Five to Four: Why Do Bare Majorities Rule on Courts?
abstract. Courts, such as the U.S. Supreme Court, make important decisions about rights by voting, and often the decision is determined by a bare majority. But the principle of majority decision (MD) for courts has not been much reflected on. There is very little scholarly literature on this and only a handful of articles proposing alternatives to MD in the judicial context. So what justifies judges’ reliance on MD? In democratic contexts, MD is usually defended on one of three grounds: (i) as a decision-procedure that is efficient; (ii) as a way of reaching the objectively best decision; or (iii) as a way of respecting the principle of political equality. However, it is difficult to see how any of these arguments really explains or justifies MD in the case of judicial decision-making. Regarding (i), other efficient procedures are available: we want to know why this sort of efficiency should be valued. On (ii), epistemic arguments for MD do not work well for the sort of bare majorities that we see in Supreme Court decision-making. And, on (iii), it is hard to rig up an argument based on political equality or political fairness that works for unelected judges as opposed to ordinary citizens or their elected representatives. After reflecting on these possibilities, the Essay concludes that, at the very least, defenders of judicial authority should be more tentative in their denunciations of democratic majoritarianism.
author. University Professor, NYU Law School and Chichele Professor of Social and Political Theory, Oxford University. Earlier versions of this were presented at workshops at the NYU School of Law and All Souls College, Oxford. I am grateful to Benjamin Fischer for research assistance, to Rachel Shalev of the Yale Law Journal for her most helpful suggestions, and to Rachel Barkow, Eyal Benvenisti, Andrew Burrows, Sujit Choudhry, Ron Davidson, Ronald Dworkin, John Ferejohn, Barry Friedman, Jeanne Fromer, Stephen Gardbaum, Bernard Grofman, Christopher Hood, Simon Hornblower, Daniel Hulsebosch, Sam Issacharoff, Lewis Kornhauser, Nicola Lacey, Daryl Levinson, Erin Murphy, Tom Nagel, Jonathan Nash, Bill Nelson, Pasquale Pasquino, Peter Pulzer, Adam Samaha, Carol Sanger, Re’em Segev, John Vickers, James Walmsley, and Kenji Yoshino for comments and criticisms.
Why, in most appellate courts, are important issues of law settled by majority decision? Why, when judges disagree, do they use the same simple method of “counting heads” that is used in electoral and legislative politics? Some scholars call this the problem of “judicial majoritarianism,”1 though that phrase is also used (by Barry Friedman and others) to describe the inclination of judges to follow majority opinion in the wider society.2 In this Essay I am not interested in “judicial majoritarianism” in Friedman’s sense. What I want to address is the decision-procedure used internally in our appellate courts.
Judges vote when they disagree and—as we all know—many important U.S. Supreme Court cases are settled by a vote of five to four among the Justices, even when the Court is reviewing legislation and deciding whether to overturn the result of a majority vote among elected representatives. Consider, for example, Citizens United v. Federal Election Commission,3 which overturned, in part, a piece of federal legislation, the Bipartisan Campaign Reform Act of 2002.4 The judicial vote to overturn it was 5-4; the legislative votes to enact it were 240-189 in the House of Representatives5 and 60-40 in the Senate.6 Majorities, everywhere you look. This leads to my question: why is bare majority decision7—I am going to abbreviate it as MD—an appropriate principle to use in an institution that is supposed to be curing or mitigating the defects of majoritarianism?8 Of course, it is not only in constitutional cases that majorities rule in court. It is pretty much universal among multi-member judicial panels, in private law appeals as well as in public law, at least in our tradition. It is simply how judges decide. But that is not an answer; it is an indication of how pervasively the question arises.
So, why is MD used in judicial decision-making? And why do people put up with it? Let me say at once that my pressing these questions is not intended as a way of discrediting judicial decision-making, not even with respect to questions of judicial review. (There are ample grounds for opposing judicial review of legislation whatever decision-procedure judges use.9) Even for those who favor judicial review, the absence of a clear theory of judicial MD constitutes a gap in our understanding of our most important legal institution.
I think my question is worth asking, for several reasons. First, it is worth asking simply out of interest. MD in court is something we take for granted, but it would be interesting to know whether this practice has ever been made the focus of explicit justificatory argument in the history of the modern judiciary. I suspect the answer is “no,” and I wonder why that has been the case, especially in light of the theoretical attention—much of it critical—that is paid to MD in democratic settings. When we consider electoral or representative institutions, we ask questions about the justification of MD all the time and we devote a lot of effort to elaborating and discussing the answers. Is the justification of MD in democratic politics epistemic? I mean, is MD an appropriate decision-procedure to use because it promises to get us more often to the right answer, e.g., to the election of good representatives or the making of good legislative choices?10 Or is it appropriate only as a fair procedure, i.e., one that respects the principle of political equality (one person, one vote)?11 Is MD in electoral or legislative contexts just “natural” in some sense that does not require justification? Do we use it simply because it is efficient?12 These are good questions to ask about voting in elections and in legislatures. They are good questions, too, when we turn our attention to courts.
Here’s a second reason for our interest in this question. Often defenders of judicial review say they are opposed to majoritarianism as such. But because they seem to have no trouble with MD in court, that can’t really be their position. It must be democratic majoritarianism to which they are opposed, not majoritarianism as such. Opportunistically, however, they will sometimes seize on certain difficulties with majoritarian decision-making as such in order to discredit democratic majoritarianism.
Whether or not they are disingenuous in this, we ought to take these difficulties seriously. So, for example, critics of democratic majoritarianism sometimes allege that MD can lead to incoherent decision-making (e.g., through Arrow’s paradox).13 If this is so, then maybe it is also true of appellate courts using MD (unless there is something about courts that keeps their decision-making within the parameters where the paradox does not arise). Dan Farber and Philip Frickey pointed this out in their excellent study Law and Public Choice. They said that if we think (for reasons associated with social choice paradoxes) that “chaos and incoherence are the inevitable outcomes of majority voting, then appellate courts . . . are equally bankrupt. . . . If we accept the thesis as to legislatures, we are left with nowhere to turn.”14 Justifying a given decision-procedure is partly a matter of answering what can be said against its use, so we ought to investigate whether what is said against MD in other contexts can be said against its use in court and, if so, whether such criticisms can be answered.15
Our question is also worth asking—this is my fourth point—because we can imagine other decision-rules for judicial settings and it might be worth considering why these are (mostly) dismissed out of hand for decision-making among judges on appellate courts.16 For a long time the practice of trial by jury proceeded on the basis that unanimity was required to convict a criminal defendant. Nowadays in some jurisdictions—England for example—a majority verdict is sufficient, but it still must be a heavy supermajority, something like ten to two.17 A bare majority is never sufficient (except for grand juries). Why is nothing similar envisaged for disagreements among judges rather than jurors? One might imagine a supermajority rule for judicial review. Actually, imagination is not necessary: the Nebraska Constitution ordains that the state’s “Supreme Court shall consist of seven judges” and that “[a] majority of the members sitting shall have authority to pronounce a decision except in cases involving the constitutionality of an act of the Legislature. No legislative act shall be held unconstitutional except by the concurrence of five judges.”18 The North Dakota Constitution is even more stringent: it requires four out of five justices to strike down legislation.19 These seem like good rules, embodying as they do a sort of presumption in favor of the constitutionality of legislation.20 So why is a supermajority rule not the decision-procedure on the Supreme Court of the United States?21
Fifthly, there is an anomaly in the absence of theoretical discussion, given the fact that MD seems to be accepted explicitly in our political community as the basis for appellate judicial decision-making. Other legal systems—most civil law systems, for example—present only a consensus judgment and do not allow the public expression of dissenting opinions.22 In these systems, judges do not appear to vote. Presumably there is often dissensus in their private deliberations, and maybe MD (or something like it) is used behind closed doors to determine what will be the consensus position.23 Yet the public does not perceive the judges as voting; they do not perceive the outcome as depending upon counting heads. In contrast, in common law systems we do. And in the United States, the fact that courts use MD is the crucial assumption on which the whole politics of judicial appointments turns. This is particularly so because many areas of constitutional decision are potentially unsettled—abortion, of course, is the best-known example. The issue may be decided by MD one year, but Justices of the Supreme Court come and go, and partisans hope that a bare majority on one side of a given issue may be replaced in time by a bare majority on the other side.24 So, for example, since 1973, it has been an important feature of presidential politics to try to secure judicial appointments to the Supreme Court that—on the assumption of MD—will either secure (for a while) or overturn Roe v. Wade.25 The matter is always on a knife-edge and pro-choice advocates are vividly aware that one or two conservative appointments might upset the five-four balance on which they rely. In this sense, MD explicitly frames the politics of judicial nomination and confirmation. Oddly, though, despite its explicit presence as a frame, the use of MD in court is never itself made a topic of argument. In abortion politics, for example, no pro-life faction ever argues that anti-abortion legislation in the states should be protected from federal review by a Nebraska-style rule. No pro-choice faction ever argues that a supermajority should be required to overturn a precedent of long standing. People just assume MD and argue around it.
Sixthly, it would be useful to have a good account of the use of MD in court to complement and illuminate scholarly discussion of other issues about judicial decision-making. As I shall shortly show, there is very little in the law review literature addressing the exact question that I have posed. (The few exceptions will be discussed in Part III.) But there is a considerable literature on other aspects of judicial decision-making.26 Scholars talk frequently about how judges, individually, should approach the exercise of their power on multi-member courts. When a judge disagrees with her colleagues, should she defend her position in the way that people defend their interests, or in the way a member of a political party defends a plank in the party platform, or in some other way? There are debates about whether judges ought to trade votes, compromise, logroll, and so on.27 Should judges switch their votes when there is no majority on the disposition of a case?28 There is some literature on when a high appellate court like the United Kingdom’s Supreme Court should hear appeals with a larger-than-usual panel of justices: some arguments about this focus on the possibility that MD might produce a narrow majority on a small panel (e.g., three-two on a panel of five) but a quite different outcome if the panel had been constituted from among the other available justices.29 There is also an interesting literature on the relation between judges’ voting and the reasons they give in their opinions. Should judges vote only on outcomes (and is that the vote that should be decisive) or should we be seeking consensus or a majority on the various reasons that justify their votes?30 At the moment, in common law systems, we do not insist on MD or on any decision-procedure regarding the reasons that are given for outcomes determined by courts. It seems we can afford to leave that undecided. But it is not clear whether MD for the outcome makes sense in the absence of a complementary procedure governing choice of reasons. These are all important questions of institutional design or institutional practice. But we may not be able to answer them adequately without some sense of why MD is thought appropriate for the judiciary in the first place.
There is one last reason I want to mention for our interest in this question. Paying explicit attention to the use of MD in court may redound to the credit of its use elsewhere in our political system. For example, it is tempting to disparage MD in a legislative context, where it seems like a decision-procedure fit only for the aggregation of preferences and for the bargaining and logrolling that involves. But when we see judges using it, what we are seeing is the application of MD to matters of principle as well as policy and preference, and we see the use of MD in court as the upshot of a deliberative process rather than as something that obstructs or usurps deliberation. I hope my discussion in this Essay will make us more open to the idea that majority voting is not opposed to principled deliberation but part and parcel of what principled deliberation involves in the modern political world.
Here is how I shall proceed. In Part II, I outline the meager discussion of this issue that exists in the scholarly literature. In Part III, I consider the very idea of a justification in this sort of area—what it means, what it involves—and I address the claim, which is sometimes heard, that a justification for MD is unnecessary. Part IV sets out the main lines of justification that are adduced in favor of MD in the political settings—(A) arguments from efficiency, (B) epistemic arguments, (C) arguments about fairness, and (D) an epistemic/fairness hybrid argument—and considers their applicability to judicial decision-making. Finally, in Part V, I draw some lessons for majoritarianism generally from our willingness to countenance the use of MD in court.
The question I am asking about the use of MD in court has not been much addressed in the law review literature. I may be wrong, but I don’t think there is any article-length treatment of judicial MD.31 I have occasionally raised the matter in discussions of judicial review, challenging defenders of that practice to explain why a supposedly counter-majoritarian institution uses this method of majority-voting. In The Core of the Case Against Judicial Review, I said that the topic is one that “I have never, ever heard a defender of judicial review introduce . . . into discussion himself or herself, let alone undertake to explain why it is a good idea.”32 And in a more recent piece, I said that after years of raising this question—why do judges use simple majority decision among themselves to determine whether to overturn statutes passed by majoritarian institutions?—“Ihave long since given up any expectation of an honorable answer.”33 A couple of people have attempted to respond to these provocations and I will discuss their answers in a moment.
Some have misinterpreted my challenge. It is not itself intended to discredit judicial review; it is not supposed to add anything to the critique that already exists. One set of authors, Krishnamurthi et al., has said that “[i]f Waldron’s intuition is correct, and defenders of judicial review cannot justify the use of judicial majoritarianism, then judicial review might also be without justification.”34 And their response was to show that that conclusion wouldn’t follow.35 They were right about that. A failure to explain why courts use majoritarian decision-making would not undermine judicial review. It would just leave us with a puzzle. Obviously courts have to use some method of decision; maybe MD is as good as any other; perhaps it does not need justification. (I will deal with this possibility in Part III.) However, by leaving the matter there, Krishnamurthi et al. really do not answer my question. They themselves provide nothing in the way of a justification—though they gesture in certain directions as to where a justification may be found.
In 2010, my late colleague Ronald Dworkin offered the following as an answer to the question, “Why, if majority rule is not intrinsically fair, is it appropriate on final appellate courts like the Supreme Court, which decides many very important cases by a 5-4 vote?” I am going to quote his answer at length:
The choice among checks on majoritarian procedures must of course depend on which options are available. Judicial review is an available option for checking legislative and executive decisions. It is also an available option for checking judicial review itself through a hierarchal system of appellate courts. But of course judicial review is not available to check the decision of the highest appellate court; if it were the court would not be the highest. It does not follow that if the judges in this series of reviews disagree the disagreement should be settled by a vote among them. A Supreme Court’s 5-4 decision might overrule the unanimous decisions of a great many more judges on lower courts.
But the head-counting procedure does hold on the Supreme Court itself, and it makes perfect sense to ask what alternatives, beyond judicial review, are available. We can easily imagine some. Constitutional courts might give more votes to senior judges because they have more experience. Or more votes to junior judges because they are likely better to represent popular opinion. The Supreme Court does give each justice an equal vote, but it also gives some justices much more power than others in shaping constitutional law. When the Chief Justice is in the majority, he decides the often crucial question who will write the opinion for the Court; when he is not the senior justice in the majority does. No vote decides that issue. The Court’s practice of adopting majority rule for the verdict itself can sensibly be challenged. But since judicial review is logically not an option at that stage, the choice of a majority decision procedure hardly suggests that that procedure is intrinsically fairer than a different process that includes judicial review.36
All this is interesting. But I don’t see any explanation here of why bare-majority-decision is the appropriate procedure—or even an appropriate procedure—to use. As far as I can tell, Dworkin does not even want to claim that it is: much of his account entertains possibilities for judicial decision-making that are at odds with MD (e.g., more weight to be accorded to the views of some Justices than others). Dworkin does make an important point when he says that there is nothing inconsistent about (i) opposing the use of MD-standing-alone in the legislative context in light of the availability in that context of MD-complemented-by-judicial-review, and yet (ii) embracing MD-standing-alone in the context of the highest court, where (by definition) MD-complemented-by-further-judicial-review is unavailable. That is a fair point. But it only takes care of an ad hominem argument of inconsistency. It provides no affirmative explanation of why MD-standing-alone is an appropriate procedure for courts to use.
The only other scholarly engagements with this issue that I am aware of are by Akhil Amar, in a recent book titled America’s Unwritten Constitution;37by Jed Shugerman, in an article arguing for a supermajority rule;38and by Rick Hills, in a thoughtful couple of paragraphs in a 2002 essay.39
Amar notes that “[f]rom its first day to the present day, the [U.S. Supreme] Court has routinely followed the majority-rule principle without even appearing to give the matter much thought.”40 He believes that it was just obvious to those who set up our judicial system that MD was the rule to use; it seemed natural, he says, and he cites John Locke, Benjamin Franklin, and Thomas Jefferson to that effect.41 His position seems to be that MD was taken for granted and decision-procedures were stipulated in the Constitution’s text only when there was to be a departure from MD.42
Jed Shugerman’s article is mainly an account of the inadequacy of bare-MD and an argument for a six-three supermajority rule.43 But Shugerman does consider some arguments for MD in the course of trying to show that a six-three rule would do better at promoting the values that MD is supposed to promote. He suggests that MD is based on a combination of practical efficiency and a “consensus theory of truth” among experts.44 I will examine what he says about this “consensus theory” in Section IV.B of the Essay.
Rick Hills’s contribution comes in the course of a symposium essay on Chris Eisgruber’s book Constitutional Self-Government.45 Eisgruber’s book is a defense of the role of principled judicial decision-making on constitutional matters and the appropriateness of its being able to override the majoritarian decision-making of legislators. Hills says that Eisgruber’s work is underpinned by an anti-majoritarian assumption “that there cannot be a connection between quality of argument and numbers of persons who are persuaded by an argument.”46 But, says Hills,
[t]his contrast between the weight of numbers and the quality of reasons . . . is confused: the assumption of any regime of political equality is that the arguments that persuade the largest number of adherents are the best arguments. Where political equality reigns, the weight of numbers is regarded as the best available proxy for the quality of argument. Thus, on that ultimate forum of principle, the United States Supreme Court, five votes decides a case and sets precedent. Why? Because the Justices are political equals: we assess the quality of argument by counting noses. Likewise, in every philosophy department, hiring decisions which rest largely on the “quality of reasons” in the published work of a candidate are determined by a majority or super-majority vote of the faculty (or their elected representatives on an executive committee). Why? Again, because the community is a community of equals, “counting heads” is regarded as the only acceptable way to determine quality of argument.47
This passage definitely points us towards an affirmative argument in favor of MD. Indeed, maybe there are two arguments here. In Part IV, I will tease out the difference between the argument that “the weight of numbers is . . . the best available proxy for the quality of argument” and the argument that, because judges are regarded as one another’s equals, MD is required as a matter of fairness. Hills runs these two lines together in an interesting way, suggesting that fairness might require us to treat the weight of numbers as if it is the best proxy for the quality of argument (even if it is not). We will examine this possibility as well, in Section IV.D below.
I am not sure why there has been so little discussion in the literature of the issue of MD in court. Perhaps it is because the use of this decision-procedure strikes people as so obvious that it needs little discussion. But in other contexts legal theorists are not deterred from discussing the obvious; that’s what philosophers do. Or it may be because the use of MD in this context is something of an embarrassment to people who predicate their support for the empowerment of judges on the perceived disrepute of MD in other contexts. I indicated at the beginning of Part II that this might be the reason for scholarly silence on the issue. Or it may be just happenstance. This is a question that has been hiding for some time in plain sight. It is now time to come to terms with it explicitly.
What is the justification for the use of MD in multi-member judicial panels? My question talks of “the justification for the use of MD.” That suggests a sort of single official justification. But there may not be anything of the sort: neither the U.S. Constitution nor the British Judicature Act makes any reference to the use of MD. Perhaps it is better to ask, can the use of MD in court be justified (whether our officials have signed up for such a justification or not)? And we may find that there is more than one line of possible justificatory argument.
Our question assumes that the use of MD requires an argument. Occasionally one finds suggestions in the literature that it does not—that it is simply something obvious and beyond justification.48 Hannah Arendt, in On Revolution, stated that “the principle of majority is inherent in the very process of decision-making” and is “likely to be adopted almost automatically in all types of deliberative councils and assemblies.”49 The paucity of scholarly discussions may suggest that this is the prevailing view in the legal academy: MD needs no justification. But really that won’t do, not for me at any rate. It is the job of philosophers to question the obvious, and to try to come up with arguments for things that practical people take for granted, whether they “need” a justification or not. Sometimes the request for justification is answered with a rhetorical question—“What’s the alternative?”—with the implication that MD is the least bad decision-procedure (after minority rule, supermajority, and unanimity are rejected). But even if one accepts this as a starting point, one wants to know more: least bad on what dimensions? What is the reason that MD is better than minority rule and is that the same reason that makes it better than unanimity or a supermajority requirement?
Throughout this Essay, I will argue that the demand for justification is connected to the issue of legitimacy. Decision-procedures like MD operate in circumstances of disagreement50: some people in society believe fervently that a given issue, currently before a court, should be resolved one way; others believe with equal fervor that it should be resolved the other way. One side is going to lose as a result of the court’s decision and which side that is may be affected by the decision-procedure that is used. The legitimacy of a decision-procedure is partly a matter of whether support can be mobilized for the decisions that are made under its auspices. But legitimacy also has a more focused aspect: it has to reconcile the losing party in particular to the decision that has been made. The decision-procedure has to carry with it some sense of justification that enables the losing party to say, “It is right for me to accept this decision, not because I agree with it on the merits, but because of some characteristic of the process by which it was made.”
What characteristic can that be? Some theorists say that MD is “natural.” This is not the same as a claim that the use of MD is obvious. A claim of naturalness does not imply that MD is beyond justification; rather it points towards a justification of a certain kind. Akhil Amar believes the American Founders must have thought MD was natural. He cites John Locke (whom he says the Founders must have studied on this point):
[I]n Assemblies impowered to act by positive Laws where no number is set by that positive Law which impowers them, the act of the Majority passes for the act of the whole, and of course determines, as having by the Law of Nature and Reason, the power of the whole.51
But Locke’s naturalistic case for this is unsatisfactory, depending as it does on a claim that each “[b]ody . . . move[s],” by a sort of physical necessity, “that way whither the greatest force carries it, which is the consent of the majority.”52 There is little reason to suppose that we can analogize the consent given by each individual to an equal quantum of physical force, certainly not for the purposes of a normative argument.53 Anyway, elsewhere in political theory—for example, when democracy is being discussed—MD is not treated as obvious or natural. Its use is regarded as an open question.54 There is an active and considerable literature on MD in political theory, and it is surely worth considering how that literature bears on its use in this particular domain.55
Having said that, we must be careful not to think that arguments that purport to make sense of MD in an electoral or legislative context can necessarily be transferred to the judicial context. Some may; some may not. Here are a couple of examples that do not survive the transfer.
It is sometimes said that MD is a sort of proxy for combat in the political arena. In an essay entitled Minimalist Conception of Democracy: A Defense, Adam Przeworski says this: “Voting constitutes ‘flexing muscles’: a reading of chances in the eventual war. If all men are equally strong (or armed) then the distribution of vote is a proxy for the outcome of war.”56 The party that wins an election, says Przeworski, is in a position to “inform the losers—‘Here is the distribution of force: if you disobey the instructions conveyed by the results of the election, I will be more likely to beat you than you will be able to beat me in a violent confrontation.’”57 This is a pretty threadbare argument even for electoral politics.58 It makes no sense at all in the judicial context. Judges use MD to decide issues not just for themselves but (in our system) for more than 300 million people. It is imaginable, I guess, that if some of these issues are not decided in an orderly manner, they will have to be decided by fighting. (This happened with some important issues about race in the United States 150 years ago.) But if fighting does break out, on an issue like abortion for example, it will not be confined to the nine Justices of the U.S. Supreme Court. And so the existence of a majority (one way or the other) among the Justices is no indication at all of who is likely to win a civil war about abortion in the country at large.
Here is another example of an argument for MD that cannot survive the transition from electoral to judicial politics. Suppose that voters are self-interested: they vote their pocketbook; they vote for policies that they think will bring them, personally, a greater balance of happiness over suffering. And suppose, too, that they are well enough informed to make this calculation wisely. (Strong assumptions, I know, but common in positive political theory.) Then the use of MD to make political decisions in an electorate of this kind may be justified on utilitarian grounds.59 The majority view corresponds to the greatest balance of happiness over suffering in society as a whole. It is a rough correspondence at best.60 But, rough or not, it will not survive the transition to judicial politics. For not even the staunchest rational choice theorist thinks that judges vote primarily for what will promote their personal benefit.61 And even if they did, there is no conceivable ethical theory that could make the greatest happiness of the greatest number of judges into an ideal for the whole society.62
The failure of the utilitarian argument to survive the trip from electoral to judicial politics reminds us of a couple of important features of the judicial terrain. First, if the use of MD is to be justified on this terrain, it has to be justified on the basis that a majority of votes represents a majority of opinions rather than a majority of interests. Judges end up with different views, not just different preferences, on the matters of principle and interpretation they are called upon to address. So we can’t use anything like a calculus of interests to address the problem posed by judicial dissensus. It is not that sort of problem.
Second point: judges are supposed to be experts in the areas on which they disagree. We are not talking about the use of MD to resolve disagreements among amateurs. Both the judges in the majority and the judges in the minority are supposed to know what they are talking about. In this regard, we should bear in mind other features of judicial decision-making. Judges don’t just “come up with” a view on the matter before them and vote in its favor. They are trained in the law. They hear extensive oral arguments back and forth and they read volumes of written submissions and precedential cases. They deliberate thoughtfully both among themselves and each in the solitude of his or her own chambers. They (or their clerks) write elaborate essays to spell out their reasons for adopting one or another view. And yet, as often as not, they still disagree—quite sharply in our judiciary. So the question is why—after all this training and argument and thought and deliberation and looking things up and testing them out—the simple procedure of “counting heads” that MD involves is the appropriate way to resolve these disagreements in these circumstances.
How is MD justified in political theory generally, and can any of these justifications be applied to its use in courts? I shall consider four possible lines of justification: (A) arguments about efficiency, (B) epistemic arguments, (C) arguments from fairness, and (D) Rick Hills’s epistemic/fairness hybrid argument.
The first kind of argument is the most common. When I raise questions about MD in courts, some of my colleagues claim it needs no justification. When I say that that is not good enough,63 they retreat to the view that MD is justified on grounds of its decisional efficiency. They say: MD lowers decision-costs compared with any other method; it is decisive; and it is easy to apply.
Those who say this are usually puzzled by my continued refusal to accept what they say as all that needs to be said for jurisprudential purposes. But more does need to be said. For there are other decision-procedures we could use that would be equally if not more efficient. A coin could be tossed as soon as an appeal comes before a multi-membered court (either in all cases or when the court reveals itself as divided). The outcome could be determined in that way. This would lower or eliminate decision-costs; it would be decisive; and it is even easier to apply than MD. My interlocutors respond by saying, “Don’t be silly. Of course we can’t just toss a coin.” And if they are asked why, they swallow their exasperation and say, “Coin-tossing would be ridiculously unresponsive to the merits of the appeal and the arguments and precedents put forward,” or they say, “Coin-tossing would be ludicrously unfair, and you know it.” And they conclude, “So, be realistic: MD is the most efficient of the acceptable methods of decision-making.”
I think this trajectory of call-and-response indicates that there is indeed much more to be said about MD in the courtroom than that it is efficient. We must ask ourselves, “How and why exactly is MD properly responsive to the merits of the case, to precedents and arguments, etc., in a way that coin-tossing is not?” Or we have to ask, “How and why exactly is MD a fairer (e.g., more respectful) decision-procedure than coin-tossing?” These are the questions we need to answer, and only answers to these questions will illuminate the problems that motivated our inquiry (in Part I above). They take us, respectively, to the second and third lines of argument: (B) the epistemic argument and (C) the argument based on fairness.
I cannot emphasize how important it is to get beyond the lazy response that MD is just more efficient. The crucial thing is to establish why it is legitimate for us to decide cases on the basis of this procedure. The legitimacy issue is particularly important in regard to judicial decisions that overturn legislation, for there a question seems to be raised about MD in one context that is not raised about its use in another context. Discussion in terms of fairness and responsiveness to the merits might help address the issue of legitimacy, but a stubborn insistence on efficiency as the only criterion suggests misleadingly that the problem of legitimacy here is simply a matter of cost.
Having said all that, I don’t deny the importance of efficiency. All I am saying is that it cannot be the whole story. I have said that efficiency includes, among other things, decisiveness: the ability to settle a contested matter with some finality. It is possible to interpret this particular element of efficiency in a way that gives MD an advantage over coin-tossing. If a given issue is settled by coin-tossing, everybody involved knows that applying the same procedure again—even if it is applied again immediately following the first settlement—might yield a different outcome. By contrast, and as a matter of political sociology, the division of the decisional body into majority and minority is likely to be more stable than that and so the first settlement is more likely to “stick.” Of course majorities can come undone; perhaps this is true mostly of large assemblies,64 though, even on a small judicial panel, a key member—a sort of Justice Kennedy figure—may waver. And maybe this is an advantage. We want a degree of settlement in the law but not absolute settlement.65 I have heard it said that settling contested matters by judicial decision on a multi-member court using MD provides the modicum of finality and settlement that the rule of law requires, but it does not preclude the prospect that the matter may be revisited sometime in the future and reversed. This prospect would be much dimmer if judicial decisions were protected by supermajority requirements against variance. MD may be valuable, in other words, precisely because it affords an optimal combination of decisiveness and non-finality.66 But there may be other efficient decision-procedures that combine decisiveness and non-finality in this way. We still need to understand why it is important to have the particular kind of decisiveness that MD secures balanced with the particular kind of non-finality that MD makes available.
Let us now consider the idea that MD may be an epistemically reliable way of getting at or near the truth, a way of responding better than any other available method to the objective merits of the arguments presented before a court. Rick Hills invited us to consider the proposition that “the arguments that persuade the largest number of adherents are the best arguments.”67 Certainly there is something intuitive about this, especially when the voting constituency consists of experts; and its intuitive appeal is reflected in ancient doctrine.68 Who is not persuaded by the slogan, “Four out of five dentists choose Colgate for their families”?69 But it is surprisingly hard to account for the intuition, once it is put under pressure.
Let’s take it step by step. That any given expert is persuaded by a proposition in his or her area of expertise is surely evidence, though not conclusive evidence, that the proposition is true. This we can accept. That any given expert denies a proposition in his or her area of expertise is surely evidence that the proposition is false. This we have to accept also. So far so good. But we have not begun to grapple yet with majorities and minorities. We can perhaps advance a step or two further. If a majority among experts is overwhelming (say four-one like the dentists or five-two like the Nebraska Supreme Court striking down a statute), then the contrary view of one or two experts can be dismissed as an aberration. No expert is infallible, and for any expert there is some small chance that he or she might have made a mistake. If there is something short of a dissensus, we may want to ask ourselves, “Which is more likely, that the one or two outliers have made a mistake or that all the members of the large and expert majority have made a mistake?” We may think it is much less likely that the majority view is mistaken. (Perhaps this is how something like judicial MD began. Maybe the earliest cases of dissensus on judicial panels were cases of an overwhelming majority of judges versus one or two dissenters.)
But it is not at all clear that an expertise theory can survive for cases in which there is a bare majority, say five to four.70 Five experts believe that a certain proposition is true and four believe it is false. I don’t think the dynamics of expertise and fallibility help us at all in this situation.71 Which is more likely—that the five fallible experts are mistaken or that the four fallible experts are mistaken? Who knows? Would we be better off, epistemically, tossing a coin if there were no budging the five to four vote?72
Perhaps I should not be so absolutist about this. Maybe it is better to say that the argument from expertise diminishes for these cases, not that it disappears altogether. We might still be better off betting that the majority is right, particularly over a large number of instances, even when the majority margin is very small. But I am not convinced that this is so. The expertise argument with which we began was not about bare numbers: it was about overwhelming majorities versus outliers. That is why I doubt that it has any application at all when majorities cease to be overwhelming and dissenters cease to be outliers.
Can the position be retrieved? Can we associate an epistemic argument with the weight of numbers as such, even the bare weight of numbers? In recent years, political theorists have become interested in something called “the Jury Theorem” associated with an eighteenth century theorist, the Marquis de Condorcet.73 As a matter of arithmetic, Condorcet proved that if a group like a jury faces a binary choice (say, guilty or not guilty), and if each of the individual jurors is more likely than not to arrive at the right answer when he votes (that is, if the probability of his getting the right answer is greater than 0.5)—I shall call this individual competence—then the likelihood that a majority of the jurors will reach the right answer is greater than the likelihood of any one of them getting the right answer.74 Moreover, the likelihood that a majority will get the right answer increases as group size increases. So, for example, if there are three jurors, each with an individual competence of 0.6, the chance that a majority of them will be right is 0.648. (And that latter chance increases quite sharply as group size increases.)75
Can we apply Condorcet’s Jury Theorem to the judiciary? Maybe. The judges are experts, so let us say that each of them surely has a greater than 0.5 chance of coming up with the right answer in a given case. And if that is true, then the Condorcet arithmetic ensures that the chance that a majority of judges—even a bare majority of judges—is right will be quite high. Indeed it will be higher, the higher up the judicial hierarchy you go: as we move from one federal district court judge, to a three-judge panel on a court of appeals, and then to a nine-judge panel on the Supreme Court. Even if the expertise did not increase on the way up—and we hope and expect that it does through the seniority and experience of judges in the higher courts, and through deliberation and the hearing of submissions76—the Jury Theorem would seem to justify the use of MD on the multi-judge panels.77 QED?
I have to say that I am uncomfortable with this result. There is something gimmicky about it. Consider first that there is nothing epistemological in Condorcet’s Jury Theorem, which could justify MD as truth-enhancing. It is just arithmetic. It works best for an urn that happens to contain 60 black balls and 40 white balls: the chance that a majority of three balls drawn at random from the urn will be black is 0.648. There is nothing about truth or knowledge in that. Certainly the theorem can be applied to cases where truth is at issue. It can be applied to cases where anything is at issue. If there is a better than 0.5 chance that a given opinion will be X then the chance that a majority of three opinions will be X is greater than that, and the chance that a majority of opinions will be X increases as the number of opinions given increases. The value of X may be “true” or “high-quality”; or the value of X may be “long-winded,” “melodious,” “printed in pink,” or “beginning with the definite article.” The result still holds. Condorcet’s theorem has nothing to do with objective truth or right answers; those terms just happen to be within the domain of its arithmetic applicability.
Defenders of MD might respond that we should just take the epistemic enhancement where we find it and not worry about how mindlessly Condorcet’s theorem generates the result. If the assumption of our application of the theorem is correct—namely, that judges really are (individually) more likely to get their decisions right than to get them wrong—then MD applied to judicial panels will enhance the quality of their decisional output, just as Condorcet predicted. It is an arithmetical certainty.
Whether this furnishes us with a justification that will satisfy the demands of political legitimacy is another matter. It depends in no small part on how the losers in a political struggle regard the competence of the judges who voted against them. It is important to remember that most people do not approach decisions by the Supreme Court using dispassionate language of objectivity and expertise in the way that a philosopher might. They certainly do not attribute undifferentiated levels of expertise to the Justices. The liberals I know think that at least four of the Justices currently on the U.S. Supreme Court are likely (unerringly and maybe deliberately) to come up with the wrong answer on most important issues. And the conservatives I know think the same about at least four other Justices. Now let us suppose that the Court has made a decision by a bare majority (five to four) about some controversial issue like abortion. How should people think about this exercise in judicial majoritarianism? Specifically, how should people who oppose the decision that the Court has made think about this method of making a decision? Bear in mind that judicial decision-making has long since left the era of consensus behind, certainly in the United States. It is bitterly divisive. If we want to justify a decision-procedure in these circumstances of controversy, we must understand that we are doing it not as an academic exercise, but in order to confer some legitimacy on the decision in the eyes of people who would otherwise bitterly oppose it.
So let us suppose that five liberal Justices have voted to uphold some pro-choice position. Can the fact that this was determined by MD possibly make the decision legitimate in the eyes of conservative pro-life advocates? Specifically, can Condorcet’s Jury Theorem contribute anything to the decision’s legitimacy in their eyes? I think not. The pro-life citizens will be convinced that the liberal Justices are incompetent on this issue (willfully or for some other reason): they will think that the chances that liberal Justices will come up with the right answer on abortion is well below 0.5 (probably they think it is something approaching zero). And since, ex hypothesi, there are five liberal Justices out of nine involved in this decision, it is likely that our pro-life citizens will attribute to the panel as a whole an average individual competence lower than 0.5. In which case, no Condorcetian arithmetic will be able to persuade them to accept the legitimacy of bare MD as a way of making this decision. (Remember: when average individual competence falls below 0.5, the Condorcet effect goes into reverse.) At best, they will end up where we were a page or two ago: when five alleged experts line up against four alleged experts on a matter like this, who knows where the truth lies? At worst, they will be where Justice Scalia was in Planned Parenthood of Southeastern Pennsylvania v. Casey: all talk of judicial expertise and objectivity in this area of basic values is nonsense.78 And if this is the likely reaction of the pro-life citizenry to pro-choice decisions on the Court, it is also likely to be the reaction of the pro-choice citizenry to pro-life decisions. On any controversial decision by the Court taken on the basis of bare-majority-decision, it is likely to be the reaction of the very group that most needs to be persuaded that this method of decision-making is legitimate.
In democratic theory, the most powerful case that can be made for MD is that it is required as a matter of fairness to all those who participate in the social choice. Fairness can be understood formally and informally.
Informally, people may be persuaded that MD is fair because, although they are losers this time around, they may be winners in the next political cycle when their faction is in the ascendancy and when their party is in a position to nominate judges and get their nominees confirmed.79 Alexis de Tocqueville argued in general along these lines—“every American discovers a kind of personal interest in obeying the laws because the man who today does not belong to the majority may tomorrow be among its ranks.”80 Similar sentiments might be applied, albeit indirectly, to the judicial context, though the timeline is longer and more haphazard than the regularized turn-taking among electoral majorities on which de Tocqueville’s point was predicated.
Formally, we may defend MD as a way of respecting political participants as equals. If a society faces a binary option, and each of its members votes one way or the other, then the advantage of MD is that it is decisive (except in the rare case of a tie); it is neutral between the options (and thus fair to the various supporters of the various options); it gives as much weight as possible to each individual’s vote, in the direction in which that vote points (so it is fair to each voter); and it gives no greater weight to any one individual’s vote than to the vote of any other individual (so that it is fair in the sense of equality). That MD alone satisfies these four conditions—decisiveness, neutrality, highest positive weight, and equality—in the circumstances of a binary choice is a well-established theorem in the theory of social choice.81 And that is why we should support its use.82
In the democratic context, these four conditions seem plausible. In the midst of political disagreement in society, we want a decision-procedure that will be decisive but not biased towards any particular political point of view. We want it to be fair in that sense. Respect for individual citizens in the context of a democracy demands that weight be given to nothing but individuals’ opinions. But democratic equality—one man, one vote—insists that the opinions of all should be treated equally. These requirements of fairness seem to argue for the use of MD to make social choices, not because of any epistemic hypothesis, but simply because this is a procedure that respects people’s opinions and respects them as equals, as they are entitled to be respected on democratic assumptions. How do these conditions fare when we move to judicial decision-making?
That we want a decisive decision-procedure in court is clear enough: an appeal must be decided one way or the other (even if its value as a precedent is later revisited). This is the efficiency argument over again. But the applicability of the other conditions is not so clear.
The Nebraska rule mentioned in Part I invites us to abandon neutrality, at least in cases of judicial review of legislation: instead of being neutral between the claim that a given statute is constitutionally invalid and the claim that it does not violate the constitution, Nebraska operates instead with a presumption of constitutionality.83 That is the default position unless a supermajority (at least five judges out of seven) can be assembled. It is by no means a silly position. A presumption of constitutionality used to be part of American constitutional doctrine,84 and for a while people toyed with the proposition that the unconstitutionality of a piece of legislation could not be thought of as established beyond reasonable doubt if four Justices believed it was constitutional.85 However, quite early on, jurists realized that the two issues of reasonable doubt and substantive dissent could be drawn apart: the issue on which the bench divided five to four might be exactly the issue of whether there was reasonable doubt about a given law’s constitutionality.86 Anyway, I am not urging the adoption of the Nebraska rule;87 I am just showing that there is room for doubt about the condition of neutrality that MD presupposes.
The other condition whose application might be thought problematic in the judicial context is the condition of equality. Each judge’s opinion counts positively in the direction in which it points—that sounds sensible. But on what basis do we insist, as Hills puts it, that “the Justices are political equals”?88 We saw Dworkin imagining a system in which more votes were given to senior judges because they had more experience, or one in which more votes were given to junior judges because they were likely better to represent popular opinion.89 As I have said elsewhere, the members of the Supreme Court are ranked by seniority, and the public commonly ranks them by their virtue, learning, and effectiveness, not to mention their politics. They have an order of precedence in their dealings with one another and so on.90 But in the authority accorded to their opinions, the rule seems to be that they are equal.91 When they disagree, the fact that the Chief Justice or a senior Justice takes one side or the other makes no difference to the weight accorded to his vote.92 Why is this? In the electoral context, the assumption of political equality is fundamental, and the fierceness of our democratic insistence upon it gives a sharp and powerful edge to the fairness argument in favor of MD. “One person, one vote” matters enormously and most of us reject out of hand the plural voting system—for example, more votes for university graduates—that John Stuart Mill envisaged.93 But we might not be so offended by a similar proposal for judges. The normative principle of political equality does not seem to be applicable to the members of a judicial panel in the same way that it is applicable to the body of citizens.
Perhaps there is a mistake here. We should not be comparing the proposition of equality for judges (on which judicial MD is predicated) with the very fundamental proposition of political equality for citizens; perhaps we should be comparing equality for judges with equality for legislators. How does that comparison come out? Well, legislators have greater or less seniority, greater or less influence in their caucuses and on the floor of the house or the Senate. But when it comes to a vote on legislation, each legislator’s vote is equal. Why? Because in a properly apportioned polity, granting equal weight to the votes of each legislator is a rough and indirect way of respecting the equality of their constituents.94 Political equality for legislators thus conveys something of our fierce and fundamental adherence to political equality for citizens.
Maybe something similar can be said for judges, though it is a lot more indirect than the legislative case. Each Justice represents an appointment by the President, and the President of course is elected (indirectly) by the people.95 So if we were to give greater weight to one judge than to another, we would be disrespecting the President who appointed the latter judge, the senators who ratified the appointment, and the voters who elected those officials. Conservatives might like the idea of assigning less weight to the votes of Justice Ginsburg than to the votes of Justice Scalia, but doing that would be a way of according less respect to those who voted for Bill Clinton in 1992 than to those who voted for Ronald Reagan in 1984. This is a pretty indirect argument, but it is the best I can come up with (and maybe something similar can be rigged up for elected judges).
I don’t doubt that we could simply announce that we propose to treat each judge on a given panel as the equal of every other judge on that panel, and just leave it at that. And that could be the equality condition that underpins the use of MD. But remember how much work MD has to do so far as legitimacy is concerned. An awful lot hangs on the use of MD in court, especially when decisions are being made on matters of fundamental principle. And for the time being, we require passionate advocates for one side or the other to put up with a constitutional position that they regard as an abomination simply because one more Justice voted for it than voted for the contrary view.96 That is a lot of weight to put on a decision-procedure, and it is especially a lot of weight to put on a procedure when one of its leading assumptions is just a stipulation. In the case of electoral or legislative politics, we can invoke a powerful ethical premise to convince electoral or legislative losers to put up with adverse outcomes:
It is an ethical premise of democracy, derived from belief in human equality (i.e., the equality of men as human beings to seek self-development, happiness, God, and fellowship), that each citizen, whether he be learned or barely literate, rich or poor, has the right to have his vote for elected officials counted equally with others.97
We plead with the losers in the name of fairness to their fellow citizens. But what do we offer the losers in five-four judicial decisions? I worry that the equivalent equality plea in the case of judicial MD is based on much weaker or perhaps even non-existent grounds.
I don’t want to end this discussion of possible grounds for MD without considering the hybrid view intimated by Rick Hills, which I outlined in Part II.98 Though Hills talks a lot about MD as a way of “assess[ing] the quality of argument,” I think his theory is really more like a fairness argument than an epistemic one. His position is that “[b]ecause the Justices are political equals[,] we assess the quality of argument by counting noses.”99 A pure fairness argument would say that “because the Justices are political equals, we count noses.” But Hills’s view is that because the Justices are political equals, we act as though nose-counting were a way of determining epistemic quality. I do not want to put too much weight on what might have been quite a loose formulation—and what follows is rather tentative—but perhaps what Hills means is this: we treat the Justices formally as one another’s equals, not in the way that we treat citizens as one another’s equals, but in the deferential way in which experts might be treated as one another’s equals. Citizens are to be respected as equals in the way that, say, stakeholders are to be respected as equals; to treat one citizen as less than an equal is to act as though it does not matter that he too has a stake in what is going on around here. The principle for citizens is quod omnes tangit ab omnibus decidentur: what affects all should be decided by all. But that is not the theory of judicial equality. We treat the judges on a panel as one another’s equals because of what they represent: the law. We—that is, those who are to be ruled by them—defer to them equally as experts in the law. Though in reality they may differ in their expertise, in court we are supposed to respect them as we would respect the law itself that they represent. It is a sort of artifice: each judge stands for the law and stands for it equally.
This is an interesting way of thinking about the equality that is supposed to underpin the fairness argument. We saw earlier that there is really nothing in judicial decision-making equivalent to the political equality that serves as a normative foundation for electoral and legislative majoritarianism.100 But maybe we can proceed with something like the following. We might say that a fairness argument for MD in court can be predicated upon the idea of an equal deference to each of the judges, that is, to the judicial office embodied, in a purely formal sense, by each of them. We might even go so far as to say it is equal deference to the law as such, of which (in theory) each of the judges is supposed to be a mouthpiece.101 Whether the principle of this deference can survive the sort of skepticism about expertise that we saw wrecking the application of the Condorcet argument is another question.102 I do not mean skepticism about judicial expertise as such; I mean skepticism about the significance of a very narrow majority among Justices, to each of whom we have reason to defer. Hills intimates a good account of the reason for individual deference, but it is still not clear that it adds up to a hybrid argument for deference to a narrow majority.
I have not been able to come up with any decisive and powerful argument for the use of MD in court. The efficiency argument is incomplete at best. The epistemic argument falters when MD seems to support decision by a bare majority. The best argument in favor of MD—the fairness argument—is somewhat weaker than the equivalent argument in an electoral or legislative context. The hybrid argument is interesting, but its rationale for a normative principle of judicial equality rests on an admitted fiction and may be vulnerable in exactly the circumstances, fraught with controversy, in which the Court’s decision-procedure has to do its most important legitimizing work.
I want to end with a provocation. Perhaps it is no surprise that American legal scholars have not come up with a convincing defense of the use of MD in court. Many of them have gotten so used to saying mean and disparaging things about majoritarianism and majority voting (among the citizenry or in the legislature) in their arguments for judicial review that they find themselves a bit tongue-tied when forced to say something about the use of exactly the same decision-procedure among judges on appellate courts. If we go around saying that a commitment to the use of MD in politics involves a “crude statistical view of democracy,”103 we will probably want to avoid drawing attention to the fact that ultimately nothing but numbers determines how the Supreme Court, which is supposedly a “forum of principle,”104 makes its decisions.105 Statistics don’t cease to be statistics just because the numbers are lower and the voters wear robes. That 5 votes beat 4 on a court is as crude and statistical as the proposition that 218 votes beat 217 in the House of Representatives.
Once we face up to this point, it might be a good idea to change the terms in which the question of judicial review is usually stated. Judicial review is said to pose a “counter-majoritarian difficulty”106 and of course we all know what that means. But it is, strictly speaking, inaccurate. Entrusting final decisions about important legislation to courts does not involve abandoning or rejecting majoritarianism. It reveals, instead, what Dennis Baker and Rainer Knopff call a preference for courtroom majoritarianism over legislative majoritarianism.107 What changes is the constituency of people whose votes will be counted on the matter: the votes of just nine unelected judges will be counted (to determine a simple majority) rather than the votes of a hundred elected senators or 435 elected representatives or millions of ordinary voters in a California-style plebiscite. It is a counter-democratic difficulty, not a counter-majoritarian difficulty.108 Our practice of referring certain matters to the courts for final decision reflects a distrust of democratic decision-making. It is a distrust of persons: we don’t trust ordinary voters or their representatives on certain matters; we prefer the judges. It is not a distrust of MD, for that is a principle we continue to deploy.
I am not saying all this to embarrass defenders of judicial review. I do believe, however, that a more open and accurate characterization of the decision-procedure that judges use may help us get clear about a few things that are relevant to that debate.
First of all, there is nothing inherently incompatible between the use of MD and the addressing of issues of principle. When we contrast judicial decision-making with majoritarianism, the implication is that majoritarianism is appropriate for bread-and-butter issues or grubby pork barrels, where crude and inarticulate interests are involved, but that issues of principle should be decided by a more elevated procedure. We must abandon that characterization. The fact that courts address matters of principle by voting tells us that there is nothing inherently inappropriate about these issues being decided in institutional contexts that are more notorious for their majoritarianism.
Secondly, frank recognition of the role that MD plays and of the circumstances of judicial decision-making that call for a decision-procedure of this kind may help us think more tolerantly about disagreement. Part of understanding that matters of principle have to be dealt with in this way involves owning up to the fact that reasonable disagreement is possible on these matters. Judges disagree about rights—there is no way around that—and although they are experts in constitutional law, there is no denying that their disagreements with each other are in many ways just like the disagreements, on these matters, that ordinary citizens have with one another. On abortion, affirmative action, campaign finance, or the juvenile death penalty—on all these issues—reasonable people disagree. That even the judges, at the end of the day, have no choice but to count heads on these issues shows us that. Acknowledging this, we might be a little more tolerant of our disagreements with one another—not relaxing our opposition necessarily, but refraining from characterizing our opponents’ views as positions that are in some sense vicious, corrupt or beyond the pale.
Thirdly, acknowledging the use of MD in court might make us more judicious in the way we use the phrase “the tyranny of the majority.” It’s a phrase that rolls easily off the tongue when we are discussing “the counter-majoritarian difficulty.” But once we see that there is no getting away from majorities and minorities, we might be more careful how we use it. Of course we must acknowledge the possibility that a majority decision may be tyrannical, whether it is a majority decision by a court or a majority decision by a legislature: Dred Scott109 was tyrannical and so were the Fugitive Slave Acts.110 But, in either context, tyranny of the majority does not happen every time someone loses a majority vote. Judicial MD helps us see that: a judicial minority may fervently disagree with the majority on the court, but in almost every case they—I mean the dissenting Justices—do not suffer from anything that can be called the tyranny of the majority. Justice Scalia does not suffer under the tyranny of the majority in Dickerson111 or in Casey.112 Once we realize this, we may want to say the same thing about being a member of a losing faction in democratic decision-making: it does not necessarily mean suffering under the tyranny of the majority. It may in cases where members of the decisional minority are also members of a topical minority group who suffer disadvantage as a result of the majority decision.113 But even in cases like that, the imposition of the disadvantage on the minority by a majority decision is not necessarily tyrannical. The disadvantage might be fair and appropriate, or its fairness and appropriateness might be a matter of dispute; it is not tyrannical just because it is a disadvantage, and it is not tyranny of the majority just because it was imposed through MD.
My fourth and final point takes us out of the judicial context, but suggests ways in which judicial MD might help us in democratic theory generally.114 Facing up honestly to the use of MD in court might help us develop more realistic accounts of what is known in the trade as “deliberative democracy.” Democratic theorists these days are quite keen on the idea of a democracy where citizens address the major issues facing the polity in a thoughtful and impartial spirit, not focusing exclusively on their own pocketbooks: citizens discuss the issues of the day, presenting their opinions to others and holding themselves open to persuasion and correction when the opinions of others are presented to them.115 But it is evident that deliberation does not always yield consensus, though consensus (on the truth) might be thought of as its telos: sometimes deliberation aggravates dissensus.116 So there is a question about how deliberative democrats should think about decision-making in the face of disagreement. The problem is that, as things stand, deliberative democracy and majority voting seem like odd bedfellows. There is something embarrassing about voting in a deliberative context—or at least that is the impression we are given; voting seems like an admission of deliberative failure, for it shows that a discussion based on the merits has failed to resolve the issue. When those who write about deliberation turn their attention to voting, the sense of distaste is almost palpable.117 I think what we need is a theory of democracy that makes voting the natural culmination of deliberation, rather than something indicating that deliberation has been in some sense inadequate. We need a theory of deliberation that dovetails with voting, not a theory of deliberation that is embarrassed by it—a theory that explains why it is reasonable to require people to submit to MD not just their self-interest, but their most impartial, their most earnest, their most high-minded, and their best-thought-through convictions about what justice, rights, or the common good require.
I suggest that we should take voting on the Supreme Court as our clue for the development of a more general theory that reconciles voting and deliberation. For there is surely no doubt that the Supreme Court is a deliberative body, and that it does not cease to be so when its members disagree with one another, even though their disagreement means that, at the end of their deliberation, the matter before them has to be determined by a vote. This, I say, we should regard as our clue—for it indicates that in principle there is nothing incompatible between deliberation, disagreement, and voting. If the combination makes sense in the courtroom, then maybe it also makes sense at the level of a more general theory of deliberative democracy.
These changes in the way we think about judicial decision-making, judicial review, and deliberation in the face of disagreement are the intended normative pay-off of this Essay. This has been an exercise in exploration rather than advocacy or denunciation. I do think it is a pity that there is not more discussion of MD in court, and it is interesting how difficult it is to transpose into the judicial context arguments for MD that seem to work reasonably well in electoral and legislative politics.
I have mentioned several times the possibility of instituting a supermajority decision-rule for striking down legislation. No doubt it is healthy to think about the possibility of a supermajority rule and to come up with reasons why it might be a good or a bad idea. But it is not the aim of this Essay to advocate such a rule: even Jed Shugerman said that his discussion of it was primarily by way of thought experiment.118 Experiments in thought can help fill out gaps in our understanding and it is primarily with that in mind that I have undertaken this discussion of the use of majoritarian methods by judges.