The Yale Law Journal

December 2014

The Limits of Enumeration

abstract. According to a well-known principle of constitutional interpretation here identified as the “internal-limits canon,” the powers of Congress must always be construed as authorizing less legislation than a general police power would. This Article argues that the internal-limits canon is unsound. Whether the powers of Congress would in practice authorize any legislation that a police power would authorize is a matter of contingency: it depends on the relationship between the powers and the social world at a given time. There is no reason why, at a given time, the powers cannot turn out to authorize any legislation that a police power would. This Article explains why setting aside the internal-limits canon is consistent with the interests of federalism, with fidelity to the Founding design, and with the text of the Constitution.

author. Theodore J. St. Antoine Collegiate Professor, the University of Michigan Law School. For comments on previous drafts, I thank Evan Caminker, Richard Friedman, Abbe Gluck, Kelly Hagen, Scott Hershovitz, Don Herzog, William Novak, Kermit Roosevelt, and Dayna Zolle. Thanks also to Akhil Amar, Samuel Bagenstos, Nicholas Bagley, Scott Bloomberg, Jessica Bulman-Pozen, Kristina Daugirdas, Monica Hakimi, Rick Hills, Gerard Magliocca, Charles Primus, Romana Primus, Adam Pritchard, Eli Temkin, and David Uhlmann, as well as the participants in faculty workshops at the University of Michigan Law School and the University of Notre Dame Law School. The title of this Article is of course reminiscent of Gil Seinfeld’s Article I, Article III, and the Limits of Enumeration, 108 Mich. L. Rev. 1389 (2010); I thank Seinfeld for his long-running and productively skeptical engagement with the idea presented here. Research for this project was funded in part by the Cook Endowment.


“The enumeration presupposes something not enumerated.”

—Chief Justice John Marshall, Gibbons v. Ogden1

The federal government is a government of limited and enumerated powers. Every law student learns this formula. And so close on its heels that it sometimes seems to be the same idea, another principle follows: there are things Congress cannot do, even without reference to affirmative prohibitions like those in the Bill of Rights. For ease of reference, we can call the first idea the enumeration principle, and, for reasons to be explained just below, we can call the second idea the internal-limits canon. So long as it is properly understood, the enumeration principle is a sound tenet of American constitutional law. But the internal-limits canon is not. The purpose of this Article is to explain why the internal-limits canon, for all its familiarity and broad acceptance, is wrong.

In referring to the idea under consideration as the internal-limits canon, I draw on a useful typology that divides limits on congressional power into three kinds.2 Internal limits are the boundaries of Congress’s powers taken on their own terms. For example, the power to govern the District of Columbia3 can be used to write a fire code for the District of Columbia, but it cannot be used to write a fire code for Delaware. This limit is “internal” to the power itself, meaning that the limit inheres in the definition of the power. External limits, in contrast, are affirmative prohibitions that prevent Congress from doing things that would otherwise be permissible exercises of its powers. Thus, the Fifteenth Amendment prevents Congress from conducting whites-only elections in the District of Columbia, despite Congress’s power to govern the District. The rule against racial discrimination in voting is not conceptually part of the power to govern the District; before the adoption of the Fifteenth Amendment, Congress could use that power to conduct racially restrictive elections. The Fifteenth Amendment creates a separate constitutional rule that pushes back against the grant of power and thus limits that power “externally.”4 Finally, there are process limits, such as the bicameral legislature, the requirement of presidential presentment, and frequent democratic elections. Unlike external limits, process limits do not place particular substantive outcomes wholly out of reach. But they raise the cost of federal action, thus diminishing the likelihood that Congress will do any particular thing, especially any particular thing that might arouse substantial opposition.5

Process limits and external limits are consequential forces constraining modern federal governance. Internal limits are not. Indeed, for much of the twentieth century, many people suspected that internal limits had lost all practical significance.6 Judicial doctrine constrained Congress on the basis of prohibitions like those in the Bill of Rights, but broad constructions of the Commerce Clause made it hard to identify enforceable limits on Congress short of those affirmative prohibitions.7 At the level of principle, though, the idea that the Constitution demands a meaningful set of internal limits lived on.8 Defenders of federal statutes have always needed to answer the question, “If Congress can do that, what can’t Congress do, other than the things the Constitution specifically forbids?”9 That question played a famously large role in National Federation of Independent Business v. Sebelius (NFIB).10 And in the wake of NFIB, constitutional lawyers wonder whether the internal-limits canon—that is, the principle that the powers of Congress must be construed as meaningfully constrained by internal limits—might be deployed in seriously consequential ways.11

Now is the time, therefore, for a frontal debunking of the internal-limits canon. It is my aim in this Article to show, despite longstanding orthodoxy to the contrary, that Congress’s powers might in practice authorize the enactment of any legislation that would be justified by a grant of general regulatory authority. “Might” is an important part of the claim. In my view, whether the powers of Congress have as great a scope in practice as a general police power is a matter of contingency, not a matter of principle. The question can only be answered by examining the powers and applying them sensibly to the social world. But in the course of that analysis, no constitutional principle bars the conclusion that Congress’s enumerated powers in practice authorize as much as a police power would.

Measured by the conventions of constitutional discourse, rejecting the internal-limits canon would be a radical step. A familiar trope among constitutional lawyers would deem it an obvious mistake. “Of course the powers of Congress are inherently limiting,” this argument says. “After all, the powers of Congress are specifically enumerated in the Constitution. If Congress had general legislative power, the Constitution would have said that, rather than providing a list of particular powers. That’s what Chief Justice John Marshall meant in Gibbons v. Ogden when he said that the enumeration presupposes something not enumerated.”12

Now it happens that this familiar way of invoking Marshall’s Gibbons dictum may not get the great Chief Justice quite right, in part because it misses a nuance in the word “presupposes.”13 But more importantly, it isn’t true that enumerations of specific authorities are always more limiting than general authorizations would be. Yes, specific lists are probably specific for a reason, at least most of the time, and “Congress can do these eighteen things”14 might seem like a funny way of authorizing Congress to legislate however it thinks best, subject to the limits of the political process and the affirmative prohibitions specified elsewhere in the Constitution. But as a conceptual rule, the claim that enumerated authorizations are always more limiting in practice than general authorizations is too sweeping. Consider this example: is “you can have chocolate, vanilla, or strawberry ice cream for dessert” more limiting in practice than the general authorization “you can have ice cream for dessert”? The answer on any given day might be yes or it might be no. It depends on the contents of the freezer.15

Obviously this enumeration differs from the Article I enumeration in many ways, but it should bring the basic point into view: whether a list of specific authorizations is in practice more limiting than a general authorization depends on facts about that particular enumeration and the circumstances in which it is applied. So the mere fact that the Constitution includes an enumeration of congressional powers cannot demonstrate that the internal limits of those powers leave Congress unable to regulate something that a police power would let Congress reach. If the powers of Congress must be construed as collectively less extensive than a police power, it has to be for some other reason.

Prevailing constitutional opinion furnishes three such reasons, sounding in the traditional categories of text, history, and structure. As a textual matter, the argument runs, Article I and the Tenth Amendment both indicate that the enumerated powers are internally limited.16 Historically, the Founders saw the enumeration as a device for limiting Congress.17 Structurally, a federal police power would let Congress eclipse the state governments and destroy the federal system.18 These arguments have some plausibility, and they enjoy a long pedigree in constitutional thought. But on their merits, they are less than compelling.

The textual grounding for the internal-limits canon, I suggest, is powerful if one already believes that congressional power must be internally limited. But without that presumption, the text is more easily read to permit Congress’s enumerated powers to go wherever they might lead.19 As a matter of history, most of the Founders did see enumeration as a strategy for limiting (and invigorating) the federal government. But enumerating the powers of Congress was only one of the Founders’ strategies for limiting federal power, and fidelity to their design does not require forcing that strategy to do the job if it does not work very well and other constitutional strategies are more up to the task.20 Finally, the federal structure of American government has long been maintained not by internal limits on Congress’s powers but by a combination of external limits, process limits, and the practical conditions that shape interactions between federal and state officials.21 There is no reason to believe that these devices deliver optimal federalism, partly because there is no reason to believe that any set of tools could yield that outcome. But there is also no reason to think that a better brand of federalism would result if some consequential set of internal limits were added to the mix. In sum, internal limits are not mandated by the text of the Constitution, not required by fidelity to the Founding, and neither necessary nor materially helpful for promoting federalism.

So am I saying that Congress is authorized to do whatever it wants? Of course not. For one thing, the whole panoply of external constitutional limits is firmly in place. Congress may not establish Christianity,22 abolish jury trial,23 commandeer a state legislature,24 or unilaterally combine the two Carolinas into a single state.25 Nor am I arguing that the commerce power (or the union of that power and the rest of Congress’s powers) authorizes all possible legislation except what the external limits prohibit. In other words, I am not arguing that the Constitution confers the equivalent of plenary power on Congress. It might, or it might not, depending on the best constructions of many different powers and the relationship between those powers and the social world at any given time. My argument takes no position on whether the Constitution authorizes Congress to do whatever a national government with a police power could do. Instead, my argument is that the answer to that question is a matter of contingency, rather than a categorical “no.” In the course of analyzing the scope of any congressional power, I contend, one should not exclude an otherwise reasonable construction on the grounds that it would leave Congress constrained only by process limits and affirmative prohibitions.26

My argument also says nothing about the wisdom, as opposed to the permissibility, of federal legislation. State and local decision making is often better than central decision making, and Congress is often well advised to leave issues in the hands of local officials. Indeed, it might be sensible to say that Congress should prefer local decision making except where some reason suggests that regulation be federal, albeit on the understanding that there are many reasons why federal regulation is sometimes the right choice.27 States today exercise a great deal of consequential governing authority. That is a healthy condition, and for reasons explained in this Article it would remain the case even if the internal-limits canon disappeared. So the argument here is not that all law should be federal, nor even that more law should be federal than currently is. Congress should consider the virtues of local decision making before enacting legislation, and judges should disallow laws that violate principles of federalism by contravening external constitutional limits on congressional power, some of which are associated with the Tenth Amendment.28 But the worry that sustaining a given law would make it impossible to identify meaningful internal limits on congressional power is not a sufficient reason to deem that law invalid.

These limitations on the scope of my argument should not conceal the importance of rejecting the internal-limits canon. As noted above, constitutional practice long featured a disjuncture between the official theory of a limiting enumeration and a de facto settlement whereby internal limits did virtually nothing to constrain federal law.29 From the New Deal until United States v. Lopez30 and United States v. Morrison,31 the Supreme Court enforced no internal limits. Even after Lopez and Morrison, workarounds like the reenacted Gun-Free School Zones Act32 and the Court’s decision in Gonzales v. Raich33 seemed to indicate that Congress could pursue pretty much any regulatory project for which it had the political will, assuming no transgression of external constitutional limits. Some commentators saw the practical reality of plenary congressional power as cause for alarm,34 and others regarded it as far less troubling.35 But even thinkers basically content to let Congress exercise general legislative power have been mostly inclined to let that arrangement persist as a fact in tension with official norms, rather than offering direct justificatory arguments in its support based on traditional sources of constitutional authority.36 More or less across the ideological spectrum, theorists and practitioners offer at least pro forma affirmation of the internal-limits canon.37 As a result, the canon has persisted as a stock idea in constitutional law. To paraphrase Justice Robert Jackson, it lies about like a loaded weapon.38

So long as mainstream constitutional decision makers lacked the inclination to invalidate important federal legislation on internal-limits grounds, this consensus at the level of principle had little significance at the level of practice. But in the controversy over the Affordable Care Act (ACA),39 the internal-limits canon threatened to become the vehicle for an enormously consequential statutory invalidation—one that would in principle have been subject to a workaround but which in practice would have killed a major legislative program.40 Critics of the ACA demanded to know what Congress couldn’t do, short of bypassing external limits, if the individual mandate was valid law.41 Supporters of the ACA lacked the option of saying, “Well, maybe nothing, and that’s all right.” That response would have identified the ACA’s supporters as constitutional heretics. But it shouldn’t have: the internal-limits canon is not entitled to the persuasive force it now enjoys.

In the post-NFIB world, it is important to explain why not, lest the canon facilitate the invalidation of important legislation that should rightfully be deemed constitutional. To be sure, it hasn’t happened yet. NFIB upheld the Affordable Care Act, and the internal-limits canon might still continue its long career of solemn invocations followed by little or nothing in the way of consequences. But in the wake of NFIB, some mainstream constitutional thinkers have begun reviving old ideas about internal limits.42 The question of whether we stand on the brink of a constitutional gestalt shift has been squarely posed.43 In the next phase of discussions about federal power, the widespread sense that everyone accepts the internal-limits canon will distort the analysis and tilt the playing field. To prevent that distortion, it is important for theorists and practitioners to recognize that traditional sources of constitutional authority might, as a practical matter, authorize Congress to make any law not prohibited by some external limit—and that such a conclusion should not be worrisome.

In Part I of this Article, I lay out the logic of the internal-limits canon. In Part II, I explain why federalism does not require congressional power to be internally limited. In Part III, I explain why fidelity to the Founding does not require the internal-limits canon. In Part IV, I explain why the text of the Constitution does not require it either.

Finally, a comment on the Article’s organization. Constitutional analyses standardly discuss arguments from text, history, and structure in precisely that order: text, history, structure. The body of this Article deliberately reverses the sequence: structure, history, text. The point of this unconventional ordering is to enable readers to assess my analysis of the Constitution’s text with a clear understanding of the structural and historical analyses that make the textual reading sensible. After all, constitutional interpreters generally (and reasonably) read ambiguous texts so as to render them sensible in light of considerations about history and structure. An interpreter’s sense of the relevant history and structure will push him toward some possible readings of text rather than others. Indeed, a textual reading that seems natural or intuitive given one set of assumptions about history and structure might seem forced and implausible given another set. In the past, when interpreters have overwhelmingly read the Constitution’s text to support the internal-limits canon, they have not done so on the basis of the text simpliciter; they have done so while approaching the text through a set of assumptions about history and structure. In what follows, I lay out a structural account of the role of internal limits within American federalism and a historical account of internal limits in the Founding design; both accounts are intended to correct prevailing misconceptions. With better understandings of history and structure, a better reading of the text comes more clearly into view.

I. two ideas

A. The Enumeration Principle

The Constitution’s enumeration of congressional powers is generally associated with the project of ensuring a federal government that is vigorous but limited. The virtues of limited government as a general matter have been extensively catalogued, and there is little need to offer a comprehensive recapitulation here. Briefly, limited government at both the local and the national level is necessary for individual liberty and for the many forms of human flourishing that individual liberty enables.44 Limitations on central power in particular preserve space for meaningful autonomy at the state and local levels and therefore for a range of benefits that the literature on federalism has made familiar.45 The constitutional system accordingly needs to consider which decisions should be made centrally and which should be made locally, as well as what limits there are to what any government may do.

Enumerating congressional powers was one of several Founding-era strategies for pursuing these ends. The foremost strategy was that of process limits, which is to say that the whole structure of power and office-holding that the Constitution created is properly understood as a set of devices for constraining the federal government as well as empowering it. For example, frequent elections were expected to keep Congress from enacting oppressive legislation, and a Senate composed of ambassadors from the state legislatures was expected to ensure that the federal government respected the prerogatives of state governments. The Founders also made use of external limits: both in the original Constitution and in the Bill of Rights, they specified affirmative prohibitions that Congress could not transgress.46 And most relevant for purposes of the present discussion, the Founders deployed a strategy of internal limits by providing that Congress would have only those powers affirmatively given to it. If Congress has only particular powers, the intuition runs, then its jurisdiction is narrower than if it possessed a general police power.

The idea that Congress has only those powers that are affirmatively given to it is fundamental in constitutional law, and constitutional lawyers typically invoke that idea by saying that the federal government is a government of enumerated powers. But there is some uncertainty about whether the best interpretation of the enumeration principle understands the phrase “enumerated powers” literally or as a term of art. In a literal sense, the “enumerated powers” of Congress are those listed in Article I, Section 8, as well as those listed in many other parts of the Constitution.47 It is commonly said that Congress can act only on the basis of its enumerated powers.48 But constitutional practice has been more complex, because the Supreme Court has periodically recognized congressional powers arising implicitly from the constitutional structure rather than from any express grant of particular authority. Examples include implicit powers in foreign affairs,49 the now-defunct power to enforce the Fugitive Slave Clause,50 and, at the limit, the power to do all things necessary to protect the federal government from destruction.51 If these examples and others like them are to be taken seriously, then the frequent statement that Congress can act only on the basis of its enumerated powers should be understood, if not as an error, then either as an approximation or as a statement that uses the term “enumerated” in a non-literal way—perhaps as a synonym for “delegated,” rather than a synonym for “articulated expressly.”52 If so, it is more accurate to say that Congress can act only on the basis of its delegated powers, which is a larger set than its enumerated powers.53 Note that the Tenth Amendment speaks this language: it reserves the “powers not delegated to the United States,” not the “powers not enumerated.”54

Someone adhering to a stronger form of the enumeration principle—whether due to a theory of textualism, a commitment to limiting Congress, a sense of fealty to a traditional maxim of constitutional law, or any combination of these and other factors—could of course deny that Congress has ever legitimately exercised unenumerated powers. Most simply, all the cases recognizing such powers might be dismissed as wrongly decided. But it is not necessary to go that far. The less destabilizing alternative is to argue that the powers the Court has described as implicit rather than enumerated really are contained within the enumerated powers, correctly understood.55 For example, the power to safeguard presidential elections, treated as an implicit congressional power in Burroughs v. United States,56 could be reinterpreted as a power necessary and proper for carrying into execution all of the functions of the President as conferred in Article II and therefore as falling within the power enumerated in the Necessary and Proper Clause.57 Such an effort, if comprehensive and persuasive, might collapse the distinction between enumerated and delegated powers, at least in the present, and perhaps also in the future, if that mode of thinking persisted. It would of course remain the case that, as a historical matter, constitutional decision makers have not categorically limited the powers of Congress to those that they believed to be enumerated in the Constitution.58 But it is always open to supporters of a strict enumerated-powers doctrine to say that the Court has sometimes stumbled. Or, more charitably, that the Justices who purported to recognize unenumerated powers actually builded better than they knew,59 deciding cases correctly even while not quite articulating the reasons why. On that interpretation, the enumeration principle is literal and straightforward: Congress simply may not do anything that does not fall within the powers expressly given to it by the text of the Constitution.

B. The Internal-Limits Canon

For present purposes, it does not matter whether “enumeration” should be understood literally or as a term of art. Either way, the enumeration principle differentiates the basis for congressional authority from that of the general legislative power that states enjoy.60 Whether or not express enumeration is the only way in which the Constitution delegates power to Congress, Congress can only do those things that it is affirmatively authorized to do. And on that theory, Supreme Court Justices standardly reject constructions of congressional power that seem tantamount to affording Congress plenary legislative authority. To permit constructions that broad would eliminate the role of internal limits as a meaningful part of the system. Thus, Chief Justice Rehnquist wrote in United States v. Lopez that reading the commerce power to authorize Congress to regulate the mere possession of firearms would leave the Court “hard pressed to posit any activity by an individual that Congress is without power to regulate.”61 The same trope featured centrally in the argument over Section 5000A of the Affordable Care Act, commonly called the individual mandate. According to the joint dissenting opinion in NFIB, upholding the mandate would “extend federal power to virtually all human activity.”62 And that cannot be right, because “the proposition that the Federal Government cannot do everything is a fundamental precept.”63

These opinions articulate the internal-limits canon by insisting that Congress’s powers cannot be construed in a way that would let Congress regulate without limit. A more careful specification of the idea requires an important clarification, because neither sustaining the Gun-Free School Zones Act nor upholding the individual mandate would suggest that there are no limits on federal power. Even if Congress can require people to buy insurance, it cannot prohibit Buddhism64 or commandeer state legislatures65 or operate segregated schools in the District of Columbia.66 The Justices are presumably always aware of such external limits, even if they sometimes write as if an absence of meaningful internal limits is the same as the absence of any limits at all.67 To avoid exaggeration, therefore, the internal-limits canon should be rendered this way: Congress’s powers cannot be construed in a way that would permit Congress to regulate anything at all unless blocked by an external limit.68

Note too that proponents of the internal-limits canon insist not merely that individual powers of Congress have internal limits but that the powers of Congress taken as a whole have such limits. Obviously individual powers have internal limits. The power to govern the District of Columbia cannot be used to govern Delaware, and the power to punish pirates cannot be used to fix the day on which the Electoral College votes. The issue is whether all the powers collectively face internal limits—that is, whether some potential legislation must lie beyond any of Congress’s powers, even without considering external limits. After all, if most of Congress’s powers had internal limits but at least one (say, the commerce power) had none, then internal limits would not in practice constrain Congress’s ability to legislate: Congress could simply use the commerce power to do anything that it would be prevented from doing by the internal limits of other powers. Similarly, even if every individual power of Congress (including the commerce power) had internal limits, internal limits would not constrain the total reach of Congress if everything beyond the limits of any given power were within the scope of some other power. Consider, as a simplifying analogy, a legislature with seven enumerated powers, each of which authorized legislation on a different day of the week. Each power would have a clear internal limit. The Monday power would not authorize legislation on Tuesday, and so on. But these internal limits would do nothing to constrain the legislature’s reach, because no matter what the legislature wished to do (and no matter when), it would have a power adequate for the task. Accordingly, the internal-limits canon is a proposition about the scope of all of Congress’s powers taken as a whole, not a proposition about the limits on those powers taken separately. It directs that some imaginable laws must lie beyond Congress’s power, even before considering the constraints imposed by external limits.

C. The Internal-Limits Canon as Non Sequitur

The enumeration principle is related to the internal-limits canon, and it is easy to conflate the two ideas. Carefully considered, however, the two are not the same. And it is a mistake to think that the first requires the second.

The enumeration principle provides a criterion for determining what powers Congress is entitled to exercise. Not to belabor the point, and subject to the wrinkle about whether “enumerate” means “delegate” or “articulate expressly,” the principle provides that Congress may exercise only those powers that the Constitution enumerates. I take that principle to be a valid rule of constitutional interpretation. But without more, the enumeration principle does not address the scope of any particular congressional power, nor does it address the scope of all those powers combined. As Chief Justice Marshall put the point, “This government is acknowledged by all, to be one of enumerated powers. . . . But the question respecting the extent of the powers actually granted, is perpetually arising . . . .”69

The internal-limits canon is different. It states a rule about the extent of congressional powers. That rule, as already explained, is that the powers of Congress may not be construed in a way that would permit Congress to pass any and all laws that it could pass if it had a general police power.

The standard wisdom in constitutional law holds that the enumeration principle implies the internal-limits canon. The powers of Congress are enumerated, the reasoning goes, and although that principle does not precisely define the scope of each power, it does mean that there is an overall limit on their scope, because there would be no point in enumerating particular powers if in practice the power of Congress is general.70 That inference might seem plausible, but it is in fact a non sequitur, and it is at that inference that my argument is aimed.

Congress has only those powers affirmatively given to it by the Constitution. But those particular powers might in practice enable Congress to do all the things that Congress could do if the Constitution gave Congress a police power, much as my authorizing my son to eat chocolate, vanilla, or strawberry ice cream would, under certain circumstances, give him a mandate as broad as the one he would have if I just authorized him to eat ice cream.71 So yes, the federal government is a government of enumerated powers. It just may not follow that the set of things Congress can do is thereby narrowed. The burden of the rest of this Article is to explain why this view is consistent with constitutional structure, history, and text.

II. structure

The dominant structural rationale for the internal-limits canon sounds in federalism.72 According to the standard account, internal limits on congressional power are essential in order to preserve the role of state and local government. There is little question about the worthiness of this goal, and the literature on federalism canvasses many reasons why. Local decision making is often better informed about local problems than federal decision making.73 Depending on one’s conception of democracy, local decision making may also be more democratic.74 Decision making in smaller polities means more opportunity for individual citizens to experience civic engagement, leadership, and political responsibility.75 Differentiated decision making in different states (and different localities) creates regulatory diversity, and regulatory diversity can satisfy the preferences of more citizens than uniform regulation can, assuming that the subject matter is not one that requires wide coordination and assuming also at least some correlation between the aggregate preferences of different state populations and the regulatory schemes those states adopt.76 Regulatory diversity may also provide the laboratories-of-democracy benefit, as the ability to see different legal rules operate increases the store of knowledge available to future policymakers.77 For all these reasons, federalism—or more particularly, the local decision making that is one side of federalism—is valuable.78

A. The Limits of Internal Limits

But are internal limits on the powers of Congress a necessary mechanism—or even a particularly helpful mechanism—for securing these benefits of federalism in the context of the American constitutional regime? The conventional answer is yes, and on an apparently straightforward rationale: internal limits mean less federal law,79 and less federal law means more space for state autonomy. It is a powerful idea, at least on the surface, in part because it seems like simple common sense. But the dynamics of modern federalism are more complicated than the conventional answer assumes. For one thing, the allocation of decision making among federal and state authorities is not always a zero-sum game. Sometimes federal regulation displaces local regulation, but sometimes federal law empowers state policymakers more than it constrains them.80 At least as importantly, internal limits might have little practical capacity to reduce the amount of federal law. This is so both because Congress can usually work around internal limits and because the conditions of modern regulation often mean that the elimination of one federal law leaves a subject matter regulated by some other federal law, rather than making the subject matter free from federal regulation entirely.

Some of these dynamics can be illustrated by reference to the most prominent internal limits articulated in recent case law: the economic-noneconomic limit on the commerce power that the Court imposed in Lopez and the action-inaction limit that five Justices endorsed in NFIB. Each of these limits makes it possible to describe laws that Congress’s commerce power would not authorize, and the Justices keenest on enforcing internal limits in Lopez and NFIB described these limits as required by the internal-limits canon. But neither limit cultivates the substantive virtues of vesting decision-making power in states rather than in the federal government all that well, because neither limit creates a significant policymaking space in which states can operate free from federal interference. Indeed, these limits did not even prevent the implementation of the substance of the federal regulations to which they were applied. After Lopez held that the Gun-Free School Zones Act could not be sustained as a law regulating an economic activity with substantial effects on interstate commerce,81 Congress reenacted the substance of the Act as a regulation of instrumentalities—here, firearms—that move in interstate commerce,82 and the courts have uniformly upheld the reenacted version.83 In NFIB, a majority of the Justices took the position that the individual mandate is not a valid exercise of the commerce power,84 but a different majority upheld the mandate as a valid exercise of the taxing power.85 Indeed, even the Justices who concluded that the mandate was not a valid exercise of the taxing power agreed that Congress could enact the substance of the mandate using its taxing power; their contention was only that Congress had not actually done it that way.86 If the internal limits announced in Lopez and NFIB cannot even prevent the implementation of the substantive regulations at issue in those two cases, it seems unlikely that they can clear away or forestall any great amount of federal law.87

The unreliability of internal limits as tools for increasing the available space for autonomous state decision making results partly from the mismatch between a formal set of limitations on Congress and the internal-limits canon’s goal of creating substantive areas of policymaking into which Congress cannot intrude. As is typical of formal rules, these internal limits can be evaded by recourse to other formal rules, and, in this area, the evasions enable Congress to reach its intended regulatory targets by formally different routes. The problem runs deep: as Justice Ginsburg suggested in NFIB, even if the action-inaction distinction could prevent Congress from passing a law compelling Americans to eat broccoli, Congress could create the same substantive regime by prohibiting the purchase or consumption of any food other than broccoli, except by persons who had already eaten their broccoli.88

This is not to say that internal limits are completely incapable of preventing federal regulation. Sometimes workarounds are costly. A given Congress might be willing to enact the individual mandate but not to create a single-payer system. In such a case, an internal limit blocking the individual mandate might result in a meaningful regulatory difference. Congress would still be authorized to enact sweeping healthcare reform, but it might choose not to, assuming that it could not identify some other workaround less costly than the single-payer approach. Moreover, if an enacting Congress does not foresee that its statute might be invalidated on the basis of an internal limit, the choice of whether to enact a workaround will probably lie with a different Congress that might not be as invested in the substance of the legislation. Congress did not reinstate the civil remedy of the Violence Against Women Act (VAWA) after the Court struck it down in United States v. Morrison,89 even though the move that worked after Lopez was available there as well, at least in substantial part. Imagine a federal statute creating a civil remedy for crimes of gender-motivated violence committed with an instrumentality of interstate commerce. The fact that Congress did not pass such a statute may reflect mostly the differing policy preferences of Congress and the President in 1994, when VAWA was passed, and in 2001, after Morrison was decided.

If the aftermath of every invalidation of a federal statute on internal-limits grounds followed Lopez’s script rather than Morrison’s, there would be less need to recognize the flaws in the internal-limits canon. Congressional workarounds reinstating the substance of the initial law would consume time and effort, but the end-state regulatory environment would be unaffected. But reinstatements by workaround are not automatic, as the Morrison example illustrates. Yes, Congress can legislate more or less as it thinks best, subject to external limits, but only if its policy preferences remain constant from the time of a statute’s initial passage to the time of the workaround, and only if Congress is willing to legislate twice.

Viewed sympathetically, this dynamic might transform internal limits into a species of process limit, one that permits Congress to pursue its chosen program so long as it makes the heightened effort required to do so. If properly tailored to protect the interests of federalism, internal limits might have some value as process limits, albeit by becoming something rather different from what constitutional lawyers have traditionally taken internal limits to be. But given the frequency with which Congress could design workarounds in cases where it knew ex ante how internal-limit rules would apply, the heightened process requirements might come mostly as a matter of ex post surprise. Had the 1994 Congress known that Lopez was coming, it might have enacted VAWA with some appropriate workaround. Similarly, although Congress in 2009 and 2010 preferred mandates to a single-payer health insurance system, Congress would likely have been pushed toward some solution that did not risk invalidation on action-inaction grounds—whether single-payer or something else—if the action-inaction distinction had been a clear part of commerce doctrine prior to the passage of the ACA. As a result, any limiting effects of internal limits might have a somewhat arbitrary shape, tracking accidents of timing more than any consistent logic about what policies should be made locally rather than centrally. Perhaps such limits could create a hodgepodge of disallowed federal statutes, or applications of federal statutes,90 and perhaps the regulatory space opened thereby would add a bit to the sum total of regulatory space available for state policymaking. But it seems doubtful that that additional space would be large or well targeted enough to add meaningfully to the robust, extensive practice of state and local decision making that exists today under a regime in which internal limits play almost no role.

These explanations of why internal limits might not fulfill the promise of enhancing federalism are partly a function of currently prevailing conditions and attitudes about federal legislation. In other words, the fact that internal limits are today poor tools for ensuring substantive areas of state policymaking might not mean that such rules could never be useful for that purpose. Once upon a time, a set of internal limits seemed to protect substantial state policymaking space. Consider the famous distinction between direct and indirect effects in commerce doctrine.91 As every law student learns, that distinction eventually came to seem absurd, and the boundary it policed disintegrated.92 But at an earlier time, that distinction and others like it seemed to work pretty well as frameworks for circumscribing the reach of federal power. Why could such formalisms succeed at some times and not at others? Three reasons—or perhaps three faces of the same reason—supply much of the answer.

First, as Lawrence Lessig has observed, these doctrines worked when they did, not simply because earlier judges were more sympathetic to metaphysical distinctions like the one between direct and indirect effects (though that may to some extent have been the case) but also because those judges as a group had a widely shared sense of how to distinguish between federal and state spheres of regulation.93 When judges implemented doctrines like the direct-indirect distinction in light of that shared sense, they reached a more or less stable set of results. Second, the sense of federalism that Lessig identifies as widely shared among judges of a certain era was also shared by many legislators, such that Congress was simply less inclined to regulate pervasively than it was later on. Congressional restraint meant that autonomous spheres of state regulation were a normal part of the landscape and that any judicially imposed limits on Congress reinforced that reality. (In a sparsely regulated world, the invalidation of a single law might leave the relevant subject matter unregulated. In today’s densely regulated world, the invalidation of a single law is more likely to mean that some other law becomes the operative regulation.) Third, Congress’s disinclination to regulate pervasively meant that the judiciary could police the federal-state balance without constant conflict with the elected branches. Persistent conflict sometimes sharpens the questions at issue, and in the absence of constant conflict judges were perhaps more able to get by with a shared-sense jurisprudence than they would have been in the face of pressure to articulate clear rules about what was and was not permissible.

These conditions did not last forever. Eventually, and partly due to the logic of living in an increasingly interconnected world, more and more members of the decision-making class came to question prevailing intuitions about the exceptional nature of federal governance. The Sixteenth Amendment signaled an appetite for a larger federal regulatory agenda and supplied the means for carrying one out.94 Congress became more willing to legislate and to legislate pervasively. In the judiciary, the erosion of an intuitive consensus about the limits of federal power meant that doctrinal distinctions like the one between direct and indirect effects could no longer maintain the old limits. Absent shared substantive understandings, the doctrinal formulas came to seem amorphous, arbitrary, or both. And the separate-spheres vision famously known as “dual federalism”95 gave way to other conceptions (marble-cake federalism,96 picket-fence federalism97) built on the recognition that modern federalism involves two sets of officials working on common subject matter, rather than two sets of officials operating in different policy domains.

Even in the era of dual federalism, though, internal limits never did the work of delineating separate spheres all by themselves. Instead, internal limits were supplemented as needed by a judicial willingness to declare forthrightly that certain spheres of substantive regulation were reserved to the states. For example, when the Supreme Court in United States v. Butler struck down a federal subsidy provided under the Agricultural Adjustment Act, the majority opinion discussed the question presented as a matter of the extent of Congress’s spending power, but then at a critical moment declared that the dispositive consideration was not to be found in the best construction of the Spending Clause. Instead, the subsidy program was unconstitutional because it invaded a reserved right of the states, namely the right to control agricultural production.98 Similarly, the Court in Hammer v. Dagenhart struck down the Child Labor Act’s rules prohibiting certain interstate shipments of goods produced with child labor not because prohibiting interstate shipments was not within the commerce power—the Court conceded that the transportation of property was as much commerce as purchase and sale—but because this restriction on interstate shipping would impinge on the inherently local matter of manufacturing.99 In these cases, the Court discussed the scope of Congress’s spending and commerce powers, but the considerations that invalidated the laws at issue were not conceptual accounts of “spending” and “commerce.” They were accounts of what regulatory subject matters must be reserved to the states. If the question is what sort of judicial doctrine would help protect local decision making, that reserved-sphere approach makes a good deal of sense: a robust doctrine of reserved subject matters can limit congressional authority in ways that preserve meaningful and independent policymaking spaces for state governments. But that solution cannot vindicate the idea that there must be internal limits on Congress’s enumerated powers, because it is a solution based on external limits rather than internal ones. Just like the Free Exercise Clause, a rule that Congress may not regulate agriculture or manufacturing cross-cuts the enumerated powers, blocking congressional action on the basis of a constitutional concept arising somewhere outside the delegation of powers to Congress.

To be sure, one could try to present reserved-sphere rules as internal limits by reading them into the interpretation of particular congressional powers. One could say, for example, that the commerce power does not authorize Congress to regulate agriculture because “commerce” and “agriculture” are two different things. But unless one also ensured that Congress’s other powers could not be used to regulate “agriculture,” this limitation on the commerce power would not prevent Congress from regulating in that area with some other tool. And if one purported to read the same internal limit into every one of the enumerated powers—if one maintained, in other words, that the proper interpretation of every power granted to Congress anywhere in the Constitution happened to exclude any power to make regulations affecting agriculture—then observers might wonder whether the real operating force was an external limit prohibiting Congress from regulating agriculture, rather than a remarkable confluence of many different internal limits.

In principle, either answer is possible. Maybe the work is really being done by an external limit, or maybe there are simply spheres of social life that lie beyond the internal limits of all of Congress’s powers. The latter possibility is, of course, the one that the internal-limits canon insists upon. But in deciding which possibility is more likely, it is worth noticing that at least some of the Court’s key decisions purporting to enforce internal limits referred to external-limit considerations at analytically critical junctures and that the reasoning of those cases is more easily comprehensible if we take external limits to be doing work. If the Court cannot explain why a spending program is unconstitutional without invoking the states’ reserved control over agriculture, and if the Court cannot explain why a law formally regulating the interstate shipment of commercial goods is unconstitutional without invoking the inherent right of states to control manufacturing, then internal limits are not the sole motive force of those decisions. Even before the New Deal, external limits did much of the work.100

B. Federalism Without Internal Limits

Nothing in this analysis suggests the demise of federalism. State governments today exercise considerable sway over a broad swath of important policymaking domains. Differentiated state decision making is more than robust enough to deliver substantial regulatory diversity, not to mention immensely greater opportunities for civic engagement and political leadership than would be possible if American political decision making were fully centralized. So the inability of internal limits to protect enough autonomous state policymaking to make federalism worthwhile has not nullified those virtues of federalism. State decision making is simply perpetuated by other mechanisms.

A few of these mechanisms are external limits on Congress’s powers. Perhaps most fundamentally, Congress may not unilaterally terminate101 or reconfigure102 existing states. Congress may not dictate the location of a state capital103 or tax a state’s own tax revenue.104 State sovereign immunity doctrines are external limits, albeit defeasible through the exercise of congressional powers rooted in constitutional provisions postdating the Eleventh Amendment.105 The anticommandeering rules are external limits that help states engage in policymaking by limiting the federal government’s ability to force the hands of state decision makers.106

Obviously, these limits are useful only if there are substantive areas in which states can make policy, and other mechanisms help to ensure that such areas exist. Some of these mechanisms are process limits, albeit not quite the set of process limits that the Founders imagined. Many of those limits never worked the way the Founders imagined they would,107 and others have been repudiated by constitutional amendment108 or undermined by changing practices and attitudes over time.109 Nonetheless, process limits remain important aspects of American governance, in ways that include both formal and informal interactions between federal decision makers and their state counterparts. For a variety of reasons that prior scholarship has canvassed, state and local officials have substantial influence in the shaping of federal law and federal regulations, and they regularly deploy that influence to prevent federal authority from unduly contravening local interests or sidelining state decision makers.110 One should not romanticize or overstate these mechanisms.111 The idea that political-process mechanisms will always do right by the interests of federalism would be a just-so story, as would the idea that the political process always does right by any other kind of interest. But one need not subscribe to the idea that process federalism cures all ills to recognize that it has some important effects.

A second set of mechanisms—perhaps overlapping with the first112—falls within the rubric known as “cooperative federalism.”113 It is a normal feature of modern American governance that Congress works with states, rather than around them, when engaging in important regulatory projects. Federal statutory schemes addressing social security,114 the environment,115 health care,116 education,117 transportation,118 crime control,119 and many other topics rely on the states as powerful players in deciding what will actually happen and who will get what.120 The Affordable Care Act provides a prominent contemporary example. Congress made the American Health Benefit Exchanges state-based institutions, which means that state officials can decide how to implement the certification criteria that will determine what plans are finally offered to the public in their states.121 Some members of Congress believed that the policy of the ACA would be best carried out by a national health insurance exchange, but others took it as a matter of critical importance that the exchange be put in state hands, and the latter view prevailed.122 The ACA also dramatically expanded Medicaid, and state bureaucracies are the ones with the expertise, local knowledge, and boots on the ground that are needed to make the envisioned expansion real. To be sure, many states have to date declined to participate in these ventures,123 and that development has thrown a fair amount of decision making back to the federal government. For present purposes, though, what matters is precisely how unusual (and indeed unanticipated) it is for states to opt out on this scale. The ACA is a limiting case, involving the greatest policy conflict in decades. Normal federal governance involves a high degree of state involvement as a matter of routine, as noted above with respect to policy areas from crime control to highway management to the environment and social security.

The normal pattern is neither an accident nor a matter of congressional grace. It is a structural aspect of modern American government, one that has emerged in light of considerations about the capabilities of states and the limitations—practical, rather than legal—of Congress.124 More than twenty percent of all “federal” nondefense spending for 2011 was spent by state and local governments administering cooperative federalism programs.125 Implementation of the ACA is driving the figure higher. Even if Congress wanted to, the federal government could not tomorrow (or next year, or in five years) displace the states from their roles in governance under this system, in part because it could not simply summon into existence the personnel and institutional capacity that would be necessary for doing so. To return again to the ACA example, the threat that state opt-out has posed to the successful implementation of the Act is predicated on the reality that Congress needs the states to make things happen.

Given that reality, state governments have considerable latitude to make decisions about the public policies that federal statutory schemes represent. Cooperative federalism is cooperative (or uncooperative)126 rather than dictatorial: the states are not neutral conveyor belts for the implementation of federal programs. State and local officials negotiate, bargain, modify, and sometimes undermine federal policy, and their opportunity to do all of those things yields the range of benefits that makes federalism valuable in the first place: local knowledge brought to bear on local questions, local responsibility for those decisions, regulatory diversity, broadened opportunities for civic engagement and political leadership, and so on.127 State officials’ choices within these schemes are constrained by boundaries set at the national level, but that is true of any system of state autonomy that respects the principle of federal supremacy—or even just constitutional supremacy.

The pertinent question is whether the policymaking discretion that state and local officials exercise is consequential enough to make those officials meaningful decision makers rather than ministerial instruction-takers, and of that there can be little doubt. Recall that many state officials wanted responsibility for the healthcare exchanges, because they understood the enormous power that responsibility confers. As many have noticed, states that want to impede the federal policy embodied in the Affordable Care Act have had ample opportunity to do so. Note, too, by way of general analogy, that local decision makers are often consequential even though most localities formally have no existence or power whatsoever except that which the state chooses to grant.128 If even the fully hierarchical relationship of states and cities does not preclude local governments from delivering many of the benefits of federalism—local knowledge, local accountability, preference matching, opportunities for civic engagement and political leadership—then it should be clear that state (and local) governments can deliver many of these benefits even when working in environments structured by federal legislation. States are, after all, much more independent of the federal government, even in an era of active federal regulation, than localities are of states.

Despite the thick reality and enormous importance of cooperative federalism, constitutional lawyers sometimes suspect that whatever “federalism” is involved in these schemes is not real federalism in the constitutional sense.129 Real federalism, the intuition runs, does not reside in complex regulatory systems that are, at bottom, creations of Congress. On this view, the federalism that the Constitution ordains requires a more thorough separation between that which is national and that which is local130—a separation more consistent with the values of early Americans who distrusted central authority.131 And that separation, the Supreme Court has explained, stands or falls with the internal-limits canon.132 As Abbe Gluck has put the point, though, the idea that that kind of separation can deliver the benefits of federalism “depends on what no longer exists,” namely “significant areas of regulation that are reserved to the states and into which federal lawmaking may not tread.”133 In a world where dual federalism is gone and federal regulation is pervasive, disallowing this or that law as beyond Congress’s powers will never recreate the kind of separation that once made it sensible to think of state policymaking as thoroughly independent of Congress, rather than as constantly interacting with federal governance.

The older vision of separate-sphere federalism has a strong hold on American imaginations. Its simplicity and historical resonance make it appealing, especially when contrasted with the bureaucratic technicality that cooperative federalism embodies.134 Fundamental constitutional intuitions are usually conceptions that well-socialized Americans appreciate without specialized professional training. We learn them as part of our civic education long before we arrive at law school. So it is a great disadvantage of cooperative federalism that even the most creative Schoolhouse Rock writer would find it a challenging concept to convey. Depicting the mechanisms of cooperative federalism can require multi-page interlocking organizational charts, and the romantic national-identity aspects of federalism tend not to resonate in discussions about, say, the criteria for federal funding of state-initiated roadside lighting projects.135 But if the question is how American governance actually works—a question that should concern practical people, and a question that courts should keep in mind when exercising their considerable disruptive power—then cooperative federalism is a central part of the answer. It is what the bulk of federalism looks like in the modern Republic, and it provides many of the benefits that make federalism valuable.

No one should think that cooperative federalism provides those benefits in exactly the optimal ways or in exactly the optimal quantities. Neither cooperative federalism nor process federalism nor any other kind of federalism—including a federalism based on internal limits—could be expected to produce that outcome. But process federalism and cooperative federalism yield the benefits that make federalism valuable more robustly than any system of internal limits has in a long time, if indeed a system of internal limits ever yielded those benefits without the support of an underlying theory of separate spheres that courts were willing to enforce through external limits on federal power.

C. Attitudinal Formation and the Internal-Limits Canon

There remains at least one more concern about the damage that discarding the internal-limits canon might do to federalism. Part of what maintains the federalist dynamics of interaction between state and federal officials is a set of ideas about what each government is supposed to do. Process federalism and cooperative federalism are maintained partly by practical conditions and partly by a set of attitudes, and the two are mutually reinforcing. In other words, these mechanisms persist in part because the players who deliberate and negotiate and decide are inclined to consider local decision making valuable. To be sure, people’s ideas about which government should do what are far from uniform, and there is constant renegotiation of certain boundaries, as there is in most complex relationships. The system could not possibly rely on everyone’s sharing precisely, or even close to precisely, the same set of views about the proper roles of state and federal government. Still, a certain amount of shared sense among officials and the public to which they respond is an important element of the system, and a wholesale shift in attitudes might cause existing dynamics to unravel. Not quickly: the limits on federal power are rooted in practical considerations as well as attitudinal ones, so it would take a fair amount of institutional reconfiguration to overcome those limits even if federal officials wanted to do so. Indeed, the difficulty of the project is probably one important force in deterring any such ambition. But in principle, and over time, a broad and deep shift in attitudes could facilitate large changes in constitutional dynamics.

Might official stories like the one embodied in the internal-limits canon play a role in preventing such an evolution? We form our constitutional expectations in part by hearing how the system is described.136 An articulated principle under which the powers of Congress are inherently limiting might teach Americans—including both officeholders and at least some of the civically literate citizens to whom they respond—to think twice about whether a given project is appropriate for the exercise of central power. If the internal-limits canon were repudiated, one of the influences supporting that consciousness would disappear. And that, one might worry, could lead to the disintegration of what is now a relatively healthy federal system.

One can only speculate as to the gravity of this concern. I expect that different people will have different intuitions, just as people through the centuries have had different intuitions about whether preserving social order and interpersonal decency requires public affirmation of orthodox religious beliefs, whether metaphysically sound or otherwise. The internal-limits canon is salient within official ideas about federalism, but that does not tell us how much weight it bears in the attitudinal architecture. Maybe it is important. But maybe it is marginal, or even superfluous. Many other influences also contribute to the attitudes necessary for maintaining a limited federal government, some conceptual and some practical. To take just one set of examples, consider the many canons of federal statutory interpretation that embody and reinforce the idea that federal governance should not impinge too much on state governance. Absent clear statements to the contrary, federal statutes are not to be read to intrude on traditional state criminal jurisdiction, or to abrogate state sovereign immunity, or to preempt state law, or, speaking more generally, to alter an existing balance between federal and state power.137 Would the loss of the internal-limits canon provoke the loss of these other canons as well, or would these other canons and a host of other forces go on teaching the basic lesson even without being supplemented by the particular practice of speaking (emptily) about internal limits? And if we were to try to figure out how removing this particular piece of the discourse might shift attitudes, we should not exclude the possibility that the internal-limits canon might undermine as well as enhance the idea that the federal government has a limited regulatory role. Everyone who knows the internal-limits canon also knows that in practice the federal government has long seemed to enjoy the equivalent of general legislative power, or very nearly so. Might the takeaway message for at least some audiences be that core tenets of federalism are quaint fables that practical people should not take seriously? Or more broadly, that constitutional law is an enterprise in which we say one thing and do another? If so, might these attitudes be at least as damaging as letting go of the internal-limits canon would be?

There is no way to measure the net discursive effect of the internal-limits canon on the self-limiting tendencies of federal officials. But in the absence of knowledge, there are reasons for skepticism about how much attitudinal damage would result if the canon were abandoned. Maybe the formula’s very familiarity leads us to overestimate not just its cogency but also its importance as an attitudinal prompt. Given that uncertainty, it seems prudent to avoid making this concern into too strong a reason for holding on to the canon, especially at a time when taking the canon seriously might cause important distortions in constitutional decision making.


The practical value of the internal-limits canon is supposed to be its capacity to limit central governance in a way that fosters the benefits of federalism. But we should not be confident that it is actually a useful means to that end. Internal limits might not be able to do very much either to limit the federal government or to enhance meaningful state decision making. It is therefore fortunate that the constitutional system contains other kinds of limits. Indeed, the system contains many of those limits precisely because the Founders collectively foresaw that internal limits might not get the job done. They insisted on alternatives.

III. history

Constitutional history is much broader than the history of the Founding period, and normative constitutional practice often appears different when viewed through the lens of one period in history rather than another. One way to understand the argument against the internal-limits canon is as an argument that takes the most recent century of constitutional law as a source of authoritative practices.138 That said, it is a fact about American constitutional thought that the Founding period occupies a privileged position among historical eras. And it is in the Founding that the historical argument in favor of the internal-limits canon is firmly grounded. According to that argument, internal limits were a critical part of the Founding design, and fidelity to that design requires the continued operation of internal limits as meaningful constraints on Congress.

There is no serious doubt that most of the Founders expected internal limits on congressional powers to constrain the federal government.139 But conventional wisdom about internal limits in the Founding design makes at least two important errors. The first, addressed in Part III.A, concerns the importance of internal limits relative to other kinds of limits in constraining the federal government. Following an argument famously advanced by Hamilton and others at the time of ratification, many leading figures have noted that the original Constitution contained an enumeration of congressional powers but no Bill of Rights and reasoned that the Founding generation saw the enumeration as the most important mechanism for limiting federal power.140 This line of thinking is a mistake. The Hamiltonian argument was not the well-considered theory of the Founding generation. It was a talking point that most of that generation dismissed as implausible.

Second, the conventional approach implicitly treats the enumeration of congressional powers as a matter of independent principle, rather than as a strategic choice intended to preserve local decision making and individual rights. The better reading is that the Founders saw enumeration as a means to those ends, not as somehow valuable in itself. And as I explain below, fidelity to choices about means sometimes differs from fidelity to choices about ends. If the idea of a government limited by its enumerated powers had been a matter of independent value to the Founders—that is, if the Founders would have insisted on internal limits as a mechanism for constraining federal power even with the knowledge that other kinds of limits would be equally effective, or more so, at preserving local decision making and protecting individual rights—then the internal-limits canon might be part of what fidelity to their design required. But on the understanding that the choice to enumerate Congress’s powers is better seen as a means to those ends, modern decision makers can be faithful to the Founding design even if Congress is not in practice meaningfully constrained by internal limits—provided, of course, that Congress exercises only the powers delegated to it, and provided also that local decision making and individual rights are protected by other means within the constitutional design. This second point is the focus of Part III.B.

A. How Important Were Internal Limits to the Founders?

Within American constitutional culture, a canonical story teaches that the Founders considered internal limits more important than external ones. According to that story, the delegates at the Constitutional Convention believed that the enumeration of congressional powers would limit the federal government. Indeed, the story continues, the Founders were so confident in the mechanism of enumeration that they considered a Bill of Rights unnecessary, or even counterproductive, because specifying affirmative prohibitions might mislead people into thinking that Congress was not confined to its enumerated powers.141 More than one important figure at the Founding articulated this idea about the enumeration and a Bill of Rights: Madison and James Wilson, for example, both prominently advanced the claim.142 But for modern audiences the idea is most closely associated with Hamilton, whose exposition of the argument in Federalist 84 is perhaps its most canonical expression.143

Given the enormously important status that modern Americans afford to the Bill of Rights, the normal role of this story is to make the enumeration and its internal limits seem essential. After all, the reasoning runs, the drafters of the Constitution considered the system of internal limits even more important than express guarantees of free speech, free religious exercise, and so forth. In NFIB, both Chief Justice Roberts and the joint dissent invoked this canonical story as a way of making the point, on the authority of the Founders, that internal limits must play a central role in the constitutional design.144 But the canonical status of this story notwithstanding, it is a mistake to think that fidelity to the Founding design requires operative internal limits. To begin to see why, it may help to think critically about the canonical story. Two points are particularly worth noting.

First, the Convention’s omission of a Bill of Rights does not demonstrate that the delegates regarded enumeration as the chief mechanism for constraining Congress.145 To most of the delegates, the most important mechanisms for constraining Congress were neither external limits nor internal limits but process limits. The arrangements to which the Convention paid attention at length and in detail concerned the composition of and relationships among decision-making institutions: popularly elected House and state-appointed Senate, single-member executive chosen with a state-based mechanism, and so forth.146 The prevailing wisdom held that elections and state governments would keep federal power in check, and if that didn’t work, then the states would use their resources—political, financial, persuasive, even perhaps military—to rally public resistance against central authority.147 Some of these expectations turned out to be chimerical.148 Others were vindicated, albeit to varying degrees and unevenly over time.149 But regardless of how one judges the successes and failures of the attempt to check federal power by process mechanisms, it remains the case that the Convention invested most heavily in this strategy. By comparison, the attention paid to enumeration and internal limits was slight. Prior to the appointment of the Committee of Detail, there was no deep engagement with questions about whether this or that power should be included among the powers of Congress, and the draft enumeration that the Committee presented on August 6 was largely accepted by the full Convention, albeit with emendations.150 To be sure, none of this demonstrates that the Convention regarded the enumeration as unimportant. Given the incomplete nature of the historical evidence, the delegates may have paid more attention to the enumeration than the surviving records reflect. But the available record strongly suggests that the delegates’ greatest focus was on other mechanisms—principally process mechanisms—for limiting federal legislation.

Second, no matter what the Convention delegates may have thought, the broader public decisively rejected the idea that the enumeration would limit Congress well enough to make a Bill of Rights unnecessary. Yes, people like Hamilton, Madison, and Wilson defended their work with that argument.151 But they utterly failed to persuade the public. Some contemporaries dismissed the claim that enumeration would suffice as just a rationalizing afterthought—an idea grasped at to parry Bill of Rights objections to the Constitution, rather than an authentic and central piece of the Convention’s plan.152 Jefferson told Madison directly that he considered the idea a ruse, one that might bamboozle a credulous audience but which on its merits should not be taken seriously.153 After all, it was not obscure even in 1788 that the powers to tax, to regulate commerce, to raise armies, and so forth could be deployed oppressively unless affirmatively limited.154 So the cry for a Bill of Rights continued unabated.155 The inadequacy of the draft Constitution’s limits on federal power was a common and insistently pressed criticism during the ratification debates,156 and the creation of vigorous external limits was one of the first orders of business once the new system was up and running. Acting pragmatically, many delegates to the state ratifying conventions chose to ratify the Constitution and fix the problem immediately afterwards rather than insisting that the whole arduous process be repeated from its beginning. But any idea that the Founding generation trusted Hamilton’s famous argument about internal limits is belied by the first great fact about constitutional reform in the new Republic: the quick passage and ratification of the first ten Amendments. In short, the most important feature of the Founding generation’s relationship to the idea that the enumeration would be sufficient for limiting Congress is this one: they didn’t buy it.

One could take the view that the Constitution’s defenders during the ratification debates were correct when they said that internal limits would do the work, regardless of how they came to that view and even though they were unable to persuade their contemporaries. The idea might then be worth taking seriously despite its rejection in its own time. But if so, the reason for thinking this conception worthy of our respect is not the legal authority of the Founding.157 It is the first-order merits of the idea. And to conclude that the idea was a good one, we would have to believe that the system of internal limits actually would have sufficed for protecting individual rights and maintaining the substantive virtues of federalism more generally. Experience does not offer much support for that view. As already described, internal limits do very little to constrain the scope of congressional regulation. One might hypothesize that the ineffectiveness of internal limits is due to the existence of too many external ones: if the public had acquiesced in the Convention’s design, the idea would run, then a combination of process limits and internal limits would have been forced to do the work, and they would have been adequate for the task. Or one could argue that the inconsequentiality of internal limits is the result of regrettable decision making by officials—judicial and otherwise—who have failed to apply Article I properly, rather than an inherent feature of the constitutional design. But each of these possibilities is speculation at best. On the first score, we cannot know how American constitutionalism would have developed in the absence of as significant an occurrence as the adoption of the Bill of Rights. On the second, much of the wisdom in sound constitutional design is the correct anticipation of how officials will behave. At some point, the fact that officials do not implement Article I in genuinely constraining ways indicates that Article I is not a usefully constraining mechanism within this system of government.

Given how little Article I’s internal limits restrain Congress, the Founding generation seems wise to have rejected the claim that enumerating Congress’s powers would be sufficient to limit the federal government. Yes, things could be different if things had been different. But it is also possible that resting on a system of internal limits would always have been risky, and that at some point those limits would have fallen short, just as the Founders feared. So there is a substantial irony in play when constitutional lawyers invoke the idea that the Founders gave pride of place to internal limits. Such invocations are probably meant to celebrate the wisdom of the Founders—a wisdom, the idea either implicitly or explicitly runs, that may be lost on those who look to external limits to do the important work.158 The greater wisdom of the Founding generation, though, lay in its refusal to rest on a system of internal limits. It does not show respect for the Founders to associate them with a flawed idea that they were prescient enough to discard as overly optimistic. We would show them more respect by associating them with the better ideas that they in fact endorsed.

B. Enumeration as a Means

The Founders rejected the idea that the enumeration would do all the work, but they did approve a system in which internal limits played a role. The crucial concern about enumerated powers and fidelity to the Founding, therefore, is not whether the enumeration should be sufficient for limiting Congress even in the absence of external limits and process limits. It is whether internal limits must operate as an essential feature of the system—whether it is necessary, in order to honor the choices of the Founders, for the internal limits of Congress’s powers, taken as a whole, to meaningfully limit the scope of federal legislation.

Because the concern of this Article is with the internal-limits canon rather than with the scope of particular congressional powers, the relevant question here requires looking to something more general than the original meanings of the Commerce Clause, the Taxing Clause, the Necessary and Proper Clause, and so forth. As already explained, I am not arguing for any particular reading of those or any other congressional powers. I am arguing against the idea that all of those powers must be interpreted to conform with a principle that exists at the level of general design—that is, that there are things that Congress cannot regulate even apart from the Constitution’s affirmative prohibitions. And also as already explained, my argument is not that Congress does have the practical equivalent of a police power under the conditions of 2014. It is that Congress might turn out to have power that broad, depending on the best interpretations of its various powers and the mapping of those powers, so interpreted, onto the present social world. So if it were the case that current constitutional law should track the original meanings of relevant constitutional clauses (whether on an original-public-meaning view or otherwise) and if it were further the case that the powers of Congress would add up to less than a police power if given their individual meanings, summed together, and applied today, then it would be the case that Congress is now meaningfully constrained by internal limits on its powers. But it would be the case only as a matter of contingency. To show that it must be the case as a matter of design requires a different kind of analysis—one that goes to Founding-era ideas about the function of the enumeration as a whole, rather than to the meanings of particular grants of power within it.

At the level of design, it is still the dominant conventional view that respect for the decisions of the Founders requires the internal-limits canon. The bases for that view are not obscure. So long as we remain conscious of the difficulties involved in attributing a complex view to a large and vaguely defined group like “the Founders,” and subject to the caveat that the Founders knew better than to trust the project of limiting Congress to internal limits alone, it is reasonable to say that the Founders intended the enumeration to play a role in limiting the jurisdiction of Congress.

It does not follow, though, that fidelity to the Founders’ design requires modern decision makers to identify consequential internal limits on Congress’s powers, because the relevant question is not whether the Founders expected internal limits to do that work. It is whether that expectation creates obligations today. This question is not simply a recapitulation of a more general question about the authority of original meanings, though there are points of contact between the present concern and that larger debate. Even if original meanings can bind later generations, the Founders’ ideas about limiting congressional power could be vindicated in the ways that matter even if internal limits turned out to impose no constraints on modern federal legislation.

To see why, it is important to recognize that the Founders’ decision to enumerate congressional powers is better understood as a matter of strategy than as a matter of principle. Those two categories are not always mutually exclusive, and it is possible to construe both “strategy” and “principle” in ways that would place the internal-limits approach under either heading. My statement that enumeration is better understood as a strategy simply means that the Founders understood enumeration as a means to an end, or a set of ends, rather than as a matter of independent value. The point of enumerating Congress’s powers was to help secure an adequately empowered but properly limited central government.159 The idea was that Congress should have the ability to do what the nation needs done, but it should neither deny individual rights nor imperil local decision making. Enumeration was a means to those ends. Madison spoke of the matter that way.160 So did Hamilton.161 So did James Wilson,162 Edmund Randolph,163 John Rutledge,164 and Oliver Ellsworth,165 who collectively constituted four-fifths of the committee that actually drafted the enumeration.166 The point here is not, of course, that every statement about the enumeration by every important Founding figure clearly articulates a means-ends orientation. The Founders were many people saying many things—including, here and there, that the enumeration would not limit Congress.167 The point is simply that understanding enumeration as a strategy rather than as a matter of independent normative value is true to an important way that many key Founders discussed the matter, such that characterizing the enumeration that way recovers a Founding-era conception rather than inventing a new one. Had the Convention delegates believed that an enumeration would not help the new government strike a good balance between federal and state power, they might well have done without one.

As things have turned out, the enumeration is not very helpful to the cause of striking such a balance. The internal-limits strategy has for a long time done little meaningful work. So it is fortunate that the Founders did not put all their eggs in the internal-limits basket. As already noted, the Convention was concerned primarily with process limits, and the broader public insisted on additional external limits precisely because it did not trust the argument that internal ones would be adequate. Today, it is those other kinds of limits that do the work of protecting individual rights and preserving robust roles for state and local governance. The question, then, is whether fidelity requires modern decision makers to use a strategy that does not work particularly well when other strategies the Founding generation blessed are available to do the work better.

The answer depends on whether there was an original commitment to the strategy as such, independent of its usefulness for achieving a set of goals. If so, fidelity to the Founders might require maintaining internal limits. But if not, modern decision makers can be perfectly faithful to the Founding design even if this particular strategy does no work at all, so long as they respect the commitments that the Founders did make and use other mechanisms to protect federalism and individual rights.

Consider the following hypothetical scenario as a relevantly parallel case. Charles’s last will and testament instructed his granddaughter Charlotte to use her inheritance to attend Radcliffe College. The will explained that this course of action was intended to help Charlotte achieve a financially secure adulthood. When Charles died in 1950, he had a clear expectation about the mechanism by which this strategy would succeed: Charlotte would meet and marry a Harvard man. But if Charlotte were to attend Radcliffe, and then law school, and then make a career as an attorney, she could both comply with her grandfather’s instruction and fulfill his ultimate purpose, whether or not she married rich, and whether or not attending Radcliffe had been a necessary step in the process. (Presumably she could have become a successful lawyer even if she had gone to college elsewhere.)

Has Charlotte been faithful to her grandfather’s design? She used the money to go to Radcliffe, not to tour Europe, and not even to attend Smith or Mount Holyoke or Bryn Mawr. She parlayed her education into a lucrative adulthood, rather than becoming a starving artist or a social worker or a missionary. So she both followed the specific directive in the will and achieved her grandfather’s articulated end. Without knowing more, that seems like enough to describe her actions as faithful to his design, even though the mechanism by which she achieved his end was different from the one he imagined.

If we sense that Charlotte subverted her grandfather’s plan while adhering to its formal requirements, it is probably because we think that Charles understood Charlotte’s marrying a wealthy man (or at least a man with good prospects) as a matter of independent value—that he would not have been content for his granddaughter to find financial security by dint of her own career. One can certainly imagine Charles as having held that view. But we cannot establish that he held that view on the basis of his will, the text of which does not speak of Charlotte’s marriage. Nor would it settle the question if Charlotte reported that her grandfather had told her before his death that he hoped she would marry well. Maybe he was simply talking about what seemed to him a natural solution, given the world he knew, rather than specifying that no other solution would be acceptable.

In the absence of a clear showing that Charles regarded marriage as the indispensable mechanism, it neither helps Charlotte nor flatters Charles to doubt that Charlotte adhered faithfully to the terms of her grandfather’s bequest. After all, we cannot doubt Charlotte’s fidelity without casting Charles as unattractively stubborn. If we remember Charles as a sensible man or even a wise one, we would attribute to him the understanding that he might not foresee everything about the future, as well as the desire for his granddaughter to do well in ways that made sense in light of the world she lived in, even if he would also want her to respect the terms of the gift he made. On that understanding, it would make sense to conclude that Charles would be satisfied with Charlotte’s decisions. She both adhered to the rule he laid down and realized the end that he wished for her. Not in exactly the way he expected, but in a way consistent with her obligations to him.

I take it as given that Congress can legislate only on the basis of powers delegated to it under the Constitution, just as Charlotte must use her inheritance to attend Radcliffe rather than Smith. The text of each governing instrument makes that a rule.168 But unless the Founders were committed to overall internal limits as a matter of value rather than as a means to an end, fidelity does not require that internal limits be the mechanism limiting Congress any more than fidelity to Charles requires Charlotte’s marrying someone she met in college to be the mechanism by which she attains financial security. The question, then, is whether we should understand the Founders as committed to internal limits as a value-based matter rather than as a strategic one. Why, if at all, should we think that the Founders regarded the enumeration of internally limited powers as a mechanism that must do meaningful work in limiting Congress, rather than as a mechanism they expected to do a fair amount of that work?

In part for reasons that I will explain in more depth in Part IV, the answer cannot be the text of the Constitution. To be sure, the Tenth Amendment and the structure of Article I, Section 8 both imply that the Founders intended the enumerated powers as less than a general grant of regulatory authority. But neither text requires understanding that intention as a binding commitment for the construction of the enumerated powers, rather than as an expectation about how the system would function. Note that the Constitution in many respects reflects certain expectations without demanding that those expectations be made real. The Post Office Clause,169 for example, reflects the Founders’ expectation that people would communicate via written documents that would be physically carried from place to place. But no constitutional principle would be violated if that system of communication were to disappear at some point after the coming of the Internet. It would simply be the case that the Framers had certain expectations about how the system would function that turned out, after a time, to be at variance with the world’s actual conditions.170 So yes, (most of) the Framers expected the enumeration to help preserve state autonomy and protect individual rights. To adapt Chief Justice Marshall’s formula from Gibbons, they “presupposed” it,171 just as the Post Office Clause presupposed the utility of a postal system and just as Grandfather Charles presupposed that marriage was a woman’s ticket to financial comfort. Article I was written with the expectation that the enumerated powers would be limiting, and the Tenth Amendment was written with the expectation that the powers delegated to the United States would have narrower scope than a grant of general legislative power would. But as both the Post Office Clause and Grandfather Charles’s will illustrate, a text can be written in light of certain expectations without embodying a demand that those expectations be realized.

If it is not clear that the Founders thought of enumeration as more than a means to an end, then there are good reasons not to attribute that idea to them. First, and as noted above, leading Founders discussed the matter in instrumental terms.172 Second, an instrumental view of enumeration is more consistent with the image of the Founders as “practical statesmen, not metaphysical philosophers.”173 The idea of enumeration as a means is straightforwardly practical: to protect state decision making and individual rights, give Congress only a limited set of powers. Finally, it bears remembering that one of the best modern value-based reasons for prizing the system of enumeration is one that would have had no purchase for the Founders. Today, affirming the role of the enumeration displays loyalty to a tradition we inherit, at least at the level of discourse and professed principle. The Founders approached the issue free of any such consideration. For them, authority and tradition did not much bear on how best to understand Congress, and their opportunity to confront the question as a matter of practical wisdom was commensurately greater.174

If the Founders regarded enumeration largely as a practical strategy, then it matters when thinking about fidelity to the Founders that the enumeration strategy is not practical. Maybe it was practical once upon a time: internal limits seemed175 meaningful in the Founders’ own lifetimes and for some period thereafter.176 But as the Founders foresaw, the system of internal limits could not be trusted to do the job by itself, and it has not worked well in a very long time. It shows no disrespect to the Founders to say that one of their strategies was effective for fewer than two hundred years, especially when we also remember that they were thoughtful enough to include other strategies that continue to be effective. Perhaps an undying commitment to a strategy that no longer works displays a certain kind of loyalty, but it is not a particularly helpful kind. Nor, assuming the Founders to have been reasonable and intelligent people, is it the kind of loyalty that the Founders would have wanted their successors to show.

On the generally warranted assumption that the leading Founders were sophisticated about human behavior, we should take for granted that they understood that not all of their strategies would succeed forever in quite the way they initially imagined. Their insistence on including external limits testifies to their recognition—or at least their suspicion—that enumeration might not be adequate for checking Congress, either in their own day or in the foreseeable future. And so long as external limits and process limits do the work of preserving state decision making and protecting individual rights, the system remains faithful to its Founding design. To say otherwise—to insist that the enumeration do meaningful work even when other mechanisms get the job done—is to mistake the object of constitutional fidelity. It confuses a practical choice with a value choice and a tool with the purpose it is supposed to serve.

IV. text

My argument about what fidelity to the Founding design requires should not be confused with a suggestion that modern decision makers are free to adopt any view of congressional power that is at some high level of generality consistent with the Founders’ wish that state and local governments wield considerable decision-making authority. Constitutional law is more constraining than that; the Constitution frequently says what officials must do and how they must do it, rather than simply what their purposes must be. For example, the process for constitutional amendment under Article V requires the assent of three-fourths of the states, and we cannot interpret Article V to permit amendment with only three-fifths of the states, or with states whose combined populations comprise three-fourths of the national population whether or not those states also comprise three-fourths of the states, even if we decided that one of those alternatives would be a more sensible threshold for vindicating the Founders’ purposes in creating an amendment process than the one Article V now provides.177 Sometimes the text of the Constitution just prescribes rules, the contravention of which is unconstitutional. So even if neither federalism nor the vision of the Founders requires the internal-limits canon, the canon might still be mandatory in light of the text of the Constitution.

In my view, the text of the Constitution does not so require. But as noted before, I am not arguing that the Constitution confers the equivalent of plenary power on Congress. It might, or it might not, depending on the best constructions of many different powers and the relationship between those powers and the social world at any given time. My argument leaves room for disagreement about the meaning of the Constitution’s power-conferring clauses, including the Commerce Clause. If it turned out that on the best reading of all of the powers, Congress possessed less than general regulatory authority, it would be true that the powers of Congress are subject to overall internal limits. But the conclusion would have been established without the influence of a fallacious rule—the internal-limits canon—forbidding any construction on which Congress turned out to have plenary power in practice. My argument about the text of the Constitution takes no position on whether the sum total of Congress’s powers is equivalent in scope to a police power. Instead, my argument is that the text is best read to make the answer to that question a matter of contingency, rather than a categorical no.

The traditional view is that Article I and the Tenth Amendment do require the categorical no that the internal-limits canon directs. From a certain perspective, those readings of the relevant texts are sensible. But what makes them sensible is the preconception that internal limits are necessary for maintaining federalism, for respecting the intentions of the Founders, or for some combination of the two.178 If those preconceptions were correct, then it would be easy to enlist the text in support. As I have already argued, however, those preconceptions should be set aside. Without those preconceptions, neither Article I nor the Tenth Amendment requires that internal limits do meaningful work.

A. The Tenth Amendment

The Tenth Amendment reads as follows: “The powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people.”179 This sentence is often read to confirm the idea that the powers of Congress are, collectively, less than a general grant of regulatory authority. The logic of that reading is easy to reconstruct. First, the Amendment speaks of “powers not delegated by the Constitution to the United States,” which implies that there are such powers. Next, the Amendment directs that such powers (except those that are “prohibited . . . to the States”) are “reserved to the States . . . or to the people,” meaning that they cannot be exercised by the federal government. Accordingly, the Tenth Amendment provides that Congress can exercise those powers delegated to the federal government and no others (again subject to the exception for powers prohibited to the states). Unless the Amendment states a rule with no applications, therefore, the powers delegated to the federal government must encompass less than general regulatory authority. Assuming that an Amendment should not be interpreted to be devoid of applications, the principle that the powers of Congress are only a subset of all possible legislative powers is prescribed by the Tenth Amendment.

The first thing to notice about this argument is what it concedes: the proposition that some powers must be withheld from Congress is not fully specified in the words of the Tenth Amendment. It rests also on an assumption about how the Tenth Amendment is supposed to operate. Everyone recognizes this feature of the reading, including the Supreme Court, even in some of its most state-protective moments. In New York v. United States, for example, the Court explained that the Tenth Amendment’s limit on the powers of Congress “is not derived from the text of the Tenth Amendment itself.”180 After all, the Tenth Amendment does not say, “The powers delegated to Congress must be construed, collectively, as less than a grant of general jurisdiction.” Instead, the Tenth Amendment supplies a rule applicable in cases where the delegated powers would not authorize federal legislation, and we infer from the existence of such a rule that there must be limits to what the delegated powers can reach. We make that inference not because the words of the text require it but because of something we think we know about how such a text is supposed to operate. What we think we know is that constitutional clauses must have applications. People do not bother to amend constitutions with provisions that do no work.

The foregoing argument traffics in an idea about the authors and ratifiers of the Tenth Amendment—specifically, that they must have meant their text to have applications. But the contention that it makes no sense to read the Tenth Amendment in a way that deprives its stated rule of all applications need not point to any evidence about the actual views of particular eighteenth-century Americans, other than whatever such evidence might be furnished by the Tenth Amendment itself. The idea that these words must have applications is an instance of an idea that might prevail in textual interpretation generally: that enacted texts should not be construed to have no meaning.181 Whatever we know (or cannot know) about the Founders, this line of thinking maintains, it makes no sense to approach text in the Constitution as if it were meaningless. The fact that this argument is easily made based on a commonsense reading of the text alone probably accounts for some of its appeal. But whether we understand the argument as “purely textual” or as some combination of textualist and originalist, the proposition that the Tenth Amendment’s words must have applications is a common idea about how that text must be read. And whichever way we understand the argument, it has the same basic flaw. It confuses a proposition about the motivations of constitution-makers with a proposition about the functioning of constitutions.

The confusion is easily stated. One can reasonably assume that people do not write rules that they know will never have applications. But it does not follow that every rule people write has, or in the future will have, applications. Sometimes people write rules providing for situations that might arise, just in case. Sometimes the world turns out to be different from what the writers of a rule expected, such that an anticipated set of applications fails to materialize or dissolves over time. This last possibility is especially plausible for rules that remain in place for long periods of time, because the circumstances for which a rule was written might be characteristic of the world the authors knew but not of the world at all later times. So even if we assume that the authors of rules expect those rules to have applications, it would be a mistake to assume that every rule people write will turn out to have applications forever. To be sure, it is prudent, upon discovering that a rule has no applications, to ask whether the rule is being read correctly. But it is not at all prudent to take a discovery that a rule has no present applications as proof that the rule is being wrongly read.

Consider an example that carries none of the baggage of debates about constitutional theory. Imagine the Wolverine Summer Day Camp, located in southeastern Michigan and founded in 1970. The bylaws of the camp, adopted at the time of its founding, contain the following rule: “On days when the temperature is not forecast to exceed ninety degrees Fahrenheit, campers will spend the day outdoors.” It is reasonable to infer from the existence of this rule that the authors of the bylaws expected there to be some summer days with temperatures above ninety degrees and other summer days with maximum temperatures below ninety degrees. In a given year, however, it might turn out that the temperature was forecast to exceed ninety degrees on every day of the summer camping season. If that happened, the rule would have no applications. Campers might spend the entire summer inside, and if they did, the rule would not have been violated. (Whether the rule should be revised at that point is a separate question.)

When the bylaws were adopted in 1970, it probably seemed far-fetched to think that southeastern Michigan could have a summer in which the temperature exceeded ninety degrees on every single day. But in 2014, it is easier to imagine that circumstance becoming a reality. Indeed, it is easier to imagine it becoming a reality on a permanent basis, or at least to imagine that days when the rule could be invoked would become exceedingly rare. If that happened, the rule would lie dormant. But it would not be violated. It would simply testify that the people who wrote the rule felt it important to address a situation that no longer presented itself.

As a matter of textual interpretation, the Tenth Amendment is properly analyzed in the same way. For a long time after its adoption in 1791, the text of the Amendment addressed a situation that arose in practice. Later, that situation dissolved, either entirely or nearly so. After that change in circumstances, the fact that the rule stated in the Tenth Amendment has virtually no applications does not mean that the Tenth Amendment is being violated. The rule simply does not come into play. It poses no problem for this understanding to say that the Founders must have expected the Amendment to have some function. Of course they did—and indeed, the Amendment’s stated rule did have applications when they wrote it. But there is no reason to think that a sensible construction of a text can only be one that ensures the text has applications always and forever. A clause can be perfectly worth adopting even if it only has consequences for a hundred years.

It is worth being clear about two limits of the present argument, each of which concerns the subtle but crucial distinction between the text and the operative content of a given constitutional provision. First, even if the rule stated in the text of the Tenth Amendment has no contemporary applications, it need not follow that the Tenth Amendment has no contemporary applications, because a constitutional provision as applied sometimes has force different from what an untutored reader of the text might expect.182 And indeed, I am not arguing that the Tenth Amendment has no applications in modern constitutional law. As a matter of doctrine, the anticommandeering rule of New York v. United States183 and the sovereign-immunity rule of Alden v. Maine184 are applications of the Tenth Amendment, at least in part. These cases did not apply the rule stated in the text of the Tenth Amendment, as the Court in both cases acknowledged.185 They applied a related but not identical proposition about the value of state sovereignty, one that is best understood as arising from some combination of a theory of constitutional structure and a view of Founding-era history. Nonetheless, constitutional practice usually associates constitutional doctrines with particular clauses, even when the text of those clauses does not quite state the relevant doctrines.186 We unhesitatingly apply the First Amendment against the federal executive, despite its being addressed to “Congress,” because we associate the First Amendment with the general idea that the government must respect the freedoms of speech and religion.187 There is a similar association between the Tenth Amendment and the general idea of state sovereignty as a limit on federal power, and that association allows Alden and New York to be classified as applications of the Tenth Amendment.188

The Tenth Amendment rule that I have argued may have no present applications is the rule prescribed by the literal text of the Tenth Amendment, not the more robust rule invoked in Alden and New York. So if the Court has construed the Tenth Amendment correctly, it remains true even today that the Amendment continues to do work. It just might not do the work of directing results in cases where Congress has enacted legislation exceeding the internal limits of its delegated powers, because that situation might no longer arise. Instead, the Tenth Amendment now has applications as a source of external limits on congressional power. New York and Alden limit what Congress can do under any of its Article I powers, and they do so on the basis of something outside of the enumeration rather than by the terms of any clause conferring power on Congress. Currently prevailing doctrine on commandeering and sovereign immunity is controversial, and I do not mean to take a position here on whether New York and Alden were rightly decided. But whatever the merits of these particular decisions, the existence of judicially enforceable external limits on congressional power is entirely consistent with the argument of this Article. My point about the Tenth Amendment is simply that its text does not compel the conclusion that there must be internal limits.

Second, I am not arguing that the Tenth Amendment cannot possibly be read to support the internal-limits canon. The text of the Tenth Amendment does not require the powers of Congress to be construed as internally limited, and in my view no sound principle of constitutional interpretation forbids reading that text to mean what it says. But again, it often happens that a constitutional provision is taken to mean something different from what is stated in its text. The Tenth Amendment itself is a prominent example, as just discussed: the Supreme Court already takes that provision to embody a proposition of federalism that is related to, rather than contained within, its text.189 So if the substantive idea that the powers of Congress must be construed as internally limited is a valid constitutional principle—because federalism requires it, or because the Founders’ intentions require it, or for any other reason—then it would be consistent with the way American constitutional interpretation works to read the Tenth Amendment to represent that idea. But if neither federalism nor fidelity requires the powers of Congress to have meaningful internal limits, then there is no reason to (mis)read the Amendment as making such limits mandatory. On its own, the text of the Tenth Amendment contains no such rule.

B. Article I

The other main textual arguments for the necessity of internal limits focus on Article I, Section 8.190 Again, the relevant arguments are not close readings of the several power-conferring clauses of Section 8, deployed in combination (and in combination with readings of the Constitution’s many other clauses conferring power on Congress) to show that the sum total of the powers leaves Congress unable to legislate on some set of topics. That kind of argument might or might not be successful. But even if such an argument successfully revealed limits on Congress’s overall power to legislate, it would not be germane to the validity of the internal-limits canon. The canon is a rule for construing the powers of Congress (“Under no circumstances may you read these powers such that they end up covering all possible subjects of legislation”), not a description of something that one would discover after construing the powers without the influence of that rule (“Hey, I read all these grants of power and thought about what they add up to collectively, and it turns out that they don’t exhaust all possible subjects of legislation”). A textual argument for the internal-limits canon would have to support the rule itself in a way conceptually prior to the interpretation and summation of many individual powers. And since nothing in Article I states the internal-limits canon directly, the Article I arguments for the internal-limits canon are not close readings of particular constitutional passages. They are readings of the structure of Section 8, considered as a whole.

To be precise, there are two textual arguments from Article I, both building from the foundation that Section 8 confers a long list of particular powers. First, there is the argument against generality. That argument, in short, is that it makes no sense to read a list of particular powers as equivalent to a grant of general power, because a list of specifics is a strange way to denote something general. If Article I were designed to confer general legislative power, it could just say so.191 Second, there is the argument against redundancy. This argument holds that it makes no sense to read one power in a list of powers—say, the commerce power—so expansively that its scope approaches that of a police power. To do so would make most of the rest of the list redundant. And it is odd, this argument maintains, to construe a text with eighteen different clauses in a way that makes most of those clauses superfluous.192

The two arguments have much in common, including the assumption that a reasonable inference about how a text was expected to function is also a reasonable inference about how the text will function. That assumption is out of place here, just as it was with respect to the Tenth Amendment.

Consider first the argument against generality. Chief Justice Marshall’s dictum that an enumeration presupposes something not enumerated is traditionally mobilized to support this idea: a specific list is something that people write when they want to distinguish those things that are on the list from other things that are not on the list.193 The basic logic here is that of the familiar interpretive canon expressio unius est exclusio alterius: the specification of one thing is the exclusion of another. As a general matter, expressio unius is a sensible interpretive intuition, so long as it is treated as rebuttable where the particulars so direct. So it is easy to note that the Constitution provides a long list of specific congressional powers, recite expressio unius, quote Chief Justice Marshall, and rest confident in the view that the internal-limits canon must be correct.194

But things are not so simple. For starters, this thought process bundles the enumeration principle and the internal-limits canon closely together and in so doing obscures a slippery transition between two steps of reasoning. The first step, which is the one that corresponds to the enumeration principle, goes like this: the Constitution specifies many powers of Congress, and we take that to mean that Congress may act only on the basis of those powers. But the second step is more problematic. It is the inference that the powers specified cannot authorize any and all legislation that Congress might pass under the circumstances of a given time. That second step is the one that purports to establish the internal-limits canon, and it does not follow automatically from the step before, because there is no general rule by which all enumerations of authority must exclude some action that the authorized party might take.

Suppose that I am leaving instructions for a friend who will stay in my house while I am on vacation. I might write as follows: “You can use my shower, or you can use the kids’ shower, or you can use the guest shower.” If my house has three showers, then this enumeration does not limit my friend’s choices at all, and I do not intend it to. If my friend has a certain sort of analytic mind, he might read the list, notice that the house has only three showers, and wonder why I bothered to write a longer sentence than “You can use any shower in the house.” But if he also has basic common sense, he will not long be bothered by the question, because he will generate perfectly adequate explanations. Maybe I wrote the list that way to make sure he knows what all of his options are. Or maybe I thought that giving permission to use the showers in general might still leave him wondering whether I preferred that he use a particular one, so it made sense to emphasize his equal privilege to use any of the three.

The enumeration of powers in Article I differs from the shower enumeration in that the powers granted in Article I were not intended to be exhaustive. When I list all three showers in my house, I know that I am conferring authority that is in practice as broad as the authority I would confer by writing, “Use any shower you like.” And I take it as given that the Founders—or most of them, anyway195—understood the enumeration to give Congress less authority than a general police power would. To appropriate the terms of the Gibbons dictum, they presupposed something not enumerated. But it is important to think carefully about what it means to presuppose. A presupposition is an assumption, but it is not always a requirement. The bylaws of the Wolverine Summer Day Camp presupposed summer days with temperatures below ninety degrees; Grandfather Charles presupposed that Charlotte’s financial welfare as an adult would be a function of her marriage. But neither the camp bylaws nor Charles’s will required that those presuppositions actually describe the world of the future, and indeed the idea of imposing such requirements would be nonsensical. The point, of course, is that because the authors of texts are not omniscient about the circumstances in which those texts will be applied, textual rules are sometimes applied in situations that lack some of the features that the authors presupposed. But textual rules often make sense even when some of their authors’ presuppositions turn out to be inapplicable, such that implementing the rules as written would still vindicate the purposes for which they were adopted.

Imagine that the Simpsons run out of milk. Marge sends Homer to the Kwik-E-Mart with a note that says, “Here are the kinds of milk you may buy: 2%, 1%, and skim. Do not buy any milk that is not on this list.” It is clear that Homer is not supposed to buy whole milk, or chocolate milk, or Duff-brand beer milk. But if Homer arrives at the Kwik-E-Mart and finds the place sold out of all milk other than 2%, 1%, and skim, he may buy any kind of milk in the store. He should not say, “Well, the shopping list clearly communicates to me that I can’t buy just any milk in the store. If I were authorized to buy any milk in the store, it wouldn’t make sense for the shopping list to be written as an enumeration of specific authorized purchases. So let me figure out what I must refrain from buying.” Instead, Homer should understand that the list is written the way it is because Marge had an expectation—a presupposition—that was not borne out in practice. And he should carry out his shopping just as he would have, on these facts, if the shopping list had said, “Buy any milk in the store”—not because Marge authorized Homer to buy any and all milk, but because in the applicable circumstances the scope of the limited authorization she gave does not exclude any practically available course of action.196

None of these examples is exactly like Article I. But they should dispel the idea that sound textual interpretation necessarily precludes reading an enumeration of particular powers to have the same effect as a grant of general power. In the absence of some reason extrinsic to the text to read the enumeration as limiting, the fact that the text is written as an enumeration does not require reading it that way.

To be sure, such extrinsic reasons might exist. But establishing the existence of such a reason would require reference to something beyond the text, and it is not clear what that something would be. For reasons described in Part II, the demands of American federalism do not furnish such a reason, because internal limits are neither necessary nor particularly helpful for a healthy federal structure. For reasons described in Part III, a desire to be faithful to the Founding design does not furnish such a reason either. And in the absence of a good reason to treat the presupposition as a condition for the exercise of the rule, there is no reason to read the text as contemplating a smaller grant of power to Congress than its words indicate. Whether the power-granting texts are properly read to confer less power in practice than would flow from a general grant of authority depends on the real-world relationship between the powers granted and the circumstances under which Congress acts.

Consider next the argument against redundancy. It is reasonable to presume that people writing lists of rules do not intend most of those rules to be redundant. But that is a far cry from saying that a reading on which some rules are redundant is necessarily a misreading of the text. Some documents include redundancies for the same reasons that speech includes redundancies. Sometimes redundancies are unintentional: writers, like speakers, are not always precise. And sometimes redundant words are included intentionally, perhaps for emphasis rather than for substantive meaning. (Under Article II, Section 1, the President swears to “preserve, protect and defend the Constitution.”197 These are not three separate duties that the President undertakes.) Moreover, the Constitution was written by many people working partly in cooperation and partly at cross-purposes, and anyone who has drafted a complex document in a large committee knows that spare and elegant composition is often costly to obtain. (“I know that § 12(g) is unnecessary in light of § 4(a), but Johnson is attached to § 12(g), so I’ll keep my mouth shut so we can keep working.”) To be sure, if one’s reading of a text causes parts of the text to do no work, it is wise to consider whether a different reading might be better. But the conclusion that parts of a text are simply redundant cannot be categorically excluded.198 If you doubt it, try for the next twenty-four hours to read every text you see as if it had no unnecessary clauses. Try to read just one carefully edited law review article that way.

Once the general point that even carefully crafted texts have redundancies is recognized, it is easy to spot redundancies in the Constitution—and not only in the clause specifying the Presidential Oath. Consider Article I, Section 8, Clause 6, which states that Congress has the power to provide for the punishment of counterfeiting the securities and current coin of the United States.199 What does this clause add to the clauses granting Congress the power to issue securities and coin in the first place? Do we think that in the absence of Clause 6 Congress could coin money but would be powerless against counterfeiters? That seems unlikely, both because it seems contrary to common sense and because it is sharply at odds with the way other parts of Section 8 are read. No constitutional clause expressly grants Congress the power to punish people who fail to pay their taxes, or people who violate copyrights, but nobody doubts that Congress has those punishment powers, whether as inherent incidents of the underlying powers to tax200 and to grant copyrights201 or as means necessary and proper for carrying those powers into execution.202 If the Constitution did not expressly specify congressional power to punish counterfeiting, the power would be there just the same, exactly as it is for all parallel congressional powers. So Clause 6 does no independent work.

One could reasonably accept all of the foregoing points and still think that reading the Commerce Clause as a grant of nearly general regulatory authority creates too much redundancy to be credible. In other words, the causes of redundancy that I identify above might explain a bit of extraneous text here and there, but they cannot plausibly explain why most of a text—here, most of the eighteen clauses of Section 8—is extraneous. That may be. But for reasons that largely overlap with the reasons why a set of enumerated powers can turn out to be indistinguishable in practice from a grant of general authority, the amount of a document that can reasonably be read as redundant can change as the circumstances in which the document is applied change.

Distinguish, therefore, between original redundancy and acquired redundancy. Original redundancy is redundancy that exists in a document on the day the document is written. For the reasons given above, the best reading of a document might find a bit of original redundancy. That said, it would be strange for a document—or at least a document that is intended as a carefully crafted and practically operative set of rules—to contain enormous redundancies as an original matter. Acquired redundancy, however, is the redundancy that exists in practice when the document is applied at some later time, after relevant conditions have changed. Imagine, for example, a New York game-preservation statute adopted in 1800 and providing as follows: “No deer shall be killed (1) within one day’s overland travel from Albany; or (2) within one day’s overland travel from New York City.” (The drafters figured that protecting deer near large population centers would reduce nonessential hunting and preserve a food supply needed in rural areas.) The statute contains little if any original redundancy. In 1800, most or even all places that lay within one day’s overland travel of one of those cities were not within one day’s overland travel of the other. Today, however, the two rules would be completely redundant. Or consider a treaty between Britain and France, signed in 1820, providing that neither country would maintain a naval presence “off the coast of the United States, or that of Texas.” After 1845, the second provision would be completely redundant. But it would be a mistake at that point to read the treaty as if the second provision carried some meaning not entirely contained within the first.

A reading of Article I, Section 8 on which the list of enumerated powers were shot through with original redundancies would likely be a bad reading. There is language here and there that could have been removed without changing the substantive import of the section even in 1787,203 but by and large it is reasonable to think that the drafters expected the several enumerated powers to do different things. It is perfectly plausible, however, to think that the enumerated powers might feature a great deal of acquired redundancy.

This is not to say that change over time necessarily justifies reading constitutional language as redundant. How much redundancy the Commerce Clause creates in the rest of Section 8 depends on how broadly the Commerce Clause applies, and as already noted I am not arguing for any particular construction of that Clause. I am simply pointing out that the appropriate constructions of the enumerated powers, whatever they might be, need not be limited to avoid redundancy. The idea that every piece of a text must be construed to carry its own distinctive meaning may arise from a well-intentioned effort to have the text make sense as a whole, but the principle of interpretation that such an effort seeks to vindicate is simply misplaced.204 If a reading produces redundancies, it is worth thinking twice about whether one is reading correctly. But if one can give a reasonable account of the redundancy in question, and if the reading otherwise makes sense, there is no good reason to rule out readings simply because they make certain pieces of text extraneous.

For all these reasons, the text of Article I, Section 8 does not require the internal-limits canon. To be sure, there is an available reading of that text that would support the idea that the enumeration is limiting if there were good extratextual reasons to understand the enumeration that way. Much the same is true of the text of the Tenth Amendment. But in the absence of good extratextual reasons for the internal-limits canon, the text of the Constitution does not require it.


For a long time, constitutional law has featured a stable paradox whereby courts articulated the internal-limits canon but permitted Congress to legislate as it thought proper, within the external limits of constitutional law. It is not clear whether that arrangement will remain stable. If it will not, the choice is between revising our practice to match a theory and revising that theory to match our practice. In my view, the practice is sensible and should prevail. In that respect, my argument is conservative, in the classical or Burkean sense: it seeks to defend the stability of a longstanding practice against the threat posed by reformers wielding an abstraction. As measured by the conventions of constitutional discourse, my argument is radical, because it recommends the abandonment of a traditionally orthodox idea. It is the radicalism of looking at what we are already doing, indeed at what we have done for a long time, and deciding to be at peace.