Introduction
H.L.A. Hart’s The Concept of Law offers what continues to be an irresistible working assumption about the structure of law. In the ordinary case, it is difficult to argue with his claim that legal rules owe their status as law to meta-rules (such as constitutional provisions delineating lawmaking authority) that there is a consensus among legal officials to accord the status of meta-rules.Footnote 1 Yet suppose the content of the law were determined exclusively in that way by empirical consensus regarding meta-rules like the “rule of recognition” Hart famously took to separate the legal from the non-legal in any given legal system.Footnote 2 We would then seem forced to conclude that judges must be deceitful or confused in the many instances in which they employ non-consensus meta-rules, such as contested methods of legal interpretation defended at least partially on grounds that are not empirical but conceptual, and sometimes even moral. Ronald Dworkin characterized this challenge to the conventionality of law as the problem of “theoretical disagreement” about the law—i.e., disagreement about the “grounds of law,” or the standards that determine what is and is not law.Footnote 3
Like Dworkin, contemporary theorists who have taken up the issue of theoretical disagreement about the law have almost invariably seen it as a challenge for legal positivism in particular.Footnote 4 In this article, my first aim and contribution is to reframe the significance of theoretical disagreement, casting it not as a thorn in the side of legal positivism in particular but as a component of a broader challenge for jurisprudence: how to refine and reconcile three theses that should appear plausible, important, and in tension, to all thoughtful observers of law, positivist or not. (1) Conventionality: the content of the law is determined, presumptively if not definitively, by meta-rules of law whose status as meta-rules arises from a consensus among relevant legal actors to treat them as having that status. (2) Disagreement: judges have theoretical disagreements about the law—i.e., disagreements about such meta-rules of law as legal interpretive methods, which they do not attempt to resolve merely by reference to explicit or implicit empirical consensus. (3) Fidelity: judges’ theoretical disagreements can be in good faith, reasonable, and legally resolvable; they need not indicate that judges are systematically confused about the foundations of the law or that they are deceptively making law while pretending to find it.
My second agenda, beyond this reframing, is to generate a novel view of faithful theoretical disagreement by trying to identify a more robust and ecumenical reconciliation of Conventionality, Disagreement, and Fidelity than has yet been found.Footnote 5 A reconciliation of those three theses is more “robust,” in my sense of the word, if it shows how non-trivial versions of all three theses can be true simultaneously, rather than how two of the theses can be saved at the expense of a third. By a more “ecumenical” reconciliation, I mean one that is not designed to serve the partisan agenda of legal positivists or their critics but instead avoids taking a stance on their longstanding dispute as far as possible. The leading voices in the academic conversation around theoretical disagreement have thus far treated it as a wedge issue in the debate over legal positivism. Perhaps partly for that reason, they have each settled for a partial reconciliation of Conventionality, Disagreement, and Fidelity that favors two of the three at the expense of a third. They have been, to my mind at least, too quick to accept unduly deflationary conclusions about, respectively, the conventionality of law (Dworkin), the nature of theoretical disagreement (Scott Shapiro), and the extent of judicial fidelity to the law (Brian Leiter).Footnote 6
This article unfolds as follows. In Section I, I use a critical overview of recent work on law’s conventionality to chart a middle ground between overstating and understanding the conventionality of law. I demonstrate that even legal positivists should and do acknowledge limits to the conventionality of law. But I also review reasons for thinking Dworkin’s alternative theory of law presupposed but obscured law’s conventionality and in that way generated an unbalanced and misleading portrait of the structure of law.
In Section II, I discuss the problem of how to identify the law when the guidance offered by consensus standards runs out and theoretical disagreement comes to seem not only salient but inevitable. I contend that Shapiro, the leading positivist who has done the most to reconstruct and try to solve the problem of theoretical disagreement, ultimately ends up explaining it away, rather than explaining it. His provocative view is essentially that for “theoretical” disagreement to be legally resolvable it must be resolvable by standards given by contingent facts of legal practice—even if those facts are abstract and complex. Much of what initially appears to be theoretical disagreement may indeed be resolvable by reference to empirically given standards, in the way Shapiro implies. But Shapiro’s implicit commitment to the idea that to be legally resolvable theoretical disagreement must be resolvable by reference to empirically given standards seems to reflect an attempt to insulate legal positivism from critique rather than an attempt to understand the scope and character of judges’ interpretive disagreements on their own terms. Dworkin, by contrast, developed his account of theoretical disagreement as part of a critique of legal positivism. Though he took theoretical disagreement more seriously than Shapiro has, Dworkin was too quick to assume that theoretical disagreement must be moral in nature. He thereby obscured the more straightforward idea that theoretical disagreement is simply conceptual disagreement about the law, which may or may not be moral in nature.
In Section III, I offer an account of how theoretical disagreement may be faithful to the law when the legal guidance of consensus standards runs out. I argue that judges may, do, and should extract legal guidance from theoretical reflection upon the nature of law and its distinctive virtues—by which I mean something close to what Lon Fuller called the “internal morality of law.”Footnote 7 They may have good faith, legally resolvable “theoretical disagreements” about the standards that separate the legal from the non-legal, because such standards are a function not only of contingent empirical truths about consensus among legal officials but also of non-contingent conceptual truths about law’s nature and distinctive virtues—and judges no less than philosophers have reasonable disagreements about those conceptual truths. Moreover, because it is an open question (about which philosophers have disagreed) whether we need to engage in moral theory to identify law’s distinctive virtues, my account of Fidelity in the face of both Conventionality and Disagreement is ecumenical: the reconciliation does not immediately require us to take a stance on the truth or falsity of legal positivism.
Finally, in Section IV, I raise and respond to three likely lines of objection to my account of the possibility and nature of good faith, reasonable, genuinely theoretical disagreements.
I. Conventionality
Legal theorists of all stripes have seen “theoretical disagreement” as a problem for legal positivists in particular. The reason is straightforward. They have supposed that legal positivists (but not their critics) believe the “grounds of law” or “criteria of legal validity” are social facts that have their validating status simply in virtue of a consensus among relevant legal actors to treat them as having that status.Footnote 8 As Shapiro notes, reconstructing Dworkin: “positivists have maintained that the criteria of legal validity are determined by convention and consensus. But debates over interpretive methodology demonstrate that no such convention or consensus exists.”Footnote 9 If we suppose that laws gain their status as law in part because they satisfy certain moral principles, then theoretical disagreement appears to be a natural outgrowth of the pervasiveness of reasonable ethical disagreement in the modern world.Footnote 10 In that sense, critics of legal positivism arguably have an easier time than legal positivists accounting for theoretical disagreement about the law.
However, the problem that theoretical disagreement poses for legal theory is how to reconcile it with the role that conventionality (broadly construed) seems to play in making law possible and determining its content. And both legal positivists and their critics vary in the extent to which they believe law is conventional. On the one hand, as I will show in the discussion to come, many leading positivists (including Shapiro) have offered persuasive reasons for rejecting the idea that law depends upon a “coordination convention” (a type of convention that philosophers have analyzed rigorously and which may be understood as the focal case of a convention) thus raising the question of in what ways and to what extent law is conventional. On the other hand, critics of positivism can, should, and sometimes do emphasize more than Dworkin both that law requires extensive convergence of behaviour and that one of its most fundamental and distinctive benefits is the degree of social coordination such convergence makes possible.Footnote 11
A. The Critique of Pure Coordination
Though it is difficult to deny that law is conventional in an important sense, it is not easy to identify in precisely what sense. In recent years, legal philosophers have made a persuasive case for rejecting the idea that law is based on the kind of narrow, paradigmatic form of convention that philosophers have labelled a “coordination convention.”Footnote 12
Loosely speaking, a convention is a pattern of coordinated activity. An extremely broad range of activities are conventional in this loose sense—activities ranging from those that are arbitrary save for coordination (say, driving and walking on the right, rather than left side of the median) all the way to activities that are far from arbitrary, yet still substantially coordinated (say, religious practices).Footnote 13 While noting that conventions come in degrees, David Lewis offered a seminal analysis of focal cases of convention that have come to be called “coordination conventions.” The kind of conventions he analyzed are not just coordinated activities but a carefully defined set of them: roughly, behavioural regularities to which individuals adhere, expecting others to adhere, and each doing so on the conditional preference that he or she adhere given that others adhere.Footnote 14 More recently, Andrei Marmor has extracted from Lewis and emphasized an important idea about the rules of conventions that applies both to coordination conventions and an even broader set of conventions beyond them. The idea is that the rules need not be ones about which their adherents are “indifferent” but only ones that are “arbitrary” for them in the sense that “the reason to prefer one of the potential rules is not stronger than the reason to follow the rule that is actually followed by others.”Footnote 15 Marmor characterizes this as meaning partly that there are “compliance-dependent reasons” for following the conventional rule.Footnote 16
The distinction Marmor highlights between “indifferent” and “arbitrary” should ease one potential qualm about the putative conventionality of law. Clearly the judges, government officials, and citizens who accept the American legal order and contribute to its continued existence are not indifferent between being in the American legal order and other actual or hypothetical legal orders. They need not be indifferent, however, for their consensus acceptance of standards separating what is and is not law to constitute a coordination convention.
However, as Leslie Green notes, “in the case of a mere [coordination] convention the fact of common practice is not just necessary but also sufficient to motivate conformity.”Footnote 17 Yet in the case of acceptance of standards of legal validity, it is not clear to what extent reasons of coordination are likely sufficient. As Green observes, unlike in the case of say, spelling, about which “it just makes no sense to suppose that there are converging convention-independent standards of correctness,” different individuals may each follow, for example, “a practice rule that prohibits murder, not because they are playing follow-the-leader, but because each thinks murder wrong, and thus independently comes to the same conclusion as the others.”Footnote 18 Moreover, as he suggests, legal officials may identify reasons for following particular legal “rules of recognition” that are not only distinct but “even incompatible” with one another.Footnote 19 So although coordination conventions can give individuals reasons to follow a rule such as a rule of recognition,Footnote 20 these are compliance-dependent reasons which do not include many of the important reasons why officials may follow a legal system’s rule or rules of recognition.Footnote 21
If a legal system is sufficiently unjust or there is another better one for which to fight, officials may follow one rule of recognition for non-coordinative reasons, and coordinative reasons might be insufficient to get them to follow another. Shapiro speculates: “If most officials suddenly abandoned the United States Constitution, this would not lead all others to similar action. Some would resign in protest, while others would continue applying the rules validly enacted under the United States Constitution.”Footnote 22 All this reflects what we already saw in Lewis’s and Marmor’s accounts of conventions, and which Green also puts well: “in any [mere] coordination problem there can be no radical conflict of interest, i.e., conflict so great that it overwhelms the benefits of coordination.”Footnote 23
B. The Need for Cooperation
Although leading legal positivists reject the idea that the rules of recognition or ultimate determinants of law should be understood on the model of a coordination convention, it remains obvious that a substantial degree of consensus among legal officials is necessary for the existence of a legal system. Given the apparent inadequacy of coordination convention accounts for explaining the foundations of law and legal systems, legal philosophers have recently proposed alternative accounts that, while loosely conventional, allow for a greater range of reasons why legal officials might adhere to shared rules of recognition. Most notably, a handful of prominent philosophers of law have explored the possibility that the practices at the root of law may not be solely coordinative conventions, but rather more substantively oriented (while still somewhat coordinative) forms of “joint intentional activity” or “shared cooperative activity.”Footnote 24
Scott Shapiro has done more than any other legal philosopher to develop a conception of legal positivism on such a basis. Applying the work of Michael Bratman, philosopher of action and intention, to law, Shapiro has argued that law ultimately rests on a joint form of activity characterized by a certain level of commitment to the activity and cooperation with other participants in the activity.Footnote 25 As Jules Coleman has noted, a good, simple example of joint activity comes from philosopher Margaret Gilbert, who has written at length about the seemingly simple case of taking a walk with another person.Footnote 26 As they both suggest, that is an activity in which individuals are not merely coordinating, but are pursuing a shared goal, which gives them reasons to respond to one another’s behaviour in certain ways.
A legal system can withstand a fair amount of disagreement among such legal officials as judges about highly abstract propositions of legal validation—as is attested by methodological disagreement about legal interpretation. This tolerance is possible in part because there is substantial overlap in the concrete implications of different sets of legal interpretive methods.Footnote 27 Yet substantial convergence is necessary, even if it arises not in the first instance because of coordination or cooperation, and even if the convergence is primarily onto lower-level legal rules regulating behaviour rather than lofty constitutional principles or meta-rules regulating the legal system. Leslie Green, who has rejected the idea that law is a mere coordination convention, has nevertheless recognized that “[a]s Postema has argued, officials could not succeed in coordinating mass behaviour if they did not, to a significant extent, attempt to coordinate not only among themselves, but also with society at large.”Footnote 28
The attraction of cooperative over merely coordinative accounts of law lies in the ability of the former to allow for a broader range of reasons why judges, officials, and citizens may adhere to the same rules of recognition separating what is and is not law. Matthew Adler has suggested, however, that even cooperative accounts of law (such as those that have been endorsed by Shapiro, Coleman, and Kutz) may imply that legal systems require and depend for their content upon convergences of attitude and behaviour that are more arbitrary than the convergences of attitude and behaviour which we sometimes observe—or at least can imagine—in actual legal systems.Footnote 29 In his view:
A striking implication of group-sensitive positivism [of the kind suggested by Shapiro, Coleman, and Kutz] is that a society of committed textualists—where each official uncompromisingly accepts the text of the 1787 Constitution, together with some interpretive method, as fundamental law, regardless of whether other officials accept a different text or method—would not be a full-fledged legal system.Footnote 30
To be sure, if officials were dedicated to some supremely good set of rules and principles, they might cling to it tenaciously even if their colleagues should jump ship en masse. However, the kind of system Adler imagines is strikingly similar to the kind of “pre-legal” order of primary rules Hart discussed in order to illuminate the problems that a legal order with both primary and secondary rules helps to solve.Footnote 31 More specifically, a system of the sort Adler imagines would be fragile. In a snapshot in time, officials would appear committed to the same rules of recognition. But they would not be committed to coordinating with one another to achieve consensus on rules of recognition should the nature of some judges’ commitments change over time.
C. Why Dworkin’s “Integrity” Cannot Replace (and Must Presuppose) Conventionality
Dworkin was dismissive of what he called conventionalist accounts of law.Footnote 32 He exhorted judges to understand the law not as merely the “preinterpretive” “law” which everyone recognizes as law, in some sense, but rather as the best moral reconstruction of the materials that legal participants and observers would, by consensus, recognize as legal materials to be interpreted.Footnote 33
Some of what he appeared to deny regarding law’s conventionality he may have accepted (in spite of himself) but simply downplayed and obscured. For he insisted that judges must weigh the competing moral merits of only those potential legal interpretations that first meet the “threshold test” of “fit” with the legal materials to which the interpretations would apply.Footnote 34 However, Dworkin’s conception of what it means for an interpretation to “fit” the legal materials to which it applies appears to have been both more and less demanding than what we would ordinarily understand “fit” to require of a legal interpretation. He seems to have thought that the legal materials an interpretation should be judged to fit or not fit are not the materials bearing upon that provision, narrowly and formally construed, but rather the underlying principles of the whole surrounding area of legal doctrine and indeed of the entire legal system.Footnote 35 He appears to have thought this because he believed that the importance of fit arose not from the nature of interpretation, per se, but rather from his substantive commitment to “law as integrity” according to which “preinterpretive” law should be read by judges in such a way as to make it as principled as possible.Footnote 36
To be fair to Dworkin, he sought to quell concern that his theory might therefore give specific provisions of law very different content than they would seem to have according to conventional understanding. He did so in several ways. First, he treated “compartmentalization [a]s a feature of legal practice no competent interpretation can ignore.”Footnote 37 Second, he recognized that “constraints of fairness and process” will sometimes require that judges temper the extent to which they pursue the “integrity” of “consistency in principle”Footnote 38 for law and require them to settle for what he called “inclusive” rather than “pure” integrity.Footnote 39
Still, in downplaying and obscuring the conventionality of law, Dworkin misrepresented the structure of law as whole, and in general. He not only failed to offer a persuasive explanation of the way in which social facts about convergence among legal officials constrain the content of the law. He also claimed, implausibly, that “[l]aw as integrity explains and justifies easy cases as well as hard ones; it also shows why they are easy.”Footnote 40 The contestable moral importance of “integrity” might well provide a moral justification for judges to follow the dictates of “easy” law given by convergent practice. But surely it cannot provide a compelling explanation of why easy cases are easy, legally speaking. Far from explaining why “easy” cases are easy, law as integrity can only presuppose that certain easy cases are easy and threaten to make them difficult.Footnote 41 That is because even in easy cases, the requirements of “consistency in principle” are often neither clear nor straightforward but matters of reasonable disagreement. Often a legal result is clear and undeniable but its consistency in principle with surrounding law is unclear and contestable.
In sum, we need not endorse legal positivism to recognize the substantial dependence of law on convergent legal practice. Even non-positivists who believe that moral rules and principles constrain what the law can require or allow must recognize that social facts about convergence in practice substantially constrain, if not uniquely determine, the content of the law.Footnote 42
II. Disagreement
A. When Consensus Ends, Legal Reasoning Begins
Focused as he was on the most difficult problems of legal interpretation and adjudication, Dworkin offered an unbalanced portrait of law that foregrounded the exotic and backgrounded the quotidian. But he was right to resist the temptation to suppose that in the difficult legal cases that make theoretical disagreement salient there must simply be “no right answer” legally speaking.Footnote 43 When, as is often the case, existing practice has not yet explicitly converged to generate a unique determinate answer to a legal question, judges do not abandon legal reasoning. In fact, it is only because, and to the extent that, practice has not yet explicitly converged on a unique determinate answer to a legal question that legal reasoning, interpretation, and judgment arise at all. As Richard Fallon has observed, “one does not need to ‘interpret’ when one knows immediately and unreflectively how to go on.”Footnote 44 Conversely, as Joseph Raz has noted, “interpretation is possible only when the meaning of what is interpreted is not obvious.”Footnote 45 Legal reasoning is precisely a matter of using general legal principles and interpretive methods to resolve specific legal questions whose answers cannot be straightforwardly determined by applying rules whose legal validity is clearly indicated by their acceptance as a matter of consensus.
Even when legal practice has not converged directly on a rule or set of principles that unambiguously answers a legal question, it may still indirectly determine a unique legal answer if it sufficiently constrains the general interpretive methods and legal principles to which judges may appeal in answering the question. The legal meaning of a phrase in a statute may seem ambiguous on its face or on initial inspection, and it may be that no legal actors have yet taken a position on its legal meaning—let alone converged in their interpretations of it to such an extent that there is no longer any real question of its legal meaning. Yet convergent practice may still limit the general interpretive methods and principles of legal reasoning to such an extent that any combination of those methods and principles would speak decisively in favor of one reading of the legal provision in question.Footnote 46 In that case, we should say that the legal meaning of the provision remains determined by social facts of practice, albeit in an especially indirect manner. Moreover, American judges have, in fact, converged to a substantial extent on a limited number of familiar legal principles and interpretive methods that they treat as relevant to the resolution of disputable points of law. For instance, Philip Bobbitt has offered a useful classification of five widely accepted kinds of constitutional argument: historical, textual, doctrinal, prudential, and structural.Footnote 47
Yet there are obvious and important limits to the extent to which unique determinate answers to legal questions arise from judges’ convergent acceptance of a range of familiar legal principles and interpretive methods. First, as many legal theorists have pointed out—including the legal realists in particular—many of the accepted canons of legal interpretation stand in tension, if not outright conflict, with one another.Footnote 48 Second, as we have seen, a salient fact about judicial practice, which Dworkin rightly emphasized, is that judges appear to have not only “empirical disagreements” about what the law is in practice but also “theoretical disagreements” about what the law is even in principle, because they disagree about the very interpretive methods and principles by which to answer any question about what the law requires or allows. For instance, some judges believe that it is inappropriate to consider legislative intent or purpose in interpreting statutes, while others believe that it is wrong to take the meaning given by text and statutory structure as decisive if that meaning clearly conflicts with the intent or purpose of the lawmakers who crafted the statute.
Dworkin argued that theoretical disagreements among judges push us toward the conclusion that the law is not ultimately determined by social facts of convergent practice but rather by the results of constructive interpretation—i.e., examining a set of “preinterpretive” materials and attempting to reconstruct them “in the best light” by adopting the construction of them that best combines “fit” with the preinterpretive materials and “justification,” in the sense of resulting moral justifiability or value attributed to the object.Footnote 49 Before turning to Dworkin’s account of theoretical disagreement, formulated in opposition to conventionalism about law, let us first examine the leading alternative account of theoretical disagreement put forth by legal positivist Scott Shapiro. Whatever other merits it may have, it is at least designed to harmonize with a broadly conventionalist view of law.
B. Shapiro’s Reduction of Theoretical Disagreement to Empirical Disagreement
In his reconstruction and response to the problem of theoretical disagreement, Shapiro makes the following observation. When judges have not only first-order disagreements about propositions of law but also second-order disagreements about the interpretive methods by which to resolve those first-order disagreements, they may still share third-order standards by which to adjudicate their second-order disagreements regarding competing interpretive methods. Shapiro writes: “The commitment to the social foundations of law, I have tried to show, can be satisfied in the absence of a specific convention about proper interpretive methodology just in case a consensus exists about the factors that ultimately determine interpretive methodology.”Footnote 50 To illustrate his point, Shapiro argues that even if a legal system has not directly settled all disputes over the interpretive methods for judges to employ, it may have indirectly settled or at least narrowed them, because the structure of specific legal institutions may have implications for the appropriate scope of judicial authority and reasoning.Footnote 51 Shapiro implies that if theoretical disagreements have unique determinate legal resolutions they do so only in virtue of higher-order convergence (within a specific legal system) on shared standards for judging interpretive methods. Hence, he does not so much explain theoretical disagreement as explain it away. In Shapiro’s view, theoretical disagreements have genuinely legal resolutions only when there are standards for resolving them given by facts about convergence in legal practice—facts that are ultimately empirical, even if also complex and abstract.
Shapiro is right to note that disagreements about interpretive methods can be explained at least in part by disagreements among judges and academic lawyers about complex, contingent social facts regarding the degree of trust a legal system implicitly accords to judges, and the relative degrees of cooperation and contestation involved in the making of laws that judges must interpret.Footnote 52 For example, Justice Scalia’s and Dworkin’s competing ideas about interpretation plausibly rested in part on divergent views about the scope of the role that their legal system has entrusted to judges.Footnote 53
Yet judges and legal theorists do not generally defend their favored interpretive theories merely by emphasizing the explicit and implicit footholds they already have in the structure of existing legal institutions and in the behavior of legal officials. As evidence of this, consider, for instance, that several prominent legal scholars have recently put forth the idea of defending originalism on predominantly empirical grounds, and framed doing so as an innovation for constitutional theory.Footnote 54 Most often judges and legal theorists defend their views, and attract adherents, on the grounds that their way of understanding the law is more faithful to attractive ideals, and not just to barren social facts about institutional structure and convergence in legal officials’ attitudes and behaviour.Footnote 55 Judges are generally less self-conscious than academics. But consider one of the more self-conscious jurists in recent memory: former Associate Justice of the Supreme Court of the United States, Antonin Scalia, who had a distinctive theory that clearly informed his practice. He argued for his textualist, originalist approach to constitutional interpretation in part by invoking considerations of political morality, such as democratic legitimacy, as against what he observed of ground-level judicial practice.Footnote 56
It is important not to move too quickly here, however. Arguments like Scalia’s democratic case for constitutional originalism do not obviously depend on truths of political morality. Often proponents of such arguments are best understood as purporting to derive interpretive methods not from the political values they judge to be morally best but rather from those political values they judge to be immanent within the constitutional order, and to give coherence and purpose to its rules, principles, and institutional structures.
Making the point more concrete helps show how far it can render Shapiro’s reductionist view of theoretical disagreement compatible with what we observe of judges’ interpretive disagreements. So let us consider constitutional interpretation in light of just one aspect of the immanent political morality of a constitutional order: the conception of representative democracy it encodes. Two judges may agree that their legal system is structured around an ideal of representative democracy and that this ideal should inform the way they interpret the constitution. Yet they may disagree about the conception of representative democracy encoded in the constitutional order in four salient ways.Footnote 57
First, judges may agree that their constitutional order encodes a particular conception of representative democracy (call it conception A), but disagree about which interpretative method would be most consistent with that conception of representative democracy.
Second, one judge may think the constitutional order encodes conception A of representative democracy while another judge thinks it encodes conception B.
Third, one judge may think the constitutional order encodes a particular conception of representative democracy. Another judge, however, may think the constitutional order encodes merely a general idea of representative democracy, forcing judges to weigh and choose among competing moral ideals of representative democracy to resolve uncertainties about how to interpret the constitution. We can illustrate the point using Shapiro’s framework of trust and distrust. The second judge may believe that there are strands of the American ideal of representative democracy that both empower and disempower judges vis-à-vis other branches of government and the constitution’s authors. Hence, she may believe the extent of her license to update constitutional provisions in light of citizens’ evolving ideals ultimately depends in part upon the compatibility of different conceptions of the judicial role with the best moral conception of representative democracy and not just with America’s immanent conception of representative democracy.
Finally, two judges may agree that the constitutional order commits judges to interpret the law in the light of the morally best conception of representative democracy but disagree about which specific conception fits that description.
In cases one through three, the disagreement may be resolvable on grounds that are best described as “empirical” because they are dependent upon contingent facts about the practices of the specific legal system at issue. There may be an empirical answer to the question of which conception of representative democracy the constitutional order encodes. Moreover, it may be that the conception encoded is one whose conceptual contours (a) can be derived from empirical facts about legal practice and (b) dictate a unique, determinate answer to whatever concrete questions of legal interpretation are at issue. However, suppose the constitutional order’s immanent conception of representative democracy is not precise enough to do whatever interpretive work is needed from it, and judges are left to draw partly upon a moral ideal of representative democracy, as in case four. In that case, disagreement they have about the best moral ideal of representative democracy will be genuinely “theoretical” in the sense that its resolution will derive from moral truths that transcend—rather than depend upon—contingent facts about practices in the legal system at issue.
Shapiro recognizes the possibility that empirical facts determining the law might direct judges to moral theory and in that way provoke putatively theoretical disagreement about the law. But he has a subtle way of deflecting the idea that the possibility of such disagreement reveals the content of the law to be potentially dependent upon non-empirical, theoretical truths of political morality. Shapiro is a so-called “exclusive” (as opposed to “inclusive”) legal positivist. This means he rejects not only Dworkin’s claim that law necessarily depends on morality in the way Dworkin suggests, but also the idea that law may contingently depend on morality in virtue of social facts that make it so dependent.Footnote 58 Shapiro recognizes that there is a sense in which judges can be—and indeed apparently are—required by positive law to take morality into account in their decisions. He simply follows Raz in arguing that this must be merely analogous to the way in which American judges can sometimes be required by American law to take foreign law into account in resolving cases—i.e., not because the foreign law is itself part of domestic law.Footnote 59 In Shapiro’s view, “in hard cases, where the pedigreed primary norms run out, American judges are simply under a legal obligation to apply extra-legal standards.”Footnote 60
Presumably, then, Shapiro would acknowledge that the convergent, higher-order standards governing disputes about competing interpretive methods may, as a matter of fact, instruct judges to consider political morality in deciding which interpretive methods to adopt. Shapiro appears committed simply to insisting that if (a) the plans of a legal system are such that judges must engage in moral reasoning to select the “correct” interpretive methods and (b) a legal question arises that cannot be resolved by reference to an “incompletely theorized agreement,”Footnote 61 but instead requires a judge to apply contested interpretive methods, then (c) there is a sense in which there is no unique determinate legal answer to the question. As Shapiro puts it:
… [J]udicial decisions can be legally regulated under one description but legally unregulated under another…. Under one description, “deciding for the morally entitled party,” the decision is legally mandated; under another, “deciding for the promisee,” it is legally unregulated.Footnote 62
Shapiro thinks it important to make such a qualification lest we “end up describing the law as providing far sharper guidance than it has in fact provided.”Footnote 63
Though I cannot address the issue in detail here, there are good reasons to doubt the truth of exclusive positivism.Footnote 64 If we reject exclusive positivism, it is open to us to allow that social facts determining the content of law can also make the law dependent upon considerations of political morality such as, for example, the best moral reconstruction of the ideals of representative democracy. And doing so gives us one way of reconciling theoretical disagreement with a broadly conventionalist understanding of the law: treating theoretical disagreement as prompted by empirical facts about legal institutions and practices that implicitly demand we employ theoretical reflection on ideals of political morality to resolve otherwise ambiguous points of law.
That said, there is an important conceptual distinction between what judges ought to do, all things considered, and what judges ought to do, if they are to be faithful to the law. In cases in which empirical facts about legal institutions and practices point judges to political morality, even exclusive legal positivists such as Shapiro can recognize that there is a sense in which judges are more faithful to the law when they answer those questions of political morality correctly. On the other hand, this is legal fidelity in an attenuated sense. If it were the only form of fidelity to law that judges engaged in genuinely theoretical disagreement might claim for themselves, then we should give up on the hope of a robust reconciliation of Conventionality, Disagreement, and Fidelity. Conventionality would in that case trump, leaving genuinely theoretical disagreement to be peripheral to the law and potentially faithful to it in only a peripheral sense.
Hence, I do not claim that a robust reconciliation of Conventionality, Disagreement, and Fidelity can be derived from an attempt to rescue (from exclusive legal positivists) the bona fide legality of legally required backstop moral reasoning. Instead, I hope to show that a robust reconciliation of Conventionality, Disagreement, and Fidelity can be found if we are not too quick to assume that as soon as empirical consensus runs out, so too does law, and all that is left is morality. In particular, we should not discount the possibility that when legal facts determined by empirical consensus run out, we may still be able to extract legal guidance from conceptual truths about law and its distinctive virtues before having to resort to backstop moral reasoning.
C. Dworkin’s Leap from Theoretical Disagreement to Political Morality
Ironically, in my view, despite their disagreement about whether to characterize backstop moral reasoning as genuinely legal or not, Dworkin and his critics make the same mistake of resorting to such reasoning too quickly. They do so because they both assume that if there is no fact of the matter about law dictated (albeit perhaps implicitly) by contingent, convergent practice, then moral reasoning is the next and last resort of a judge tasked with either identifying—or, failing that, making—law. But that is not so. Although Dworkin fares better than his critics on this score, neither pays adequate attention to the possibility of extracting legal guidance from the distinctive virtues of law before resorting to more capacious ethical theorizing to find or make law.
We can read Dworkin as taking “integrity” to be the distinctive virtue of law.Footnote 65 But he elevated integrity to the exclusion of other important, distinctively legal virtues, such as the predictability in governance it produces.Footnote 66 Moreover, although concern for integrity perhaps can and should help judges resolve difficult cases, Dworkin uses it first (and arguably foremost) as a rationale simply for accepting the meaning of some legal provisions as “easy” (because only one interpretation of them “fits” the landscape of legal principle that surrounds them) and turning other “easy” cases into difficult ones by “encourag[ing] a judge to be wide-ranging and imaginative in his search for coherence with fundamental principle.”Footnote 67 Dworkin contended that “[h]ard cases arise, for any judge, when his threshold test [of “fit” as determined by the demands of integrity] does not discriminate between two or more interpretations of some statute or line of cases.”Footnote 68 Rather than emphasize the extent to which distinctively legal virtues in and of themselves might be used to resolve cases in which multiple interpretations are consistent with the empirical facts of legal practice, Dworkin emphasized instead the extent to which such cases prompt a need for more general and expansive moral theory: “in hard cases judges must make controversial judgments of political morality whichever conception of law they hold.”Footnote 69
In sum, though Dworkin saw law’s content as dependent upon its distinctive virtues in a way that many positivists have refused to, he obscured the force of that position by blurring the line between law’s distinctive virtues and the many substantive values that it ought to serve as a matter of political morality but that are not distinctive of law. In the next section, I will offer a sketch of the guidance judges can extract from law’s distinctive virtues, which I believe are close in structure to (even if not the same in substance as) the kinds of virtues Lon Fuller famously called “the internal morality of law.” Moreover, I will argue that by understanding theoretical disagreement as reasonable disagreement about the nature and distinctive virtues of law, we can better understand how theoretical disagreement can be in good faith, genuinely legal in nature, and even about the content of existing law—and all this consistent with the idea that law is broadly conventional and even that legal positivism may be true.
III. Fidelity
A. Theoretical Disagreement as Conceptual Disagreement about Law
Legal systems may—and, I believe, often do—include implicit rules or principles instructing judges to take into account such distinctive virtues of law as generality and predictability in government in their efforts to resolve legal issues otherwise indeterminate within some range.Footnote 70 And if such virtues are contingently picked out by the social rules and principles of a legal system as determinants of the content of the law, then no one but (perhaps) the exclusive legal positivist should be expected to object to saying that such virtues may dictate legally determined answers to legal questions.Footnote 71
However, if specific distinctive virtues of law affect the requirements of judicial fidelity to law only because convergent practice has picked them out as determinants of the law, then the only legally relevant disagreements about those virtues and their implications in particular cases must be complex empirical disagreements about (a) which virtues have gained the requisite acceptance within the particular legal system at issue and (b) how best to realize those virtues in particular cases. And I am particularly interested in precisely those disagreements about law that are genuinely theoretical in the sense that they are not resolved by empirical facts about convergent practice.
Suppose, for example, that there is a dispute about whether to apply the rule of lenity to a criminal statute. Further assume that in the legal system in question, the rule of lenity has not yet achieved consensus acceptance. Nor, let us assume, is there a direct consensus among judges to interpret ambiguous law in the light of specific rule of law values such as predictability in the state’s use of coercion.
As a matter of empirical fact, judges in the legal system might nevertheless be bound to interpret ambiguous laws in a way that accords with rule of law values, and this might in turn dictate that they apply the rule of lenity. As legal systems develop more complex divisions of labor and seek more precisely to delineate the bounds of the judicial role, more concrete rules and principles like the rule of lenity are likely to emerge which preemptively decide how judges should interpret the law so as best to realize law’s distinctive virtues. In that sense, theoretical disagreement about the law (understood as disagreement about the implications of rule of law values for the content of the law) has undoubtedly diminished, and is likely to continue to diminish, along with the historical phenomenon of the increasing formalization and codification of positive law.
But however advanced the state of narrowly positive law may be in anticipating and channeling uses of rule of law values to resolve otherwise uncertain points of law, gaps in narrowly positive law are bound to remain. And my claim is that when they arise, judges should follow a strategy of filling them in by resorting to theoretical reflection on law and its distinctive virtues—before opening their reasoning to more general considerations of political morality that are not distinctively legal in nature. The theoretical disagreements judges will have about rule of law values (which narrowly positive law has not preemptively resolved) will have legal resolutions in a more robust, less attenuated sense than disagreements among judges that have devolved to backstop moral reasoning of a generic and not distinctively legal kind.
In short, theoretical disagreements appear genuinely theoretical, yet susceptible to legal resolution, if they are understood as disagreements about the best way to conceptualize law and its distinctive virtues—disagreements therefore resolvable by reference to the standards for judging superior and inferior accounts of the distinctive virtues of law and their relative importance in the ideal of the rule of law.Footnote 72 Even if the law appears indeterminate, leaving judges to create new law to resolve its indeterminacy, still, if it is possible in doing so for judges to promote the distinctive virtues of law, then their resulting “lawmaking” remains open to praise or criticism according to its adherence to law (even if not to the laws in a narrower sense).Footnote 73 Even judges who are on the wrong side of theoretical disagreements may be acting in good faith, because they may be taking as a guide a reasonable, albeit inferior, conception of law and its distinctive virtues. Whereas fidelity applies to the identification and application of the law because judges can do better or worse at identifying and applying the law that exists (rather than the law they think ought to exist), fidelity applies to judicial “lawmaking” because the laws judges make to resolve the indeterminacies of existing law can be better or worse qua law—i.e., they can be more or less successful at promoting the distinctive virtues of law.
I now want to make an even stronger point, but one I contend is the same point more boldly unveiled. When contingent social facts of convergent legal practice fail to dictate a unique legal answer, but judges extract sufficient guidance to resolve the issue by considering what disposition would make the law most faithful to the rule of law and its distinctive virtues, judges are simply following the law. Raz has rightly acknowledged that “[i]n cases of [legal] indeterminacy there is often no clear divide between application and innovation,” and “on most occasions the reasoning justifying law-making decisions is similar to and continuous with decisions interpreting and applying law.”Footnote 74 However, precisely because judges are subject to the requirement of fidelity to law in not only the application but also the “creation” of law, Raz overstates the difference between the two activities when he claims that “[i]t is the difference between rules imposed by authority which the courts are bound to apply … whether they like them or not and rules which they follow because they judge them to be suitable.”Footnote 75 For the same reason, Shapiro is wrong to distinguish sharply between “legal reasoning” and “judicial decision making” and insist that “when the law has gaps or is inconsistent, a judge who is obligated to rule cannot employ legal reasoning, and therefore has no choice but to rely on policy arguments in order to discharge his duty.”Footnote 76
We might say that judges are bound both in the application and “creation” of law by the authority of law—in the first case, by the authority of the laws and in the second case, by the rule of law and its distinctive virtues. But when judges extract a unique answer to a legal question by determining that the rule of law requires that the answer be X rather than Y, it seems artificial to say that simply because the resolution arises from conceptual truths about law as such rather than empirical truths about the law, it is therefore something that judges “create” rather than find and apply. Arguably that thought is implicit in the plausible, if controversial claim of Dworkin and Waldron that “disagreements about what constitutes law in [hard] cases and about what the Rule of Law requires amount to the same disagreement.”Footnote 77
In many cases, it will be difficult to know whether judges trying to make their decisions reflect the distinctive virtues of law are doing so because they understand this to be part of fidelity to law as a conceptual matter or because they understand those virtues of law to be principles that are the subject of convergent acceptance by officials in their legal system. The most obvious cases in which it may be charitable and illuminating to understand judges as appealing to putative distinctive virtues of law on theoretical rather than empirical grounds will be ones in which judges distinguish themselves with non-consensus (and, if they are lucky, pathbreaking) styles of interpretation. The view I have defended implies that whether their cutting edge interpretive approaches should be judged legal successes or failures depends not—or not only—on whether they eventually enter the mainstream of judicial practice, or on whether they make what passes for law morally better than it otherwise would be, but on how the application of their methods bolsters the rule of law by promoting law’s distinctive virtues. The innovators of judicial history—whether they be John Marshall or Oliver Wendell Holmes, Jr., Earl Warren or Antonin Scalia—need not be unfaithful to the law simply because they buck mainstream trends. But whatever moral virtues can be claimed for their interpretive visions should not be taken to make them good jurists, except to the extent that those moral virtues are distinctively legal virtues.
B. Conceptual Disagreement about Law: From Dworkin to Positivists to Judges
I have suggested that the content of the law is a function not only of contingent social facts but also of non-contingent facts about the nature of law and its distinctive virtues. It is critics of legal positivism who are generally thought to insist that law’s content depends on non-empirical truths—namely moral truths. But even prominent legal positivists accept a version—albeit an implicit and negative version—of the idea that the content of the law regarding a particular issue in a given legal system depends upon non-contingent conceptual truths about law and its functions, if not its distinctive virtues.
For instance, Shapiro argues that laws are plans and the content of a plan cannot depend upon an independent resolution of the very issue that a plan attempts to resolve; hence, the content of the law cannot depend on the resolution of contestable matters it attempts to resolve.Footnote 78 Put concretely, if I need to resolve the question of which movie to watch in order to know which movie I plan to watch, then I do not yet have a plan about which movie to watch. Likewise, Shapiro claims, if I need to appeal to moral rules to know which moral rules the law binds me to respect, then there is no preexisting law that already settles which moral rules I am bound to respect.Footnote 79 Raz makes a very similar claim, which also implies that what law we have depends upon the nature and distinctive function of law: “The law to be law must be capable of guiding behaviour, however inefficiently.”Footnote 80 In sum, as we have seen, it is precisely Raz’s and Shapiro’s belief that law must be able to serve its function of providing guidance separate from the guidance of morality that leads them to conclude that the law cannot include or depend upon moral principles.
Even if the distinctive virtues of law sometimes shape what the law requires, permits, and empowers us to do, it is a further question whether understanding the nature and implications of such virtues requires us to engage in moral reasoning, making morality a determinant of the content of the law.Footnote 81 However, we need not take a stance on that issue now. In fact, for present purposes, it is a good thing to the extent that it remains a matter of reasonable dispute. For the reasonableness of disagreement about whether law’s distinctive virtues can be identified without reference to morality implies that both positivists and their critics can consistently endorse my view of faithful theoretical disagreement. Non-positivists surely can, because they allow that the content of the law may depend in part upon morality. But positivists can too, provided they maintain that law’s distinctive virtues can be identified without reference to morality.
C. Recovering Dworkin’s Obscured Insight
I believe Dworkin was committed to the idea that judges should interpret the law in light of law’s distinctive virtuesFootnote 82 but that he obscured this interesting fact by adopting a conception of the nature and virtues of law that is so moralized as to blur the distinction between the requirements of law and the requirements of political morality more broadly. If one accepts Dworkin’s moralized conception of law and its distinctive virtues, then when the lower-level details are fleshed out, the principled reconciliation I have sketched of conventionality and disagreement in the law will lead to a view of law more moralized than conventionalist. However, there are good reasons not to accept “law as integrity.” One of the most important is that it seems radically to underestimate the extent to which the rule of law depends on broadly conventional practice—and morally must, if it is to realize its distinctive virtues as a way of structuring society. The path I have opened makes it possible to recognize the conventionality of law more than Dworkin did while still making sense of the idea that judicial theoretical disagreement can be a reasonable, good faith response to the imperative of fidelity to law rather than necessarily a sign of confusion or bad faith.
To account for good faith theoretical disagreement, one need not accept a specific conception of law and its distinctive virtues—neither Dworkin’s or Waldron’s, nor Raz’s or Shapiro’s. To the contrary: theoretical disagreement appears all the more reasonable and likely to be in good faith on the assumption that conceptions of law such as theirs remain the subject of reasonable, good faith contestation among judges and lawyers no less than philosophers.Footnote 83 On that score, Dworkin was correct—whether or not he was also right in his substantive claim that “law as integrity” offers us the best account of law and its distinctive virtues (or in his methodological claim that in order to arrive upon that conception, we need to resort to political morality).Footnote 84
IV. Objections
Before concluding, let me briefly raise and respond to three salient objections to the reconciliation of Conventionality, Disagreement, and Fidelity I have articulated and tentatively defended.
Objection 1: The argument that the content of the law may depend on law’s distinctive virtues depends upon conflating ideal and non-ideal senses of law. Legal positivists have drawn several distinctions that might be thought to limit or preclude the dependence of the content of the law on the distinctive virtues of law in general. One is the distinction between functional law, which fulfills the functions or realizes the aims of law, and nonfunctional law, which fails to do so.Footnote 85 Unsurprisingly, positivists have frequently supposed that whether a rule or principle has the status of law does not depend upon whether it functions well qua law, or successfully realizes law’s distinctive virtues. Raz and Shapiro, for instance, insist that in order actually to be law, a putative law must be the sort of thing that in principle could fulfill its function of creating special reasons for action distinct from those provided by general moral principles—not necessarily a thing that actually does fulfill that function.Footnote 86 Positivists have similarly distinguished between the existence of law, in the sense of having rules and principles of a legal system, and the existence of the “rule of law,” in the sense of a legal system that lives up to law’s distinctive virtues.Footnote 87 And as Waldron has suggested, legal positivists have often thought that whether we have law, and what our law is, are questions independent of how successfully our legal system realizes the rule of law.Footnote 88 Finally, Raz has distinguished between law in the sense of instances of the concept of law and law in the sense of what passes for law in a particular legal system.Footnote 89 Raz has used this distinction to make a notable claim that stands opposed to the view I have defended in this article:
Judicial decisions in American courts are vulnerable to the charge that they are wrong as a matter of American law. But it is irrelevant to their justification that they conform, if we can make sense of the notion, with the correct theory of the nature of law….
… The point is that [judges’] duty (under the system in whose courts they sit) is to judge in accordance with the rules of that system, and it matters not at all whether these rules are legal ones.Footnote 90
Perhaps there is something correct in the claim that there is a sense in which a clock is no less a clock for the fact that it is broken, and a law no less a law for the fact that it fails to perform well the distinctive functions of law and realize its distinctive virtues.Footnote 91 Yet legal positivists often imply that fidelity to law always and only requires adhering to the laws, even when they do not yield unique determinate answers to legal questions and even when they fail to realize law’s distinctive virtues. I defend here only the limited claim that fidelity to law may sometimes require judges to supplement narrowly positive law, fixed by convergent legal practice but indeterminate within some range, by considering what legal resolutions to difficult cases would best promote the distinctive virtues of law. But it is also plausible, albeit more controversial, to believe that judges may sometimes be most faithful to law by resisting convergent practice when it involves the acceptance of hostility to the rule of law, in the form of, say, ad hoc decrees, bills of attainder, or ex post facto punishment. Fuller was on to something important when he claimed that judges need not reach outside law to morality—but can appeal to law itself—to resist those forms of injustice in the law brought about through “laws” profoundly at odds with the rule of law.Footnote 92 Finally, though I can accept Raz’s distinction between fidelity to law as such and fidelity to our law or the law, I cannot accept the use to which he puts it in the passage from which I have quoted. If there is a difference between the reasons for caring about judges’ fidelity to the law qua law and the reasons for caring about judges’ fidelity to the rule of law, it is extraordinarily subtle—enough so that its significance is beyond this author’s present comprehension. Why, then, think that judges have an obligation of fidelity only to the law, construed in narrowly positivist fashion, but not to the rule of law?
Objection 2: Perhaps law has no distinctive virtues. We have not been offered an account of what it would mean for a virtue to be “distinctive” to law. A second way to object to the view I have articulated would be to argue that law does not have distinctive virtues.Footnote 93 I believe that claim is false, and clearly so, but also that it merits several responses. Initially, it is important to notice that the idea that law has no distinctive virtues is compatible with the idea that theoretical disagreement about the law can be reasonable, good faith disagreement about law’s distinctive virtues, so long as it can be reasonable to believe that law does in fact have distinctive virtues. (That said, if a belief that law has distinctive virtues can be reasonable but is ultimately mistaken, then although they are not vulnerable to charges of bad faith, judges engaged in theoretical disagreement understood in the manner I have proposed are guilty of confusion—albeit reasonable confusion).
Still, the idea of virtues “distinctive” to law calls for clarification. Are law’s “distinctive” virtues merely those that are “intrinsic” to it, in the sense that law necessarily promotes them to at least some extent? To the contrary: I have chosen the term “distinctive” precisely to avoid a problem that arises when one focuses exclusively on law’s “intrinsic” virtues. As Frederick Schauer has pointed out, it is sometimes true that some of the qualities of a class of entities that are of greatest importance in understanding the class and distinguishing it from other classes are not exhibited by all entities in the class.Footnote 94 To appropriate his helpful example, though an animal’s ability to fly is neither necessary nor sufficient to make it a bird, we could not adequately describe what I would call the “distinctive” qualities of birds without mentioning their ability to fly.Footnote 95 I will not attempt here to provide a set of conditions necessary and sufficient to qualify a trait of an entity—such as a virtue of the law—as distinctive to it. I need not do so to make my point.
Objection 3: Fidelity to law’s distinctive virtues should not be equated with fidelity to law. Nevertheless, the way I have defined distinctive virtues may seem to give rise to a third objection to the idea that judges are more or less faithful to law depending upon the extent of their allegiance to law’s distinctive virtues. On the one hand, law does not necessarily realize its own apparently distinctive virtues and, on the other hand, its apparently distinctive virtues can be promoted by some means other than law. On the one hand, laws may be relatively ad hoc, unclear, and unpredictable. On the other hand, traditions and customs, for instance, might also promote such values as generality and predictability in governance without themselves being or becoming laws. One might wonder, therefore, whether we should say that when judges attempt to fill apparent gaps in the content of the law by reference to such virtues they are not only avoiding infidelity to law but attaining affirmative fidelity to law.
Yet should we say that promoting the distinctive virtues of an institution or practice is a means of being faithful to it only if those virtues are not only distinctive of the institution or practice but also uniquely realizable through it? I see no reason to adopt such a restricted view and good reason to reject it. Perhaps promoting the distinctive virtues of an institution or practice is an especially faithful act when those virtues cannot be realized other than through that institution or practice. But when we are concerned about whether and why we should be faithful to an institution or practice, we are concerned about whether and why we ought to follow it even when its dictates and guidance come into conflict with other reasons for action. And the reasons in such a case for faithfully following the institution or practice lie in the virtues that set it apart from salient alternatives—and not solely in those virtues that it claims exclusively (which after all may be none).
Conclusion
The aims of this article have been twofold: first, to place the problem of theoretical disagreement in a wider frame than the debate between legal positivists and their critics and, second, to begin to draw a novel and ecumenical view of theoretical disagreement out of an attempt to wrestle with it in that broader frame—a broader frame which consists in the tension among Conventionality, Disagreement, and Fidelity. I have argued that a plausible reconciliation of law’s conventionality and faithful theoretical disagreement about the law lies in the view that the content of the law is a function not only of contingent empirical facts about the convergent practice of legal officials such as judges but also of non-contingent conceptual truths about the nature of law and its distinctive virtues. Judges disagree about interpretive methods and principles not (or not only) because they disagree about what vague or ambiguous laws ought to mean, all things considered, but because they disagree about what such laws ought to mean in order that decisions made under them should uphold the rule of law.
