Rights, Mini-Publics, and Judicial Review

Abstract

Landmark Supreme Court rulings determine American law by adjudicating among competing reasonable interpretations of basic political rights. Jeremy Waldron argues that this practice is democratically illegitimate because what determines the content of basic rights is a bare majority vote of an unelected, democratically unaccountable, elitist body of nine judges. I argue that Waldron's democratic critique of judicial review has implications for real-world reform, but not the implications he thinks it has. He argues that systems of legislative supremacy over the judiciary are democratically preferable to the American one. I provide reasons why his argument is unsound and explain that, properly construed, Waldron's premises support implementing a system where ordinary citizens chosen by lottery participate in a deliberative mini-public to vote on which reasonable interpretation of a basic political right will become the law of the land.

Introduction

According to legal theorist John Ely (1980: 4–5), the ‘central problem’ of constitutional judicial review is that ‘a body that is not elected or otherwise politically responsible in any significant way is telling the people's elected representatives that they cannot govern as they'd like.’ Defenses of the practice show how judicial review is a crucial bulwark against majoritarian infringements on minority rights. But these defenses do not justify judicial review as a procedure for interpreting basic political rights where the contents of those rights are subject to reasonable disagreement. In these cases, there are multiple possible rulings that plausibly respect legal precedent and interpret basic rights in light of liberal principles of justice (Waldron 2014: 1727). A rough criterion that a ruling is subject to reasonable disagreement is that the court issues a divided ruling, decided by bare majority vote, where ordinary citizens are deeply riven over whether the decision is correct. Ordinary citizens respond so strongly because the court is perceived as ruling not on some legal technicality but on matters of basic justice. Thus, some divided rulings, like Wal-Mart v. Dukes et al. (5–4), turn on legal issues that promote little popular discussion. Others, such as Citizens United v. FEC (5–4) or Roe v. Wade (7–2), result in prolonged and widespread discussion and protest. The present analysis concerns the democratic legitimacy of these latter kinds of cases, where matters of basic justice are decided for the polity by an elitist body whose membership is determined by arbitrary factors such as whether a previous justice retires or dies shortly before or after a presidential election—for example, when conservative Amy Coney Barrett swiftly replaced Ruth Bader Ginsburg during the final months of Donald Trump's presidency.

This paper evaluates Jeremy Waldron's influential formulation of the charge that judicial review is democratically illegitimate. I show that although many of Waldron's premises survive scrutiny, his conclusion does not. He concludes that legal systems that grant elected legislatures authority over the judiciary are democratically more legitimate than the American system of judicial supremacy for deciding among reasonable interpretations of basic political rights. I argue that Waldron's own premises lead elsewhere: that rather than instituting a system of legislative supremacy, legal systems like that of the United States should, in some circumstances, replace the vote by nine members of the court with majority decision by a randomly selected body of ordinary citizens. This ‘lottocratic’ proposal would transfer the court's authoritative power over issues subject to reasonable disagreement to a deliberative mini-public representing the diversity of informed reasonable views citizens have in the wider polity. Here, the value of diversity resides in its contribution to the legitimacy of democratic procedures when agreed-upon criteria for assessing a single outcome as uniquely correct are not available.

The paper proceeds as follows. In section 1, I spell out Waldron's argument against judicial review in what I take to be its most defensible form. That form is modest in its aims, not critiquing the practice of judicial review per se, but rather arguing for a principle of legislative deference on issues where the only way to adjudicate among competing reasonable interpretations is through a bare-majority vote. In section 2, I show that this principle of legislative deference does not follow from Waldron's egalitarian critique of judicial review, as elected legislators—even idealized representatives who care only about justice and operate within well-functioning institutions—have moral reason to set aside their individual judgments of which interpretation of a basic political right is uniquely correct, instead adopting and advancing a party platform. I argue in section 3 that the institutional remedy for this problem is to implement the lottocratic proposal described above. I conclude in section 4. The present analysis has two main aims. First, it shows how institutional facts, such as the tendency of elected representatives to form political parties, is relevant to the democratic critique of judicial review even in the idealizing conditions Waldron assumes. Second, it shows how Waldron's challenge to judicial review, despite being unsound, nevertheless has real-world implications for institutional reform.

1. Waldron's Challenge from Reasonable Disagreement

Waldron develops and defends his critique of constitutional judicial review across several papers and books. That critique takes stronger and weaker forms: he can be read as attacking the democratic legitimacy of the practice of judicial review itself or as more modestly defending a doctrine of judicial restraint that holds much of judicial review to be legitimate. And while his argument always highlights the persistence of disagreement about justice in free societies, sometimes his focus is on reasonable disagreement, and sometimes it also includes illiberal, irrational disagreement. The version of Waldron's challenge I focus on is both modest in its aims—viewing much of judicial review as democratically legitimate—and presupposes citizens’ disagreements are reasonable, in a sense I will explain. I depart from Enoch (2006: 24) who thinks nothing in Waldron's critique depends on the reasonableness of disagreement about justice. I start by presenting that challenge reconstructed in full and then proceed to explain its key moves.

1. 1. Basic political rights require interpretation.

2. 2. Some basic political rights admit of a range of reasonable interpretations, i ₁ . . . i _N.

3. 3. There is no appropriate impartial way to adjudicate among these competing reasonable interpretations except through some sort of vote.

4. 4. One voting system, V ₁, is more legitimate than another voting system, V ₂, for adjudicating the dispute among reasonable interpretations when V ₁ better satisfies the principle of equal judgment than V ₂.

5. 5. Majority decision of elected representative assemblies better satisfies the principle of equal judgment than majority decision of unelected judicial bodies.

   ∴ Majority decision of elected representative assemblies is more legitimate than majority decision of unelected judicial bodies for deciding which reasonable interpretation of a political right will be binding on all.

In what follows, I explain the concepts of reasonable interpretation of political rights, the principle of equal judgment, and legitimacy. I will ultimately endorse premises (1)–(4) but reject (5) and the conclusion in the next section.

Legal systems have evolving divisions of labor between legislatures and the judiciary for interpreting the law. This need for interpretation arises from various factors, including the vagueness of statutes or the emergent clash of legal principles. Interpretations resolve a general right into a series of specific Hohfeldian incidents, including permission-rights to engage in specific actions and claim-rights against interference. For example, a general right to speech requires interpretation to specify what is legally prosecutable hate speech or seditious speech. There is a very fine line between the interpretation of existing law to meet pressing needs and the effective creation of new law through interpretive acts. The American Supreme Court's rulings on the constitutionality of existing law often straddle this line. While the court is an unelected body and so does not represent ordinary citizens’ interests, the practice of judicial review, as Kramer (2012: 625) argues, is generally recognized as an institution that preserves the authority of ‘the People’ and makes decisions on its behalf. But this authority becomes questionable when the court issues verdicts on the constitutionality of existing law—which always involve acts of interpretation—that are divisive even among members of the court. In instances where the court itself is deeply divided, it becomes unclear whether the appropriate division of interpretative labor between legislatures and the judiciary should place deciding power with an unelected, small body of judges.

Deep division in the court often reflects conflicts among reasonable interpretations of political rights. According to Dworkin's (1986) influential theory of judicial interpretation, judges inevitably rely on a theory of justice to guide their interpretation of existing bodies of law when considerations of precedent are not decisive. Disagreements about justice are part of what Waldron (1999: 160) calls the ‘circumstances of politics’, much as modest scarcity and reasonable pluralism are a part of what Rawls (2005: 66) calls the ‘circumstances of justice’. Citizens, including judges, can reach conflicting judgments about the requirements of justice even when each blamelessly reasons for an impartial perspective. Joint deliberation, where discussants can raise challenges to others’ views, does not guarantee achieving consensus on these matters. Disagreement can survive considered discussion even though all participants are deeply committed to liberal values, have reached conclusions about the correct interpretation of political rights on the basis of blameless reasoning, and have duly considered serious objections to their views. There is reasonable disagreement about the proper interpretation of basic political rights when discussion satisfies these conditions. I will call what results from this discussion the set of reasonable interpretations of basic political rights, referring to that set as i ₁ . . . i _N.

My characterization of the reasonable set involves a slight departure from Waldron's own argument, which I now clarify and defend. Waldron (2006: 1360) is concerned with reasonable disagreements that arise in well-functioning judicial and democratic institutions, where ordinary citizens care about minority rights. I insist, whereas Waldron does not, that reasonable disagreement arises when citizens have blamelessly reasoned from an impartial perspective to their conclusion regarding which member of the set i ₁ . . . i _N is correct. To do this, citizens must satisfy informational, inferential, and deliberative conditions. Borrowing some conditions from Gaus (2011: 236–37), I highlight the following requirements. Citizens must have all the information they need to determine which interpretation of basic rights is correct; they must make valid inferences from their initially held set of beliefs; and they must have subjected those initial beliefs to scrutiny, considering a range of objections from peers who reject those beliefs. These conditions can be filled out in many ways. Satisfying them ensures that disagreements about the correct interpretation of basic political rights are not due to irrationality, lack of information, or lack of engagement with one's disagreeing peers. In short, we will call this the deliberative assumption, and citizens who satisfy it deliberative citizens.

This deliberative assumption helps bring Waldron's challenge to judicial review into sharpest relief. We are concerned with legitimacy challenges that arise when disagreements about justice and basic rights persist even under optimal conditions for achieving consensus. Waldron's challenge is not simply that a small, unelected body of judges gives unequal say to the range of views endorsed by the irrational or the uninformed. Rather, it is that this body fails to give equal say to the range of views endorsed by citizens with reasoning capacities and values at least equal to the judges’ own. For these disagreements—we are not concerned with disagreements that do not satisfy the deliberative assumption—there is no impartial criterion for adjudicating the dispute over which interpretation is uniquely correct. Disputants are not guilty of fallacious reasoning or ignoring relevant considerations. Each reaches her reasonable conclusion about justice on the basis of standards that include criteria for impartially weighing competing considerations. We cannot deem one of these reasonable views as uniquely correct without privileging one of the sets of criteria that are in dispute, thus begging the question at issue. Accordingly, we cannot select one member from the reasonable set on the basis of its substantive merits alone.

Instead, we must rely on some voting procedure. Citizens disagree over which voting procedure is best, much as they disagree over which member of the set of reasonable interpretations is uniquely correct. Waldron, as I interpret him, is not interested in showing that citizens will agree that one voting method is best for this problem of social choice. Rather, he aims to show that citizens will agree on a principle for comparing voting methods, and he focuses on a single aspect of voting procedures to which that forms. There are different rules (plurality, Condorcet, Borda) as well as different voting bodies and agenda-setting procedures for determining which options appear on the ballot. Waldron's focus is on whose vote ought to count, comparing elected representatives with unelected judges. He argues that all democratic citizens should accept a principle of political equality as guiding comparative assessments of voting systems when no single voting outcome can decisively be shown to be best. This principle is comparative because it does not define an optimal voting system for selecting a single interpretation from the reasonable set. It can only compare two voting systems, V₁ and V₂, which are both feasibly implementable. That principle compares two voting systems not in terms of their justice, but of their legitimacy, as I explain below.

The relevant principle of political equality is that ‘each person [has] the greatest say possible compatible with an equal say for each of the others’ (Waldron 2006: 1388–89). Waldron (2006: 1389) disambiguates this principle into two parts: ‘equal voice and equal decisional authority’. He does not subject this principle of equal say or its disambiguations to much scrutiny, taking as obvious what they require. I formulate that principle as follows.

Principle of equal judgment: The judgment of each deliberative citizen has the greatest say possible in determining which reasonable interpretation of political rights will be binding on all compatible with an equal say for each of the other deliberative citizens.¹

Deliberative citizens will reach conflicting judgments about which member of the reasonable set of interpretations, i ₁ . . . i _N, is uniquely correct. The aggregate of all of these judgments for all deliberative citizens is the set i ₁. . . i _N. This principle requires that deliberative citizens who reach conflicting judgments about this set have equal say in determining which interpretation will be coercively imposed on all. Citizens have equal say when they can voice their support for what they regard as just policy in the public sphere with the aim of convincing others—compatible with others doing likewise. Additionally, citizens must have some chance to vote for what they regard as just policy during elections. It hardly makes sense to say that citizens have equal say when their preferred option has no chance of appearing on the ballot because another party has agenda-setting control.

How do we apply the principle of equal judgment when comparing two voting systems, V₁ and V₂? The principle asks us to compare two systems in terms of their relative responsiveness to the diversity of judgments of deliberative citizens regarding which interpretation of basic political rights is uniquely correct. As a rough heuristic, we can say that the principle favors V₁ over V₂ when V₁ gives voice to more of these judgments than does V₂. This heuristic does not ask us to simply compare the number of people who can vote in V₁ as opposed to V₂. As I argue in the next section, it is possible that a large representative assembly in fact voices fewer judgments of deliberative citizens than does a small, unelected body of judges. Therefore, we cannot infer that because the House of Representatives has 435 members and the Supreme Court only 9, that the principle of equal judgment favors the former. This inference does not hold even when we assume that all voters are morally motivated. As I will show, elected representatives can have moral reason not to vote for the interpretation of political rights they regard as uniquely correct.

The principle of equal judgment tells us which of two voting procedures is more legitimate. While many theorists follow Simmons (1999: 746) in analyzing legitimacy as ‘a complex moral right’ to impose binding duties on subjects possessed by states or governments, Waldron instead treats legitimacy as a property of specific political decision-making procedures. He identifies ‘the theory of legitimacy’ with ‘a process-based response’ to citizens who ask why they should comply with a specific decision, given that had another procedure been used, another decision would have been selected (Waldron 2006: 1387; 2014: 1706–7) We are to imagine a specific deliberative citizen who judges that interpretation i₂ is uniquely correct. She also justifiably believes that i₂ would be enacted and collectively imposed on all citizens were V₂ used as the procedure for deciding the issue. Instead, society employs procedure V₁, deciding to enact and impose i₁ as binding on all. A Waldronian theory of legitimacy for i₁ and V₁ consists of giving to citizens disappointed in the result reasons (i) why the judgments of V₁'s members are privileged in decision-making over those of everyone else in the political community, and specifically over those who would have been counted in V₂, and (ii) why a procedure that gives greater weight to i₂ (i.e., V₂) was not used (Waldron 2006: 1387, 1390). A theory of legitimacy gives these disappointed citizens reasons to accept as binding society's decision that i₁ will be implemented because it was selected by majority decision via V₁ even though some deliberative citizens judge i₁ to be the wrong interpretation of a basic political right. When they have reason to see i₁ as legitimate, they ought to comply with its dictates and not interfere with its implementation even though they may remain convinced that i₂ is correct and that society made a grave error in enacting i₁ and that i₁ ought to be democratically overturned.

In what follows, I accept premises (1)–(4) and challenge (5) in the next section. I take premises (1) and (2) to be fairly uncontroversial and to need no defense: any liberal theorist can find reason to accept them. Premise (3) is controversial; yet, it follows from Waldron's characterization of the circumstances of politics, which the present analysis accepts out of charity. Premise (4) is also subject to challenge. Specifically, the principle of equal judgment may seem to offer an inappropriate basis for evaluating judicial review because that principle does not seem to acknowledge liberal constraints on the exercise of political power. Thus, Christiano (2008: 284–88) has argued that the principle of equal judgment could justify infringement of minority and individual rights. This challenge to the principle is misguided. The disagreements that concern us here involve citizens’ disputes over which of a set of liberal interpretations of basic rights will be binding on all. We appeal to the equal judgment of deliberative citizens when foundational liberal principles are indecisive in selecting among members of set i _1. . .i _N, each of which represents, by some citizens’ lights, the best attempt to make sense of liberal justice's demands. However, when foundational liberal principles are decisive, we can appeal to those principles directly in justifying the collectively imposed choice of interpretation. In either case, those liberal principles constrain the justified use of political power.

2. Rejecting Waldron's Comparison

Doherty and Pevnick (2014) view Waldron's argument as an exercise in ideal theory that fails to issue recommendations that can guide real-world reform. They argue that ‘the assumptions that guide Waldron's inquiry do not accurately reflect the conditions that we face and are therefore inappropriate for questions of institutional design’ (2014: 96). They single out various of Waldron's assumptions as empirically dubious, including Waldron's (2006: 1360) characterization of what he calls the ‘core case’ against judicial review, which presumes a general commitment shared among citizens, their representatives, and judges to respecting minority rights. They also convincingly undermine the reasons Waldron offers supporting (5) above (Doherty and Pevnick 2014: 89–92). Nevertheless, I do not think their various criticisms are sufficient to show that Waldron's challenge lacks implications for real-world reform. Instead, what they reveal is that Waldron's premise (5) is inadequately defended. This leaves open the possibility that premise (5) is true for reasons other than those Waldron emphasizes. In this section, I offer what I take to be the least controversial argument supporting (5)—one Waldron himself does not offer—and show why it is unsound, at least for two-party systems.

I call this the intuitive argument for (5) and show why it is not vulnerable to the objections Doherty and Pevnick raise. The intuitive argument is modest in its aim, relying on the fact that Waldron's argument is comparative. He does not identify the optimal procedure for adjudicating disputes among reasonable interpretations of basic rights; rather, he offers a defense and application of criteria for comparing the relative legitimacy of two given procedures. He is interested in comparing two procedures Western liberal democracies rely on for adjudicating disputes among reasonable interpretations of basic political rights: majority voting by elected representative assemblies and majority voting by unelected judges. Because these procedures both employ majority voting, Waldron need not offer reasons why the principle of equal judgment favors majority voting over, say, supermajority rules. For that argument to be sound, Waldron does not need the claims Doherty and Pevnick rightfully show he does not establish. Yet, I will show that even the relatively weak presuppositions of the intuitive argument are unsound when we consider two-party systems like that of the United States and that the argument fails to establish that (5) holds of these systems. Nevertheless, even though we lack sufficient reason to accept (5), Waldron's challenge, expressed in premises (1)–(4), has real-world implications for institutional reform.

The intuitive argument for (5) goes as follows. Assume the relevant voters are deliberative citizens: they care about minority rights and have reasoned with the necessary time and care to the conclusion about which interpretation in the set i _1. . .i _N they judge to be correct. Now we appeal to the intuitively compelling size principle, which states the following.

Size Principle: A larger, more diverse, body of deliberative citizen voters will do a better job than a smaller group in giving equal say to the diversity of judgments regarding which member of set i _1. . .i _N is uniquely correct.²

We cannot trivially infer (5) from the size principle; while elected representative assemblies are typically much larger than unelected judicial bodies, we also need reason to think they are more diverse. We need to establish the diversity claim on which a representative assembly would duly consider in its deliberations and voting a wider range of members of the set i _1. . .i _N than would an unelected court of nine. Diversity is not desirable here because of its epistemic benefits, so we do not rely on work like Hong and Page (2004) to establish the epistemic properties of diverse groups. Instead, establishing the diversity claim helps us answer the legitimacy challenge a citizen could raise who views the voting procedure as selecting the wrong outcome. That citizen asks: ‘Why use this procedure rather than another, especially when it selected the wrong result?’ The answer we aim to provide is: ‘Because this procedure was responsive to a more diverse set of judgments, including the one you regard as correct, than an alternative that would have selected your favored outcome’.³

According to the intuitive argument, the diversity claim follows from the nature of political representation itself. Normative theorists of political representation follow Pitkin (1972) in distinguishing two views of the relationship between representatives and their constituencies, the ‘trustee model’ and the ‘delegate model’. Trustees exercise their own judgment when deciding which policies to support, are less responsive to voter sanctions, and aim to advance the good of the whole political community, not just of their electorate. Delegates, in contrast, vote in accordance with their electorate's judgments and are highly responsive to electoral sanctions. As Jane Mansbridge notes, on neither model will representatives’ judgments diverge too much from those of their constituencies. Trustees, who view themselves as entirely unaccountable to their electorates, will, she writes, ‘see themselves as ‘like’ their constituents, in demographic characteristics, political attitudes, or both’ (Mansbridge 2011: 623). And delegates will implement what their voters want.

Thus, both normative theories view political representatives as to some extent reflecting the value judgments of their constituency. Admittedly, this is an imprecise claim. But paired with the size principle, it provides the framework of the intuitive argument for (5). The general voting public holds a wider range of judgments regarding which member of set i _1. . .i _N is uniquely correct than does a small, unelected body of judges, all of whom have similar educational backgrounds, and some of whom may have served decades on the court. Both normative theories of political representation hold that elected representatives to some extent reflect the underlying diversity of views of their constituencies. We lack reason to think the same of unelected judicial bodies. Our fairly weak reason to think representative assemblies are more diverse than unelected judicial bodies seems to be sufficient to bridge the gap between the size principle and (5).

This intuitive argument for (5) does not rely on any of the assumptions Doherty and Pevnick criticize. It does not appeal to results like May's theorem for privileging majority voting over alternative voting rules, nor does it require a strong thesis that representative assemblies in fact reflect to a significant degree the diversity of judgments in the wider political community. To bridge the gap from the size principle to (5), the intuitive argument requires only the modest comparative claim that representative assemblies reflect this diversity better than judicial bodies. I have suggested that normative theories of political representation give us reason to think they do better reflect diversity. Despite the intuitive argument's prima facie plausibility, it does not cross the bridge from the size principle to (5), as I now argue. Problems for the argument arise when we consider the role of party politics in representative assemblies.

Waldron says little about party politics in his critique of judicial review, besides assuming that ‘legislators’ party affiliations are key to their taking a view that ranges more broadly than the interests and opinions of their immediate constituents’ (Waldron 2006: 1361). Waldron treats this assumption as innocuous for his argument. But it is not. Under certain conditions, elected representatives can have moral reason to consider a more restricted range of proposals in the set i _1. . .i _N while deciding which to implement than would a small body of unelected judges. Waldron cannot rule these conditions out a priori even when assuming that representatives vote in accordance with their beliefs about justice. The size principle thus does not give us reason to favor deliberation by large elected representative assemblies over small, unelected judicial bodies.

The intuitive argument for (5) fails because elected representatives can have moral reason to form and vote in accordance with a majority winning political platform, where this platform is the result of mutually beneficial vote trading. According to what is called Duverger's law, which Riker (1982) illuminatingly defends, simple majority voting systems tend to develop into two-party systems. Under simple majority rule, representatives self-sort into two parties both due to pressures from the electorate, who are reluctant to ‘waste’ their vote on third-party candidates, and in order to increase the probability of passing their preferred legislation. I focus on this second motive for self-sorting. Political parties allow representatives with disparate values and interests to present a united face with a single platform of principles and values for rank-ordering options on a ballot. The party's rank-ordering of options may not be identical to any single representative's private rank-ordering: the two may even sometimes sharply conflict. Nevertheless, individual representatives can better achieve their moral goals in a simple majority system by voting in accordance with their party's platform rather than for the option they privately regard as best. This is possible when that party platform is the result of vote trading, or logrolling. According to John Thrasher (2016: 616), vote trading ‘occurs when representative A trades a vote on issue x to representative B, so that B will vote in A's favour on issue y’. Thrasher argues that individual representatives have a ‘weak duty’ to engage in vote trading when doing so advances their constituency's interests. This weak duty survives even when we consider representatives who, like deliberative citizens, vote in accordance with their judgments of justice. These representatives can engage in what each regards from the perspective of justice as mutually beneficial vote trades because, in the circumstances of politics, they disagree about two things: first, how to rank-order options on a single ballot and, second, how important different issues across ballots are with respect to each other. Let me give a simple example.

Consider majority decision votes among five representatives, A, B, C, D, and E. They have two upcoming ballots. Ballot 1 proposes legislation to address social justice issues and consists of options i₁, i₂, and i₃. Ballot 2 proposes legislation for economic distributive justice issues and has options i₄, i₅, and i₆. We are assuming that all options on the ballot belong in the reasonable set of interpretations of basic political rights; thus, both ballots address policy proposals that fall within the sphere of Waldron's challenge. The following tables give representatives’ rank-orderings of the options on the two ballots, with 1^st as best.

Option ranking for Ballot 1 (Social Justice)

Option ranking for Ballot 2 (Distributive Justice)

Absent additional assumptions, we cannot say which option on each ballot majority decision would select when voters are only moved by considerations of justice. One defensible assumption is that votes are open, not secret. In open voting, it can become public knowledge both how individuals rank the specific options and which single option they vote for. The anonymity of secret ballots provides important protections against intimidation and the possibility of corrupt vote trading and vote buying. But in the settings of Ballots 1 and 2, secret ballots can result in no single option receiving bare majority support when each representative selects the option they regard as best. Insofar as we presume that voters are deliberative citizens, the unscrupulous tactics that secret ballots protect against will not be in play. With open ballots, representatives can engage in open deliberation concerning the merits of the options on each ballot, ensuring that their option-rankings duly reflect the strengths and weaknesses of each proposal. Additionally, open ballots enable voters to reach an agreement on passing some option via majority decision. Sometimes it is desirable that no legislation be passed if no single option secures a majority. That is not the case here. By hypothesis, each option on the ballot represents a deliberative citizens’ judgment regarding what just legislation requires. And representatives face the problem of selecting among these options because circumstances call for an interpretation of political rights. Consensus via majority decision on any option is, presumably, better than no consensus at all. Open ballots help ensure that consensus results.

There are two scenarios to consider regarding an open ballot. In the first, we assume that each voter seeks to maximize her moral aims on each individual ballot given the other voters’ preferences. Thus, E could maximize her aims on Ballot 1 by noting that her most preferred option, i₃, cannot secure a majority vote, but that she can prove decisive in winning a majority for her second most preferred option, i₂. The same goes for Ballot 2, with E providing the decisive vote for i₅. In the second scenario, voters seek to maximize their moral aims over time and across ballots, using public knowledge of others’ option-rankings to engage in vote trading. Voters A–E would be willing to engage in vote trading in order to secure a majority win for their most preferred option on the justice issue they regard as most important. Say that A and B both regard distributive justice issues as far more important than social justice, whereas E believes the opposite. E may agree to vote for i₄ on Ballot 2 on condition that A and B vote for i₃ on Ballot 1. In doing so, all three representatives advance what they individually regard as the correct conception of justice even though this involves A and E voting, on different ballots, for what they regard as the least just outcome. Should A, B, and E find regular opportunities to engage in mutually beneficial vote trades, they would have reason to form a standing coalition that advances A's and B's views on social justice and E's views on distributive justice.

This second scenario shows why the intuitive argument from the size principle to (5) fails. Deliberative citizens, faced with open ballots will best promote their individual moral aims over time by forming coalitions and voting in accordance with a party platform to secure bare majorities for the issues they regard as most important. Unelected judges, on the other hand, have a duty to consider the legality of each case that comes before them and the implications of their ruling without regard to future potential rulings. The Supreme Court recognizes this duty with its strong informal norms against discussing cases for purposes of logrolling and coalition formation (Liptak 2010).

Because of these different incentives, we cannot infer that because an elected body of representatives is more diverse than a smaller body of unelected judges the former will consider a more diverse array of judgments regarding which member of set i ₁ . . . i _N is uniquely correct. Pressures toward coalition formation may, in fact, result in the former body considering fewer members of that set in open debate. In the above example where A, B, and E form a coalition together, one option on each ballot—i₁ and i₆—does not receive an equal hearing. We cannot infer a priori that the group as a whole will give equal say to the options the group's individual members judge to be uniquely correct. This is because members sometimes have reason to silence their individual judgments the better to achieve their moral aims across ballots as a party.

Nevertheless, it could be argued that these pressures toward coalition building and maintenance give us reason to accept (5). Voting in accordance with a party platform offers representatives a better way to advance their moral aims over time than does voting on each ballot in favor of the correct option. Should elected officials create that platform in a way that respects the principle of equal judgment, the case for (5) may yet stand. The idea is that diverse judgments regarding justice receive equal say not on the single ballot but behind the scenes, in the coalition-building process that results in a unified party platform. For this possibility to support (5), it must be the case that diverse judgments about justice plausibly receive something like equal say in the actual process of creating and sustaining a party platform. It is an empirical question to what extent different national legislatures satisfy this principle. The reality seems to be that in many countries, including the United States, the process of building and sustaining party platforms is controlled by political elites and by special interests.⁴ To whatever extent considerations of justice arise in this process, it is unlikely a diverse array of such judgments receive equal hearing. Some voices are amplified, others silenced. Absent extensive further investigation of these and additional factors, we simply do not know whether a specific two-party electoral representative system better respects the principle of equal judgment than does a democratically unaccountable body of nine judges.

The intuitive argument for (5) fails to show that the premise is true in any real-world setting, seemingly leaving Waldron's challenge without any implications for guiding real-world institutional reform. Yet, there could be some third institutional arrangement, not yet considered, that is decisively preferable to both elected representative assemblies and judicial bodies according to the principle of equal judgment. Were this so, a revised form of Waldron's challenge would have real-world implications for reform, recommending that political institutions adopt this third approach as more democratically legitimate than the other two. In the next section, I argue that lottocratic mini-publics fit this description.

3. Lottocracy and Judicial Review

The relevant challenge to judicial review arises when an unelected body of nine judges determines the law of the land by a split vote, of 5–4 or 7–2, say. A citizen, disappointed in the result, asks: why should these nine determine the laws to which I am subject, rather than some other authoritative body that would give a better hearing to my favored view, especially when this split vote is the result of a series of arbitrary appointments, determined by factors such as who is president when a previous member dies? Waldron and I agree that this citizen voices a valid democratic complaint that should be answered. Where we disagree is that I do not think there is good reason to believe representative assemblies are comparatively more legitimate in the relevant sense than unelected judicial bodies. I now turn to a third institutional form—majority decision by lottocratic (or sortition-based) mini-publics—and argue that in pairwise comparisons with either majority-decision by elected representative legislature or by an unelected judiciary, this third form fares better on the egalitarian dimension of the principle of equal judgment.

Lottocracies grant decision-making powers to ordinary citizens who are chosen via random selection, rather than via election or official appointment. In recent years they have received renewed interest from democratic theorists, such as Guerrero (2014) and Landemore (2020), as providing a possible remedy to representative democracy's vulnerability to capture by special interests. In both theoretical design and implemented form, lottocracies vary along several dimensions. They can be widely inclusive and participatory, with thousands or millions of participants, such as Ackerman and Fishkin's (2008) ‘deliberation day’. Or they can be exclusive and deliberative, including perhaps a hundred members or fewer who can engage with each other in face-to-face (or electronically mediated) reasoned debate, such as with the crowdsourcing phase in Iceland's 2010–2013 constitutional convention (Landemore 2020: 161–62). They can be local in focus, such as using mini-publics to plan rebuilding the World Trade Center site after the 9/11 attacks (Goodin and Dryzek 2006: 229–30). Or they can be national, deliberating over the merits of constitutional reform, as with France's Great National Debate in 2019. They can be temporary bodies that convene to address a single policy issue, or semipermanent bodies where members are empowered to decide several issues—such as the permanent assembly Abizadeh (2021) proposes to replace Canada's Senate. They can be tasked with information gathering and careful agenda setting for the purposes of crafting detailed, complex policy. Or they can adjudicate among a given menu of options provided by an independent board of legal experts, as was the case with the Icelandic constitutional convention (Landemore 2020: 171). Finally, they can have merely advisory powers, recommending to other legislators which policy to enact, such as Fishkin's deliberative polls, which have only a ‘recommending force’ for elected legislators by indicating where public opinion resides (Fishkin 2009: 134). Or they can have authoritative powers to decide directly on enactment. The institution I defend as responsive to Waldron's challenge is a lottocratic mini-public that is national, exclusive/deliberative, single-issue, and it assigns members limited information-gathering duties and grants the body's verdicts authoritative force. My brief sketch imagines this body as part of the American legal system, supplementing or replacing the Supreme Court's adjudicative powers on issues where the court is riven by reasonable disagreement, all in a way to enhance the democratic legitimacy of the resulting decision.

The Supreme Court itself could be entrusted with the power to convene this deliberative body as needed. Many of the Supreme Court's decisions are unanimous, or nearly so, and even among those decisions decided by bare majority vote, only a few rest on controversial interpretations of moral issues about which ordinary citizens have developed views (Turberville 2018). Deliberative bodies that supplement or replace the court's decision need be convened only for those cases where the court itself is bitterly divided and where individual justices justifiably expect that their decision, whatever it may be, will be divisive in the general population. I cannot say how often conditions for convening a lottocratic body would arise. I suspect the court's reputation as an antidemocratic, elitist institution arises from polarizing landmark cases that come about only a few times each decade. The court's normal functioning does little to excite controversy. And it is stably regarded as the most trustworthy branch of government: since 1970, around 30 percent of respondents to the General Social Survey indicate ‘a great deal of confidence’ in the Supreme Court, the other branches being accorded far lower levels of trust (Smith et al. 2018). The proposed lottocratic body is a tool for countering the perception of the court as exercising arbitrary political power through its rulings. Because justices care deeply about the perceived legitimacy of the court as an impartial interpreter of the law, we can expect them to avail themselves of this tool as needed when they face a specific case that both divides the court itself and the general public.

As noted, the proposed body would be single-issue. This has two benefits: first, single-issue voters do not have the opportunity to engage in logrolling; second, focusing on a single issue minimizes the epistemic burdens on voters who must become informed and update their beliefs about justice in response to reasoned debate. (The extreme costs of scaling up this updating process to include millions of citizens makes mini-publics preferable over popular referenda.) I have argued that logrolling can prevent a diversity of judgments regarding what justice requires from receiving equal hearing in discussion. Single-issue voters have no incentive to silence their judgments in this way because they lack the opportunity to advance their conception of justice across ballots. Unlike elected representatives, single-issue voters have reason to regard each case on its own merits, much as justices do. And unlike a small body of judges, a suitably designed lottocratic mini-public can use the size principle to create a setting wherein voters with a representative diversity of backgrounds can advance, defend, and revise their beliefs about justice to the point where each reaches his or her own blamelessly held conclusion about which interpretation of a political right is uniquely correct.

For this setting to work, randomly selected participants must be a representative sample from the general population along some cluster of features that can serve as proxy measures for diverse judgments of justice. Plausibly, a person's race, gender, socioeconomic background, geographic background, and education level all affect her perspective on the requirements of justice. After a representative sample is chosen, other conditions must also be satisfied: for example, the deliberative setting must empower participants who may be shy or insecure to voice their epistemically blameless judgments without undue deference to more charismatic members. I cannot offer a full account of how the lottocratic body would select a representative sample, nor which criteria the court should prioritize when evaluating a body as sufficiently representative. This would depend on the constitutional matter at stake. Issues concerning economic justice would perhaps require prioritizing representing a diversity of socioeconomic classes; racial justice would call for including members of several distinct ethnic groups. In all cases, the aim is to avoid privileging any one perspective on justice over others and to take no stand in the selection process on which conception of justice is correct. If these features can serve as proxy measures in the relevant sense, then a representative sample along these five dimensions—which are not intended to be exhaustive—would go far toward capturing the underlying diversity of views about justice in the general population.

Lottocratic selection does not aim to perfectly capture the full range of views on justice. We do not require that some member of the body endorse each of the reasonable interpretations i _1. . .i _N as uniquely correct. To answer Waldron's legitimacy challenge, all we need is to find a voting body that would consider a more diverse range of views than either the Supreme Court's nine justices or a legislative body subject to party discipline. Inevitably, some reasonable views—particularly those only a small segment of the population regard as uniquely correct—will be left out. The body serves its purpose if its members debate, vote, and select the interpretation that most citizens in the polity would endorse as uniquely correct if they were transformed into deliberative citizens. This makes the process democratic in two senses: it obeys a democratic majoritarian principle, and it serves as a discovery mechanism for determining which view the majority of citizens would regard as correct if they gave the matter sufficient attention.

Voters are not expected to join the body with judgments of justice fully formed. Their diverse backgrounds serve as a proxy measure for the range of conceptions of justice they find most compelling. But a goal of group deliberation in the lottocratic body is to refine those undeveloped views. Voters would have responsibility for updating their beliefs, through discussion with other members of the deliberative body and with nonvoting advisers to reach their conclusion about a single issue where those beliefs are subjected to rigorous empirical and moral scrutiny. Through this process, voters become what I have called deliberative citizens. The aim of discussion is not to achieve consensus on the issue: we assume that deliberative citizens will blamelessly reach conflicting conclusions about the requirements of justice and that their disagreement must be settled by a vote. It is important to recall that the justices convene this mini-public only when the relevant issues are highly controversial among citizens deeply committed to liberal values. Insofar as a representative sample of voters from the general public have become deliberative citizens on the issue, and they vote in accordance with their judgment of what liberal justice requires, the outcome—whatever that may be—will result from a procedure that respects the principle of equal judgment. Therefore, someone who regards that vote as selecting the ‘wrong’ outcome can be told that the procedure used gave a diversity of views, including the complainants’ favored one, equal hearing, determining which option to impose on all by fair majority decision. A citizen who accepts that her favored view received equal hearing and who continues to object to the decision would be rejecting the principle of equal judgment as a democratic basis for adjudicating disputes over the correct interpretation of liberal rights. What to say to such a citizen falls outside the scope of the present analysis.

The proposed lottocratic body forms to vote on a given menu of options, where this menu aims to represent as fully as is feasible the set of reasonable options i _1. . .i _N. This may sound like a daunting project; yet, it is work the Supreme Court already performs when preparing their decisions. In the American system, justices both justify their majority-winning decisions and offer dissents, which sometimes offer a competing reasonable interpretation on a divisive issue. There is an existing legal infrastructure of judges, lawyers, and clerks who hold a diversity of views and have the incentive to provide the correct interpretation of a basic legal right in light of both legal precedent and considerations of justice. This existing infrastructure can offer the lottocratic body a menu of reasonable alternatives and expert advice for adding options to that menu on the basis of suggestions from the assembly itself. Prior to voting, the task of the lottocratic body is twofold. First, its members must aid the judiciary in the construction of a full set of feasibly implementable reasonable interpretations of the political right at issue. Second, they must understand both the justifications for and the expected implications of implementing each interpretation. These tasks proceed simultaneously as assembly members discuss among themselves and with the judiciary's representatives the merits of each alternative and explore the remaining logical space of the reasonable set. No expert legal competence or moral expertise is required on the ordinary citizens’ part beyond a capacity to understand basic legal and moral arguments and to rank-order a set of reasonable alternatives in accordance with the citizens’ beliefs about justice. These beliefs will be refined and updated in the deliberative process, but, due to the circumstances of politics, a diversity of views about what justice requires will survive scrutiny.⁵

It is important that the lottocratic assembly has some authoritative power in determining the outcome of the judiciary's decisions. One way to implement this power is to have the lottocratic body's majority decision replace the court's, such that if the court votes 5-4 in favor of i₁ over i₂, and if the assembly votes in majority favor of i₂, then the court's vote is disregarded and i₂ is implemented. (There are several other possible implementations, all subject to different trade-offs among citizen empowerment, institutional efficiency, and perhaps, in unusual cases, the protection of minority rights.) Granting the assembly this authoritative power is necessary to address the relevant democratic challenge against judicial review. That challenge, recall, is not against judicial review per se. Rather, it demands reasons legitimating a procedure that empowers an unelected and democratically unaccountable body to issue rulings that determine the content of citizens’ basic political rights, where the correct interpretation of these rights is in reasonable dispute. On the present analysis, this demand is appropriately issued because a principle of democratic equality, the principle of equal judgment, requires that the procedure for adjudicating disputes among reasonable interpretations be one that is responsive to a diversity of views regarding which interpretation is best. The Supreme Court does not provide that responsiveness because its members do not adequately reflect the range of considered, reflective views about justice ordinary citizens harbor, where these views bear on questions of constitutionality.

In order to answer the democratic challenge to judicial review, we must identify and implement a procedure that adjudicates disputes among reasonable interpretations while doing the best job, out of a range of feasible alternative procedures, of giving the diversity of reasonable views equal say. Majority decision by a single-issue lottocratic mini-public does better on this front than does majority decision by either an elected representative assembly or an unelected judicial body. Instead of rule by elites, as with judicial review, or rule by parties, as with elected legislatures, the proposed lottocratic body has citizens ruled by the judgment of their better selves. Following Lafont (2015: 51), we may worry that ordinary citizens would regard decisions by their ‘better selves’ as no more authoritative than decision by a panel of judges. Ordinary, nondeliberative citizens may see the proposed deliberative mini-public as swapping one group of experts for another. And they would be correct. But the democratic challenge to judicial review, to which we are responding, is not directed toward rule by experts per se. Assuming the lottocratic body has authoritative powers and does not merely serve an advisory role, we are swapping rule by nondiverse elites for rule by diverse deliberative citizens, who are actual citizens transformed into experts on the relevant issue. This makes our procedures responsive to the democratic challenge to judicial review.

4. Conclusion

Many of the egalitarian critiques of constitutional judicial review fail. I have focused on one version of that critique, found in Waldron's work, and argued that while it does not succeed on its own terms, it nevertheless has implications for real-world political reform. Well-intentioned, epistemically blameless citizens disagree about basic matters of justice, and these disagreements can persist even in the ideal deliberative setting for achieving consensus about such matters. Diverse societies need some procedure for adjudicating these disputes about justice when they arise in a judicial or legislative setting. The American legal system places the final say on these disputes with the Supreme Court, an unelected and democratically unaccountable body of nine judges, where citizens have no reason to regard members of this body as endorsing a representative diversity of considered views about justice. I have argued that Waldron's premises lead to the conclusion that the most democratically legitimate legal order we can feasibly implement would radically revise the practice of judicial review, delegating some of its authoritative powers to a lottocratic mini-public of ordinary citizens. Were some version of this proposal put into place, citizens who at present justifiably view some of the Supreme Court's decisions as elitist, extremely unjust, and undemocratic would no longer have reason to view decisions they disagree with as illegitimate.