A recent, and characteristically illuminating, article by Professor Henry Monaghan confidently announces that "[p]recedent is, of course, part of our understanding of what law is."1 As a descriptive matter, Professor Monaghan is entirely correct. Legal analysis by lawyers, courts, and academics typically begins and ends with precedent. Law students are meticulously trained in the art of reading, applying, and distinguishing cases. Court pinions, including Supreme Court opinions, on constitutional matters frequently consist entirely of discussions of past decisions, without so much as a reference to the Constitution itself.2 Even in this era of law-and-metatheory, case analysis is still the mainstay of legal scholarship. Legal actors disagree, of course, about such important matters as the precise weight to be afforded precedents in particular cases, the appropriate way to determine a precedent's scope, the extent to which precedent actually constrains the judicial process, and the circumstances under which precedents should be distinguished or overruled. Yet almost everyone, at all points on the political and jurisprudential spectrums, agrees to some extent with Professor Monaghan that prior

* The panel was moderated by Douglas Ginsburg, Judge, United States Court of Appeals for the District of Columbia Circuit.

** Associate Professor, Northwestern University School of Law. This article is a slightly edited version of remarks delivered at Harvard Law School on March 13, 1993, at the Twelfth Annual National Federalist Society Symposium on Law and Public Policy. I am profoundly grateful to my fellow panelists Akhil Reed Amar, Charles Fried, Douglas H. Ginsburg, and Frederick Schauer for their incisive and courteous comments on my presentation. I have not revised my remarks after the fact to take account of their criticisms (nor have I recanted my heresies), but they have provided me with much food for thought. I am also grateful to Anthony D'Amato, Peter B. McCutchen, and participants at a faculty colloquium at Northwestern University School of Law for helpful comments on an earlier draft.

1. Henry P. Monaghan, Stare Decisis and Constitutional Adjudication, 88 COLUM. L. REV. 723, 748 (1988).

2. See Gary Lawson, An Interpretivist Agenda, 15 HARV. J.L. & PUB. POL'Y 157, 157 n.2 (1992) (noting that during the October 1986 Supreme Court Term, 34 of the 54 cases involving interpretation of the Bill of Rights either directly or through incorporation did not even quote the supposedly relevant constitutional provision).

judicial decisions are a critical aspect of law itself. The authority of precedent is a postulate of our legal system. It is uncontroversial. It is warm and fuzzy. And, in some of its most familiar applications, it is unconstitutional.

One should be clear about the scope of this claim. In the circumstances that I describe in Part I below, the practice of following precedent is not merely nonobligatory or a bad idea; it is affirmatively inconsistent with the federal Constitution. My modest tasks here are to persuade you to take this suggestion seriously and to focus attention on the questions that need to be answered, and the assumptions that need to be made and defended, by advocates of precedent.


My constitutional argument against precedent is subject to four threshold qualifications. First, the argument addresses only the use of precedent by federal courts. The extent to which my argument carries over to any particular state court depends on the proper interpretation of the relevant state constitution.

Second, the argument applies only to horizontal precedent that is, the use of precedent by a court at the same, or a higher, level of the judicial hierarchy than the court whose decision is being put forward as a precedent.3 The question whether inferior federal courts are permitted, or obligated, to follow the precedents of superior courts raises more complicated issues, and I want to put it aside for the moment.4

Third, the argument put forward here is limited to cases involving interpretation of the federal Constitution. I believe that the analysis is fully generalizable to cases involving statutory interpretation, but I want to limit myself here to the simpler setting of constitutional adjudication. My argument emphatically does not extend to common-law adjudication, which has been the principal subject of sophisticated scholarly analyses of precedent.5

3. Following Larry Alexander, I henceforth call a court whose decision is invoked as a precedent the "precedent court" See Larry Alexander, Constrained by Precedent, 63 S. CAL. L. REV. 1, 6 (1989).

4. Some very brief preliminary thoughts on this question appear in Steven G. Calabresi & Gary Lawson, Equity and Hierarchy: Reflections on the Harris Execution, 102 YALE LJ. 255, 276 n. 106 (1992).

5. In this respect, my approach is precisely the opposite of that taken by Professor Schauer in his seminal article on precedent See Frederick Schauer, Precedent, 39 STAN. L. REV. 571 (1987). Professor Schauer's article explores reasoning by precedent as a general social phenomenon and points out the numerous ways in which our thinking about pre-

Fourth, and finally, my argument considers only the practice of treating prior judicial decisions as legally authoritative merely by virtue of their status as prior judicial decisions. Courts are free to give weight, and even decisive weight, to prior decisions because of the persuasiveness of their reasoning just as they may give weight, and even decisive weight, to persuasive arguments in briefs, law review articles, or newspaper columns. Precedents are not worth less than these sources, but neither are they worth more.


THE GOOSE. . . .

So why would anyone believe such a silly thing as that precedent is unconstitutional, much less offer such a suggestion to one of the most high-powered panels ever to grace a Federalist Society conference? The more interesting question, really, is why the practice of following precedent has gotten such an easy ride. The prima facie case against horizontal precedent in federal constitutional adjudication is elegantly simple, devastatingly powerful, and strikingly familiar from other contexts.

Consider Marbury v. Madison6 and the constitutional case for judicial review of legislation.7 Suppose that the Congress, employing the formalities for lawmaking specified in the Constitution,8 enacts a statute. A person whose rights are affected by that statute challenges its application in a federal judicial proceeding on the ground that the statute violates the Constitution and therefore cannot be given legal effect. The court, which is called upon to exercise its Article III "judicial Power"9 by deciding the case, agrees that the statute violates the Constitution. The court

cedent in the law can be illuminated by considering our use of precedent in non-legal contexts. I instead argue that federal constitutional adjudication has a unique decisionmaking structure that precludes straightforward application of general principles of precedential reasoning. Thus, unlike Professor Schauer, I aim to narrow rather than broaden our focus regarding precedent.

6. 5 U.S. (1 Cranch) 137 (1803).

7. I recognize that there may also be a non-constitutional, or extra-constitutional, case for judicial review, just as there may be a non-constitutional case for precedent. My concern here, however, is solely with the prescriptions contained in the Constitution. I leave the prescriptions to be drawn from moral philosophy to moral philosophers two of the best of whom are on this panel.

8. See U.S. CONST. art. I, § 7 (describing the bicameral voting and presidential presentment requirements for valid lawmaking).

9. U.S. CONST. art. III, § 1 ("The judicial Power of the United States, shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish.").

must now determine what effect, if any, to give to the statute in the case before it.

Marbury's eminently correct answer to this problem is: "Too bad for the statute."10 According to Marbury's reasoning (and it is hopefully evident that I am invoking Marbury solely because its reasoning is persuasive), "[t]he judicial Power" is fundamentally the power to decide cases in accordance with law. The Constitution stands at the apex of our legal system as supreme law, hierarchically superior to all other claimed sources of law. An enactment that has the form of a statute, but that violates the Constitution, has no valid legal status.11 A court called upon to decide a case in accordance with law thus has not merely the power, but the obligation, to disregard such a legislative usurpation. Furthermore, the court must prefer the Constitution to the statute even when doing so will upset settled expectations, reduce judicial economy, or result in differential treatment for similarly-situated litigants.

This is the true import of Chief Justice Marshall's oft-misunderstood injunction that "[i]t is emphatically the province and duty of the judicial department to say what the law is."12 The court must decide the case in accordance with law, and a vital part of the judicial task is to determine whether a claimed source of law even one that seems prima facie to be a proper subject of judicial cognizance may be inapplicable to the case at hand because it conflicts with some hierarchically superior legal source. As Marshall aptly observed in Marbury, "[i]f two laws conflict with each other, the courts must decide on the operation of each."13 If the Constitution is truly supreme law, a legislative expression that violates the Constitution cannot repeat, cannot properly be given effect in an adjudication.

Suppose that the enacting Congress (and perhaps also the signing President) objects that it has already determined that the

10. See 5 U.S. (1 Cranch) at 177 ("It is a proposition too plain to be contested, that the constitution controls any legislative act repugnant to it. ... [A]n act of the legislature, repugnant to the constitution, is void.").

11. At least, it has no legal status once it is properly challenged. Suppose a statute really does violate the Constitution, but the adversely affected party does not raise that challenge. May the court give effect to the statute in that case? The answer is yes, so long as the statute does not unconstitutionally expand (or contract) the subject-matter jurisdiction of the court. In a party-driven system of litigation, unconstitutional statutes are voidable rather than void. I will henceforth ignore this refinement.

12. Marbury, 5 U.S. (1 Cranch) at 177 (emphasis added).

13. Id.

statute is constitutional. The court's proper answer is, "thanks, but we'll check anyway." On Marshallian premises, legislative or executive interpretations of the Constitution are no substitute for the Constitution itself. The court's job is to figure out the true meaning of the Constitution, not the meaning ascribed to the Constitution by the legislative or executive departments.14

Precisely the same analysis holds if the court is called upon to judge an exercise of the executive power. Executive action that violates the Constitution can have no valid legal status in an adjudication. It would be affirmatively wrong for a court to give such action the force and effect of law when the hierarchically supreme Constitution says otherwise. Again, the court's task is to ascertain the Constitution's true meaning, not the meaning ascribed to it by the President.15

Suppose now that a court is faced with a conflict between the Constitution on the one hand and a prior judicial decision on the other. Is there any doubt that, under the reasoning of Marbury, the court must choose the Constitution over the prior decision? If a statute, enacted with all of the majestic formalities for lawmaking prescribed by the Constitution, and stamped with the imprimatur of representative democracy, cannot legitimately be given effect in an adjudication when it conflicts with the Constitution, how can a mere judicial decision possibly have a greater legal status? If the Constitution says X and a prior judicial deci-

14. This glib statement sidesteps the question whether the prior legislative or presidential determination of constitutionality is entitled to any deference that is, what standard of proof a court should employ when deciding constitutional questions. See generally Gary Lawson, Proving the Law, 86 Nw. U. L. REV. 859 (1992) (discussing standards of proof for legal propositions). The Marshallian argument for judicial review strongly supports, though it does not compel, the conclusion that judges should exercise independent judgment when interpreting the Constitution. See Gary Lawson, Thayer Versus Marshall, 88 Nw. U. L. REV.__(forthcoming 1994) (contrasting the Marshallian argument with James Bradley Thayer's argument for deference to the constitutional judgments of the political departments). Whatever the appropriate standard of proof may be, however, once the court determines in accordance with that standard that a statute is unconstitutional, the court has an unconditional obligation to act on that judgment in the case at hand.

15. By the same token, Congress and the President are not bound by judicial interpretations of the Constitution but must make their own independent judgments. It is increasingly recognized that Marbury's argument for judicial review leads to this "departmentalist" view that each department of the national government has a co-equal power and obligation to interpret the Constitution. See generally THE FEDERALIST SOCIETY, WHO SPEAKS FOR THE CONSTITUTION? THE DEBATE OVER INTERPRETIVE AUTHORITY (1992) (collecting writings on departmentalism). Judge Frank Easterbrook has elegantly described the Marshallian constitutional rule of interpretation as "'Every man for himself.' Each official owes the same duty to the [legal] hierarchy and must make his own decisions." Frank H. Easterbrook, Presidential Review, 40 CASE W. RES. L. REV. 905, 919-20 (1990).

sion says Y, a court has not merely the power, but the obligation, to prefer the Constitution. Furthermore, if courts must search for the true meaning of the Constitution, rather than the meaning ascribed to it by the Congress or the President, there is no apparent reason why they must not also prefer the document's true meaning to the meaning ascribed to it by a precedent court. Thus, the case for judicial review of legislative or executive action is precisely coterminous with the case for judicial review of prior judicial action. What's sauce for the legislative or executive goose is also sauce for the judicial gander. At least as a prima facie matter, the reasoning of Marbury thoroughly de-legitimizes precedent.


If the argument against precedent is this simple, why does it tend to generate bemused laughter rather than universal assent?16 Different people have defended the use of precedent on different bases, but at least three classes of arguments might be thought to support the use of precedent. None of these classes of arguments defeats the prima facie case against precedent, but two of them deserve careful attention.

The class of pro-precedent arguments that does not deserve careful attention involves the claim that following precedent serves important prudential interests, such as stability, predictability, judicial economy, fairness, and legitimacy.17 Even if the practice of following precedent in fact promotes these interests,18 that would at most establish that a well-crafted constitution would permit, or require, courts to follow precedent. I have no strong view, and do not mean to imply one here, on how a well-crafted constitution should handle precedent. My present con-

16. Prior critics of precedent have stopped short of actually declaring the practice unconstitutional. See Charles J. Cooper, Stare Decisis: Precedent and Principle in Constitutional Adjudication, 73 CORNELL L. REV. 401 (1988); David E. Engdahl, What's in a Name? The Constitutionality of Multiple "Supreme" Courts, 66 IND. LJ. 457, 507-10 (1991); Note, James C. Rehnquist, The Power that Shall Be Vested in a Precedent: Stare Decisis, the Constitution, and the Supreme Court, 66 B.U. L. REV. 345 (1986). Judges occasionally declare their allegiance to the Constitution over precedent, see, e.g., William O. Douglas, Stare Decisis, 49 COLUM. L. REV. 735, 736 (1949), but I know of no judge who expressly renounced the use of precedent on constitutional grounds.

17. For an excellent representative argument, see Earl Maltz, The Nature of Precedent, 66 N.C. L. REV. 367, 368-72 (1988).

18. For reasons too complicated to explore here, the true practical effect of following precedent is not clear. For a profound introduction to some of the problems, see Frank H. Easterbrook, Stability and Reliability in Judicial Decisions, 73 CORNELL L. REV. 422 (1988).

cern is with the actual Constitution, however well- or ill-crafted it may be, and arguments from prudence go nowhere unless they are tied to the interpretation of some provision of the constitutional text.

A second class of pro-precedent arguments claims to find a textual provision in which to ground the practice of following precedent. One can plausibly argue that the power to follow precedent stems from the same source as the power to decide cases in the first instance: the grant of "[t]he judicial Power" in Article III.19 The Constitution's framers, the argument runs, were well aware of the established British practice of treating precedent as a source of law a practice that extended to the interpretation of written texts, such as statutes.20 To the best of my knowledge, none of the framers or ratifiers of the Constitution suggested that the Constitution in any way altered the historically accepted practice of following precedent Thus, the argument might conclude, when the Constitution authorized judges to decide cases, it must also be taken to have authorized them to use the tools traditionally employed by judges in that endeavor, including the attribution of legal effect to prior decisions.

The argument is tempting, but it sidesteps rather than rebuts the prima facie case against precedent. As John Marshall well recognized in Marbury, once it is granted that "[t]he judicial Power" is fundamentally the case-deciding power, one needs to know the sources of law that courts should employ when deciding cases and the hierarchical order of those sources in the event of a conflict among them. In the American system, the written Constitution is hierarchically superior to all other potential sources of law. Thus, courts will ordinarily give effect to other sources of law, such as statutes, that apply to the case in question, but if the statute conflicts with the Constitution, the court must disregard the statute. "The judicial Power" includes, as a structural inference from the Constitution's status as the supreme legal document, the unconditional obligation to prefer the Constitution to lesser sources of law, such as statutes. The same structural inference, whether or not it was specifically intended by the framers or ratifiers of the Constitution, must similarly limit what is other-

19. See Akhil Reed Amar, Our Forgotten Constitution: A Bicentennial Comment, 97 YALE LJ. 281, 294 n. 51 (1987); Monaghan, supra note 1, at 754.

20. See H. Jefferson Powell, The Modem Misunderstanding of Original Intent, 54 U. CHI. L. REV. 1513, 1536-37 (1987).

wise the power of a court to give legal effect to prior judicial decisions. That is, putting aside any structural inference of a power of judicial review, the grant of "[t]he judicial Power" certainly includes the power and perhaps even the duty to treat prior decisions as a source of law, just as it includes the traditional power (and perhaps duty) to treat applicable statutes as sources of law. However, if the structural inference in favor of judicial review qualifies Article III's grant of the case-deciding power where statutes (or executive actions) are concerned, it must also qualify the case-deciding power where precedents are concerned. If the Constitution is supreme law, it is supreme over all competing sources of law.

There is, however, one important respect in which Article III permits, and indeed requires, adherence to an incorrect judicial decision. The federal courts' Article III power to decide cases necessarily includes the power to decide them with finality. That is what distinguishes the case-deciding power from the power to issue advisory or recommendatory opinions. Thus, "[t]he judicial Power" includes the power to issue binding judgments that are not subject to collateral challenge by the parties, and that must be enforced by the executive against the parties, even when the executive (correctly) believes that the court has made a mistake.21 Yet while the meaning of "[t]he judicial Power" as the case-deciding power supports this notion of the finality, or enforceability, of particular judgments, it does not support a wider application of precedent outside the four corners of the particular case.22 "The judicial Power" is given full effect as the case-deciding power even if precedent is excluded as a proper source of law in the decision-making process.

I suspect, however, that most doubts about the Marshallian argument against precedent stem less from alternative constructions of Article III than from rejection of the argument's implicit

21. See David E. Engdahl, John Marshall's "Jeffersonian" Concept of Judicial Review, 42 DUKE LJ. 279, 312-31 (1992). In the American system, collateral attack on a judgment is sometimes permissible for example, if the judgment was in some respect fraudulent or issued without notice. See 18 CHARLES A. WRIGHT ET. AL., FEDERAL PRACTICE AND PROCEDURE §§ 4415, 4426 (1981) (describing exceptions to the general principles of, respectively, claim preclusion and issue preclusion). I have not fully considered whether the Constitution prescribes not merely the doctrine of res judicata but its form and details as well.

22. See Engdahl, supra note 21, at 313-14. I have not fully considered whether principles of repose other than res judicata, such as collateral estoppel or law of the case, can also be derived from Article III's grant of the case-deciding power, though I suspect that they cannot be so justified.

view of constitutional interpretation. The argument seems to assume that one can simply place a judicial decision alongside the Constitution, identify an inconsistency, and demand that a court choose the Constitution over the incorrect decision. That is not, a critic might claim, how the world of constitutional interpretation works. The critic can then endorse any of the myriad competing conceptions of interpretation that reject the Marshallian approach.

The critic is absolutely correct that my argument rests on the premise that the Constitution has, at least in principle, an objectively ascertainable meaning. Obviously, I cannot here canvas the many powerful arguments against this kind of naive right-answerism,23 so I will confine myself here to some general observations about the interplay between theories of precedent and theories of interpretation.

First, unless the critic asserts the total indeterminacy of constitutional meaning (which is an absurd assertion),24 the absence of right answers in some cases affects the practical scope of my argument but not its formal soundness. The use of precedent will still be unconstitutional in any case in which a right answer is ascertainable. This is a powerful conclusion, even if such cases are relatively rare.

Second, and much more fundamentally, using constitutional indeterminacy to defend the doctrine of precedent is like saving the bathwater by discarding the baby. I have deliberately described my argument against precedent as "Marshallian," not because it was specifically endorsed by John Marshall nor because I regard his views as authoritative, but in order to emphasize that the constitutional case against precedent is precisely coterminous with the constitutional case for judicial review. To the extent that indeterminacy undermines the argument against precedent, it also under-

23. I borrow this felicitous phrase from my fellow naive right-answerist, John Harrison. See John Harrison, Discussion: The Role of the Legislative and Executive Branches in Interpreting the Constitution, 73 CORNELL L. REV. 386, 386 (1988).

24. If all constitutional meaning is indeterminate, my constitutional argument against precedent fails, but then so does every other kind of constitutional argument, including the argument that judges have the constitutional power to decide cases. More fundamentally, the assertion suffers from the same self-referential problem that plagues all skeptical theories: There is no good reason to think that the Constitution is less intelligible than the skeptic's own utterances, so if the former's meaning is unknowable, so is the latter, and there is thus no cognizable skeptical argument to consider.

mines the case for judicial review in the first instance.25 So far as the Constitution is concerned, the cases for judicial review and against precedent stand or fall together.

Third, the argument against precedent holds for any theory of interpretation that prescribes objective right answers26 to constitutional questions, regardless of how those answers are derived, as long as the theory ascribes supreme legal status to the Constitution. It is emphatically not an argument that is uniquely employable by textualists and originalists. Suppose, for example, that you believe that the meaning of the Constitution evolves over time in keeping with modern social mores. As long as the evolved meaning of Article III includes the notion that judges are obliged to prefer the Constitution to other sources of law, the use of precedent is constitutionally forbidden, however the meaning of the Constitution is ascertained. Objective right-answerism of any sort plus Marbury equals trouble for precedent.

Fourth and finally, however, there are some non-objective forms of right-answerism that do allow room for precedent. For example, consider the view that the right constitutional answers are whatever a majority of the Supreme Court thinks are the right answers at any given time. The First Amendment thus means precisely what the judges presently think it means, and "[t]he judicial Power" means precisely what the judges presently think it means. Under this theory of interpretation, as long as judges presently say that "[t]he judicial Power" includes the power to give legal effect to prior decisions, that proposition is true by definition, and the practice of following precedent is therefore constitutionally justified. Put more generally, if you believe that constitutional meaning stems not from something external to the judicial process, but from the process of judicial interpretation itself, then you can also argue that the authority of precedent is (because the judges say so) part of the definition of a right answer. Of course, if the Constitution means whatever judges presently think it means, the justification for judicial review will be, to say the least, interesting.

Thus, in the constitutional world, precedent cannot be treated apart from a more general theory of interpretation. One cannot

25. Cf. Easterbrook, supra note 15, at 923-24 (explaining that the absence of determinate answers to constitutional questions defeats the claim of a court to disregard sources of law such as statutes).

26. By "objective" I mean "external to the act of decision-making."

adopt an interpretative theory and then simply add on a theory of precedent as icing. The chosen theory of interpretation will have implications for the meaning of Article III and will determine whether there is an interpretative standard outside the act of decision-making. Both considerations profoundly affect whether there is a sound constitutional case for precedent.


As is generally true with academic discussions of jurisprudential topics, the doctrine of constitutional stare decisis is in no real-world danger from me, John Marshall, or anyone else. The doctrine of precedent is too deeply ingrained in the legal system to permit serious inquiry into its own legitimacy. Precedent is not unique in this respect. No matter how convincingly one can demonstrate, for example, that fiat currency not immediately redeemable in precious metals is unconstitutional (and one can demonstrate it beyond a reasonable doubt), no one is going to pay attention, and litigants who raise the issue will routinely be sanctioned. Yet the legal system should at least have the decency to acknowledge openly that it has chosen political expediency over the Constitution.27 If judges are going to continue to employ precedent, there is value in acknowledging that the practice is nothing more than "a sort of intellectual adverse possession"28 and that the territory adversely possessed is nothing less than the Constitution.

27. Once that acknowledgment is made, one can then consider whether to adopt damage-control measures short of wholesale abandonment of precedent. For a fascinating argument that the irrevocable enshrinement of errors through precedent should compel the self-conscious creation of further errors in subsequent cases, see Peter B. McCutchen, Mistakes, Precedent, and the Rise of the Administrative State: Toward a Constitutional Theory of the Second Best (manuscript on file with the author).

28. Tyler Pipe Indus., Inc. v. Washington Dep't of Revenue, 483 U.S. 232, 265 (1987)(Scalia,J., dissenting).