With Stare Decisis and Constitutional Text, Jonathan Mitchell has produced what I think is the most interesting and creative textual defense (or at least partial defense) to date of the use of horizontal precedent in federal constitutional cases. Mitchell's careful analysis of the Supremacy Clause is fascinating and instructive, and he does an impeccable job of drawing out the implications of his premise that the Supremacy Clause prescribes only a very limited choice-of-law rule-a rule that does not, by its own terms, specifically elevate the Constitution above federal statutes and treaties. His innovative and intriguing framework yields four distinct conclusions about the permissible uses of precedent. In brief, under Mitchell's analysis, the Supremacy Clause forbids using precedent (1) to invalidate congressional statutes (because congressional statutes are the supreme law of the land while prior court decisions are not) or (2) to uphold constitutionally challenged state laws (because the Constitution is the supreme law of the land while prior court decisions and state-law interpretations of the Constitution are not). Yet, according to Mitchell, the Supremacy Clause does not forbid using precedent (3) to uphold congressional statutes (because both the Constitution and congressional statutes are equally supreme, and there is no constitutional mandate to prefer one to the other) or (4) to invalidate state laws (because neither prior court decisions nor state laws are supreme, and there is no constitutional mandate to prefer one to the other). The article is an eminently worthy contribution to a vibrant debate, and I am delighted to have the opportunity to respond to it-as well as to clarify some ambiguities in my own prior work on precedent.
For all of its elegance and provocativeness, however, Mitchell's analysis has one major problem: it places all of its eggs in one basket, and that basket simply cannot hold the eggs. Mitchell's argument is all about the Supremacy Clause. But the constitutional case against precedent does not depend upon the Supremacy Clause. The argument against precedent would exist without the Supremacy Clause. While the Supremacy Clause provides a modicum of support for the argument, the clause's presence in the Constitution actually has very little effect on the argument's logical structure.
To be sure, Mitchell's contrary reading of my prior work is entirely fair: there are certainly passages in that work that seem to say that the case against precedent flows from the text of the Supremacy Clause. It is no defense that I did not intend to make that claim; the baseline meaning of a communicative text is the meaning that would be perceived by a reasonable observer, and a reasonable observer could easily interpret my argument as did Mitchell. In truth, however, one can (and perhaps should) construct the constitutional case against precedent with no mention of the Supremacy Clause whatsoever. Thus, if my antiprecedent position makes me something of a rebel in the constitutional world, I am a rebel without a clause.
I take this opportunity to set forth the minimalist role of the Supremacy Clause in the constitutional case against precedent, which renders much of Mitchell's careful analysis beside the point. But nothing that I say here should detract from my admiration for this project and for Mitchell. The academy's (I trust temporary) loss is most definitely a big gain for the people of Texas.
I. Deconstructing the Supremacy Clause
The federal Constitution's Supremacy Clause provides:
This Constitution, and the Laws of the United States which shall be made in Pursuance thereof; and all Treaties made, or which shall be made, under the Authority of the United States, shall be the supreme Law of the Land; and the Judges in every State shall be bound thereby, any Thing in the Constitution or Laws of any state to the Contrary notwithstanding.
Mitchell is correct to point out that the text of this provision does not set out an express hierarchical ordering among the three named sources of federal law (the Constitution, laws made pursuant to the Constitution, and treaties made under the authority of the United States); it simply declares those three sources of federal law to be hierarchically superior to two named sources of state law (state constitutions and state laws). He is also correct that the clause's reference to federal laws made "in Pursuance" of the Constitution is most likely a reference to laws enacted according to the lawmaking procedures specified in Article I, section 7, rather than a reference to federal laws that substantively conform to wider constitutional principles. He is further correct that the Supremacy Clause says absolutely nothing about the hierarchical ordering among sources of law other than the five sources specifically named in the provision. That is, most of the key premises about the text and structure of the Supremacy Clause that drive his argument are correct. Mitchell is fundamentally wrong, however, to identify the Supremacy Clause as "the very provision that the textualist critics [most notably including yours truly] invoke for their attacks" on precedent-or at least he is wrong that such critics must invoke or rely upon the Supremacy Clause. Because his defense of precedent is rooted in the Supremacy Clause, it is actually tangential rather than responsive to the primary originalist case against precedent.
The Supremacy Clause is a very thin reed on which to base much of anything regarding the application of the Constitution. Indeed, if one actually tries to construct a general theory of precedent out of the Supremacy Clause, as Mitchell assumes that originalist critics of precedent try to do, one encounters problems even more basic than those identified by Mitchell. First, while the initial phrase in the Supremacy Clause, which identifies a hierarchy among certain state and federal legal sources, is perfectly general and is therefore addressed to all constitutional interpreters, the second phrase, which prescribes a legal effect for the preceding phrase, binds only "the Judges in every State." That latter term is an obvious reference to state court judges, which means that the pure text of the Supremacy Clause does not issue legal commands to federal judges at all, much less to federal or state legislative or executive officials (or jurors, or citizens). Accordingly, it would be absurd to construct a universal theory of precedent that binds federal officials out of a clause that does not appear to speak to those officials. Even if one stretches the term "Judges" to include federal and not just state judges, the clause would not appear to apply to judges in federal territories or the District of Columbia, since such "Judges" are not "in" any particular "State." Supreme Court Justices, for example, would be bound by the clause only when "riding circuit" in a specific state. And there is no way to stretch the term "Judges" to include other officials. Thus, a pure textualist analysis of the Supremacy Clause yields far too narrow a scope to let it serve as the basis for a grand (much less a Grand) theory of precedent.
Second, the clause elevates the three named sources of federal law only above state constitutions and state laws. It says nothing about the hierarchical status of state court decisions, state administrative regulations, or state jury decisions. Does that mean that those other exercises of state legal authority are not necessarily subordinated to supreme federal law? If the only constitutional source of conflict-of-laws principles is the text of the Supremacy Clause, an affirmative answer seems difficult to avoid.
These problems with using the Supremacy Clause as the sole basis for a broad theory of precedent stem from trying to read the Supremacy Clause as more than it is. The Supremacy Clause is a provision that deals with one specific set of conflicts among sources of law that would predictably arise once the Constitution was ratified. It does not purport to exhaust the universe of conflict-of-laws principles, nor does it even necessarily establish a new principle that would not otherwise exist in its absence. Many constitutional provisions exist solely or primarily for emphasis or clarification; it would be neither odd nor poor drafting for the Supremacy Clause to serve that function. It is clear that the Supremacy Clause cannot serve as the foundation for a comprehensive conflict-of-laws theory, and because precedent is a particular species of conflict-of-laws problems, the Supremacy Clause cannot serve as the foundation for a comprehensive theory of precedent.
II. Reconstructing Constitutional Supremacy
So if the case against precedent does not rest on the Supremacy Clause, on what does it rest? It rests, as I argued (albeit with regrettable ambiguity) almost two decades ago, on exactly the same foundation as the case in favor of judicial review. It rests on the supremacy of the Constitution over all competing sources of law, including the other sources identified by the Supremacy Clause as themselves supreme over state laws and constitutions. This constitutional supremacy does not derive textually from the Supremacy Clause. One can perhaps invoke the sequencing of sources of law within the Supremacy Clause (the Constitution, then federal laws, then federal treaties) as some modest evidence in favor of the Constitution's place at the top of the legal food chain, which is precisely how Chief Justice Marshall invoked it more than two centuries ago in Marbury v. Madison. But the supremacy of the Constitution logically precedes the Supremacy Clause. Indeed, it is the supremacy of the Constitution that makes the Supremacy Clause itself supreme over contrary statements in state laws or constitutions.
The case for constitutional supremacy set out in Marbury v. Madison remains today as good as anything that has followed it. The case flows from a series of mutually reinforcing arguments, none of which specifically invokes the Supremacy Clause. First, the very nature of a written constitution that creates a new government and sets forth its powers and limitations, in a pure and original act of legal creativity, implies that the constitution is hierarchically superior to the laws and other institutions that spring from it. As Chief Justice Marshall explained:
That the people have an original right to establish, for their future government, such principles as, in their opinion, shall most conduce to their own happiness, is the basis, on which the whole American fabric has been erected . . . . The principles, therefore, so established, are deemed fundamental. And as the authority, from which they proceed, is supreme, and can seldom act, they are designed to be permanent.
Marshall's point is implicit in the Preamble, which declares that "We the People of the United States . . . do ordain and establish this Constitution for the United States of America."
To be sure, it is not logically impossible to have a constitution that is hierarchically equivalent to, or even subordinate to, other sources of law such as ordinary legislation. The founding generation was quite familiar with constitutions that could be changed by ordinary legislation. But the federal Constitution, as a self-conscious act of political creation, is best understood as holding itself out, by its very nature, as supreme law.
Second, on a purely textual level, the Constitution specifically prescribes that officials swear oaths to uphold it. A provision of Article VI other than the Supremacy Clause declares that "[t]he Senators and Representatives before mentioned, and the Members of the several State Legislatures, and all executive and judicial Officers, both of the United States and of the several States, shall be bound by Oath or Affirmation, to support this Constitution . . . ." Article II prescribes that the president give the following oath, before taking office: "I do solemnly swear (or affirm) that I will faithfully execute the Office of President of the United States, and will to the best of my Ability, preserve, protect and defend the Constitution of the United States." Government officials are not constitutionally mandated to swear equivalent oaths to uphold laws, treaties, judicial decisions, or other legal instruments.
Third, the particular terms and structure of the Constitution, which contains some very specific provisions regulating the composition, character, and scope of the national government (and to a lesser extent state governments) suggest, as Chief Justice Marshall argued more than two centuries ago, that the Constitution is meant to control, and therefore be superior to, these other sources of law.
While there is no deductively airtight case for constitutional supremacy, on balance it seems more plausible to suppose that the Constitution is hierarchically superior to all other sources of law than to suppose the contrary. That is why Chief Justice Marshall was right in principle in Marbury to subordinate congressional statutes to the supreme Constitution (though he may have been wrong in practice about the meaning of both the relevant constitutional provision and the relevant statute in that case).
And once the supremacy of the Constitution is established-a supremacy that can be established without use of the Supremacy Clause-the prima facie case against precedent is set. The "judicial Power" is the power to decide cases in accordance with governing law. If the Constitution conflicts with any other potentially applicable source of law, such as statutes or prior judicial decisions (and I am granting that prior judicial decisions can indeed count as law in some meaningful sense), the Constitution must prevail. Accordingly, the power and duty to decide in accordance with law includes the power and duty to decide in accordance with the Constitution, even when prior congresses, prior presidents, prior courts, prior state officials, and prior scholars have said otherwise.
Mitchell's strict textualist analysis of the Supremacy Clause does not rebut the constitutional case against precedent because the analysis rests on the false premise that the Constitution does not place itself hierarchically above all competing sources of law. It is true that the raw text of the Supremacy Clause makes no such direct and explicit claim to constitutional supremacy, but that does not mean that no such constitutional claim is made. As Vasan Kesavan has aptly noted, "It is unnecessary for the Constitution to specify that it is superior to other law because it is higher law made by We the People-and the only such law."
III. Un-constructing Constitutional Deference
Of course, as a practical matter, the theoretical case against precedent only has bite if the power and duty to decide in accordance with the Constitution includes the power and duty to decide in accordance with the decisionmaker's own best judgment about the meaning of the Constitution. If the Constitution permits judges to defer to the constitutional views of Congress or the president, it is hard to see why it would not also permit them to defer to the constitutional views of prior judges. Such a practice of deference-and precedent is unquestionably a subspecies of deference-would not involve "subordinating" the Constitution to other sources of law but simply "subordinating" one's own view of the Constitution to that of others. While this is not the main line of Mitchell's defense of precedent, it is implicit in his suggestion that relying on precedent, at least to uphold the constitutionality of federal statutes, is akin to deferring to the political departments. As Mitchell says, "On this view, stare decisis is no different from a tacit expansion of the political-questions doctrine, a ruling that grows the domain of interpretive questions to be resolved authoritatively by the national political branches."
I have elsewhere argued at some length that, with a few notable exceptions, federal courts generally are not constitutionally required to defer to the constitutional views of other actors. But are they allowed to do so? Of course they are-provided that such deference is a reasonable attempt to decide cases in accordance with governing law. There are many instances in which the views of someone other than the judge deciding the case are likely to be more in accordance with the Constitution than the judge's own views. In such instances, the judge is not merely permitted but required to defer. The same is true with respect to deference to the constitutional views of prior judges: reliance on "precedent" in this epistemological sense is not merely permissible but mandatory when it is the best available means of determining the right answer to constitutional questions.
What federal judges are not permitted to do is defer to the views of other actors for reasons other than good faith attempts to get the right answer, such as theories of democracy, concerns about efficiency or decisionmaking costs, or considerations of stability or reliance. That is because the sole constitutional power of the federal judge is "[t]he judicial Power," which is quintessentially the power to decide cases in accordance with governing law. Unless the Constitution itself says otherwise, anything that involves deciding cases in accordance with something other than governing law-and the Constitution is supreme governing law when it applies-is outside the constitutionally granted power of the federal judge and is therefore impermissible. To be sure, the optimal determination of what counts as governing law does not always require the judge to determine every matter from scratch; indeed, on some occasions that kind of lone wolf approach would be affirmatively forbidden by the primary obligation to get the right answer. But any reliance on other actors must be justified, directly or indirectly, by its ability to facilitate decisionmaking according to law. Any reliance on precedent that does not ultimately trace to getting the right answer to constitutional questions exceeds the boundaries of the judiciary's authority and is itself unconstitutional. 
Does that mean that the political questions doctrine is a mistake-and that, derivatively, expanding it through precedent would also be a mistake? In many, and perhaps even most, instances, probably yes. There are limited circumstances in which the Constitution effectively prescribes decisionmaking authority over certain matters to actors other than judges. When faced with this kind of "textually demonstrable constitutional commitment of the issue to a coordinate political department," one must obey the Constitution's allocation of interpretative authority. But in the absence of such a specification, there is no warrant for deferring to the views of other actors except a good faith judgment that those other actors are better situated than is the judge to determine the right answer.
In short, with respect to interpretative authority, the constitutional rule really is: That which is not forbidden is required and that which is not required is forbidden. Reliance on precedent for any reason other than epistemological reliability is forbidden.
IV. Constructing Constitutional Proof
We can now see why Mitchell's four-part schema does not accurately describe the constitutional rules regarding precedent.
With respect to the use of precedent to uphold the constitutionality of federal statutes, Mitchell posits a constitutional equivalence between the Constitution and statutes enacted pursuant to the Constitution. There is no such equivalence. The Constitution is supreme-not because the text of the Supremacy Clause says so (it does not) but because the nature and structure of the federal Constitution gives rise to that inference. With respect to the use of precedent to invalidate federal statutes, Mitchell's conclusion that such use is improper is right, but for the wrong reason. It is not because the Constitution is mentioned in the Supremacy Clause while court decisions are not, but simply because the Constitution is hierarchically superior to court decisions-and would be hierarchically superior even in the absence of the Supremacy Clause. The same is true with respect to Mitchell's conclusion that courts may not use precedent to uphold state laws against federal constitutional challenges: the conclusion is right, but that conclusion does not depend upon the particular wording of the Supremacy Clause. And with respect to the use of precedent to invalidate state laws, Mitchell is at least partly wrong. Where the best understanding of the Constitution is that the state law is invalid, one must invalidate the state law.
But what if there were a judicial precedent that one believed was not the best understanding of the Constitution, and the state law would not be in conflict with the Constitution on that best understanding, but the state law would be in conflict with the judicial precedent? Could a court nonetheless invalidate the state law as inconsistent with precedent? I have no view on that matter. That would require establishing a conflict-of-laws hierarchy with respect to state laws and federal judicial decisions (and perhaps the "general law" on which those decisions are based). My argument establishes only that the federal Constitution trumps everything else. I have nothing interesting to say about the hierarchy below that level, involving things like state laws, general law, international law, and so forth.
But surely it cannot be the case that federal courts (and other legal actors) must determine the right answers to constitutional questions no matter the cost in terms of time and money. After all, wouldn't spending just a bit more time studying a problem always make it marginally more likely that one would reach the correct answer, so that it would never be proper to issue a decision without thinking about it some more? Doesn't there have to be some kind of decisionmaking heuristic, or at least a proxy, that decisionmakers can employ to bring their inquiries to an end?
Yes, there does, but it is not precedent (or any other consideration that tries to substitute itself for the right answer). The trick lies in defining what counts as a "right" answer. It turns out that "right" answers for ivory tower scholars are not necessarily the same things as "right" answers for real-world decisionmakers, even if all of those persons are employing essentially the same interpretative methodology. Real-world decisions always take place in a context in which someone has a burden of proof with respect to the relevant legal question and a standard of proof that he must meet in order to establish his claim. It is proper for a judge to make a legal decision when a party has satisfied (or failed to satisfy) the applicable burdens and standards of proof-and that can happen well before infinity.
The Constitution implicitly allocates burdens of proof, and it does so in precisely the opposite fashion advocated by Mitchell and, at least implicitly, by much of modern law. Mitchell argues that federal statutes come with a presumption of constitutionality while state statutes do not. That may very well be a good description of modern practice, and it may very well be a good description of the regime that Mitchell derives from the Supremacy Clause. But as a matter of original meaning, federal statutes come with a presumption of unconstitutionality, because the burden of proof initially lies with whoever claims that the federal government has the enumerated power to enact the statute in question. He who asserts must prove, and all claims of federal power logically begin with the assertion that the enumerated powers of the federal government permit the action in question. The presumption is precisely the opposite with respect to state laws. As far as the federal Constitution is concerned, state governments are limited in power only by specific provisions in the Constitution denying them certain powers; states do not need to trace their actions to affirmative authorizations in the Constitution. Thus, the burden of proof must lie on whoever raises a federal constitutional challenge to state authority. If the answer to a constitutional question seems indeterminate, then the legal ruling must go against federal power and in favor of state power.
The real question is how much evidence of constitutional meaning one must adduce in order to meet or overcome the relevant burdens of proof. The real question, in other words, is what standard of proof claimants must satisfy in order to win constitutional cases.
I raised this issue in preliminary fashion two decades ago and have said very little about the subject since that time. The reasons for that long silence have not changed. I have no answer to that question. But it is the question that one must answer.
I have not even touched on many aspects of Mitchell's argument that should attract the attention of scholars in widely varying fields for years to come. Mitchell has given us all a very rich vein to mine. At the end of the day, however, precedent is still mostly unconstitutional-for reasons having little if anything to do with the Supremacy Clause.
. Jonathan F. Mitchell, Stare Decisis and Constitutional Text, 110 Mich. L. Rev. 1 (2011).
. As Mitchell points out, defenses of precedent are almost always instrumentally rather than constitutionally grounded. The rare exceptions to that rule have tended to focus the discussion of precedent on deference or history rather than the constitutional text. See, e.g., Thomas Healy, Stare Decisis and the Constitution: Four Questions and Answers, 83 Notre Dame L. Rev. 1173 (2008) (discussing precedent in terms of deference); Lee J. Strang, An Originalist Theory of Precedent: Originalism, Nonoriginalist Precedent, and the Common Good, 36 N.M. L. Rev. 419 (2006) (discussing history).
. So that there is no mistake, I am not making any kind of normative claim about the moral authority or superiority of the Constitution over statutes, natural law, judicial decisions, or anything else. I am simply describing the best account of the Constitution's legal relationship to other sources of law within the American political order.
. It is, however, one of the main lines advanced by Professor Thomas Healy in a characteristically engaging article. See Healy, supra note 2, at 1184-88. Professor Healy deserves a longer response than I am providing here, and I might even provide one in the near future.
. Most obviously, to avoid so-called "double jeopardy" issues, federal courts are required to defer conclusively to the constitutional views of juries when those views result in acquittals in criminal cases. Less obviously, there is an argument that federal courts are constitutionally obliged to defer conclusively to the constitutional views of the parties to litigation when those parties all agree about constitutional meaning. Suppose that all of the parties to a case agree that the Constitution, for example, restricts the ability of state governments to regulate abortions, while the judge is firmly convinced that the Constitution contains no such restriction. Must the court prefer the Constitution as it really is to the fake "Constitution" about which the parties have chosen to argue? The question is actually trickier than it might seem. Article III does not grant federal courts a freestanding power to interpret the Constitution. It grants them the power to decide cases in accordance with governing law, which includes as a necessary incident the power to determine the governing law. But the principal power is the power to decide cases. What is a case? Is it the operative set of facts that give rise to the dispute among the parties, or is it the set of propositions about which the parties are arguing? If it is the latter, and if the parties agree on propositions about constitutional meaning, then those propositions are simply not part of the dispute before the court and are therefore not within the court's constitutional power to adjudicate-even if those propositions are objectively false. I have begun a preliminary exploration of these issues in the pages of this journal. See Gary Lawson, Stipulating the Law, 109 Mich. L. Rev. 1191 (2011). A full treatment of the application of those issues to Article III requires a separate article, which I hope to produce in the not-too-distant future.
. Everything that is true of the
interpretative responsibilities of federal courts is also true of the
interpretative responsibilities of the president and Congress. No federal
institution has a monopoly on, or even a preferred status with regard to,
constitutional interpretation. This means that presidents do indeed "violate
the Constitution or their Oath of Office if they choose to enforce and defend
all Acts of Congress-even when they sincerely believe that a statute contravenes
the Constitution." Mitchell, supra note 1, at 32. That "[t]he power to disregard a federal statute on
constitutional grounds is a[n] . . . implied power" does not mean that it is a "discretionary" power. Id. Far from making the Constitution a
"suicide pact," id., this dispersion
of interpretative authority is a perfectly sensible device for preserving
as is the dispersion of other governmental powers. See Gary Lawson, Interpretative
Equality As a Structural Imperative (or "Pucker Up and Settle This!"), 20 Const.
Commentary 379 (2003). As for the expectations and demands of federal
courts "that their decisions be enforced and obeyed, no matter how strongly the
political branches disapprove of their constitutional reasoning," Mitchell, supra note 1, at 29, there is a
presumption of enforceability that accompanies judgments in specific cases. But
that presumption is rebuttable if the president or Congress is firmly convinced
that the courts are wrong; and there is no presumption of
correctness-or at least no presumption not grounded in sound epistemological considerations-for the reasoning accompanying those judgments. See Lawson & Moore, supra note 23, at 1313-29. If the federal courts have stronger expectations, those expectations have no constitutional grounding.