i. Operative and Foundational Concepts

One might worry that arguments in favor of the adoption of LP by legal officials would rely on the premise that it is up to individuals to decide what to believe, or that the truth of propositions might be determined by the consequences generated by belief in them.Footnote ⁹

This is a plausible concern, but it does not affect the normative argument for LP made here. To show why, let me introduce a distinction that, though not formulated as such in legal theory so far, seems both intuitive and plausible: the distinction between foundational and operative concepts. In general, I will understand a ‘concept’ as a structured psychological entity (a mental representation) that agents employ in thought and talk.Footnote ¹⁰ By an ‘operative concept’, I understand the concept that governs representational, discursive, and inferential practices regarding an object within a specific domain.Footnote ¹¹ A ‘foundational concept’, on the other hand, is the concept that fully captures the true nature or the constitutive features of the object to which it refers.

Consider the concept woman. The foundational concept woman, if there is one, would depend on the constitutive features of women. Let me assume for the sake of argument that what is constitutive of being a woman is a cluster of features that specific tokens of the kind satisfy to some extent and that are related to each other by a form of family resemblance. ^{Footnote 12} This is, under the distinction offered above, the foundational concept of woman. At the same time, the operative concept of woman in the practice of American physicians in the 1950s might have been ‘biologically female human being.’Footnote ¹³

Questions about foundational concepts are questions about fit between the mental representation and the nature of the object or its constitutive features. The correct foundational concept will be whatever mental representation correctly captures the relevant features of the object.

Descriptive questions about operative concepts, on the other hand, are questions about social fact, such as what was the concept of woman in American medicine in 1950? This is not to say that foundational concepts are entirely independent of social facts. Plausibly, the foundational concepts of social kinds, as is the case of law, are at least partly sensitive to social facts because social kinds are grounded, at least in part, in contingent human conceptual activities.Footnote ¹⁴

The questions we might ask about operative concepts are not exclusively descriptive. Normative questions are also possible. For instance, we could ask what the best operative concept of an object would be (given certain normative considerations and constraints) in a specific domain. Such a question relies on the idea that, if things would go better (under some specification of ‘better’) if some agent adopted operative concept X in a certain context, that agent should adopt operative concept X in that context. These are questions of conceptual ethics or conceptual prescription.Footnote ¹⁵

Now some might think that this type of conceptual ethics question does not matter much. Our inferential, linguistic, and conceptual practices—our operative concept, in my terms—should simply track the true foundational concept. This natural implication, however, neglects the existence of situations where even those with equal epistemic authority disagree about the content or extension of the foundational concept. In these circumstances of disagreement between epistemic peers, the terrain is fertile for conceptual prescription, particularly if the relevant concept has practical and moral significance. This is precisely the case of the concept of law: there is intractable expert disagreement about the foundational concept, and yet the concept matters pragmatically.Footnote ¹⁶ Our main jurisprudential theories seem to equally fit convergent views and intuitions about law in specific cases. As the history of jurisprudence shows, these convergent intuitions are not sufficient to adjudicate jurisprudential disputes. The disagreement between the theories lies in the characterization of law as a general practice or institution. At that level, the debate is at a stalemate.Footnote ¹⁷ At the same time, how legal officials understand law is a matter of practical significance that can have important effects on social life.Footnote ¹⁸

Note that the fact that operative concept X is or ought to be adopted in a specific domain and for a specific purpose does not settle questions about the foundational concept. Thus, conceptual prescription about operative concepts has no direct impact on foundational concepts.Footnote ¹⁹ From this perspective, the normative argument for LP is an argument about the operative concept of law that ought to be adopted by legal officials in the context of adjudication—whatever the right view about the foundational concept of law might be, which is an independent question.

ii. Jurisprudence, Wishful Thinking, and Conceptual Prescription

Still, the project might seem somewhat dubious. Surely, we want to know what law is and what it should be—the traditional concerns of analytical and normative jurisprudence. It’s not obvious why what certain agents should think law is should matter. The task of analytical jurisprudence, for instance, might be thought to require adequately explaining the essential properties of law.Footnote ²⁰ What’s politically valuable has no impact on that question.Footnote ²¹ Answering questions about the concept of law in terms of what would result in good moral consequences seems, again, to involve a form of wishful thinking.Footnote ²²

One tempting but ultimately inadequate answer to this concern resorts to law’s ontology. According to this answer, law is a social kind, shaped by what those who participate in the practice understand it to be.Footnote ²³ Thus, it is open for people to change—in light of normative considerations—their concept of law.Footnote ²⁴ The problem with this potential response is that, even if law is a social kind, just changing the way we think about law will not change what law is.Footnote ²⁵ It would be good if X, therefore X does not hold—even when X involves a social kind.

Thus, the concern about wishful thinking persists. In Donelson’s reconstruction, instances of wishful thinking rely on the following type of premise: “[I]f p has pragmatic property x, p has the relevant epistemic property y, where x does not imply y.”Footnote ²⁶ Such a premise is obviously false. The argument for LP would fail if it relied on this premise.

However, the argument for LP does not involve this type of wishful thinking. Instead, the argument relies on the premise that: If the adoption of p by X would be good, p ought to be adopted by X, where p stands for an operative concept.Footnote ²⁷ Such an argument is valid, and it does not rely on any evidently false premises. It is not an instance of “wishful thinking.”Footnote ²⁸ This type of argument for LP would not be a description of our concept of law, but a prescription in favor of LP’s adoption by specific agents (legal officials) in a certain domain (adjudication).Footnote ²⁹ The argument is not an attempt to resolve the question about the (true) foundational concept. Instead, it is an evaluation of which operative concept that tracks our basic shared understandings is morally best,Footnote ³⁰ in terms of the impact it would have if adopted by legal officials.Footnote ³¹

Against what the wishful thinking critique holds, then, the conceptual prescription in favor of LP does not assume that the moral consequences of different concepts determine which concept of law is true.

iii. Do Legal Officials Need an Operative Concept?

Now, the argument so far has assumed that the operative concept of law indeed makes a difference. However, in their routine work, legal officials are not concerned with the nature of law, the relationship between law and morality, or the concept of law. In most cases, legal reasoning is perfectly feasible without articulated views about these issues.Footnote ³²

Still, the operation of legal reasoning is impacted by the concept of law. The concept of law can be a central part of the image of law—the complex set of beliefs regarding what legal discourse and legal reasoning are about—that determines how legal enactments ought to be understood and applied in a specific legal system.Footnote ³³ An ‘image of law’ is what, I believe, comparative lawyers have in mind when they discuss the shared philosophical and cultural commitments that underlie the surface of legal systems and are necessary to fully understand them.Footnote ³⁴ As these comparative lawyers argue, understanding legal systems requires grasping the underlying ideas that animate and differentiate legal systems,Footnote ³⁵ including their concept of law.Footnote ³⁶ The operation of legal systems, as well as the differences and similarities between them, depend on the ‘cognitive structures’ underlying those legal systems.Footnote ³⁷ These include their conceptions of law, as Atiyah and Summers note.Footnote ³⁸ In fact, even legal theorists’ conceptions are impacted by the operative concepts at work in their jurisdictions. American legal realists and Dworkin offer accounts of law that are commonsensical for the American lawyer, but completely bizarre for their German counterpart—for whom Kelsen’s theories seem much more plausible.Footnote ³⁹

The operative concept of law, then, has an impact on the operation of legal institutions. The relevant effects of any operative concept of law are produced because of its adoption by legal officials, which impacts how they go about their law-applying tasks. So, in making my argument for LP, I do not assume that legal officials adopt an articulate, explicit concept of law. Instead, I make a much weaker assumption: legal participants have, at least, an inchoate picture of what law and legal facts are, and that picture (that ‘image of law’) is reflective—albeit imperfectly—of a specific concept of law that, at least in part, explains differences between legal systems.

This explains why Liam Murphy’s pessimism about instrumental arguments about concepts of law does not affect the argument for LP, as construed here.Footnote ⁴⁰ While Murphy is right that the concept of law is part of ordinary discourse, and therefore convergence in usage as a response to normative argument is not a realistic expectation,Footnote ⁴¹ the issue is different regarding the operative concept adopted by legal officials.Footnote ⁴² It is not beyond the imaginable that theoretical arguments made by legal scholars and legal theorists can impact the formation and intellectual norms of the professional bar and of judges. The change will certainly be slow, and the causal question about how conceptual change happens is complicated. But legal officials will inevitably have at least an implicit concept of law as they carry out their activities. This concept can be impacted by legal theorists’ and legal scholars’ arguments about it—and will itself impact legal officials’ activities.

i. Moral Disagreement and Legal Validity

Contemporary social life is characterized by the fact of pluralism. People hold inconsistent views about the meaning of life, human excellence, the beginning and the end of human life, etc. This is something the liberal tradition has long been aware of.Footnote ⁴⁴ Yet people are also divided about how to live together. We do not just disagree about the good life; we also disagree about what makes a polity just or even morally acceptable.Footnote ⁴⁵

Despite these disagreements, we still need to get on with the business of personal and social life. We still need, as individuals, to execute contracts, transfer property, and get married and divorced. We also need, as citizens, to figure out how to secure, allocate, and control the expenditure of public resources, how to organize socioeconomic structures, etc. From this perspective, LP has an advantage: it does not replicate our fundamental moral and political disagreements in the context of adjudication, and instead facilitates what Alexander and Sherwin call “authoritative settlement.”Footnote ⁴⁶

Rawlsian liberals traditionally argue that a public conception of justice helps us deal with disagreement, while treating citizens with equal concern and respect.Footnote ⁴⁷ The prospects of that project seem dubious today. In practice, even a minimal and consensus-oriented conception of public reason might fail to give people a shared language to articulate their differences.Footnote ⁴⁸ Perfectionist liberals, Catholic conservatives, and radical socialists simply cannot accept the basic assumptions and premises of public reason liberalism. Perhaps this shows that (some of) these views are morally problematic.Footnote ⁴⁹ But we need to coordinate our behavior and resolve our legal disputes with people who hold views we find problematic. In solving this problem, a relatively amoralistic conception of legal facts serves legal officials well as a modus vivendi.

Moral truth—even political, not metaphysical, moral truth—never appears “in propria persona, but only … in the form of somebody’s controversial [view].”Footnote ⁵⁰ Moreover, those controversial views are not just views we disagree with; they are views that we might find abhorrent, misogynistic, racist, totalitarian, morally bankrupt, etc., and that celebrate precisely the political decisions that we despise.Footnote ⁵¹ The concern, when faced with these views, is not that we will not find shared public commitments to freedom and equality, but the prospect of complete disintegration of social life. That prospect makes it urgent to channel disagreement. The requirement is more urgent than—and, in some sense, prior to—finding a common conception of justice over which those holding different views of the good life can overlap. The most urgent task is to avoid a life that could be “nasty, brutish, and short.”Footnote ⁵² This is, as Williams wrote, “the first political question.”Footnote ⁵³

For that purpose, an operative concept of law that sees the determination of legal facts as substantively connected to morality is poorly suited. The view clashes with the diversity of moral opinions and convictions, and with the need to secure a minimal degree of settlement of everyday disputes without replicating more abstract disagreements—precisely, one of the central purposes of legal adjudication.Footnote ⁵⁴

This suggests that a Hobbesian approach—an approach that sees the legal system as autonomous, with its own criteria of justice—might be a good fit for legal officials.Footnote ⁵⁵ That approach highlights the distinction between judgments about the moral standing of a rule and its legal validity, and attempts to prevent discoordination and unilateralism in contexts of disagreement.Footnote ⁵⁶ While we ideally want law to be just, wise, efficient, etc., we first need it to act as a framework for decision-making that allows for ordered, coordinated social interaction.Footnote ⁵⁷ For this purpose, the identification of legal facts ought to isolate, rather than replicate, our underlying disagreements.Footnote ⁵⁸ Law is morally valuable not only when its content is meritorious, but also, and primarily, because of its ability to solve coordination problems independently of our moral judgments about its content.Footnote ⁵⁹ The moral reasons for positive law and its ability to coordinate behavior are, at the same time, reasons for law whose content can go against (any specific legal official’s view of) morality.Footnote ⁶⁰

As Mark Greenberg has noted, even if law aims to settle disagreement, it also aims at other important goals, such as ensuring that state coercion is consistent with past political decisions, serving the public good, and improving our moral situation.Footnote ⁶¹ The crucial point here, however, is that unresolved moral disagreement can be an obstacle to achieving justice, to using coercion in justifiable ways, and to using law as an instrument of valuable ends. In this sense, legal adjudication’s ability to settle disagreement is prior to law’s achievement of other valuable things.Footnote ⁶² We cannot reach valuable social outcomes through law if those in charge of applying it do not agree about what it requires.

One potential response to this argument is that LP would not prevent interpreters from importing their moral convictions into law. First, when interpreters face hard cases, both the nonpositivist and the positivist can accept that the legal interpreter ought to resort to moral principles—the only difference between them is whether they would say that those moral norms are part of the law.Footnote ⁶³ Moreover, given the presence of moral language and moral concepts in law, even the routine application of legal standards might require moral evaluation. LP would only change the description of what’s going on.

There are a couple of potential responses to this counterargument. First, we should reject the notion that the adoption of LP would merely change how legal officials see what they are doing. Courts are law-applying institutions, and—for better or for worse—the idea that they cannot legitimately engage as a general matter in ‘legislating from the bench,’ but should rather apply the law, is an established notion in contemporary democratic regimes. There are strawman versions of this idea: the declaratory theory of the common law, the brooding omnipresence in the sky, etc.Footnote ⁶⁴ Yet a more nuanced version of this idea seems quite intuitive both for laypeople and legal insiders: even if judges can engage in development of the law and innovate at the margin, they cannot assume a general legislative authority in the context of adjudication.Footnote ⁶⁵ Thus, assuming most legal officials see adjudication as a venue that, generally, ought to be tasked with the application of law, then the adoption of LP will have an impact. A conscientious judge that adopted the positivist view, for instance, would be motivated to decide hard cases, as much as possible, on the basis of legal reasons (understood in terms of LP).

Second, even if LP only changed the terms of description, it would have the virtue of clearly differentiating between situations where legal officials are arguing or resolving disputes by applying pre-existing authoritative sources of law and situations where they do so by resorting to moral considerations. This might become an important source of political accountability regarding judicial lawmaking, and it helps us deal with moral disagreement directly as what it is, rather than as a rarified form of legal argument.

Still, it is certainly true that moral language is present in many legal norms. LP is compatible with the law requiring judges to engage in moral reasoning. It is only incompatible with, first, the claim that judges need to engage in moral reasoning in order to determine the law—that law is necessarily dependent on moral facts, no matter what the relevant legal sources demand—and, second, the notion that when the law requires judges to engage in moral reasoning it is incorporating moral facts or morality as such into the grounds of law. The presence of moral language and moral reasoning in law is not the same as the incorporation of morality into law.Footnote ⁶⁶ And it is certainly not evidence of the necessary incorporation of morality into law, as NP claims. In these circumstances, LP is compatible with legal sources directing interpreters to engage in moral reasoning. Whether that is the case depends, under LP, on positive law.

In hard cases, some might fear that LP threatens to become a ‘formalist’ view.Footnote ⁶⁷ Sometimes, the counterclaim goes, the law just runs out, and we should be clear-eyed about this instead of hiding moral choices behind a supposedly legal basis. This sounds right, but as I have argued, LP might actually generate clarity in this regard. In any case, we should question how frequently law runs out and how commonly hard cases are cases where there is no law—as opposed to cases where the interpreter does not like the conventionally derived upshots of legal materials.Footnote ⁶⁸ The usual jurisprudential focus on constitutional law—as opposed to large areas of private law, civil procedure, and so on—distorts, moreover, our understanding of this phenomenon.Footnote ⁶⁹

Importantly, genuine gaps only exist when there is no law to be applied. Even in cases that seem facially unregulated, as Kelsen explains, positive law can contain general closure rules.Footnote ⁷⁰ A good example is the basic principle in criminal law that every behavior not expressly mandated or prohibited is allowed.Footnote ⁷¹ This closure rule shows that even “when the law is silent” it can provide a solution.Footnote ⁷² Gaps in the communicative content of legal materials do not imply gaps in legal content.Footnote ⁷³ Legally regulated cases need not be expressly regulated cases. There will still be some genuine gaps, with no closure rule to apply. In these circumstances, when the legal constraints run out, LP has the virtue of transparency: it tells us that legal officials are in fact engaging in ordinary moral reasoning.

In some regulated cases, the legally prescribed outcome might be morally objectionable, given the interpreters’ value judgments and preferences. Yet here one must recall that what someone sees as a morally objectionable outcome is, on someone else’s view, a moral achievement. In these circumstances, an impartial, coordinated outcome—the validity of which is derived exclusively from social facts—might be morally preferable, even if substantively objectionable (for a particular legal official).

ii. Majority Rule

In democratic regimes, positive law is an instrument of governance through rules deliberately enacted by elected legislatures.Footnote ⁷⁴ LP and democracy have a complicated relationship.Footnote ⁷⁵ Here, I will make a narrow argument that leaves the question about that relationship relatively open: LP offers a conception of law that is consistent with relatively familiar ideas associated to majority rule.

While there is no necessary connection between LP and democracy, the adoption of LP by legal officials can be a valuable complement to a majoritarian procedure for enacting law.Footnote ⁷⁶ If democracy is valuable, and majority rule is at least an important part of democracy, that LP can contribute to majority rule better than NP is a reason in favor of LP.

Democrats are typically concerned with what happens before the enactment of formal laws: the procedures through which proposals are brought to the public’s attention, discussed, reformed, and approved as law.Footnote ⁷⁷ Yet early positivists were acutely aware of the fragility of authoritative legal governance after enactment. Bentham—though not always an adherent of democracyFootnote ⁷⁸ —was concerned with the power of judges and their use of natural law arguments to trump legislative enactments.Footnote ⁷⁹ His fear was that ‘Judge & Co.’ would undermine the possibility of governance through enacted law.Footnote ⁸⁰ LP is consistent with this concern, and empowers those with formal lawmaking powers at the expense of interpreters.

Precisely who holds the power to enact law is a—perhaps, the—crucial question for majoritarian democracy.Footnote ⁸¹ It was also a crucial question for classical legal positivists like AustinFootnote ⁸² and Bentham.Footnote ⁸³ Because LP sees questions about the concept of law as ultimately connected to historical and social facts, the adoption of LP by legal officials, within a democracy, gives more control to the democratically elected legislature. It gives force to democratic concerns about the source of the rules that govern social life.Footnote ⁸⁴

Vague moral language, when present in legal norms, invites moral deliberation by lawyers, judges, and laypeople, and this can sometimes be valuable in a democratic regime.Footnote ⁸⁵ LP does not reject the possibility that duly enacted legal sources might contain moral language and require moral reasoning. The crucial difference here between NP and LP is that LP always allows agents with formal lawmaking powers to take intentional, direct action to redirect practical deliberation. Under LP, whether moral deliberation by legal officials is good or not is itself something that ends up being decided, in a democracy, by the citizens’ representatives and their enactment of open-ended standards. LP is committed to a concept of law that is exclusively connected to social facts—it does not reject the possibility that moral language and moral reasoning might be part of the relevant legal sources, validated by the relevant social facts.Footnote ⁸⁶ NP, in contrast, makes it always potentially legitimate for legal interpreters to engage in moral reasoning, whatever the content of the relevant legal sources is.

The implementation of majority rule requires those who interpret democratically enacted sources to possess self-restraint. The interpreters ought to resist the tendency to take legal interpretation as an exercise in moral and political reflection.Footnote ⁸⁷ LP encourages that self-restraint, and therefore encourages a form of fidelity to majority rule. Respect to majoritarian decision-making is, in this argument, reinterpreted as respect within legal interpretation to sources enacted through a majoritarian process.Footnote ⁸⁸

Many theorists argue that democracy requires much more than majority rule and includes respect for substantive political rights and values.Footnote ⁸⁹ However, even those who hold such views would accept that majority rule in the generality of cases is at least one ingredient of a democratic regime. From that perspective, my argument certainly has less bite for this type of democrat than for a scrupulous majoritarian: it is just an argument about how NP undermines one of the ingredients of a democratic regime. Still, that is sufficient for my purpose, which is not showing that democracy requires LP, but that LP has one thing going for it: it is more consistent than NP with majority rule, and that is one reason that weighs in favor of LP.

iii. Injustice and Constraint

Law can be unjust—as well as inefficient, oppressive, and so on. LP is a reminder of the fact that legal validity is entirely compatible with injustice.Footnote ⁹⁰ There is a certain symmetry in the positivist view about law’s ability to coordinate behavior by avoiding moral conflict and its cold assessment that laws and legal officials’ law-applying activities should not be presumed to be morally binding just because they are legally valid.Footnote ⁹¹

For positivists, whether law ought to be obeyed is always an open moral question.Footnote ⁹² The adoption of LP is thus a source of modesty for legal officials. It leads them to see their own activities in a humble light, as something that can be deficient and that can be fixed by lawmakers.Footnote ⁹³ It also leads them to accept that whether the inputs and outputs of their activities are morally binding is an open question.

The relevance of the awareness that law might be defective should not be understated. Consider an example.Footnote ⁹⁴ You have a positivist interpreter and a nonpositivist interpreter. They both face a statute that excludes same-sex couples from marriage. Suppose both interpreters are against this statute on moral grounds, but there are no constitutional provisions that, interpreted according to the relevant interpretive standards of the jurisdiction, would allow judges to resort to those moral concerns in deciding the case. The positivist will probably say that, under the law of the jurisdiction, a same-sex couple cannot marry. The nonpositivist might say, reading constitutional or legal norms morally and applying the principles of political morality that best justify the legal system, that the couple can get married.Footnote ⁹⁵ There is something compelling about the nonpositivist view—it is, after all, telling us that under our legal system people are not excluded from an important civil institution because of their sexual orientation. It is finding the resources to temper the injustice of legal enactments.Footnote ⁹⁶ Yet there is something quite problematic going on as well. The legal official adopting NP is hiding the injustice of positive law. They are portraying positive law as something better than it is.

The nonpositivist would reply that this is a tendentious framing. After all, the nonpositivist interpretation puts itself forward as a claim about what the law on marriage actually is. The positivist is assuming that there is an unproblematic fact about the law that makes its content immoral—yet the nonpositivist is arguing that this ‘fact’ is not a fact at all, but rather the positivist’s mistaken legal interpretation. The nonpositivist judge who is reaching the conclusion that the law is immoral will, because of this circumstance, have a reason to go back and check their interpretation again.Footnote ⁹⁷

Perhaps so. Yet this claim about the process of identifying legal facts is one that empowers interpreters. We better hope—as my example suggests—that we get the interpreters with the right moral views. Given the risk that we might not, a more risk-averse strategy in a democratic regime might be to disempower interpreters. An amoralized view of law also produces a less rosy self-conception by legal officials—one that highlights the way in which legal institutions fall short of the ideal. It shows the victims of injustice exactly where in the positive law the injustices might lie, and exactly what needs to be reformed. It also leads legal officials to adopt a more morally austere conception of their own activities.

Now one dimension along which the nonpositivist concept seems to fare better is that it allows the victims of injustice to achieve a just outcome within legal adjudication. But this is a highly contingent advantage. In a democratic regime, it turns on the relative merits of the moral views adopted by legal interpreters and those of the citizenry or their democratic representatives. Perhaps there are times and places where a nonpositivist approach is not a danger but a safeguard (I return to this below). This is an empirical and historical question, but we have examples of how things can go terribly wrong when jurists see law as indissolubly connected to their moral conceptions.Footnote ⁹⁸ NP can also be supplemented with moral views that dramatically diverge from, say, liberal egalitarian ideas.Footnote ⁹⁹ The empowerment of interpreters to bring their moral views into legal interpretation is neutral vis-à-vis the substantive correctness of those views.

iv. Integrity

In Dworkin’s formulation, integrity requires the state to “act on a single, coherent set of principles even when its citizens are divided about … justice.”Footnote ¹⁰⁰ According to this view, we should strive to see political institutions, as much as possible, as speaking with one voice reflecting a coherent set of principles.Footnote ¹⁰¹

Societies need determinate decisions on controversial issues. Integrity requires attempting to read these multiple decisions as upholding one view that counts as the view of the community.Footnote ¹⁰² That view might not be morally optimal—but there is still value, according to Dworkin, in a certain coherence about the commitments that inform our legal institutions.Footnote ¹⁰³ Law provides a shared framework for deliberation, which can coordinate behavior and constrain state coercionFootnote ¹⁰⁴ under a consistent set of standards.Footnote ¹⁰⁵

We should question whether the adoption of NP by legal officials advances the cause of integrity. There are two concerns here: one of stability and one of legitimacy. The problem of stability is that, under moralized approaches to legal reasoning, the interpreter is asked to become “Hercules.”Footnote ¹⁰⁶ They must come up with the principles that best justify the legal system and test legal materials against those principles. However, in politically diverse and heterogeneous societies people disagree about morality. A legal world packed with Herculean interpreters tends towards moral fragmentation. Every interpreter is authorized to draw their own implications from legal materials.Footnote ¹⁰⁷ The metaphor of Dworkin’s jurisprudential approach as a “protestant” view asking each agent to identify the principles that justify law,Footnote ¹⁰⁸ “severally and independently,”Footnote ¹⁰⁹ is in this aspect perfectly apt. Dworkin’s protestant jurisprudence tends towards an ever-increasing plurality of legal interpretations.Footnote ¹¹⁰ The protestant approach to legal interpretation undermines law’s coordinating function and, particularly, its ability to speak with one voice—its integrity.Footnote ¹¹¹

Law’s integrity depends on legal officials’ ability to overcome moral disagreement and adjudicate disputes under a unified framework that does not simply replicate underlying disagreements.Footnote ¹¹² There is thus a certain self-defeating character to Dworkin’s view here: in the social world we inhabit, characterized by political disagreement, integrity is undermined by Dworkin’s protestant approach.

Perhaps, in a society with greater agreement, integrity and protestant interpretation could coexist. Here, a legitimacy issue arises. If the interpreters with political or institutional salience agree, then their dominant interpretation would, as Marmor argues, crowd out other views and undermine moral pluralism.Footnote ¹¹³ This concern leads Marmor to doubt whether integrity is a desirable ideal. I make a more limited claim. Assuming integrity is valuable, the only way in which law can speak with one voice, while respecting moral pluralism and ideological diversity, is for legal officials to resist seeing the determination of its content in terms of morality. Integrity might require legal officials to avoid trying to answer questions of political morality.Footnote ¹¹⁴

v. The Circumstances of Non-Positivism

As I have argued, there are multiple reasons in favor of the adoption of LP. But which operative concept of law will have better consequences if adopted is ultimately an empirical and contextual question. My argument is not that one size fits all. The argument is compatible with the claim that, in some circumstances, NP might be a better concept than LP. But those circumstances are sufficiently rare that they confirm the general case for LP.

If legal officials adopted NP, they would engage in moral reasoning for determining the law in force, no matter what the relevant legal sources require. Legal officials might be better at moral reasoning (in terms of their ability to reach correct moral decisions) than legislators and other (non-judicial) lawmakers; the population of legal officials might be on average equally adept at moral reasoning as legislators and other lawmakers; or they might be worse at it than legislators and other lawmakers.Footnote ¹¹⁵ The third situation is one in which LP would seem to be better than NP, and so I will ignore it here. The argument for NP would have to assume (i) that legal officials are more likely to achieve correct moral outcomes than legislators and other lawmakers, or (ii) that all of these agents are equally likely to reach morally correct outcomes.

The first assumption is problematic, for a very simple reason. Because we disagree about morality, we would also disagree about how to evaluate moral expertise and what counts as a morally correct outcome. Thus, even if legal officials were abler than legislators and other lawmakers to reach morally correct outcomes, we would not be able to agree on this.

Despite this disagreement on what counts as morally correct outcomes, perhaps we could still agree on the second assumption: that, on average, legal officials and other agents are equally likely to achieve morally correct outcomes, whatever they happen to be. This second assumption might be right, but it is irrelevant. If legal officials are as likely, on average, to reach correct outcomes as legislators and other lawmakers, then there is no reason to think that NP would produce morally superior results in specific cases.

The nonpositivist could argue that the relevant comparison is not between legal officials and other lawmakers, but rather between the institutional mechanisms and structures for making the relevant decisions: concrete, case-by-case decision-making in judicial and legal venues versus general, prospective decision making in a legislative venue. From that perspective, the proponent of NP could argue that the first type of venue is more likely to achieve morally correct outcomes than the second one, even if we disagree about how to characterize what would make for a morally correct outcome. But here we face similar problems of disagreement—just at a different level. We disagree morally and politically about whether the judicial venue is, in fact, better than a legislative venue at making moral decisions.Footnote ¹¹⁶ Like other morally disputed questions, this is also one that, under LP, positive law could answer authoritatively.

There are conditions, however, under which NP could be better than LP. If other state institutions are not responsive to the moral concerns and views of the majority of citizens, then legal officials might actually bring legal outcomes closer to those concerns. One problem with this argument is that, if the population of legal officials is morally diverse, this is an advantage that would be obtained at the cost of a loss in predictability and certainty. But when legal officials are a relatively homogeneous population and the outcomes of their moral reflection tend to converge, this problem tends to disappear. Thus, in circumstances under which

then NP is preferable to LP. The situations in which these four conditions obtain seem extremely rare—so rare that they reinforce that, in most cases, LP is the best choice.

i. Effacement and Esoterism

How can it be rational or legitimate for legal officials to adopt an operative concept of law that is at odds with what they think is ultimately true at the foundational level? There are two concerns at stake here. One concern is effacement.Footnote ¹³⁷ The other is what I will call Government-House jurisprudence.

Regarding the first concern, let me assume a judge believes that NP is the correct foundational concept of law. If I am right about the normative argument for them to adopt LP as the operative concept of law, then it seems that I am arguing that the foundational concept should be effaced: legal officials should believe in a different operative concept, LP, because belief in NP would be morally worse.Footnote ¹³⁸ Law thus seems to become a practice where legal officials ought to adopt a concept they might think is false, because belief in what is true would bring about bad moral consequences. My argument seems to require a bizarre and perhaps illegitimate form of psychological compartmentalization.

This concern, while reasonable, is misplaced. The institutional locus of NP need not be courts. For instance, a judge might believe that there is such a thing as natural law, and that natural law constrains the legitimate content of positive law, but also believe that, for moral considerations like the ones explored in Part II, judges are not in charge of enforcing natural law, and they would be better off, as judges, conceptualizing the law along the lines of LP. Questions about the content and bindingness of natural law are independent, as Robert George argues, from questions about the authority to give natural law a practical effect.Footnote ¹³⁹ A judge who believes that, for the reasons I have explored above, they are better off adopting a positivist conception when they act in a judicial capacity, is not effacing their belief in the truth of natural law but displacing its application to a different setting. In other words, that judge would be accepting that the operative concept of law that ought to be adopted in adjudication might be different from the foundational concept of law.

The argument for LP—as construed here—is a moral argument for the concept of law that should be used by legal officials in interpreting and applying the law. These agents can believe NP or OSV is true and also believe that, because of moral considerations (which NP might see as relevant, and OSV directly tells us are relevant, in the legal domain), the concept of law they ought to use or presuppose in legal reasoning within the legal system is LP. There is no effacement here, because views like NP and OSV at the foundational level might not require moral considerations to impact legal reasoning directly and without mediation at the operative level. Of course, moral considerations could impact the way in which legal participants think and talk about law directly, and Dworkinians would argue precisely that. However, the premise at issue here is whether this must be the case. That will depend on the relevant normative considerations. What operative concept of law certain agents ought to adopt is itself a normative question, and one that can perfectly coexist with the foundational views those agents might have about the nature of law.

Again, this does not mean that anything goes. The operative concept adopted must at least fit some basic convergent intuitions about what makes law as a matter of fact. As I have explained, the main jurisprudential views on offer are, at this level, on a par. They all seem to equally fit convergent intuitions about law. Their disagreement is about what to make of law as a general type of social practice and how to characterize it. Thus, the adoption of LP does not require assuming jurisprudential skepticism, but simply recognizing the fact of reasonable jurisprudential disagreement.

Consider the case of a judge who believes in the truth of OSV. That judge can perfectly say that while law is part of the larger moral enterprise, LP offers a better view of this corner of political morality, and a better account of how they ought to determine the content of law. That is precisely the type of discussion that Dworkin would have considered adequate, and compatible with OSV.Footnote ¹⁴⁰ Our judge can legitimately believe that, while law is ultimately a branch of political morality, in their activities they ought to adopt a positivist concept of law because of the relevant moral reasons.

But this takes us to the second concern, which is specific to judges, given their formal authority. According to that concern, which we could call Government-House jurisprudence,Footnote ¹⁴¹ judges persuaded that they should adopt LP even though they believe it does not reflect the truth about the foundational concept of law would be involved in deception. They will present their legal analyses and conclusions—which will lead to coercive enforcement—as flowing from law, determined in a positivist manner, even though they might think, for instance, that law is, at the foundational level, continuous with political morality. This truth about law will thus become esoteric: judges who believe a non-positivist view is correct will hide it from the “ignorant majority” when they engage in legal reasoning adopting a positivist operative concept.Footnote ¹⁴² Judges will artificially act as if law is a distinct, limited normative domain exclusively grounded in social facts, even though they believe that—at the foundational level—it is not.

The response here is closely aligned with the response to the effacement concern. Not just judges, but everyone—or, at least, everyone with the disposition to be interested in these questions—can understand that, while law is ultimately connected to larger questions of political morality, the balance of moral considerations suggests that legal officials engaged in legal reasoning ought to adopt a concept of law that is artificially separated from those larger moral questions. What the normative argument for LP requires from legal officials is not deceit but restraint. There is no reason why such restraint cannot be publicly acknowledged as such.

ii. Second-Order Disagreements

The second objection is the following. My argument rests on some empirical assumptions. It is a general argument in favor of LP, which is open to the possibility that NP could be a preferable alternative in certain contexts. But even granting all the relevant empirical assumptions, the defender of NP might still disagree with me about how to weigh different consequences. The problem is not just about evaluating different states of affairs but also about the criteria we can use to assess different states of affairs. Under this argument, even if LP and NP would produce the consequences I describe above, we could disagree about the moral assessment of those consequences.

For example, one could argue that NP will make it harder for law to predictably guide behavior.Footnote ¹⁴³ This, however, is not unambiguously bad. There might be moral benefits to some vagueness in the content of legal mandates, and the optimal degree of clear guidance by law is an open moral question. Uncertainty might have important moral benefits.Footnote ¹⁴⁴ And similar second-order disagreements might apply to the considerations I offer above. For instance, we could disagree about the relevance of moral disagreement and how much legal adjudication should or should not replicate it; we could disagree (as I have already recognized) about the relevance of majority rule; we could disagree about the value of integrity; etc.

This is entirely correct, but it is not obvious to me that the defender of NP can reject wholesale all the second-order evaluations on which my argument for LP rests. My argument rests on familiar and ecumenical values—some of which are even emphasized by nonpositivists. We should be concerned with the impact of the concept of law on the attainment of those values in any given society. Perhaps one or another of these values, or the specific way in which I evaluate the relevant consequences for any one of them, might not be shared by the nonpositivist critic. But wholesale rejection of the values and the weighing of the relevant consequences seems implausible. While implausible, it is certainly true that the critic might reject all the values I refer to and how I weigh the relevant consequences—and to that extent my argument will certainly not be compelling to the critic. At that point, however, I am inclined to think this is a reason for the critic to revise their moral commitments rather than to reject my argument.

Certainly, other values beyond the ones I have mentioned above matter as well. Fairness, solidarity, freedom, and other moral values are significant too. The issue here, though, is that these substantive and thick moral aims cannot be achieved in the context of widespread and uncoordinated moral disagreement. In that context, as I have argued above, the most urgent task is to allow for coordinated solutions that can avoid the conditions of the state of nature. The resolution of disagreement is, in this sense, prior to the achievement of these more ambitious moral goals.

iii. The Irrelevance Objection

There is an objection that is pessimistic about the prospects of any argument for adopting any concept of law: whatever conception of law legal officials adopt will lack any significant political implications.Footnote ¹⁴⁵ Soper, for instance, imagines two potential citizens, Gandhi and Eichmann. Gandhi is ideally conscientious: he tests every legal norm by reference to critical morality. Eichmann is the exact opposite.Footnote ¹⁴⁶ Soper replicates the model regarding judges: we ought to imagine a good judge who tries to achieve what morality demands despite the immorality of positive law, and an evil judge who pursues morally problematic ends independently of the content of the law. All of these agents behave exactly the same way whether they adopt NP or LP.

Soper’s argument seems valid. But the implication he derives from it is unwarranted. For the implication—which operative concept of law is adopted is practically irrelevant—to be true, it would have to be the case that all citizens in any given jurisdiction are like either Gandhi or Eichmann, and all judges are either good or evil. Yet most legal officials and citizens are like neither of these pairs. For most legal officials and citizens in reasonably well-ordered societies, law makes a practical difference (unlike for Gandhi and our imagined judges), without fully preempting moral evaluation (unlike for Eichmann). These agents want to know what behaviors the law requires and what would happen if they do not comply, because these are relevant inputs for their decisions. For these agents, what counts as law matters—and what counts as law turns on the concept of law they adopt.

There is a concern here, though, that I am misinterpreting LP and NP as theories of adjudication rather than as views about the concept of law. According to this objection, a judge that adopted LP might still think that legal facts are not the only relevant consideration for deciding specific cases. A judge might adopt LP and still ask whether, morally, they should just apply the law. LP might not impact the judge’s theory of adjudication, which is a separate normative question, distinct from the question about the concept of law.Footnote ¹⁴⁷ A positivist judge can be a judicial activist, and a non-positivist judge can be a formalist, an incrementalist, etc. The concept of law only tells legal operators what counts as law, not what they ought to do with it.

This is, in fact, a standard position within the positivist tradition. One of the great points of that tradition is, in Hart’s words, that it separates questions about the legal status of norms and questions about whether they should be obeyed.Footnote ¹⁴⁸ The issue here is that LP does not exhaust the conceptual toolbox of judges and lawyers—in other words, our legal culture does not lack answers to Hart’s separate question. There are familiar ideas, as I have mentioned, within that toolbox—ideas according to which judges’ primary responsibility is to apply the law. If judges think they should apply the law, then LP as an account of what counts as law secures the goods and goals I mentioned in Part II.Footnote ¹⁴⁹ Perhaps at this point the critic might want to argue that the idea that judges ought to apply the law is implausible, should be discarded, or is absurd. I believe they would be mistaken,Footnote ¹⁵⁰ but that is a separate discussion from whether—given the social world we inhabit, where the view that judges’ obligation is generally to apply the law is common and, I would add, plausible—the adoption of LP would be a good idea.

iv. The Eliminativist Objection

One final and related concern here is an eliminativist worry. Why should we debate the concept of law instead of asking the moral question about what judges and legal actors ought to do directly?

The objection can be framed from the perspective of OSV (which, as I explained, is ultimately an eliminativist position). Legal officials, according to the objection, could very well go about their activities by simply having substantive views about democracy, rights, legislative authority, the correct use of judicial power, and so on. There is no need for a concept of law.Footnote ¹⁵¹

The response, to my mind, is relatively simple. The eliminativist is right that we could replace discussion about the concept of law with discussion about what to their mind are the important, substantive questions. Thus, instead of ‘judges and lawyers should adopt LP,’ one could argue that ‘judges and lawyers should decide disputes and argue for their clients on the basis of legal norms derived from authoritative legal sources whose authority is determined exclusively on the basis of questions of social fact,’ or something along those lines. To my mind, this is just a reductio of eliminativism: it shows how eliminativism forces us to paraphrase ordinary discourse about law and replace it with increasingly convoluted normative speech.Footnote ¹⁵² So, indeed, there is no necessity here: judges can decide disputes without a concept of law, and all that we might want to say in terms of judges being generally supposed to apply the law, and how they should see and determine that law, can be said without reference to a concept of law. However, can does not imply must. Concepts are useful tools. They allow us to coordinate and interact with each other, to interpret the social world, and to impact it in different ways.Footnote ¹⁵³ We can think and talk like the eliminativist wants us to. But we do not need to. An eliminativist persuaded by my moral argument can replace all my talk about the concept of law with their preferred paraphrases. That would make our discourse more convoluted, and legal officials’ conceptualization of their tasks more complicated. In other words, eliminativism forces us to make the normative argument in different terms, but doesn’t change its substance.