I. Introduction
The argument from theoretical disagreement (ATD) is a deep and unresolved objection to legal positivism, which canonically conceives of law as a conventional social practice. On Hart’s orthodox view, competent legal practitioners—paradigmatically, legal officials—conventionally recognize valid legal rules in common.Footnote 1 A familiar implication of this view is that practitioners’ shared recognition of valid law is guided by agreement on a set of substantive interpretive criteria—that is, by an interpretive convention specifying the grounds of law. Ronald Dworkin famously objected that this cannot be right, since judges and lawyers frequently and persistently disagree about which interpretive criteria ground legal validity.Footnote 2
The ATD presses this objection into a challenge to positivism’s conventionality thesis.Footnote 3 If law is grounded in shared interpretive criteria, then legal officials should agree on those criteria.Footnote 4 Yet in practice, disagreement over the grounds of law is ubiquitous and often deep, extending not only to which criteria apply but to what justifies their application.Footnote 5 The persistence of such disagreement undermines the claim that law is conventionally recognized by virtue of agreement on shared criteria of validity.
In this article, I argue that this conclusion rests on a misdescription of legal practice. When we correctly describe what practitioners are actually up to, the supposed problem largely dissolves. Valid law is conventionally recognized despite widespread theoretical disagreement because legal practice exhibits rationally interdependent attitudes in connection with the use of closure procedures, not because practitioners converge on substantive interpretive criteria. Rational interdependence here implies only that part of a practitioner’s reason for recognizing a law as valid is that others, acting in their institutional roles, have done so as well; no endorsement of their interpretive grounds or deference to their interpretive methodology is implied.
The article proceeds as follows. Section II clarifies how the problem of the grounds of law is practical rather than metaphysical and specifies a precise version of the ATD directed at positivism’s canonical conventionality thesis. Section III surveys and criticizes existing positivist responses to the ATD. Section IV advances a procedural account of legal conventions, showing how closure procedures, rather than substantive interpretive agreements, explain practitioners’ conventional recognition of valid law. Section V explores the implications of this account for legal reasoning and judicial dissent, arguing that it yields a more accurate and complete picture of legal practice than Dworkinian and existing positivist alternatives.
II. The grounds of law, orthodox positivism, and theoretical disagreement
What are the grounds of law? This deceptively simple question has long occupied the center of modern jurisprudence. Dworkin famously distinguished between “propositions of law”—statements of what the law permits, prohibits, or requires—and the “grounds of law,” i.e., considerations in virtue of which such propositions are true or false.Footnote 6 Hart similarly characterized legal validity in terms of “criteria of validity” for identifying a legal system’s primary rules.Footnote 7
Contemporary debates about the grounds of law often proceed on a metaphysical register. Here, the dispute concerns the ultimate determinants of legal validity—for example, whether the grounds of law ultimately consist only in social facts, as positivists maintain, or whether they necessarily incorporate moral considerations, as nonpositivists contend.Footnote 8 The ATD, however, draws its intuitive force not from this metaphysical conception of the grounds of law, but from a practical one. On this practical reading, the grounds of law consist in the interpretive criteria that legal practitioners advance in opinions, briefs, and judgments when arguing about the validity of legal rules.
These practical criteria—plain meaning, legislative intent, precedent, constitutional principle, rational purpose, and so on—constitute the everyday currency of legal argument. They are what judges and lawyers invoke, contest, and disagree about in practice. Philosophers, too, frequently debate their significance. Distinguishing this practical conception of the grounds of law from a metaphysical conception is essential because the ATD appears intractable only when the two are run together.
According to positivism writ large, legal validity is socially constructed. The orthodox view maintains a conventionality thesis according to which law is based on social convention, rather than morality, and more specifically, in agreement on criteria of validity. Framed in these terms, the ATD argues that if legal officials disagree about the grounds of law, then legal validity cannot be settled by convention; conversely, if legal validity is settled by convention, then officials cannot disagree about the grounds of law. Yet such disagreement appears to be widespread and persistent. The conclusion drawn is that positivism cannot be true.Footnote 9
On a metaphysical reading, this argument threatens to collapse into a familiar stalemate. Positivists and nonpositivists presuppose different conceptions of law’s nature, and their disagreement—largely underdetermined by legal practice—admits of no decisive resolution. On a practical reading, however, the problem is formidable but not intractable. Here, “grounds of law” and “criteria of validity” refer not to law’s ultimate truth-makers, but to the interpretive methodologies practitioners advance within institutional practice.Footnote 10
For example, consider the legal proposition that a party found to be fifty percent or more at fault for an accident cannot recover compensation. Some practitioners might argue that this proposition is true in virtue of the plain meaning of statutory text duly enacted by the legislature. Others might appeal to legislative intent or to the rational purposes of tort law. On the orthodox positivist picture, when legal officials agree on such grounds, the agreed grounds function as the legal system’s criteria of validity, guiding the conventional recognition of valid law.
Hart’s account of the rule of recognition embodies this orthodox picture. The rule of recognition is a social rule by which officials identify valid legal rules.Footnote 11 Because it is a “secondary rule”—fundamentally a conventional social practice among legal officials and not itself a primary legal rule—it is more accurately construed as a conventional practice of valid rule recognition than the somewhat misleadingly labeled “rule of recognition.”Footnote 12 On this view, a conventional practice of recognizing valid law exists if and only if competent practitioners converge on the criteria that constitute the relevant test for legal validity.Footnote 13
Dworkin’s formulation of the ATD presses precisely this point. If judges disagree about the content of the test for valid law—if they disagree about the grounds and criteria of validity—then he argues, it is not uncertain whether they have a conventional practice, it is rather certain that they do not have one.Footnote 14 The force of this argument derives from the practical, not the metaphysical, conception of the grounds of law.Footnote 15 Whatever the ultimate truth-makers of legal propositions may be, Dworkin’s claim is that legal practitioners cannot sustain a conventional practice while disagreeing about interpretive criteria.
Of course, not all legal disagreements are theoretical disagreements over the grounds of law. Legal practitioners may instead have empirical disagreements or application disputes—disputes about how to apply shared interpretive criteria in particular cases. In such disagreements, there is no dispute over the grounds of law. Rather, the disagreement concerns the proper application of a shared standard.Footnote 16
Bostock v. Clayton County provides a familiar illustration.Footnote 17 There, the justices disagreed about the truth of the legal proposition that it is unlawful under Title VII of the Civil Rights Act of 1964 for an employer to fire an employee for being homosexual or transgender. Yet all agreed that the relevant criterion of validity was the plain meaning of the statutory text. For the majority—but not the minority—the plain meaning of “sex” applied to sexual orientation and gender identity. There was, however, no substantial disagreement about the interpretive criterion itself.
Theoretical disagreement, by contrast, arises when legal practitioners disagree about the criteria of legal validity. Judges may dispute whether the plain meaning of a statute controls or whether to avoid the absurd results of its plain meaning. In such cases, disagreement does not concern how to apply a shared standard, but which standard governs. Practitioners may even agree on the plain meaning of a legal text while disagreeing about its relevance to legal validity.Footnote 18
These disagreements pose a challenge for orthodox positivism, which requires widespread agreement among legal officials on criteria of validity. On the orthodox picture, the content of these shared criteria determines which laws are valid and which are not.Footnote 19 Hence the view requires a substantive interpretive convention: absent agreement on the content of the criteria for assessing validity, there is no conventional “test” for valid law, and thus no valid law on the matter.
Because interpretive disagreements frequently occur when courts decide difficult or controversial cases, the account raises the problem of nonregulation. That is, it risks admitting that extensive reaches of legal practice are unregulated by valid law, a counterintuitive result.Footnote 20 Moreover, in the absence of agreement on criteria of validity, the account also raises the problem of error. Parties to a dispute cannot genuinely disagree about the validity of a law when there is no valid law about which to disagree—rather, they must be mistaken, incoherent, or disingenuous.Footnote 21
In response, orthodox positivism has tended to embrace what Dworkin called the “plain fact” view of the grounds of law, according to which theoretical disagreement is either illusory or rare, because judges and lawyers largely agree about the criteria of legal validity.Footnote 22 Traditionally, the view is committed to the following claims: (a) legal conventions require agreement on criteria of validity, (b) there is widespread agreement on such criteria, and (c) legal validity can be, and is, settled by convention. Dworkin rejects this picture, arguing instead that interpretive disagreement is widespread and that legal practice is therefore nonconventional.Footnote 23
A. The argument from theoretical disagreement
The ATD does not challenge positivism by disputing the metaphysical grounds of law. Rather, it targets positivism’s claim that legal validity is conventionally recognized in common by legal practitioners. What is ultimately at stake is the viability of legal positivism in the face of pervasive interpretive disagreement: if positivism requires practitioners, especially judges and officials, to agree on criteria of validity, persistent disagreement undermines its core claim.
Originally formulated by Dworkin and further articulated by Kenneth Himma, Scott Shapiro, and Dale Smith, the ATD is commonly reconstructed as follows:Footnote 24 (1) If positivism is true, then competent legal practitioners share criteria of legal validity—a social rule of recognition (of valid legal rules). (2) A social rule of recognition requires legal practitioners to agree on criteria of validity. (3) There is disagreement among legal practitioners on criteria of validity. (4) Therefore, positivism is false.
This formulation captures the challenge in broad outline. However, the ATD can be stated with greater precision by disambiguating several terms. Accordingly, I restate the argument to preserve its intuitive force while setting up the analysis to follow:
Premise 1: Positivism requires a conventional practice of valid rule recognition among competent legal practitioners.
Premise 2: A conventional practice of valid rule recognition requires widespread theoretical agreement among practitioners on interpretive criteria.
Premise 3: There is widespread theoretical disagreement among practitioners on interpretive criteria.
Conclusion: Positivism is false.
The ATD actually turns on a more specific version of Premise 2—capturing a particular instance of what that premise generalizes:
Premise 2*: Forming a legal opinion based on interpretive criteria over which there is active theoretical disagreement with other practitioners precludes simultaneously engaging with them in a conventional practice of valid rule recognition.
Positivists and nonpositivists alike accept Premise 2*. They disagree primarily about how pervasive theoretical disagreement is and about its implications. If such disagreement is widespread, as nonpositivists contend, then legal conventions appear impossible. If it is marginal or rare, as orthodox positivists often maintain, it is thought to pose little threat to the social conventionality of law.Footnote 25
Premise 2* is the true crux of the ATD. Its force does not derive from an intractable metaphysical dispute beyond resolution by evidence from legal practice, but rather in its reliance on strong evidence from legal practice. If correct, it leaves positivism with no convincing reply. Notably, Premise 2* itself rests on two further assumptions:
Premise A: Conventions require special agreements among the parties.
Premise B: Legal interpretation is the exclusive mode of valid legal rule recognition.
Together, these assumptions yield the conclusion that a conventional practice of valid rule recognition requires special agreement on interpretive criteria (i.e., “theoretical agreement”). In Section IV, I argue that both Premise A and Premise B are false. As a result, Premises 2* and 2 fail, and the ATD misdescribes how legal practice actually functions.
The appeal to “special agreement” in Premise A requires brief clarification. When both Dworkin and orthodox positivists maintain that legal conventions require agreement on interpretive criteria, at least two distinct senses of agreement are in play. In one sense, agreement consists in an explicit compact or deliberate arrangement on shared standards, as in a constitutional convention or other formally articulated settlement. In another sense, agreement is implicit, consisting in a convergence of interpretive methodologies across practitioners, not by compact but by a prior matching or synchronization of independent interpretive convictions.
Although positivists and nonpositivists disagree about whether implicit agreement suffices for convention, both assume that some form of special agreement—explicit or implicit—on interpretive criteria is a necessary condition for a conventional practice of valid rule recognition. It is this assumption that the account I develop in Section IV ultimately rejects. Before turning to that account, the next section surveys how orthodox and nonorthodox positivist theories have sought to accommodate or deflect the challenge posed by theoretical disagreement.
III. The orthodox positivist response: application and error
Orthodox positivism responds to the ATD by accepting Premise 2 while denying Premise 3. On this view, apparent disagreements over a legal system’s rule of recognition concern the application of shared criteria of validity rather than disagreement about their content.Footnote 26 If challenging disputes can be reduced to application disagreements, they pose little threat to positivism; genuine theoretical disagreement, if it occurs at all, would mark only marginal gaps in otherwise settled conventions.Footnote 27
This response faces several difficulties. First, it implies that in most cases of principled disagreement, the shared content of the rule of recognition must be overly broad and nonspecific to be the subject of special agreement between the disputants—something like a principle of interpreting the law “properly” or in the “right” or “best” way. Even if a convention with this shared content obtained, it would provide little or no support for positivism framed in such abstract moral terms.Footnote 28
Second, not all interpretive disagreements can be explained in terms of application disputes—instead, many entail applications of disparate criteria. For instance, in TVA v. Hill, a case emphasized by Dworkin, Justice Powell’s dissent disagreed with Justice Burger’s majority opinion over criteria of validity.Footnote 29 The majority argued that the plain meaning of the Endangered Species Act required upholding an injunction on completing an expensive dam project to save a small, endangered fish unless there was evidence that Congress did not intend that result. Call this the “plain meaning criterion.”Footnote 30 Powell’s dissenting opinion argued that the Act must accord with “some modicum of common sense and the public weal” and that an absurd result must be avoided unless there was evidence that Congress did intend the injunction. Call this the “absurd results criterion.”Footnote 31
Here, practitioners genuinely disagreed about how to understand valid law.Footnote 32 This was not an application dispute because they explicitly affirmed different criteria. So, Burger and Powell had no explicit agreement on grounds of law that could constitute a conventional practice of valid rule recognition. It could be argued that they implicitly applied the same criteria, i.e., that they were both analyzing the plain meaning, looking for empirical evidence of legislative intent, and weighing absurd results.Footnote 33 But not all cases of interpretive disagreement involve overlapping references to similar criteria such as legislative intent and, even in this case, the distinction between their criteria makes a difference.
Burger looked for (and found) the absence of legislative intent to subvert the plain meaning. Powell looked for (and could not find) the presence of legislative intent to embrace the plain meaning. Where the evidence of legislative intent is ambiguous, indeterminate, or missing, one of them will adopt the plain meaning and the other will avoid the absurd result. Though similar, they applied different criteria, with different results, to the same set of circumstances. So, they also did not have an implicit agreement on the grounds of law that could constitute a conventional practice of valid rule recognition. There was simply no special agreement between them about which interpretive criteria to apply.
Accordingly, although orthodox positivists plausibly contend that many disputes concern the application of shared criteria, they falter in claiming that lawyers and judges consistently or comprehensively agree on the grounds of law and that theoretical disagreement is rare or merely apparent. Taken as a whole, the reply underestimates the depth and persistence of methodological pluralism in legal practice.Footnote 34
Third, if these disagreements are not “application disputes,” the orthodox view struggles to acknowledge them as genuine disagreements about what the law is instead of what the law should be—in the absence of agreement on grounds of law, the law runs out. Because there is no valid law about which to disagree, judges who claim they are discovering, not making law, in such contexts are either incoherent, mistaken, or disingenuous. This kind of error theory, which results from accepting Premise 2* above, is the most common positivist account of theoretical disagreement.Footnote 35
The objection is that theoretical disagreement is not so much explained as explained away since orthodox positivism rejects taking such disputes at “face value.”Footnote 36 Moreover, describing a significant portion of argumentative legal practice as error, pretense, or confusion exacts a substantial cost. The more widespread theoretical disagreement is taken to be, the less plausible the orthodox response becomes. This does not decisively refute orthodox positivism, but it significantly weakens its explanatory appeal.
A. A nonorthodox positivist response: institutional design
Having seen where orthodox positivist accounts intervene, it is instructive to examine a prominent nonorthodox positivist response developed by Shapiro.Footnote 37 Shapiro accepts Premises 2, 2*, and 3 of the ATD, but denies Premise 1. He agrees that the frequency and persistence of interpretive disagreement show that legal practice is not conventional in the sense presupposed by orthodox positivism and emphasizes the institutional design of legal systems over social conventions. Importantly, Shapiro treats interpretive disagreements as genuine disputes that require explanation, not marginal anomalies to be explained away.Footnote 38
Before turning to Shapiro’s positive account, however, it is useful to clarify how deep the grounds of law are taken to be on the practical conception adopted here. On this conception, “grounds of law” and “criteria of validity” are synonyms: both designate the interpretive criteria that legal practitioners advance when arguing about the validity of legal rules. When a judge appeals to the plain meaning of statutory text, legislative intent, purposive rationales, precedent, or constitutional principle, she is advancing a ground of law in this practical sense.
Judges also, of course, offer arguments aimed at justifying the use of certain interpretive criteria over others. The question is whether such arguments themselves constitute theoretical disagreements about the grounds of law. Smith proposes a broader conception, distinguishing disputes about relatively “shallow” grounds—those corresponding to criteria of validity—from disputes about “deeper” grounds, which concern justifications for proper criteria and more general considerations about the nature of law.Footnote 39 By contrast, Shapiro distinguishes interpretive disputes, which directly concern which criteria or grounds apply, from meta-interpretive disputes, which concern the justification of interpretive methodologies.Footnote 40
On the practical conception of the grounds of law, meta-interpretive considerations are not themselves grounds of law. They may guide a practitioner’s choice of interpretive criteria, but they are not advanced to assess the truth or falsity of particular legal propositions. A textualist judge claims that a legal rule is valid because of the plain meaning of statutory text, not because of her conception of the judicial role in a democratic society.
At the same time, if a judge were to argue from the bench that laws prohibiting abortion are valid because it is morally right to prohibit abortion, or that laws permitting capital punishment are valid because it is morally permissible to impose the death penalty, or that legal validity is necessarily determined by moral principles, then such interpretive criteria would count among the practical grounds of law. In this sense, no matter the depth of practitioners’ convictions, the practical grounds of law remain “shallow.”Footnote 41
Shapiro’s response builds on the distinction between interpretive and meta-interpretive disagreement. He treats theoretical disagreements primarily as meta-interpretive disputes—disagreements about how interpretive criteria should be justified. Textualists, for example, may justify their methodology by appeal to majoritarian values and the importance of limiting judicial discretion to depart from the plain meaning of democratically enacted text.Footnote 42 Purposivists, by contrast, may invoke counter-majoritarian considerations, arguing that the judicial role in pluralistic societies requires discretion to elevate constitutional principles protecting minorities.
While these disagreements concern institutional design, they can have downstream consequences for legal interpretation and judgment. A position concerning the proper role of judges in a democratic society is neither itself a proposition of law nor a method of interpretation, but it may justify reliance on one interpretive criterion rather than another. Meta-interpretive considerations thus help explain why interpretive disagreement persists.
Shapiro’s account remains positivist because he maintains that empirical social facts ultimately settle which interpretive methodologies best harmonize with a legal system’s institutional design. While practitioners may genuinely disagree about which interpretive methodologies their legal system presupposes, Shapiro argues that there are ultimately some wrong answers to these questions, determined by the “animating ideology” of the system’s “master plan.”Footnote 43 On this basis, he criticizes both Dworkin’s constructive interpretation and Posner’s pragmatism as improper interpretive methodologies within the American legal system.Footnote 44
Shapiro’s account, however, leaves unexplained an important range of interpretive disagreements found in legal practice. Not all such disagreements presuppose disputes about how to justify interpretive criteria.Footnote 45 In general, employing an interpretive methodology does not require articulating or even working out a justification of why it is the best method relative to other methods. So, while some interpretive disagreements may be traceable to meta-interpretive disputes, many are not. Fundamentally, they are disagreements about which criteria of legal validity apply, not merely about how to apply shared criteria or how to justify disparate ones.
The alternative pursued here accepts the prevalence of theoretical disagreement but denies Premise 2 of the ATD by rejecting Premises A, B, and 2*. It denies that a special agreement on interpretive criteria is required for a conventional practice of valid rule recognition.Footnote 46 What this denial amounts to is not a revisionary proposal for how legal practitioners might reconceive their disagreements or change their attitudes toward alternate interpretive methodologies, but a descriptive claim about how legal practice already functions.
Dworkin’s ATD and existing positivist responses presuppose that special agreement on interpretive criteria is the sole mechanism by which practitioners might conventionally recognize valid law in common. In Section IV, I offer a more perspicuous reconstruction of legal practice. Practitioners in fact conventionally recognize valid law through the use of closure procedures rather than interpretive consensus, and by way of rationally interdependent attitudes rather than explicit or implicit agreement. On this view, persistent theoretical disagreement is not an anomaly for positivism to explain away, but a familiar and tractable feature of conventional legal practice once that practice is properly understood.
IV. Closure procedures and the grounds of law
Existing positivist and nonpositivist accounts alike assume that theoretical agreement on interpretive criteria is required for a conventional practice of recognizing valid law. This assumption yields a puzzling picture of legal practice: it becomes difficult to see how jurists could ever recognize as valid a law that has been judged valid on the basis of an interpretation they reject. Jurists do take strongly principled stands on matters of interpretation, particularly in controversial cases, but focusing exclusively on their interpretive attitudes presents an incomplete picture of legal practice. What it overlooks is the frequency with which jurists recognize valid law by way of closure procedures rather than by their own best-lights interpretation.
This phenomenon is especially salient in cases involving concurring and dissenting opinions. In many such cases, interpretive dissensus is so deep that there is no single interpretive rationale commanding majority support, yet none of this prevents judges—including dissenters—from recognizing the resulting holding as valid law.
Take, for example, the following list of judicial opinions filed in Parents Involved in Community Schools v. Seattle School Dist. No. 1, which is not unusual: ROBERTS, C. J., announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I, II, III-A, and III-C, in which SCALIA, KENNEDY, THOMAS, and ALITO, JJ., joined, and an opinion with respect to Parts III-B and IV, in which SCALIA, THOMAS, and ALITO, JJ., joined. THOMAS, J., filed a concurring opinion; KENNEDY, J., filed an opinion concurring in part and concurring in the judgment. STEVENS, J., filed a dissenting opinion. BREYER, J., filed a dissenting opinion, in which STEVENS, SOUTER, and GINSBURG, JJ., joined.Footnote 47
In this case, there was a majority opinion only with respect to Parts I, II, III-A, and III-C. Justice Kennedy was the decisive vote concurring in the judgment invalidating the school districts’ student assignment plans (aimed at racial integration of public schools) as violative of the Equal Protection Clause of the Fourteenth Amendment. While the Court invalidated the student assignment plans by a 5–4 vote, no single interpretive rationale united the justices concurring in the judgment. A plurality applied an anticlassification rationale; dissenters favored an antisubordination rationale; Justice Kennedy, whose vote was decisive, favored what has been called an antibalkanization rationale.Footnote 48 These represent divergent convictions about the grounds of law and thus exemplify the kind of theoretical disagreement emphasized by the ATD.
Nonetheless, the plans were invalidated as the outcome of a closure procedure, and all justices—including dissenters—recognized that outcome as settling the law, for better or worse. Breyer’s dissent, for example, states: “To invalidate the plans under review is to threaten the promise of Brown … This is a decision that the Court and the Nation will come to regret.”Footnote 49 Judges in fact routinely recognize legal validity even while rejecting the interpretive grounds on which a decision rests. As Justice Brandeis famously remarked, “it is more important that the applicable rule of law be settled than that it be settled right.”Footnote 50
A tempting explanation of the phenomenon, along orthodox positivist lines, is to construe the courts’ closure procedures themselves as the grounds of law upon which practitioners, lawyers, and judges all agree. On this view, disputes over interpretive criteria do not constitute genuine theoretical disagreements because they are disputes about what the courts should do, not about what the courts can do, i.e., not about the courts’ authority to produce valid holdings through their closure procedures.Footnote 51 Only if practitioners were to dispute what the courts, in this sense, had the authority to do (i.e. produce valid holdings) would they have a genuine theoretical disagreement over the grounds of law—a kind of dispute that is arguably much rarer.
Himma, for example, suggests that judges share a “principle of constitutional interpretation” that treats divergent interpretations of the constitution by the Supreme Court as all resulting in valid law.Footnote 52 This formulation, however, mischaracterizes the practice. The “principle” that Himma identifies is not, strictly speaking, a principle of constitutional interpretation but a statement about judicial discretion: a description of what courts are permitted to do.Footnote 53 Yet, neither the courts’ authority to produce valid holdings through closure procedures, nor judicial discretion to employ different interpretive criteria while executing them, themselves constitute interpretive criteria in the sense of the practical grounds of law introduced in Section II. A more plausible hypothesis is that such discretion and authority are normative incidents attaching to an accepted judicial role—one that also carries a responsibility to recognize the outcomes of the court’s closure procedures as valid.Footnote 54
If the court’s authority were itself an interpretive criterion, we would expect judges to argue that a legal proposition is true or false because the court has authority to say so.Footnote 55 Yet this is not what we observe. Consider, for instance, NLRB v. Noel Canning,Footnote 56 a case concerning recess appointments in which “the Court was unanimous on the judgment but divided 5–4 on the most fundamental questions.”Footnote 57 The majority relied on the rational purpose of the Recess Appointments Clause, while the minority relied on its plain meaning. Although all justices clearly accepted their own interpretive discretion and the court’s authority to invalidate the appointments, none argued that the appointments were invalid merely because the court possessed such discretion and authority.
Executing a closure procedure, then, is not equivalent to applying an interpretive criterion, and recognizing the outcome of such a procedure as valid is not itself an act of interpretation. It is therefore a mistake to infer from jurists’ common recognition of procedural outcomes that they agree on a shared set of interpretive criteria. Such inferences misdescribe the common recognition of valid law as entailing theoretical agreement on its grounds.
The puzzle is resolved by rejecting the mistaken assumption that recognizing valid law in common requires consensus on grounds of law. Judges on the same panel frequently endorse disparate interpretive criteria when judging the same law valid, and in dissent, they often criticize as incorrect the grounds of a holding they nonetheless recognize as valid. Principled interpretive disagreement does not inhibit their participation in a conventional practice recognizing valid law. As the opinions in Parents Involved, Noel Canning, and similar cases suggest, legal practice involves distinct activity types: (1) interpreting what grounds valid law, (2) judging the validity of particular laws, and (3) recognizing which laws are valid within the system. Closure procedures operate on the third level. Once this is appreciated, the persistence of theoretical disagreement no longer threatens the conventionality of law.
A. Faces of convention: interpretive agreements versus procedural conventions
The preceding section provides strong reasons to reject a picture of legal practice according to which practitioners can conventionally recognize valid law only by converging on substantive interpretive criteria. Why, then, has that picture proven so attractive to positivists and nonpositivists alike? The answer, I suggest, lies in the appeal of two assumptions introduced in Section II.A:
Premise A: Conventions require special agreements between the parties.
Premise B: Legal interpretation is the exclusive mode of valid legal rule recognition.
Taken together, these premises support the conclusion that a conventional practice of valid rule recognition requires special theoretical agreement on interpretive criteria. Dworkin, for instance, accepted Premise A, holding that the only plausible way to understand the conventional development of law would be to posit the existence of “special mini-constitutional conventions” or “special agreements to have a new set of conventions.” Because such special agreements are absent from ordinary appellate and Supreme Court judicial practice, he concluded that the law is not a conventional practice.Footnote 58
However, Premise A is false. There exist numerous and varied examples of conventional practices that do not depend on special agreement, either in the sense of an explicit compact or a prior synchronization.Footnote 59 In fact, as both Dworkin and Hart acknowledged in their debate, the defining feature of a conventional practice is that part of the reason one does something is that others do as well, whereas in a merely concurrent practice one acts in the same way as others on the basis of independent conviction.Footnote 60 But individuals may act in part because others do—and thus behave conventionally in this respect—without prior synchronization and without entering into a special agreement to do so.Footnote 61 Instead, they may exhibit rationally interdependent attitudes, such as a disposition to recognize the validity of laws judged valid by others. This kind of uptake-dependent recognition requires no special agreement—it requires only recognizing something as having a certain status because others have taken it to have that status.Footnote 62
Rational interdependence here is not a normatively loaded notion. It does not imply that it is reasonable for judges to accept the outcome of closure procedures or to defer to alternative interpretive methodologies. It means only that part of the reason a practitioner recognizes a law as valid is that others have recognized it as such. Legal practitioners do not generally exhibit rational interdependence with respect to interpretive methodologies. They do, however, exhibit such interdependence with respect to recognizing the outcome of closure procedures.
One might wonder whether such patterns of uptake-dependent recognition amount to a form of implicit agreement—either on closure procedures or on the validity of their outcomes. For present purposes, however, nothing turns on this classification, and it is less important how procedures are adopted than how their use structures recognition. For instance, it is orthogonal to the present claim whether convergence on closure procedures occurs by explicit compact, implicit synchronization (via prior training or acceptance of institutional roles), or by rationally interdependent attitudes (as when their contested scope and jurisdiction are settled by their very use).Footnote 63 The agreement presupposed by both Dworkin and orthodox positivists is agreement—explicit or implicit—on interpretive criteria. The account offered here shows that the use of closure procedures generates rationally interdependent recognition of legal validity without any such consensus on grounds of law.
Even once Premise A is rejected, many theorists remain committed to Premise B—the assumption that recognizing legal validity is fundamentally an act of interpretation. But recognizing the outcome of a closure procedure as valid is not only a form of uptake-dependent recognition that requires no special agreement but it may also arise from commitment to an institutional role rather than from independent interpretive conviction.
Consider, for example, Justice White’s dissent in Roe v. Wade, where he acknowledges that “the Court perhaps has authority to do what it does today” while criticizing its judgment as “an improvident and extravagant exercise of the power of judicial review.”Footnote 64 It would be a stretch to suggest that Justice White’s recognition that the majority’s decision results in valid law rests on any independent interpretive conviction—or even on an act of interpretation at all. He recognizes the holding as valid (in part, or even entirely, because the Court’s majority has judged it valid through the exercise of its closure procedure) while he simultaneously criticizes the proffered grounds as an incorrect interpretation.Footnote 65
The most coherent explanation of such recognition is not interpretive agreement, but commitment to an institutional role. That commitment sustains a procedure-based convention of interdependent attitudes and behavior, rather than reflecting participation in a substantive interpretive convention understood as a special theoretical agreement on interpretive criteria.Footnote 66 There is therefore no need to suppose that this practice constitutes an interpretive “criterion,” “test,” or “rule” specifying how practitioners or officials must understand valid law, nor any need to adopt a “plain fact” view on which theoretical disagreement is mainly illusory or rare.
Once Premises A and B are rejected, Premises 2* and 2 of the ATD fall with them. A procedural account of legal conventions thus provides a more accurate reconstruction of legal practice than either the Dworkinian picture or orthodox positivist accounts that model legal conventions on substantive interpretive agreement.
V. The new problem of error
The procedural view of legal conventions relies on a distinction between legal interpretation, judgment, and valid rule recognition as separate though often overlapping activity types. This distinction is supported by familiar features of legal practice in which interpretation, judgment, and recognition come apart. In Bostock, an application dispute, the justices shared interpretive criteria but diverged in judgment. In Noel Canning, they diverged on interpretive criteria but converged in judgment. In Parents Involved, Roe v. Wade, and countless other cases, dissenters diverge from the controlling opinion both in judgment and interpretive criteria while nonetheless converging on recognition of valid law as the outcome of the court’s closure procedures.
While the procedural view distinguishes recognition from interpretation and judgment, it remains consistent with the idea that the primary aim of judicial activity is the correct interpretation of the law. What it denies is that this aim structures every act of recognition. The view is therefore compatible, from the internal point of view of judges, with any interpretive methodology they may have, and poses no contradiction between recognizing the current validity of a legal rule as the outcome of a prior decision procedure and subsequently judging it invalid.Footnote 67 The question is whether this picture offers a more intelligible account of legal practice than orthodox positivist and nonpositivist alternatives.
Orthodox positivism holds that legal validity is settled by agreement among practitioners on interpretive criteria. Because such agreement does not exist, the view faces the problem of nonregulation: large portions of legal practice appear ungoverned by valid law. The procedural view avoids this difficulty by explaining how conventional recognition of valid law occurs despite pervasive interpretive disagreement. Yet it appears to generate a different, and more subtle, problem of error. Practitioners are not mistaken because there is no valid law about which to disagree; rather, they risk error, incoherence, or pretense because their interpretive claims about what grounds valid law conflict with their practices of valid rule recognition.
The new problem of error is two-fold. First, by recognizing as valid the outcome of closure procedures while repudiating the interpretive grounds on which those outcomes rest, practitioners appear to accept the fruit of a poisonous tree with one hand while wielding an axe at its trunk with the other. This looks incoherent. Dworkin’s picture avoids this tension by denying that practitioners genuinely accept the fruit, focusing instead on the hand with the axe. But this “protestant view” of legal practice is incomplete, tracking only practitioners’ interpretive attitudes while papering over their sober recognition of contrary outcomes.Footnote 68 Orthodox positivism avoids the tension by denying that the tree is poisonous at all, misrepresenting deference to outcomes as interpretive convictions.Footnote 69 Both pictures are inaccurate.
Second, and more pointedly, on the procedural view it appears that practitioners’ interpretive criteria cannot be what determine legal validity, since closure procedures perform that function instead. If so, when judges claim that a legal rule is valid because of its plain meaning, its purpose, or legislative intent, they appear mistaken, incoherent, or disingenuous if they also accept that legal validity is settled by procedural outcomes rather than by their preferred interpretive grounds. How can we best make sense of these practices?
A. The content of judicial claims about validity
Judicial opinions—including concurring and dissenting opinions—are often treated as straightforward attempts to state what presently makes law valid. That assumption, however, obscures the variety of claims judges make when offering reasons for how to decide a case. Once the procedural account of legal conventions is in view, it becomes clear that judicial reasoning aimed at influencing or justifying the exercise of closure procedures takes several distinct forms.
When judges offer reasons to decide a case in a particular way—where the final decision itself will determine valid law—their claims fall into four categories. First, judges sometimes make present-validity claims, asserting that certain reasons ground what the valid law is. Second, they make normative claims, urging that certain reasons show what the valid law should be. Third, they make historical claims, arguing that certain reasons ground what the valid law has been. Fourth, they make predictive claims, contending that certain reasons ground what the valid law will be because a higher or future court is likely to adopt them. These categories are not mutually exclusive, and judicial opinions often blend them.
The most philosophically problematic category is the first. Present-validity claims appear to conflict with judges’ procedural commitments. If valid law is determined by closure procedures, then a dissenting justice who asserts that a legal rule is invalid despite the majority’s decision appears to be making a false, incoherent, or disingenuous claim—especially when that same justice accepts the court’s judgment as settling the law. But this problem arises only if claims about validity made in dissent are read exclusively as claims about what presently grounds valid law. That reading is neither mandatory nor always charitable.
For instance, judges often use explicitly normative language in their claims about validity. They say things like “judges should not use the due process formula … to invalidate legislation offensive to their personal preferences.”Footnote 70 With the procedural account in view, these claims are about the kinds of reasons that ought or ought not to be used to justify the outcome of the court’s decision procedures.
Judges also make explicitly historical claims about validity, describing the state of the law just prior to the court’s decision. The dissenters in Dobbs, for instance, explicitly decry the overturning of longstanding precedent when they maintain that “Roe and Casey have been the law of the land for decades” but that the court now “reverses course” and overrules them because the “composition of the court has changed”Footnote 71 and it now “has the votes to discard them.”Footnote 72 Even while mounting a vigorous defense of precedent with the interpretive principle of stare decisis, they do not pretend that Roe and Casey are still valid law. They lament that millions of women “have today lost a fundamental constitutional protection.”Footnote 73 With the procedural account in view, their claims are about the kinds of reasons that justify the outcomes of the court’s prior exercise of its decision procedures.
Finally, judges frequently make explicitly predictive claims about validity, describing the state of the law to come.Footnote 74 Strongly worded dissents, especially those joined by a substantial plurality of justices, often carry precisely this anticipatory tone that the court’s current decision will be overruled by the intelligence of a higher or future court. Lower courts naturally take this predictive posture toward higher courts more often but recall that Breyer’s dissent in Parents Involved predicts that the “decision is one the Court and the Nation will come to regret.”Footnote 75 With the procedural account in view, interpretive claims in this context are about the kinds of reasons that will be used to justify the outcomes of the court’s future exercise of its decision procedures, whether or not those predictions come to pass.
Dissents of this kind do not deny the current state of valid law. Similar patterns recur throughout dissenting opinions, where judges often speak in the idiom of present validity while advancing claims that are better understood as normative, historical, or predictive.Footnote 76 Once the persuasive role of these claims is appreciated, the new problem of error is largely mitigated.Footnote 77 Judges need not be understood as asserting propositions they simultaneously reject in practice. Instead, they participate in a conventional argumentative practice in which the idiom of present validity serves as a persuasive rhetorical vehicle for claims about how closure procedures ought to be exercised, how they have been exercised, or how they are likely to be exercised in the future—in particular about the types of reasons that do or do not justify procedural outcomes.Footnote 78
But the problem is not merely that practitioners sometimes speak in the idiom of present validity while meaning something normative, predictive, or historical. It runs deeper. In Section II, the practical grounds of law—plain meaning, legislative intent, rational purpose, precedent, constitutional principle, etc.—were introduced as considerations in virtue of which practitioners argue that propositions of law are true or false and legal rules valid or invalid. An interpretive methodology just is the systematic deployment of such interpretive criteria in the service of correctly interpreting the law. But if closure procedures determine legal validity rather than interpretive considerations, then practitioners who advance interpretive criteria as what makes law valid appear to be asserting precisely the wrong kind of thing—not merely rhetorically, but constitutively.
Resolving the apparent incoherence requires a friendly reframing of the practical grounds of law. Interpretive criteria need not be considerations in virtue of which practitioners argue that legal propositions are literally true or legal rules literally valid. Instead, they are considerations in virtue of which procedural outcomes are justified or unjustified—the kinds of reasons that in general best justify the exercise and outcomes of closure procedures. With this insight, the apparent conflict largely dissolves: interpretive criteria and procedural validity no longer compete. For practitioners, validity can be determined by closure procedures while the practical grounds of law operate as considerations in virtue of which procedural outcomes are better or worse justified.
On Dworkin’s “protestant view,” interpretive disagreement precludes the conventional recognition of valid law, implying that dissenting judges must deny that certain legal outcomes are valid at all. Taken seriously, this view entails a systematic misrepresentation of legal practice in which dissenting judges in fact routinely acknowledge and act upon the settled state of the law despite their interpretive objections. Orthodox positivism avoids this conclusion only by insisting—contrary to extensive evidence of methodological pluralism—that there exists a substantive omnibus interpretive convention.Footnote 79 The cost is that large domains of legal practice are rendered ungoverned by valid law whenever interpretive agreement is absent.
By contrast, the procedural view retains greater fidelity to legal practice. By locating the relevant form of conventionality in the use of closure procedures rather than interpretive consensus, it captures how practitioners recognize valid law in common while sustaining principled disagreement about its grounds. Interpretive agreement may itself be conventional in many contexts; however, its scope is neither universal nor explanatory of the conventional recognition of legal validity amid pervasive theoretical disagreement. Moreover, by charitably reconstructing the content of judicial claims about validity, the procedural view explains how practitioners can participate coherently in legal practice without abandoning their interpretive convictions or misrepresenting the current state of the law.
VI. Conclusion
The argument from theoretical disagreement (ATD) challenges the very possibility of legal conventions. Nonpositivists contend that if legal practitioners—especially judges—disagree about the grounds of law, then legal validity cannot be settled by convention. This poses a serious difficulty for orthodox positivism, which traditionally requires broad agreement among officials on criteria of validity. This yields the twin problems of nonregulation—large areas of legal practice unregulated by valid law—and error—that practitioners cannot genuinely and coherently disagree about what the law is when there is no valid law to disagree about.
Both orthodox positivists and nonpositivist critics, however, share the assumption that legal conventions must rest on special interpretive agreements. This article has argued that this assumption is mistaken. Legal conventions need not take the form of substantive interpretive consensus. They can instead consist in rationally interdependent attitudes with respect to closure procedures through which practitioners recognize the outcomes of common decision-making processes as valid law. On this procedural view, conventionality is located not in agreement on interpretive methodology, but in the common recognition of procedural outcomes. This reconstruction resolves the problem of nonregulation by explaining how legal validity is conventionally recognized despite pervasive theoretical disagreement.
The problem of error proves more subtle. Practitioners do not err because there is no valid law to disagree about. Rather, tension arises when their interpretive claims about what grounds legal validity come apart from their practice of recognizing valid law as settled by closure procedures. Read uncharitably, this might suggest incoherence or insincerity. A more charitable reading treats many judicial claims about validity as hybrid claims, combining present-validity idioms with normative, historical, and predictive content.
Law is a complex institutional practice in which the judicial role involves both (1) arguing, through principled interpretation, to shape the exercise of decision procedures and (2) recognizing the various outcomes of those procedures as valid. By distinguishing acts of interpretation, judgment, and recognition, the procedural view clarifies how law can be both a conventional and an interpretive social practice. It preserves what is most compelling in both positivist and nonpositivist insights while avoiding their shared misdescription of legal practice.
Acknowledgements
I would like to thank Michael Bratman, Leif Wenar, Wendy Salkin, Bernadette Meyler, Ralph Richard Banks, and Richard Thompson Ford for their extremely helpful comments on earlier drafts of this paper. I am also grateful to an anonymous referee at Legal Theory for detailed and constructive feedback that significantly improved the article. Earlier versions of this paper were presented in 2025 at the Legal Studies Workshop at Stanford University and the Legal Philosophy Workshop at Amsterdam Law School. I am thankful to participants for the engaging discussions, particularly Elliot Setzer and Fabian Corver.
Competing interests
The author declares none.
Open access
