80 U Chi L Rev Dialogue 205 [Essay PDF]
Judge Robert Bork inveighed against courts that decided on the basis of their own values. Judges, he said, must employ only those principles that they could derive, define, and apply neutrally.1 By neutrally, he meant without bringing in their own conception of the good.2 Yet he also endorsed judicial elaboration of doctrines that would keep the Constitution up with the times and sparred with his then-colleague Antonin Scalia on that issue.3 Bork thought that judges could be objective and creative at the same time. Moreover, he was well aware that judges face interpretive difficulties even when they are not self-consciously making their own contribution. He believed that interpretive uncertainty could be resolved without judicial value choice and had to be if the courts were to retain their legitimacy.
Recent scholarship has explored possible responses to legal indeterminacy. One body of work, now often going under the name of “construction,” asks what judges and other interpreters do and should do in order to select among plausible meanings.4 Another line of scholarship deals with judge-made doctrine, especially constitutional doctrine.5 This Essay briefly discusses Bork’s thinking on those problems, and comments on the current issues in light of that discussion.
The first issue I will discuss is uncertainty as to the meaning of an authoritative text. To say that the highest norm in this country is the Constitution’s text is only the beginning of the analysis. For example, an interpreter seeking the original semantic meaning of a part of the Constitution must deal with the possibility that at the relevant time more than one meaning was in use by well-informed speakers of English. Sophisticated language users may have had different concepts that went by that name, each coherent and related to the others, but nevertheless distinct.
Indeterminacy of that kind can be resolved in different ways, some of which may entail normative judgments by the interpreter. One possibility is that the concept of language meaning itself contains the resources to deal with situations like this. Perhaps when there are plural meanings, the most common one is the meaning, with others as variants. Applying that principle to the facts of history may be difficult, but does not in principle require a normative judgment. But if plurality like this is handled by identifying the meaning that best fits the Constitution’s overall scheme, normative judgments are more likely to enter. And if the appropriate meaning is simply selected from among the candidates on the basis of the interpreter’s views about the best result, normative judgments are inevitable.
The problem of multiple plausible semantic meanings is important, but of limited relevance in Robert Bork’s thinking. Although Bork firmly embraced originalism, sometimes emphasizing intent and sometimes emphasizing text, his interpretive method was in fact overwhelmingly structural. He reasoned from claims about the kind of government the Constitution creates, and in particular from that government’s fundamentally democratic character. Indeed, he used that structural principle to derive the methodological principle with which he is so closely associated and that I am exploring in this Essay: the principle that judges must not decide on the basis of their own values.
Bork relied on the essentially democratic character of the Constitution to address two notorious and important instances of textual unclarity, both involving the so-called level-of-generality problem. The freedom of speech and of the press found in the First Amendment can be understood more or less abstractly. In more abstract form, it is a political or moral principle that constrains the legal rules that affect expression. In more concrete form, it is itself a body of such legal rules; for example, the rule that allows the criminal punishment of seditious libel provided that truth is a defense and jury trial is available.
According to Bork, structural considerations pointed to a particular formulation, more general than the legal rules that prevailed at the time of the framing but more specific than a broad license to convey whatever messages one wishes without adverse legal consequences. In his famous 1971 article, Bork brushed text and history aside and rested his argument entirely on structure.6 As a judge he seemed more concerned with history, but located the appropriate reading at the same level of generality and for basically the same reason. For representative democracy to succeed, debate on political issues must be substantially uninhibited by legal sanctions directed to speech.7
Democracy, and the constraint it imposes on unelected, life-tenured judges, also figured centrally in Bork’s approach to two vexed questions concerning the Equal Protection Clause. Bork defended Brown v Board of Education of Topeka8 and rejected permission for symmetrical race discrimination, as in Plessy v Ferguson,9 on the grounds that judges needed a principle of racial equality general enough to relieve them of the need to make controversial choices.10 Under Plessy, they had to decide that physical equality mattered, but that the psychological inequality that might result from separation did not,11 so that the message of inferiority conveyed by separation was merely a construction placed upon it by black people. Those were choices judges should not make.12
Having ascended to that level of generality in interpreting the Equal Protection Clause, Bork stopped, once again on grounds of democracy and judicial restraint. In 1986 he rejected the argument that the clause restricts discrimination on the basis of sexual orientation:
The intentionalist may conclude that he must enforce black and racial equality but that he has no guidance at all about any higher level of generality. He has, therefore, no warrant to displace a legislative choice that prohibits certain forms of sexual behavior. That result follows from the principle of acceptance of democratic choice where the Constitution is silent. . . . In short, the problem of levels of generality may be solved by choosing no level of generality higher than that which interpretation of the words, structure, and history of the Constitution fairly support.13
The higher the level of generality at which a constitutional limitation operates, the more it restricts subconstitutional law. If one equates subconstitutional law with democracy and believes that the Constitution regards democracy as the norm and limitation of democracy as the exception, that is a reason to resolve doubts in favor of less-general readings.
But as Bork himself explained, the Constitution both empowers and constrains lawmakers, with the constraints partly enforced by the judiciary. That is the structure he called Madisonian.14 To find that one aspect of the system, democratic choice, is primary and the other, limitation on democratic choice, is secondary, so that one should be preferred over the other in doubtful cases, is a quite subtle inference. Bork did not base his argument on any particular piece of the text, and it would be very difficult to make such an argument. The Constitution gives and limits power, but it does not say whether one kind of provision is normal while the other is, in Professor Alexander Bickel’s word, “deviant.”15
Any such inference must be based on the structure as a whole. Structural reasoning is perfectly common, but requires at least a limited form of normative reasoning, reasoning that goes beyond identifying semantic meaning. Structural reasoning proceeds from particulars to the more general principles that produce them by asking, what would have led a reasonable constitution maker to choose those particular features? Sometimes the answer will be that the particulars support no inference to a meaningful structural principle, because different constitution makers with different priorities, all reasonable, all might have produced the same design. A constitution that combines substantial legislative power with important affirmative limits thereon might reflect the view that democracy is basic and limitation should be imposed only when absolutely necessary. It also might reflect the view that democracy is to be distrusted and limitation is wholesome, but that drafting limitations well is difficult. If the designer had been led by the latter weighing of competing interests, interpreters might well say that they should be eager to extend limitations in light of understandings that had not been available to the drafters but had arisen in the course of experience.
The possibility that concrete provisions will underdetermine abstract design principles can lead in more than one direction. One response is to embrace normative interpretation and say that interpreters should attribute to the Constitution the principles that would have led them, as reasonable drafters, to produce the concrete provisions. But someone like Bork who believes that judges, at least, should not be following their own values will be very troubled by that way of deducing structural principles.
Another response is to rely on only those structural principles that any reasonable designer must have relied on. Whether many principles will satisfy this test is doubtful. Even more doubtful is whether any reasonable author of the Constitution must have thought that democracy is the rule and judicially enforced limitation is the exception. In my view, norms as broad as that one generally are constructions as that concept is now used: they will rest on both genuine constitutional principles and also the normative views of those who support the construction, because those views will establish the priority among constitutional values that is necessary to produce determinacy at such a high level of generality. A preference for nonconstitutional law over constitutional law might well appeal to construction makers who had had a bad experience with judicial review.
Robert Bork, however, did not indicate that he was proposing something that would now be called a construction. He seems to have regarded the fundamentally democratic character of the Constitution as so blindingly obvious that it hardly required any explanation. Anyone could see that limitations on democratic power enforced by the courts were a carefully cabined exception, not the basic principle. He was certainly not alone in that view. His great friend Alexander Bickel worked from the same premise, as did Bork’s colleague Professor John Hart Ely.16
Perhaps they were right; it is hard to match that trio as students of the Constitution. But not everyone will see it that way. As indicated above, I would say that a reasonable person who thought democracy fundamental could have written the Constitution—as could a reasonable person who thought limitation on democracy fundamental—so the structure does not support inferences based on one assumption or the other.
We know that Bork made structural inferences with confidence and that he inferred that the Constitution fundamentally creates an electoral democracy. We can only speculate, however, about the details of his understanding of structural reasoning. In particular, I know of no evidence bearing on his approach to the situation in which reasonable minds differ as to the design principles that can be inferred from the observed structure. If he thought that the inferences he drew were indubitable, the problem may well never have occurred to him. I suspect that he did, and that it did not.
While we do not know how far Bork would have been prepared to go with structural inferences that did not seem obviously correct to him, it is possible to identify one interesting potential answer that may seem to be consistent with his principles but that I think is not. It is tempting to think that Bork would have favored an approach to structural inference, and to the resolution of ambiguity and vagueness more generally, that produces determinate answers. If the Constitution is less determinate in its meaning, there is more room for judicial subjectivity. But the preference for democracy against judicial value judgments is itself the product of a structural inference. Once established, it might be used to guide interpretation on other issues, but it cannot be established using the premise that clear answers are preferred because they avoid judicial value judgments. As Professor Bork explained in his Constitutional Law I class, circular arguments are not wrong, but they are also not helpful, because they do not tell us anything we do not already know.
Recent scholarship has also explored the role of judge-made doctrine as a means of implementing constitutional provisions that are clear in their semantic meaning but vague in their application. Vague standards make for unpredictable and sometimes unprincipled results, as adjudicators take advantage of fuzzy edges to follow their own views of justice and sound policy. In a system with judicial precedent, the highest court can impose a certain amount of determinacy by establishing rule-like generalizations for future cases and lower courts to follow instead of directly applying the vague provision themselves.
Equal protection levels of scrutiny provide a standard example. Hardly anyone believes that the Equal Protection Clause’s semantic meaning embraces strict scrutiny for race-based classifications and intermediate scrutiny for classifications based on illegitimacy. As Justice Stevens said, “There is only one Equal Protection Clause.”17 To say that the clause denounces only one form of classification, which may for example be called invidious classification, is also to say that it is quite vague. One might think that race classifications are especially likely to be invidious, sex-based classifications somewhat less likely to be invidious, and so on. The Supreme Court can transform such generalizations, by themselves only rules of thumb, into binding rules. Through the operation of the rules of precedent, it can establish tests that lower courts (and the Court in its later cases) are to treat as conclusively identifying violations of the underlying vague standard. Of course, those tests will make mistakes, as judged by the actual, vague norm. For a court to establish such a test is therefore in effect for it to legislate, effectively replacing the Constitution’s text with a doctrinal proxy for it. Doctrine has the relation to the Constitution’s rules that rules have to their reasons.
To call the elaboration of doctrines judicial lawmaking is not just rhetoric. It emphasizes the fact that producing a doctrine routinely requires making very contestable normative judgments. Because generalizations make mistakes, choosing a generalization means choosing among different kinds of mistakes. One of the best-known trade-offs in Anglo-American law is said to justify the standard of proof in criminal cases: it is better to let ten guilty men go free in order to keep from convicting one innocent man. Others might favor a different burden, thinking that one jailed innocent is worth only five criminals set free. The equal protection tiers of scrutiny are set up so that they will err on the side of barring racial classifications, and so will convict some that are innocent, so that no guilty will go free. Whether that makes sense depends on how harmful genuinely unconstitutional racial classifications are, and how harmful the thwarting of legitimate subconstitutional decisions is.
Robert Bork, as far as I know, simply took judge-made doctrine for granted; whether he had any view on its origins, I doubt. But he did believe in it and candidly distinguished between constitutional and statutory norms and the judicially constructed tests that implement them. Almost certainly, that early awareness of judicial creativity derived from his work as an antitrust scholar. Bork was too sophisticated to believe that the semantic meaning of “restraint of trade” included a per se ban on horizontal price fixing. Before he turned to constitutional law, he called the adumbration of doctrine under the Sherman Act that “awesome task” of the judge.18
Bork accepted that some doctrines, at least, would change with the times. He was untroubled by such change and the judicial creativity that brought it about, as long as the judges always took the authoritative enactment as their “major premise.”19 Bork did not expand on the difference between major and minor premises, nor on the source or character of the latter.20
He may have believed that the distinction was just a matter of degree, and that that was enough. If there is a meaningful distinction between large and small value or policy choices, then confining courts to the latter really will reduce their policy-making role. How one might identify the difference, however, is hard to say. Equal protection scrutiny shows how doctrine can rest on a quite delicate and controversial judgment. To know whether to err on the side of finding race discrimination unconstitutional, one must know which is worse, bad race discrimination that is permitted or harmless race discrimination that is forbidden. The cost of permitting bad race discrimination, however, is not a secondary issue; it is the very thing on which the drafters of the clause, by hypothesis, based their primary decision. A court that makes that choice is doing the same work as the framers.
Considering another possible answer to this question can illuminate both Bork’s thinking and the problem of judicial implementation of authoritative texts, especially the Constitution. The value Bork found in the Sherman Act, with the help of his rejection of judicial value judgments, is a maximand: the purpose of the Sherman Act is to maximize consumer welfare.21 While it is conceivable that more than one set of antitrust doctrines would produce the most possible consumer welfare, the constraints are tight enough to make that unlikely. Bork himself praised the rigor of microeconomic analysis and very likely thought that the inferences he drew from it were essentially deductive, pointing to a single conclusion. Although judges might differ as to that answer, they would agree that in principle only one of them was right. And in Bork’s view the limited role of the judiciary meant that courts were obliged to find an interpretation of the Sherman Act that they could implement without making forbidden value judgments.22
Bork had less to say about the elaboration of constitutional doctrine, but he applied a similar principle to judge-made constitutional law. He favored the rule of Brown over that of Plessy on the grounds that the former did not require that courts tell good from bad race discrimination on the basis of their own value judgments.23 A simple ban on race discrimination, with no exception for separate but equal, was in accordance with the proper judicial role.
Brown and Plessy were already familiar territory when Bork wrote about them in 1971. He set out to blaze a new trail for the First Amendment in his concurring opinion in Ollman v Evans.24 Although he explicitly embraced judicial creativity, there is no reason to think that he had significantly changed his view of the permissible extent of judicial choice. His reasoning in Ollman suggests that he thought the courts to operate under fairly tight constraints in creating First Amendment doctrine, as they did under the Sherman Act. He assumed that the First Amendment required some practical degree of press freedom and urged a rule that in his view would achieve it.25 Bork did not ask whether his rule was only one of many, among which a judge could choose on the basis of the judge’s own values. He did not present it as such a choice, but as the best solution to the practical problem posed by the Constitution and changing circumstances. An example he gave of a similar development, the Supreme Court’s decision to treat electronic eavesdropping as a search, did not involve a judicial selection among options based on the judges’ preferences.26 That interpretation of the Fourth Amendment answered a yes-or-no question.
Quite possibly, Bork believed that judges perform a difficult but purely technical task, one in which the text’s posited values constrain their instrumental inquiry so that it has in principle one right answer. If so, he would have been untroubled by judicial creativity. He seems to have thought that, or something quite like it. Indeed, because he embraced a limited judicial role as a basic constitutional principle, Bork almost certainly believed that doctrine making was legitimate only insofar as it could be done without significant value choices. Courts could make law, but only while remaining courts.27
Bork took some things for granted. He seems simply to have assumed that courts would make lawlike doctrine, and his view about the Constitution’s fundamentally democratic character may have been more an assumption than a conclusion carefully reasoned to. His view about the judicial role and the impropriety of judicial value judgments, however, was self-consciously derived from more fundamental premises. It was not just an assumption. And having derived it with some care, he sought systematically to base his approach to judge-made-law interpretation on it.
The contemporary debates about doctrine, its sources and its justification, can profit from that kind of careful inquiry into and exposition of the judicial role. For one thing, a natural argument from vagueness to judge-made law makes an assumption that, on careful examination, proves to be more controversial than it may seem. As noted above, courts make law by setting binding precedents and can make law only insofar as they may do so.28 Judge-made law concerning vague constitutional provisions arises when rules of thumb that courts use to aid them in resolving vagueness harden into binding rules, and thereby replace the amorphous provision with a norm that is more determinate and, for that reason if no other, different. Rules of thumb turn into binding rules because of precedent: when courts take as authoritative a formulation found in a prior case, they substitute the prior case’s proxy for the rule itself.
Precedent is familiar enough, but it is a contingent feature of the legal system. If precedent were not binding, each court in each new case would apply vague constitutional provisions with nothing more than rules of thumb to guide it. That would have serious drawbacks, which is why stare decisis is so common a rule, but in my view the Constitution permits Congress to limit or eliminate the rule of stare decisis by the federal courts, because the rules of precedent are themselves subconstitutional, and subject to Congress’s power to carry into execution the judicial power.29 The claim that the Constitution in effect requires the judicial creation of rule-like doctrines through precedent should be examined and defended, not simply taken for granted. If the rules of precedent are not inevitable, then the exercise of judicial policy discretion that goes into making binding rules out of rules of thumb is not inevitable either. Whether that discretion is consistent with the courts’ constitutional role is a question to be answered on its own terms.
If courts inevitably will apply vague provisions in ways that reflect their own views about sound policy, and if in the process they create binding precedents that are more rule-like than the Constitution itself, then in effect courts will in a real sense make law. They will make law under those circumstances even when the vague provisions have, in principle, one right answer in every case. If that is how the constitutional system works, one might reasonably regard judicial lawmaking as inevitable but nevertheless unfortunate, on the grounds that the tendency to apply vague provisions in accordance with one’s own policy views comes from frail human nature, not the Constitution itself. But judicial lawmaking might be genuinely defensible, rather than regrettable even if inevitable, if some constitutional provisions are fundamentally indeterminate even in principle. Perhaps some constitutional provisions are not just vague, but genuinely incomplete. They might do nothing more than supply a value, not in the highly constraining sense that Bork seems to have assumed, but in a way that is not by itself enough to resolve particular cases, even in principle. Here the most natural model is the kind of statutory grant of authority to an administrative agency that gives the agency a goal and imposes some constraints, but still leaves the agency with important choices to make. Statutory provisions like that are quite properly likened to delegations of legislative power, precisely because they leave the agency to choose among competing values in formulating a rule that will actually decide cases. A directive that a regulatory agency set rates and ensure a fair return on investment for the regulated industry constrains the agency somewhat but is nevertheless incomplete, requiring that the agency either make substantive choices of its own or decide in a merely arbitrary fashion.
Perhaps the Constitution has provisions that are genuinely incomplete in that sense, incapable on their own of applying to concrete questions. If so, the argument that the Constitution genuinely endorses judicial lawmaking, rather than tolerating it as an unavoidable failing of judges, is plausible. That is not to say, however, that the Constitution contains any such provisions. A provision that is merely vague, and in principle complete, will function much like one that is truly incomplete when applied by courts that follow stare decisis and are used to making their own policy judgments. Provisions that are not delegations of authority to the courts can appear to be such delegations because of their functional similarity. But functional similarity is not identity, and one reason to examine one’s premises is to distinguish the similar from the same. Judges, lawyers, and scholars are thoroughly accustomed to a system in which the Supreme Court of the United States operates as if it were an agency granted power under a statute that is often vague and sometimes underdetermines concrete results. To take that metaphor for reality, and to interpret the Constitution in light of it, however, is to reason in a circle, to justify one’s assumptions on the basis of one’s assumptions. And as Robert Bork said, circular reasoning, though not false, leads only around in a circle.
† James Madison Distinguished Professor and Joseph C. Carter Jr Research Professor, University of Virginia School of Law. Law Clerk to the Honorable Robert H. Bork, United States Court of Appeals for the District of Columbia Circuit, 1982–1983.
1 See generally Robert H. Bork, Neutral Principles and Some First Amendment Problems, 47 Ind L J 1 (1971).
2 Id at 3.
3 See Ollman v Evans, 750 F2d 970, 995–97 (DC Cir 1984) (en banc) (Bork concurring) (endorsing judicial modification of doctrine over time and disagreeing with Judge Scalia); id at 1038 n 2 (Scalia dissenting) (disagreeing with Judge Bork).
4 Important work on construction in this sense includes Keith E. Whittington, Constitutional Construction: Divided Powers and Constitutional Meaning (Harvard 1999), and Lawrence B. Solum, The Interpretation-Construction Distinction, 27 Const Commen 95 (2010).
5 Leading contributions on this topic include Mitchell N. Berman, Constitutional Decision Rules, 90 Va L Rev 1 (2004), and Richard H. Fallon Jr, Implementing the Constitution (Harvard 2001).
6 See Bork, 47 Ind L J at 20–23 (cited in note 1).
7 Ollman, 750 F2d at 996–97.
8 349 US 294 (1955).
9 163 US 537 (1896).
10 See Bork, 47 Ind L J at 14–15 (cited in note 1).
11 See Plessy, 163 US at 544.
12 See Bork, 47 Ind L J at 13–15 (cited in note 1).
13 Robert H. Bork, The Constitution, Original Intent, and Economic Rights, 23 San Diego L Rev 823, 828 (1986). Bork’s statement is ambiguous. He may have meant only that the Constitution means what its enactors expected or desired it to accomplish and that they did not expect or desire to accomplish anything with respect to discrimination on the basis of sexual orientation. On that reading, displacement of legislative choice is simply a way of referring to a finding of unconstitutionality, and Bork said no more than that the Constitution means what it means. But he may have meant something significantly different, and I think that he did. Bork was addressing the argument that the text, at least, is unclear as to its level of generality; the principle that the Constitution means whatever it means will not tell anyone what it means. In response, he said that the text should be read at a level of abstraction no higher than the words, the structure, and the history will support. Why not read the text at a level of abstraction no lower than one that can be so supported? Requiring that the level of abstraction be as low as possible, however, does make sense if displacement of democratic choice is independently undesirable. (It might be undesirable either because nonconstitutional law, being easier to make or change, is more democratic than the Constitution, or because constitutional law inevitably reflects the value choices of judges.) It thus seems to me quite likely that Bork believed that the structural principle in favor of democracy provided an interpretive principle to be applied to the Equal Protection Clause.
14 Bork, 47 Ind L J at 2–3 (cited in note 1).
15 See J. Skelly Wright, Professor Bickel, the Scholarly Tradition, and the Supreme Court, 84 Harv L Rev 769, 784 (1971) (noting that Bickel regards the Supreme Court as a “‘deviant’ institution in a democracy”).
16 Bork followed Professor Charles Black in concluding that substantial protection for political speech could be deduced from the constitutional structure even without the First Amendment. See Charles L. Black Jr, Structure and Relationship in Constitutional Law 39–48 (Louisiana State 1969). Ely’s view of the Constitution is reflected in the first word of the title of his enormously influential book on constitutional theory. See John Hart Ely, Democracy and Distrust: A Theory of Judicial Review 44–48 (Harvard 1980).
17 Craig v Boren, 429 US 190, 211 (1976) (Stevens concurring).
18 “Courts charged by Congress with the maximization of consumer welfare are free to revise not only prior judge-made rules but, it would seem, rules contemplated by Congress.” Robert H. Bork, Legislative Intent and the Policy of the Sherman Act, 9 J L & Econ 7, 48 (1966). That was because “Sherman and others clearly believed that they were legislating a policy and delegating to the courts the elaboration of subsidiary rules.” Id. The “awesome task” of the judge was that of “continually creating and recreating the Sherman Act out of his understanding of economics and his conception of the requirements of the judicial process.” Id.
19 Bork, 23 San Diego L Rev at 826 (cited in note 13).
21 See Robert H. Bork, The Antitrust Paradox: A Policy at War with Itself 20 (Basic Books 1978).
22 Bork endorsed the reading of the Sherman Act that “takes a pro-consumer policy as the base rule and requires exceptions in particular cases to be made by the legislature.” Robert H. Bork, The Rule of Reason and the Per Se Concept: Price Fixing and Market Division, 74 Yale L J 775, 838–39 (1965). That understanding, he said, enables the courts “to be impersonal in an important and desirable sense.” Id at 839.
23 See Bork, 47 Ind L J at 14–15 (cited in note 1).
24 750 F2d 970, 994 (DC Cir 1984) (Bork concurring).
25 “We now face a need similar to that which courts have met in the past. [New York Times v] Sullivan, for reasons that need not detain us here, seems not to have provided in full measure the protection for the marketplace of ideas that it was designed to do.” Ollman, 750 F2d at 996 (Bork concurring).
26 Id at 995–96 (Bork concurring).
27 Bork apparently applied his principle that judges should not make value choices both to judge-created doctrine and to the identification of meaning; he did not present the difference between Plessy and Brown as a difference in judge-made rules, but rather as a difference in interpretation. As far as I know he did not consider the possibility that the limited judicial role affects the two differently.
28 Judge-made doctrine is an artifact of the rules of precedent and the appellate hierarchy within which they operate; for example, the United States Court of Appeals for the Seventh Circuit does not make law for the United States District Court for the Western District of Virginia.
29 See John Harrison, The Power of Congress over the Rules of Precedent, 50 Duke L J 503, 505 (2000) (arguing that rules of precedent in federal courts are federal common law subject to change by Congress).