it overlooks the dimension of relevance. The functional analysis offered in the previous section was not meant to identify the source of the normativity of justice. Its aim was, rather, to identify the problem-structure to which principles of justice are meant to respond. Principles of justice address a special kind of cooperative relation, one marked by the conjunction of conflict conditions and coordination conditions, and one that presents agents with an unavoidable practical problem. Once that is so, the reasons that may enter into the justification of principles of justice must themselves be sensitive to the structure of that problem. In that sense, functional analysis supplies a relevance constraint. It enables us to distinguish principles of justice from moral ideals more generally. The latter may be morally attractive, but if they cannot answer the problem at hand, they do not constitute appropriate reasons for justifying principles of justice.

So functionalism neither infers substantive values from the point of a practice nor derives normative content directly from facts. Its claim is second-order. Only those moral reasons that can adequately respond to the problem of cooperation are appropriate reasons for justifying principles of justice. The way in which practice affects justification, then, is by allowing structural need to constrain which higher-order moral principles are eligible to do justificatory work. Functionalism thus preserves the role of moral principles in justification while insisting that principles of justice must remain sensitive to practice.

5. The Priority of Politics: The Methodological Significance of Functionalism

The argument from functionalism shows that, although the structural needs internal to practice do not constitute a special source of normativity, they can nevertheless enter into the justification of principles of justice by imposing a relevance constraint. With that, we have answered what is arguably the most serious challenge facing practice-dependence. I now want to go one step further and explore the broader methodological significance of this conclusion: why functionalism supports the priority of politics over morality, and how, once so understood, it helps us locate the methodological standing of practice-dependence itself.

According to a typical ethics-first version of practice-independence, morality has justificatory priority over politics. That is, any legitimate political norm must ultimately be grounded in moral value. Practice-independent theorists often support this claim by way of a regress-style argument: If we keep asking after the normative source of a principle of justice, we will eventually arrive at the end of the chain of reasons, and there discover some foundational reason. Since there is a gap between “is” and “ought,” facts cannot be part of such a foundational reason. Principles of justice must therefore ultimately depend on some purely normative proposition expressing a certain moral value (Erman & Möller, Reference Erman and Möller2025, pp. 33–35; Nye, Reference Nye2015, pp. 1087–1088). The crucial point is that this moral value is external to practice: Its normative standing is not itself constrained by practice. And that, in turn, means that moral value always enjoys justificatory priority over political norms.

But this line of argument does not touch the functionalist claim, because showing that moral value is prior to principles of justice is not the same as showing that morality is prior to politics. Functionalism defuses the tension between politics and morality by dividing the justification of principles of justice into a two-level structure. At the first level, one identifies agents’ structural needs on the basis of the conflict conditions and coordination conditions of a cooperative relation, and thereby filters the higher-order moral principles that are appropriate to the case. At the second level, one derives a principle of justice suitable for the practice from the content of those higher-order moral principles. The first level thus consists in a political relevance constraint; only the second consists in moral justification. So even if a principle of justice does in fact depend for its justification on some higher-order moral principle, it does not follow that morality is prior to politics. On the contrary, the structure of political practice can impose a prior constraint on which higher-order moral principles are eligible to do justificatory work at all. In that sense, politics retains priority over moral value throughout the justificatory process.

More than that, the functionalist argument helps expose a characteristic strategy employed by practice-independent theorists. Practice-independence often distinguishes between normative principles, which answer the question of what justice is, and rules of regulation, which are meant to solve particular social problems (Cohen, Reference Cohen2008, p. 253; Estlund, Reference Estlund2020, pp. 9–12). Only the former, they claim, are properly principles of justice; the latter are merely managerial devices produced by applying principles of justice to concrete societies. On this view, whatever demands practice-dependent theorists make are really demands on rules of regulation rather than on principles of justice themselves. As David Estlund puts it, “Realism of that kind is an obvious constraint on practical proposals, but … an illegitimate constraint on what principles of justice might require” (Estlund, Reference Estlund2020, p. 21). If that is right, then the genuinely justice-constituting normative principles remain untouched by practice, and independent of all such facts.

Seen from that angle, however, the disagreement between practice-dependence and practice-independence threatens to collapse into a merely verbal dispute over whether principles of justice are really normative principles or rules of regulation, and so to lose its substantive normative significance altogether. The deeper problem, from the functionalist point of view, is that the practice-independent distinction severs the connection between normative principles and rules of regulation much too sharply. True, normative principles cannot be simply reduced to rules of regulation. But neither are they wholly external to them. The structural needs to which rules of regulation are meant to respond help determine what practical problem must be addressed, and thereby also what kinds of normative principles are relevant in the first place. Practice-independent theorists cannot therefore avoid every criticism of utopian theories of justice simply by invoking the distinction between principles and rules.

I want, finally, to say something about the relation between functionalism and political realism. Many practice-dependent theorists take practice-independence to be one form of political moralism (Rossi, Reference Rossi2012, p. 152; Sangiovanni, Reference Sangiovanni2008, pp. 157–158). But the connection between practice-dependence and political realism is far less clear. Some regard practice-dependence as belonging to the realist family (Rossi, Reference Rossi2012, pp. 158–162, Reference Rossi2019, pp. 640–641; Sangiovanni, Reference Sangiovanni2008, pp. 156–159); others insist that, as an heir to Rawlsian political theory, it remains importantly distinct from realism (Estlund, Reference Estlund2020, pp. 56–57; Jubb, Reference Jubb2016, pp. 94–95). So how should we understand the methodological standing of practice-dependence?

Here Jonathan Leader Maynard is helpful. On his account, political realism should be understood not as a metaethical thesis about political normativity, but as a methodological orientation. Its defining feature is that it places “empirically constitutive political realities” at the foundation of political theory, so that real political conditions themselves bear on the formation of normative theory (Leader Maynard, Reference Leader Maynard2022, p. 454). That description fits the ambitions of practice-dependence remarkably well. Functionalist practice-dependence should therefore be understood as belonging to the realist family — as a form of moderate realism.

Within that realist family, functionalism is closest to what is sometimes called the “filter view.” On this view, the political domain has distinctive constitutive features, and those features provide the context within which political values must be understood. If a value is to count as a political value, it must pass through the political filter: It must be consistent with the constitutive features of politics. Moral values cannot simply bypass those features and directly govern political activity (Sleat, Reference Sleat2016, p. 253). Functionalism accepts this general picture, but extends it in two further respects.

First, functionalism offers a more general and more open-ended account of political features, one capable of accommodating a wider range of political phenomena. Existing versions of the filter view often characterize politics as an attempt to secure order under conditions of disagreement by means of authority and legitimate coercion. On that basis, they describe the constitutive features of politics in terms of disagreement, authority, and legitimate coercion (Sleat, Reference Sleat2016, pp. 255–258). But that description itself presupposes a state-centred picture of politics focused on authoritative rule, and, in doing so, it overlooks other political possibilities. Paul Raekstad, drawing on a wide range of anthropological material, argues that stateless human societies nonetheless exhibit politics in the full sense: They display social organization, coordination, and implementation of collective action; they make use of some degree of coercion; and they engage in judgements and practices concerning the legitimate exercise of power (Raekstad, Reference Raekstad2018, pp. 154–158). Philip Pettit’s genealogical reconstruction of the state’s function points in the same direction: Responses to problems of cooperation do not begin with the state, but develop continuously through conventions, norms, law, and only then the state (Pettit, Reference Pettit2023, Chapter 1). What should therefore count as constitutive of the political is not authoritative rule under state conditions as such, but the more basic structure of the problem of cooperation. That structure is the empirical fact to which political theory ought to attend, and it is what ultimately shapes the route by which moral value enters the political domain.

Second, functionalism gives us a more precise account of the filtering process itself. Matt Sleat argues that political filtering consists not only in the compatibility of a political value with political features, but also in whether, under our political and social conditions, it can count as “a value for us” (Sleat, Reference Sleat2016, p. 260). That is, the standing of a political value depends on the presence of the right social preconditions. Functionalism’s account of structural need helps explain this more clearly. What counts as a value “for us” is not determined simply by cultural outlook or historical trajectory. It is determined, rather, by the cooperative problem agents face under specific social conditions, and by the structural needs that arise from it. Consider, for instance, the difference between ancient Rome and late eighteenth-century France. Roman citizens enjoyed a comparatively high degree of agreement regarding political identity and the ends of the community; and the maintenance of the polity depended heavily on citizens’ continuing participation in public affairs, so coordination relied on virtue, honour, and direct identification with the commonwealth. By contrast, late eighteenth-century French society existed under radically different conditions: sharper conflicts of interest, class opposition, and value pluralism, together with a state whose scale and complexity far exceeded those of ancient Rome. Both societies faced problems of cooperation, of course, but the structural needs were dramatically different. And that means that what counts as a value “for us” must itself be understood as a political value capable of responding to the cooperative problem under those here-and-now conditions.

The reason functionalism rejects the strong version of realism is that it does not take the function of principles of justice to yield a substantive normative standard by itself. In that respect, functionalist practice-dependence differs from Burelli’s political functionalism. Burelli argues that a functional analysis of politics can directly ground a non-moral normative judgement: Any political institution must be capable of making binding collective decisions, and where it cannot, it is a failed institution (Burelli, Reference Burelli2022, pp. 639–640). But the difficulty is that this standard sets the bar for political normativity much too low. It functions at most as a baseline constraint. Above that baseline, political functionalism has little more to say. Functionalist practice-dependence avoids this problem by dividing the labour between structural need and higher-order moral principles. Thus, for example, a functionalist may argue, on the basis of the structure of cooperation in modern society, that in large-scale societies where kinship institutions have weakened, a universalist norm centred on equal status is better suited to meeting people’s cooperative needs (Blunden & Lane, Reference Blunden and Lane2025, pp. 257–259). But what substantive value is to serve as the core of that universalist norm remains a matter for the substantive moral principles available in the particular context.

To sum up, the functionalist version of practice-dependence holds that social practice enters into the justification of principles of justice not because practice itself supplies independent moral content, but because the structural needs internal to cooperative relations constrain which higher-order moral principles can count as appropriate reasons for justifying principles of justice. In that way, functionalism overcomes the justificatory weakness of current versions of practice-dependence. And the broader methodological lesson is that functionalism can answer the practice-independent objection to the priority of politics, thereby laying the groundwork for a moderate realist theory of justice.