Substantive due process is among the most vilified ideas in American law. Sometimes it seems lawyers compete to find the cleverest way to ridicule it, as an “oxymoron,” a “contradiction in terms,” or a mere trick whereby judges substitute their personal political opinions for the law. Just weeks ago, Justice Clarence Thomas restated his longstanding rejection of substantive due process: “the Fourteenth Amendment’s Due Process Clause is not a secret repository of substantive guarantees against ‘unfairness.’”

I think that’s wrong. As I have argued recently in the Harvard Journal of Law and Public Policy, substantive due process is as legitimate—indeed, as crucial—a part of our Constitution as the principle of, say, separation of powers. Note that the phrase “separation of powers” doesn’t appear in the Constitution; it’s an abstract principle one infers from the structure, ideas, and history of the document. The same is true of substantive due process.

Perhaps saddest of all, it often seems that the most vocal critics of substantive due process don’t even understand how the doctrine works in the first place. My point here is to explain briefly how the Constitution’s promise that “no person shall be deprived of life, liberty, or property without due process of law” means not only that government must take certain procedural steps (hearings, trials, and so forth) when it imposes a deprivation, but also that some acts are off limits for government, “regardless of the fairness of the procedures used to implement them.”

The confusion may be partly due to sloppy language. “Substantive due process” is an epithet, coined in the New Deal era by Progressive legal scholars who opposed the theory. The judges they attacked for embracing the doctrine—people like Justice Stephen J. Field—would not have recognized the term. They just called it “due process of law.” Also confusing is that lawyers today habitually refer to the “Due Process Clause,” leaving out the most important part of that clause: the phrase “of law.” The Constitution does not just guarantee “process”; it guarantees a process of law. To understand what “substantive due process” means—even if one rejects the idea—one must not forget that what this Clause promises is that government will accord us lawful treatment; that it will only take away life, liberty, or property, in accordance with principles of lawfulness. So our first step is to ask, what does it mean for government to pledge to deal with us in a lawful manner?

Law as the opposite of arbitrariness

What is “law”? Philosophers have debated this question for many years, and while they may disagree on certain points, some basic answers have emerged: law is the use of government’s coercive powers in the service of some general principle of the public good. Aristotle famously distinguished between lawless, corrupt regimes where the people were governed with “regard only [for] the interests of the rulers”—and lawful, healthy regimes “which have a regard to the common interest.” The former is governed by specific commands to particular persons to do particular things, or by actions that lack any general purpose, or only to the benefit those who wield power. They are “despotic”—more analogous to the rule of a master over a slave. Lawful regimes, by contrast, are characterized by general rules that govern for the benefit of all. Or, as a more recent authority, Cass Sunstein, has put it, “many of the most important clauses of the Constitution,” including the Due Process Clause, are “focused on a single underlying evil: the distribution of resources or opportunities to one group rather than another solely on the ground that those favored have exercised the raw political power to obtain what they want.” This is the distinction between “the rule of law and the rule of men.”

More basically, law is the opposite of arbitrariness; it is not the self-serving use of force by those who wield it. It is not ipse dixit—not merely that someone in power has said so. And this is both a substantive and a formal criterion: if government imposes a rule without following the procedures whereby a rule obtains its official character, the final product cannot be called law; likewise, when government imposes something that it calls legislation which is nevertheless arbitrary, self-serving, ipse dixit, that, too, cannot be called law, notwithstanding any procedural formalities. As Daniel Webster famously said in his argument in the 1819 Dartmouth College case, the promise of due process of law means

that every citizen shall hold his life, liberty, property, and immunities, under the protection of the general rules which govern society. Every thing which may pass under the form of an enactment, is not, therefore, to be considered the law of the land. If this were so, acts of attainder, bills of pains and penalties, acts of confiscation, acts reversing judgments, and acts directly transferring one man’s estate to another, legislative judgments, decrees, and forfeitures, in all possible forms, would be the law of the land. Such a strange construction would render constitutional provisions of the highest importance completely inoperative and void. It would tend directly to establish the union of all powers in the legislature.

More simply, due process of law means government may not limit our freedom without good reason. What qualifies as a “good reason” is a question answered by reference to political and legal principles. Not everything the legislature promulgates is, on that account, a good reason. According citizens due process of law means to treat them, not in accordance with whatever the majority happens to desire at any particular time, or to serve the ruler’s (or rulers’) self-interest. Thus the overlap of “procedure” and “substance” is inevitable: to be treated lawfully means to be treated in accordance (procedural) with general, public principles (substantive). Let me make this last point clearer.

The logic of substantive due process

Suppose Congress passed a bill—say, a new tax—which the President vetoes. Having been vetoed, that bill does not become not a law, for procedural reasons: the shortcoming that deprives it of status as law is entirely formal, not substantive. It doesn’t depend on the bill’s content, but on rules of promulgation. If the IRS were to try to enforce the tax by punishing someone who doesn’t pay it, the punishment would be unlawful because it is unauthorized: the IRS would be enforcing something that is not law, thus depriving the citizen of liberty without due process of law.

The same approach holds where a purported law fails for substantive, instead of procedural, reasons. Thus, imagine Congress were to pass a bill, and the President were to sign it, establishing an official religion for the United States. Since the First Amendment denies Congress power to make such a law, no matter what procedural steps it takes, the resulting statute would have no claim to status as law. And if the sheriff were to arrest a dissenter for violating it, he would be depriving that person of liberty without due process of law, just as in the first hypothetical. Here, the sheriff’s purported authority for arresting the citizen fails not because of any formal shortcomings, but because the substance of the purported law is such that it cannot claim the character of law.

These two examples are relatively easy because they rest on explicit constitutional limits on government power. But the same logic holds with regard to implicit or inherent limits on government power. These implicit limits have been the sources of the major battles over substantive due process. But even if one disagrees as to what those limits might be, the argument follows a plain logical form: if the legislature passes a statute that it lacks authority to make, that statute has no standing as law, and enforcing it would violate the citizen’s right not to be deprived of life, liberty, or property except by due process of law.

Implicit limits on lawmaking authority are commonplace. In Clinton v. City of New York, for example, the Supreme Court ruled that the Constitution does not allow Congress to give the president a line-item veto, even though it does not explicitly prohibit such a thing. The reason is that the Constitution sets forth in detail the procedure for making bills into laws, and “is silent on the subject” of alternative methods. This “constitutional silence” is “equivalent to an express prohibition.” Thus the Constitution implicitly bars Congress from devising different methods for making or vetoing laws.

If there are inherent restrictions on the procedures by which a bill can become a law, then there would seem no denying that there are also implicit limits on the content of laws that can be made. If law is the opposite of arbitrariness, then the legislature cannot get around the prohibition on arbitrariness by simply labeling an arbitrary act “law.”

Arbitrators and bank robbers

Are there implicit limits on the kinds of laws the legislature can pass? Consider—as the Founding Fathers often did—some analogies to contract law. In arbitration law, an arbitrator derives his authority from the contract between the parties, but that contract can limit the arbitrator’s authority either explicitly or implicitly. As the Seventh Circuit has observed:

A suit to throw out a labor arbitrator’s award is…a suit to enforce the labor contract that contained the clause authorizing the arbitration of disputes arising out of the contract…. [T]he plaintiff normally will be pointing to implicit or explicit limits that the contract places on the arbitrator’s authority—principally that he was to interpret the contract and not go off on a frolic of his own—and arguing that the arbitrator exceeded those limits.

An arbitration decision that exceeds the implicit limits of the contract is unauthorized and has no validity.

Or consider a bank guard. Classical liberals like the Framers envisioned government as akin to a guard: the people in society, anxious to protect their resources and freedoms, “hire” the government to protect them just as the owner of a bank would hire an armed guard. But while this may prevent robbery, the bank owner now has a new problem: he has allowed someone in his bank with a gun, and the guard might give in to temptation and rob the bank himself. What happens then?

Something very like this actually happened in Sunshine Security & Detective Agency v. Wells Fargo Armored Services Corp., a Florida case in which a bank contracted with a detective agency to provide a guard, who then conspired with third parties to rob the bank. The bank sued the detective agency on respondeat superior grounds, but the court rejected this argument because the guard’s actions were “[a] classic case of an employee acting outside the scope of his employment. The subject employee was hired…to guard the bank which he, in fact, conspired to rob. In this endeavor, we think the employee was plainly off on a frolic of his own, [and] was in no way furthering the interests of his employer.”

But notice: the guard’s employment contract almost certainly included no explicit prohibition on robbing the bank. (Who would think of such a thing?) Yet the court rightly concluded that such a prohibition was implicit in the employment contract: it arose from the logic of the contract itself. The whole point of hiring him was to ensure against robbery. So when the guard betrayed these purposes, he went beyond the scope of his legitimate authority—he acted ultra vires—beyond his powers.

Where do the inherent limits on government action come from? They lie in the word “law.” Law is the use of government power in the service of a rational, general, public principle. A government action that lacks these elements is not law; it’s something else. As Randy Barnett put it in a different context, “the qualities must go in before the name ‘law’ goes on.” So a government action that does not serve a rational principle of the public good is not a law, and thus deprives persons of life, liberty, or property without due process of law.

The Constitution, too, is a law. Thus by promising that government will accord us due process of law, the Constitution not only blocks government from breaching the explicit boundaries on its authority, but also from violating inherent limits that arise from the logic of the constitutional “contract” itself. The Constitution is the employment contract by which government is hired to protect society’s bank from being robbed. Implicit in that contract is the principle that government may not rob the bank itself: may not violate individual rights, or act in its own self-interest, or serve those who have only exercised raw political power to obtain what they want. This is what Justice Samuel Chase meant when he wrote in 1798, in one of the classic statements of substantive due process, that “[t]here are certain vital principles in our free Republican governments, which will determine and over-rule an apparent and flagrant abuse of legislative power,” such as if the legislature attempted to violate the rights “for the protection whereof the government was established. An ACT of the Legislature (for I cannot call it a law) contrary to the great first principles of the social compact; cannot be considered a rightful exercise of legislative authority.” Or, as the Supreme Court explained almost a century later, “Arbitrary power…is not law, whether manifested as the decree of a personal monarch or of an impersonal multitude. [Enforcing] the limitations imposed by our constitutional law upon the action of the governments, both State and national…is the device of self-governing communities to protect the rights of individuals and minorities, as well against the power of numbers as against the violence of public agents transcending the limits of lawful authority, even when acting in the name and wielding the force of the government.”

Procedural due process is a subset of substantive due process

Critics typically differentiate “substantive” from “procedural” due process, dismissing the former as philosophical hot air, and contending that the Constitution guarantees only procedural safeguards. But if I’m right that “due process of law” is best seen as a pledge against arbitrary or unauthorized government action, then these procedural aspects fit into a larger picture: they are a subset of the overall promise of lawful treatment. The modern, process-only interpretation of the Due Process Clause, by contrast, is unable to connect the procedural requirements with any deeper principle. Why accord people any procedural formalities unless we are committed to what Justice Thomas calls a “substantive guarantee against ‘unfairness’”? Without a substantive guarantee, a coin toss would suffice as a trial.

Indeed, unless it has some substantive fairness, a “trial” wouldn’t even be a trial; it would be something else. A lynch mob is generally seen as the opposite of procedural due process, but lynch mobs do abide by some rituals—often holding sham trials before imposing their predetermined verdict. The problem with lynch mobs isn’t the lack of procedure, but the fact that their “procedures” are not fair, general, or reasonable, and are thus substantively arbitrary. By contrast, a hearing that gives real effect to these substantive guarantees is still valid even if it fails to follow some typical procedural formality. Courts call this the harmless error rule, which “focus[es] on the underlying fairness of the trial rather than on the virtually inevitable presence of immaterial error.” Procedural formalities are not alone sufficient to qualify a government act as law.

To put this point differently, legal procedures are themselves composed of substantive steps or rules. What from a distance appear to be procedural guarantees are, on close examination, actually clusters of substantive guarantees. One might think of a trial as a procedural right, but a trial is composed of certain substantive rights—the right to cross-examine witnesses, the right to be represented by an attorney, the right not to be compelled to testify against oneself. What’s more, the trial as a whole is one substantive element in the broader right to not be treated arbitrarily.

In short, it’s precisely because the Due Process Clause is a “repository of substantive guarantees against ‘unfairness,’” that we have interpreted it to require fair trials, among other things. If it didn’t guarantee against unfairness, then whatever protections it did provide would hardly be the protection of law. Citizens are entitled to procedures only because they are entitled to lawful treatment.

Two examples

I’ll close with two examples of substantive due process cases—neither of which use that term— Loan Association v. Topeka in 1876 and Lawrence v. Texas in 2003. In both cases, the Court had to decide whether the state was depriving people of property or liberty in a lawful way—that is, in the service of a general principle of public good—or whether it was depriving them of their rights in arbitrarily, without a good, public reason.

In Loan Association, the Court invalidated an attempt to take money from one group and give it to another, more politically influential group, for their own private benefit. This attempt to take with “one hand” the property of citizens and “with the other to bestow it upon favored individuals to aid private enterprises” was “none the less a robbery because it is done under the forms of law … . This is not legislation. It is a decree under legislative forms.” And if it is not legislation—if it is an arbitrary, unprincipled, self-serving exploitation of political power—then it takes property without due process of law. Likewise, in Lawrence, the state proscribed private, consensual, homosexual conduct, not to protect the general public from harm, but simply to impose a burden on a disfavored minority. The law at issue did not involve public conduct, or prostitution, or government benefits; it did not protect minors, or “persons who might be injured or coerced.” Instead, the law was what the late Prof. Peter Cicchino called “the legal enforcement of private bias, casting lawmaking as a kind of Nietzschean struggle of will, with various moral interest groups trying to gain legal enforcement of their beliefs without having to give reasons for those beliefs other than saying, ‘we believe it.’” Such an ipse dixit infringement on liberty—prohibiting private, consensual sexual activity just because the legislature wanted to—could not be characterized as due process of law.

Substantive due process makes sense: some things government does—arbitrary, irrational, self-serving exercises of political power—lack the elements that make a legislative pronouncement a “law.” These fail to fulfill the Constitution’s promise that the government will accord us due process of law. Deciding whether a challenged government act is or is not a law in this sense is fraught with normative considerations, of course, so it’s unsurprising that it would be the target of heated criticism. So, too, judges sometimes reach wrong decisions, here as in other things. But these critiques are equally true of principles like separation of powers, which also require judges to engage in complicated analyses and refer to normative and practical considerations outside the four corners of the Constitution. The task before lawyers and judges is to engage in such analyses, and lawyers are badly served by the contemporary fashion of ridiculing substantive due process, or dismissing normative deliberations as somehow not in our job description.

Timothy Sandefur tells us that the Due Process Clause is not just a promise of fair procedure, but also a substantive guarantee that “government may not limit our freedom without good reason.” In fact, he tells us that we need only understand the concept of “law” to embrace “substantive due process.” Not so fast, Mr. Sandefur.

Text

The Fifth and Fourteenth Amendments prohibit deprivations of life, liberty, or property “without due process of law.” The classic objection to the view that the Due Process Clause offers something beyond a guarantee of fair procedures is textual. To use John Hart Ely’s formulation, “ ‘substantive due process’ is a contradiction in terms – sort of like ‘green, pastel, redness.’” Sandefur tells us that Professor Ely and his fellow travelers fail to understand what the last word of the Due Process Clause means. He writes that “law” is not merely “a formal criterion”; it is also “the opposite of arbitrariness,” and requires “the use of governmental power in the service of a rational, general public principle.”

This is not the way we use the term “law” in common parlance. We ordinarily regard as “law” whatever rules have been enacted consistent with the procedural requirements for lawmaking, whether we consider them justified or not. We speak of “unjust laws,” “unfair laws,” or “arbitrary laws” without thinking we utter oxymorons. Our Constitution also uses the term “law” in this sense. Article I tells us that if a bill is passed by both houses of Congress and signed by the president, or if the president’s veto is overridden, “it shall become Law.” The only requisites for the existence of “law,” in Constitution’s view, are procedural.

To be sure, the Constitution places substantive limitations on legislative power; thus, as Sandefur notes, the First Amendment provides that “Congress shall make no law … abridging the freedom of speech, or of the press.” Even so, the First Amendment does not seem to regard a “law … abridging freedom of speech” as an oxymoron; the First Amendment treats even an enactment abridging the freedom of speech is as a “law,” albeit one beyond congressional power. In just this vein, Chief Justice Marshall famously wrote in Marbury v. Madison that “a law repugnant to the constitution is void”; but he did not write that “a law repugnant to the constitution” is not even a law. In any event, even if laws that exceed the enumerated limitations on legislative power are not properly regarded as “law,” this does not prove that the Due Process Clause is properly understood as among those limitations. The First Amendment expressly limits the scope of legislative power; the Due Process Clause does not. The text of the Due Process Clause is no better than ambiguous when it comes to “substantive due process.”

But put all this aside and assume that Sandefur is correct that a “law is the opposite of arbitrariness,” and requires “the use of government power in the service of a rational, general public principle.” Even so, this claim does not come close to justifying the doctrine of substantive due process that Sandefur seeks to defend.

Consider one of the cases that Sandefur offers as an exemplar of substantive due process—the Supreme Court’s decision in Lawrence v. Texas invaliding a statutory prohibition on consensual, same-sex sodomy. Whatever one thinks of the statute at issue in Lawrence, it seems to be anything but “arbitrary”; indeed, it seems quite plainly to use “governmental power in the service of a rational, general public principle.” Texas’s prohibition on sodomy was not enacted on a whim; it reflected a longstanding view that the law should encourage procreative relationships by, among other things, proscribing homosexual activity. As it happens, I have considerable sympathy for the outcome in Lawrence; but I do not think that it can be justified merely by branding the Texas statute “arbitrary,” or unrelated to any “rational, general principle.” Using law to encourage favored relationships and proscribe disfavored ones strikes me as a “general public principle,” as is the desire to promote procreative relationships and activities. It takes more than an embrace of Sandefur’s counterintuitive definition of “law” to justify Lawrence.

History

Sandefur claims more than unadorned text in support of his conception of substantive due process; he tells us that history is on his side. Sandefur argues that his view has deep historical roots, citing as examples Daniel Webster’s argument in the Dartmouth College case and Justice Chase’s opinion in Calder v. Bull.

More than usual caution is required whenever one encounters lawyers (or law professors) advancing historical arguments. Lawyers are, by training and temperament, advocates; they seize on evidence that seems to support their case and ignore or minimize anything else. It is not for nothing that we are often warned of the perils of “law office history.” So it is here.

As it happens, neither Webster in Dartmouth College nor Chase in Calder v. Bull invoked the Due Process Clause to support their claim of inherent limits on the legislative power. Moreover, one finds nowhere in the legislative history of the Fifth and Fourteenth Amendments any mention of the conception of due process advanced by Sandefur.

To be sure, there is some historical support for Sandefur’s view. Lord Coke, for example, seems to have regarded the English antecedents of our Due Process Clause as a limitation of legislative power. Coke’s view, however, was unorthodox. By the time of our Constitution’s framing, the generally held view was that the legislative power was supreme, and not subject to implied and judicially enforceable limitations in the name of “due process of law” or otherwise. This, for example, was the position taken in William Blackstone’s treatise, the most influential of its kind in the framing era. Similarly, the most influential early American treatise writers—such as Justice Joseph Story and Chancellor James Kent—described due process in purely procedural terms.

By the time of the Fourteenth Amendment, there were a few cases and commentaries that described due process in more substantive terms. Most notable was the Supreme Court’s 1856 decision in Dred Scott v. Sandford, holding that federal legislation prohibiting slavery in federal territories deprived slaveholders of property without due process of law. Yet, in the Court’s first opinion to treat at any length with the Fourteenth Amendment’s Due Process Clause, the 1872 decision in The Slaughter-House Cases upholding a state legislative grant of monopoly rights to a private business, only Justice Bradley, in lone dissent, expressed anything like Sandefur’s view of substantive due process. To be sure, later in the nineteenth century, the notion that due process placed substantive limitations on governmental regulatory power came into vogue, but it is far from clear that this development rested on a deeply historically rooted conception of due process rather than the ideological predilections of the era’s judges.

My point is not that history argues against Sandefur’s conception of substantive due process; history on this point, as is often the case, is muddled. Instead, my point is that Sandefur cannot cherry-pick the evidence and then claim historical support for his conception of substantive due process. History is no better than a draw.

Prudence

If we are to embrace substantive due process, perhaps a conception of the judiciary’s role in ascertaining the limits of “law” should include a measure of prudence and intellectual humility. The fact that the Supreme Court’s first embrace of substantive due process came in Dred Scott should remind us of the perils in an understanding of due process that grants the judiciary a chancellor’s foot veto over everything that it regards as limiting freedom “without good reason.” The exercise of unaccountable power is always fraught with potential for abuse. Under the Constitution, legislators are not tyrants; they are always accountable at the next election, imperfect though that remedy may sometimes be. A life-tenured judiciary, however, is accountable to no one.

I have as high a regard for lawyers as one is likely to find; I even devote my professional energy to producing more of them. Yet I tremble at the power that Sandefur would have lawyers exercise when they ascend to the bench. The question of whether a garden-variety air-quality regulation amounts to an unwarranted interference with property rights or a sensible effort to force property owners to internalize the true costs of their activities rather than shifting them to those who live downwind—who may pay those costs with their lives—can contain enormous complexities. There is little reason to expect judges to have the expertise to resolve them with anything approaching an acceptable risk of error. Even worse, the judiciary is particularly poor rectifying its mistakes. Legislators who make bad policy usually hear about their mistakes soon enough, but our system of adjudication does not require the Supreme Court to monitor the consequences of its decision, or to defend them at the next election. The precedential power of judicial decisions, moreover, is a potent inhibition on the ability of the judiciary to recognize and correct mistakes.

Deciding whether a law is supported by “good reason” is the essence of policymaking. Our Constitution guarantees a republican form of government, and in a republic, policy is made by those who are politically accountable for their decisions. Sandefur’s conception of due process of law, however, creates a judicial platonic guardianship that must approve every policy decision. To my eye, this looks more like the due process of the ayatollahs than the due process of our Constitution. Leaving assessments of the wisdom of the “law” to the next election is surely one aspect of “due process of law” in a republic. This what the Court means when it writes: “The Constitution presumes that, absent some reason to infer antipathy, even improvident decisions will eventually be rectified by the democratic process and that judicial intervention is generally unwarranted no matter how unwisely we may think a political branch has acted (Cleburne, TX v. Cleburne Living Center).”

This is not to say that I regard Sandefur’s conception of due process as entirely wrongheaded. As Americans, we are deeply attached to a conception of individual rights; we do not think that every decision should be made by majority vote. There is something to the view that certain matters are not properly left to majoritarian resolution; reserving these for the realm of individual conscience may well be an aspect of “due process of law.” Drawing the line between what is properly the business of the majority and what must be left to the conscience of the individual, however, is no easy matter. This is the difficult task of substantive due process. We should face up to those difficulties with prudence and humility. My conception of “due process” includes generous dollops of these virtues. Sandefur’s version, it seems, not so much.

Timothy Sandefur argues that substantive due process makes sense. On this point, I fully agree. He further contends that “due process of law” in both its procedural and substantive dimensions can be explained by reference to the simple principle that government may not act “arbitrarily.” On this point, I am less convinced. One considerable difficulty with Mr. Sandefur’s argument is that he views “due process of law” as a unitary concept without giving sufficient attention to the fact that the Constitution contains two separate Due Process Clauses—the Fifth Amendment Due Process Clause, adopted in 1791, and the Fourteenth Amendment Due Process Clause, adopted in 1868. As I have previously argued, the generations that framed and ratified these two Due Process Clauses almost certainly did not share a common understanding of what “due process of law” required. In order to understand the constitutional guarantee of “due process of law,” it is therefore important to view both provisions in their full historical context, with particular focus on the public understandings of each provision at the time of its respective enactment.

The English Origins of “Due Process of Law”

“Due process” originated in 1215 with the English Magna Carta, an important provision of which was that no freeman would be deprived of certain rights except “by the judgment of his peers and by the law of the land.” This guarantee was later codified by Parliament in a series of statutes, one of which replaced “law of the land” with “due process of law.” Sir Edward Coke and other influential writers on English law declared the two phrases synonymous and this view was widely accepted by American courts and commentators during the late eighteenth and early nineteenth centuries.

Consistent with its language, the law of the land provision was understood to require that deprivations of individual rights could only be accomplished in accordance with the law. As applied to the King, this required that deprivations be supported by some external source of legal authority, such as an act of Parliament, rather than by the King’s own unilateral command. As applied to courts, the provision required application of the appropriate underlying substantive law as well as procedures consistent with the traditional practices of common-law courts. Importantly, neither the “law of the land” nor the closely related “due process” concept were understood to restrain Parliament itself, which had ultimate authority to say what the “law of the land” would be. As Coke explained in his influential treatise on English law, “the power and jurisdiction of the parliament, for making of laws” was “so transcendent and absolute” that it could not “be confined … within any bounds.”

The English colonists who immigrated to America carried with them their legal traditions, including the Magna Carta law of the land provision which was incorporated in some form in declarations of rights adopted by the majority of early American colonies. A few colonies adopted separate “due process of law” or “due course of law” provisions. Each of these latter provisions was directed specifically to the procedures by which an accused could be brought to “answer” before a court of law.

“Due Process of Law” in the Fifth Amendment

It is against this background that the Fifth Amendment’s Due Process Clause must be understood. The Fifth Amendment itself attracted little public debate or commentary and the Due Process Clause received virtually no attention in the framing and ratification debates. This relative inattention is not particularly surprising. The large majority of existing states in 1791 had incorporated a law of the land equivalent in their own constitutions. And in view of the centuries of experience with the concept under English and colonial law, Americans of the Founding generation likely saw little worth discussing.

Not much is known about why the Fifth Amendment’s framers chose “due process of law” rather than the more familiar “law of the land” formulation. What is known is that Coke had equated “due process of law” with “presentment and indictment” and nearly all early commentators on the Constitution assumed the Fifth Amendment Due Process Clause carried a similar meaning. In view of this background, the most plausible interpretation of the Fifth Amendment Due Process Clause, when read in its original context, is as a requirement of fair and impartial judicial procedures consistent with those used at common law. If the provision was understood to restrain Congress at all, it was likely understood solely as a limit on Congress’s ability to authorize or require deviations from such traditional judicial procedures.

The Nineteenth-Century Transformation of “Due Process of Law”

Americans have always been innovators. American lawyers in particular have long been adept at developing creative arguments to further their clients’ interests. In the early decades after the Fifth Amendment’s adoption, a handful of American lawyers and judges began experimenting with new readings of state constitutional law of the land provisions that would restrain legislative, as well as executive and judicial, power.

The first courts in which such arguments gained traction developed two principal accounts of how law of the land provisions might restrain the legislature. The first approach focused on “law of the land” as a protection of traditional judicial procedures. Because legislatures were not courts and thus could not provide judicial procedures themselves, these courts reasoned that law of the land provisions should be read to prohibit legislatures from directly depriving people of property or other “vested” rights or interests through retrospective legislation. Instead, such “vested rights” could only be taken away by courts and then only as punishment for violating some preexisting legal duty.

The second approach focused on the word “law” and reasoned that legislative enactments that were not sufficiently law-like should not be considered the “law of the land.” In particular, because legislation was paradigmatically general in nature, this reading assumed that special legislation that benefited or burdened particular groups was not the “law of the land” and thus exceeded the legislature’s authority.

Each of these innovative readings drew upon broadly shared intuitions regarding the appropriate subjects of legislation. The innovative move reflected in these early cases was to connect such abstract principles to the specific language of the law of the land provision. The novelty of such interpretations is reflected in the fact that numerous courts at around the same time struggled to reach similar results without ever suggesting that “law of the land” or “due process of law” had anything to do with the issue. For example, Justice Samuel Chase’s famous opinion in Calder v. Bull, which Mr. Sandefur characterizes as “one of the classic statements of substantive due process,” made no mention of the Fifth Amendment Due Process Clause or of any state-law equivalent. Instead, Chase simply assumed that laws transferring property from one person to another contradicted unwritten “vital principles” and were therefore void—a view famously contested by Chase’s fellow Justice, James Iredell.

In the mid-nineteenth century, as more and more courts came to embrace Justice Iredell’s skepticism of unwritten legal principles, state courts increasingly turned to “law of the land” and “due process” provisions as a convenient textual source for reining in perceived legislative abuses involving “special” legislation or interference with “vested rights.” By the time of the Fourteenth Amendment’s adoption in 1868, courts in at least 20 of the 37 then-existing states had embraced one or both of these readings in interpreting their own state constitutions. The U.S. Supreme Court also experimented with the “vested rights” reading, most notoriously in Chief Justice Roger Taney’s declaration in Dred Scott v. Sandford that a law prohibiting slaveholders from retaining their slaves in federal territories would deprive them of property without “due process of law.”

In view of this background, it is probable that the public understanding of “due process of law” at the time of the Fourteenth Amendment’s adoption was broad enough to encompass both the “vested rights” and “general law” interpretations. Importantly, both of these interpretations went to the validity of the underlying substantive law rather than the procedures through which courts were to apply that law. There is thus a strong argument that, despite the nearly identical phrasing of the Fifth and Fourteenth Amendment Due Process Clauses, only the latter provision was originally understood to embrace a recognizable form of “substantive due process.”

The Due Process Clauses and the “Non-Arbitrariness” Principle

This brief survey of the history of “due process of law” allows us to better assess Mr. Sandefur’s claim that the concept may be reduced to a simple directive that governments not act “arbitrarily.” As a matter of the Fifth Amendment’s original meaning, this claim seems implausible. Of course, requiring the King and the courts to act in accordance with the “law of the land” and provide “due process of law” served as a useful check on certain forms of arbitrary power. But the traditional understanding of these concepts did nothing to restrain arbitrary action by the legislature. There is no evidence that a broader understanding was widely embraced by Americans of the Founding era or that the Fifth Amendment’s Due Process Clause was generally understood to limit Congress’s authority in prescribing substantive rules of law.

When it comes to the Fourteenth Amendment, things are somewhat less clear. Certainly, many American courts by 1868 had embraced the view that “due process” limited the legislature’s authority to prescribe substantive as well as procedural rules. But for the most part, these courts did not view due process as embodying an abstract “non-arbitrariness” principle. Rather, the concept was generally applied to prohibit two relatively specific forms of legislation—laws that directly deprived individuals of “vested rights” and laws that were unduly “special” or “partial” in nature. The important step of connecting “due process of law” with an abstract prohibition of “unnecessary and arbitrary” interference with individual rights had yet to be taken.

In sum, while I agree with Mr. Sandefur that at least some forms of “substantive due process” can be considered a legitimate part of the Constitution, I do not believe that the pre-enactment history of either Due Process Clause will support a reading of “due process of law” that is as broad as he suggests.

Tim Sandefur’s defense of the much-maligned idea of substantive due process is elegantly simple. In the context of the Fifth Amendment’s Due Process Clause (we will get to the Fourteenth Amendment in time), the argument runs roughly as follows:

1. At a minimum, the Fifth Amendment’s Due Process Clause embodies the principle of legality from Magna Carta, which declares that executive and judicial deprivations of life, liberty, or property must be authorized by valid sources of law;
2. Unconstitutional legislative acts (statutes) do not function as law because they are ultra vires;
3. Executive or judicial deprivations “authorized” only by unconstitutional statutes are therefore deprivations without “due process of law,” no matter what procedures are employed for those deprivations;
4. The Constitution contains both implicit and explicit limitations on congressional authority to authorize valid deprivations, including an implicit general ban on arbitrary legislative action; and so
5. Attempts by Congress to authorize substantively arbitrary deprivations of life, liberty, or property are attempts to authorize deprivations without due process of law, regardless of the procedures employed for those deprivations.

Every step in this argument is correct—sometimes even more correct than Sandefur lets on. The general ban on arbitrary federal legislative action, for example, which is likely to strike many readers as the most problematic aspect of his argument, is actually well grounded in the fiduciary obligations imposed on Congress by its status as a public agent who received delegated power.[1] That obligation is textually embodied in the requirement that laws for the implementation of federal power be “necessary and proper” for executing those powers, with “proper” serving as a standard term to describe such fiduciary obligations.[2] One of the most basic obligations of a fiduciary is to behave non-arbitrarily, so in principle Sandefur’s case for what he calls “substantive due process” against the federal government seems quite sound, at the very least in any instance in which Congress is relying on the Necessary and Proper Clause (which is most of the time). Nonetheless, I think that Sandefur is wrong to ascribe these consequences to substantive due process, or to the Due Process Clause at all.

Imagine any federal law that Sandefur thinks would fail his “substantive due process” test, such as (I presume) a hypothetical federal statute requiring interstate railroads to maintain racially segregated cars. According to Sandefur, would such a law have been constitutional if enacted on December 14, 1791—before the Bill of Rights was ratified?

I will venture to guess that his answer would be “no”—as well it should be. Sandefur packs his analysis of constitutionally legitimate federal action into the very concept of law, which gives it general application from the moment of ratification. Congress can only act through passing laws.[3] The President has the “executive Power,” which just is (at least in most contexts) the power to execute laws.[4] Federal courts have the “judicial Power,” which just is the power to decide cases in accordance with governing law. If the very concept of law is as value-laden as Sandefur would have it, then the original, unamended Constitution already forbade arbitrary federal action from June 21, 1788 onwards.

On December 15, 1791, when the Bill of Rights took effect, the due process clause codified and clarified this basic requirement of (procedural and substantive) legality but did not create it. Either no laws or, at most, very few laws or potential laws that were constitutional on December 14, 1791 suddenly became unconstitutional because of the Fifth Amendment. As with the vast majority of the provisions of the Bill of Rights, the Fifth Amendment’s due process clause is essentially a truism—though no less important or significant for having that status.[5]

Can Sandefur therefore declare victory, because an unconstitutional statute by any other name (even if that name is not “violative of substantive due process”) would still be unconstitutional? In part, yes; it does not really matter for Sandefur’s purposes whether his proposed limitations on federal power come from the phrase “of law” in the Fifth Amendment or from the concept of law as embodied in the original Constitution’s grants of power to federal agents. If he is right about the general ban on arbitrary federal action—and he is more right than wrong about that[6]—he is right without needing to make reference to the Due Process Clause at all. Sandefur is pretty much guaranteed half a loaf—or, to pick another metaphor, he is going to be able to get rid of unconstitutional federal bathwater.

The problem is with getting the other half of the loaf (or with keeping the baby). The Constitution, by empowering the various institutions of the federal government, functions as a kind of agency instrument with respect to those institutions. It is that agency relationship that imports a general requirement of non-arbitrariness into the power grants to federal actors. But the Constitution does not empower the states. It limits them in some fashions, but the states do not trace their basic legal authority to the federal Constitution. The Fifth Amendment Due Process Clause can readily be understood as a codification of one important aspect of the fiduciary limitations on federal agents. What can it mean, however, to import the same text into a provision that limits non-agent states?

This is a particularly important question, because almost every example of a violation of substantive due process raised by Sandefur involves state rather than federal action. Once one understands the Constitution as an agency instrument,[7] the argument is very straightforward; the case for arbitrariness review of (at least most) federal legislation does not apply to state actors.

Here, the case for Sandefur-style arbitrariness review must take another form. If the Fourteenth Amendment’s Due Process Clause simply reiterates the principle of legality and requires traditional procedures for deprivations, then there is no real federal constitutional case for arbitrariness (or other “substantive”) review of state laws, unless one can ground such a requirement in some other part of the Fourteenth Amendment, such as the Privileges or Immunities Clause. As described above, a principle of legality embodied in the federal Constitution has a substantive bite against the national government that it does not have against the state governments. Accordingly, the case for “substantive due process” against the states turns on the answers to three questions: (1) Did the Fourteenth Amendment Due Process Clause have a different public meaning when enacted in 1868 than did the Fifth Amendment Due Process Clause when enacted in 1791; (2) if so, did that different meaning include a general prohibition on state arbitrariness analogous to the fiduciary obligations imposed on federal actors by the Constitution, and (3) would a different public meaning in 1868 actually determine the meaning of the Fourteenth Amendment?

I am not a Fourteenth Amendment scholar, so I defer to the superior learning and wisdom of others, including some participants in this exchange, for answers to questions (1) and (2). I offer only one observation on those questions and an observation on question (3).

First, in baseball, a tie goes to the runner. In constitutional law, a tie goes to the state and goes against the federal government. That is, state laws enter constitutional analysis with a presumption of constitutionality, while federal laws enter with a presumption of unconstitutionality. That is precisely the opposite of the allocation of burdens of proof typically employed in modern law, but it is the correct allocation.[8] He who asserts must prove. The proponent of federal power must assert an affirmative grant of power to the relevant federal actor, while a proponent of state power has no such obligation; rather, the opponent of state power is necessarily asserting the affirmative existence of a federal constitutional limitation. Whatever the appropriate standard of proof may be for questions of constitutional meaning, the person asserting a limitation on state power must meet that standard.

Second, it is not obvious that the public meaning of an amendment at the time of its ratification determines its constitutional meaning. It might determine it, but it is not obvious. Article V of the Constitution declares that amendments are part of “this Constitution.” Does that make them subject as well to the interpretative conventions applicable to “this Constitution”—meaning the Constitution of 1788—including the interpretative convention that makes the hypothetical reasonable observer of 1788 the referent for meaning? If so, there is at least a plausible case that the original public meaning of 1788 determines the constitutional meaning even of amendments ratified in the distant future.[9] If that is correct, and if the 1788 meaning of the Fifth Amendment’s Due Process Clause was essentially “principle of legality plus fair procedures,” then the case for substantive due process at the state level will be very hard to make.

In any event, we should all congratulate Timothy Sandefur for writing such a though-provoking article—and for walking away with at least half a loaf. That is a lot more than a forthright proponent of substantive due process probably expects in this business.

Notes

[1] GARY LAWSON, GEOFFREY P. MILLER, ROBERT G. NATELSON & GUY I. SEIDMAN, THE ORIGINS OF THE NECESSARY AND PROPER CLAUSE (2010); Robert G. Natelson, The Constitution and the Public Trust, 52 BUFF. L. REV. 1077 (2004).

[2] Gary Lawson & Guy I. Seidman, “Necessity, Propriety, and Reasonableness,” in THE ORIGINS OF THE NECESSARY AND PROPER CLAUSE, supra note 1, at 120.

[3] To be sure, the Constitution defines as “laws” anything that is enacted pursuant to the form for legislation set out in Article I, section 7. There is thus a prima facie argument that any legal norm enacted pursuant to Article I, section 7 is law, which taken baldly would rule out Sandefur’s version of substantive due process. It would also, however, rule out any form of constitutional review, including review pursuant to explicit and express constitutional limitations. There is an intellectually respectable case for such a position, but on balance there is a stronger case for a hierarchy among legal sources, with the Constitution standing hierarchically superior to Article I, section 7 enactments. See Gary Lawson, Rebel without a Clause: The Irrelevance of Article VI to Constitutional Supremacy, 110 MICH. L. REV. FIRST IMPRESSIONS 33 (2011).

[4] There are some applications of executive power that involve lawmaking rather than law execution, such as the governance of occupied territory during wartime. But they are exceptions to a general rule.

[5] On the largely declaratory but still important status of the Bill of Rights (and the few but significant exceptions to that status, such as federal governance of territories), see Gary Lawson, A Truism with Attitude: The Tenth Amendment in Constitutional Context, 83 NOTRE DAME L. REV. 469 (2008); Gary Lawson, The Bill of Rights As an Exclamation Point, 33 U. RICHMOND L. REV. 511 (1999).

[6] I am currently working on an article that explores the extent to which various federal activities are subject to the fiduciary requirement of equal treatment (or non-arbitrariness)—or in other words the proper scope of what today would be called “federal equal protection.” I am not yet convinced that such a requirement applies to all federal action. Federal governance of territories, for example, may not be subject to that norm, though it is, after December 15, 1791, subject to the Due Process Clause. But this is all for another day.

[7] Just to be clear: I did not so understand the Constitution until just a few years ago. I am perpetually indebted to Robert G. Natelson and Guy Seidman for opening my eyes and mind.

[8] Gary Lawson, Dead Document Walking, 92 B.U. L. REV. – (2012) (forthcoming); Gary Lawson, Legal Indeterminacy: Its Cause and Cure, 19 HARV. J.L. & PUB. POL’Y 411 (1996).

[9] Gary Lawson & Guy Seidman, Originalism As a Legal Enterprise, 23 CONST. COMMENTARY 47 (2006).

I’m honored and grateful to have the chance to discuss the ideas of my article “In Defense of Substantive Due Process” with scholars like Profs. Rosenthal, Williams, and Lawson, as well as David Bernstein and Michael Rappaport at their own blogs. I can’t hope to answer all of their points, but I want in what follows to address a few of the big ones. The first relates to unenumerated rights. Prof. Bernstein, echoing many critiques of substantive due process, basically asks where judges are supposed to stop: would it be proper for a judge to say at any point “I think this law intrudes on unenumerated rights and, were I a legislator, I’d therefore oppose it, but as a judge I will defer to elected representatives”?

My answer to this is no. The point of our constitutional system is not to establish a democracy, but to limit it by imposing, among other things, the requirement that democracy act within the boundaries of law. But this is not merely a procedural requirement; to be entitled to the name “law,” a government act must comply with certain substantive values—it must be general, it must be fair, it must be public-oriented, and so forth. The Constitution deems any “democratic” action that exceeds these boundaries to be illegal, and a judge is bound to regard it as invalid. I see no basis for deference to a law that exceeds these boundaries and violates individual rights. Deference is legitimate only within the realm of legitimate action: should the age of consent be 16 or 18? Should the police station be located on this corner or on that? But there can be no legitimate action depriving people of rights—whether they be enumerated or not—and therefore it is proper for the court to invalidate such acts.

Now, you’ll note that in my opening article, I don’t really talk about unenumerated rights that much. Instead, my focus is on the requirement that government operate lawfully. Unenumerated rights are a function of that requirement, as I was again reminded when recently watching the film Gideon’s Trumpet. In one scene, Henry Fonda (playing Gideon) is asked by a fellow prisoner if he really thinks the Constitution guarantees a right to a lawyer in a criminal proceeding. He answers in unadorned words and unimpeachable logic:

“‘No state shall deprive any person of life, liberty, or property without due process of law’…. I asked them for a lawyer and they wouldn’t give me one. Means I was deprived of liberty without due process of law…. They can’t put you in jail without they give you a fair trial.”

The overlap between “procedure” and “substance” can be found in the fact that, if it’s not fundamentally fair, it isn’t really a trial at all. For a procedure to really be a trial means that it includes certain elements, and Gideon’s argument is that one of those elements is the right to a lawyer, because the contemporary trial process is so complicated that without one the whole procedure is so arbitrary that it is not really a trial at all. This is a perfectly valid argument, and one can disagree with it only by addressing the truth or falsehood of the premises—that is, by entering into the normative deliberations that the positivist seeks to avoid.

Again, I see no basis for deference on such matters. When the legislature acts within its legitimate boundaries, courts should defer, but not when it acts outside of those boundaries. This should not strike lawyers as anything new or shocking. It’s well established that courts are not entitled to refuse to exercise their jurisdiction if they have it. And it’s well established that courts do not defer when deciding their jurisdiction. So why should they defer when a law infringes on rights?

The whole notion of “deference,” I submit, is the child of a Progressive ideology that prioritizes “democracy” over liberty and individual rights. That ideology is contrary to the Constitution, which, after all, states in its very first sentence that “liberty” is a “blessing”—and limits democracy for the purpose of protecting that liberty.

Of course I don’t think judges should exceed their authority or abuse their power; just as I don’t think any official should exceed authority or abuse power. A judge who exceeds his legitimate authority is acting lawlessly, and his decisions cannot be regarded as true law, just as when a legislature exceeds its authority. But I think a robust judiciary, more willing to engage the question of legislative authority and to defend individual rights, would be healthier for American democracy than the often overly deferential judiciary that we now have. Our system was meant to balance the judiciary against the legislature, and for good reason. We do harm to that system when our judges defer to the legislature to exceed its constitutional boundaries. To do this, as Hamilton said, is to affirm that the deputy is greater than the principal, or that the legislature is superior to the people.

A response to Ryan Williams and Michael Rappaport.

Profs. Ryan Williams and Michael Rappaport express sympathy with my argument on a conceptual level, but doubt the framers had such things in mind when writing the Constitution. I think I provide strong historical evidence (as have others) that substantive due process was familiar to and respected by the founders, although they didn’t call it by that name. But it’s true that my focus is more conceptual than historical: I’m interested in what the Constitution’s language means, not in what any particular people at any particular time thought it meant.

Nevertheless, consider: the founding fathers grew up without a Constitution. Then they wrote a Constitution that lacked a Bill of Rights. It was only in 1791, when James Madison was 40 years old, that the Due Process of Law Clause became part of the Constitution. His own state had only adopted its due process clause (Virginia’s “law of the land” clause) in 1776, when he was 25. Before then, judicial protection of individual rights was done by English common law courts, as it had been done since at least Lord Coke’s day, largely under the law of the land clause of Magna Carta. In many famous decisions that were the commonplace learning of law students in the founding era, English judges applied the law of the land clause’s promise of lawful, non-arbitrary rule to protect individual rights—rights that were, of course, unenumerated. (And yes, this was done even before the 1689 English Bill of Rights.) The founders would have been very comfortable with the idea of judicial enforcement of unenumerated rights under the law of the land or due process clauses, because for much of their lives, that was all the law they knew.

Prof. Williams, of course, musters evidence in his recent Yale Law Journal article, to show that the theory of substantive due process flowered much more in the nineteenth than in the eighteenth century. There are many reasons for this. For one thing, American courts operated independently only for the last quarter of the eighteenth century, and legal recordkeeping was poor. So, too, westward expansion brought new legal issues to the forefront; rising literacy rates and the democratization of the legal profession made litigation more common. There were more states, and they were doing more things, in 1880 than in 1820. Moreover, there was a profoundly important national conversation, so to speak, going on in this era—not just over slavery, but over sovereignty, administration, tort law, the rights of women, and so forth. It’s not surprising that we find a richer source of legal precedent and a greater concentration of legal scholars on these questions during this period. Virtually all our tort law also comes from this period—that doesn’t mean that tort law somehow lacks an originalist pedigree. I wouldn’t call this “innovation,” as Prof. Williams does; I would call it exegesis, elaboration—in short, interpretation. But whatever the reason, the fact that substantive due process was enforced more in the nineteenth century than earlier doesn’t disprove the thesis that the words mean what I say they mean.

Prof. Rappaport is more severe. He thinks I’m just picking and choosing whatever I want in the historical materials. I think this is unjust. My argument is that (1) ultra vires government action is not “law” for purposes of the Due Process of Law requirement; (2) government acts are ultra vires if (among other things) they exceed either explicit or implicit limits on lawmaking; and (3) we can determine what the implicit limits on lawmaking are by inquiring into the principles of the American founding, which means referring to documents like the Declaration of Independence and the writings of the founders. It’s not my fault that they happen to have been classical liberals, nowadays called libertarians. But I think this should answer Prof. Rappaport’s question about method.

Of course, this isn’t the time or place to enter into a full debate over whether originalism is the way to go. But I’ll make one observation on that point. If my argument is correct, then it must follow of course that we are bound by it, even if the authors of the Constitution didn’t expect the words to be applied in this particular way. If the Due Process of Law Clause prohibits arbitrariness, and arbitrariness means self-serving, ipse dixit, irrational, non-general uses of government power, then it follows that that’s what the framers intended, since that’s what they said. That’s how we interpret just about any other written text; contracts, for example. We don’t consult the subjective intentions of the author—we consult their words, and we interpret those words. If the words mean X, then X is what the author intended, or at least, it’s what we’re legally bound to deem to have been their intent.

As Tara Smith has observed, the great pitfall of originalism is to get into a census of the subjective beliefs of the authors of the Constitution. Serious originalists don’t do this. Instead, they inquire into the principles—the concepts—for which the words stand, and follow those. Consider Kyllo v. United States or the recent GPS decision, Jones. To say that use of infrared detection equipment or GPS are not Fourth Amendment searches, because the founders didn’t know of these technologies, would, I think, be ludicrous. The better conclusion is what Justice Scalia said in Jones: “The concurrence begins by accusing us of applying ‘18th-century tort law.’ That is a distortion. What we apply is an 18th-century guarantee against unreasonable searches.” Cases like these are resolved by determining the meaning of the word “search”—in other words, that we seek out the principle, not the specific applications, when understanding a legal rule. In applying the Due Process Clause, we are applying an 18th century guarantee against arbitrary or lawless government action—we are in the business of interpreting and applying the principle, not of ascertaining the subjective understandings of an individual or a group. Or, again to pick and choose from Lord Coke, “though a man can tell the law, yet if he know not the reason thereof, he shall soone forget his superficial knowledge. But when he findeth the right reason of the law, and so bringeth it to his natural reason, that he comprehendeth it as his own, this will not only serve him for the understanding of that particular case, but of many others: for cognitio legis est copulata et complicata.” Another term for inquiring into the right reason of the law is “seeking very hard to find out how a reasonable person at the time would have understood constitutional clauses in context.” I believe this is just what I’m doing—just what Daniel Webster did—just what the Lord Coke did—when determining that government acts that are arbitrary, self-serving, ipse dixit, and so forth, are not “law” and are therefore barred by the law of the land and due process of law clauses.

Now, I’m persuaded that the history does show that the founders believed, and rightly, that this clause barred government from acting in an arbitrary, self-serving, fundamentally unfair, ipse dixit, irrational manner—that is, from acting lawlessly. But I admit I’m more concerned with the “rightly” part of that sentence than with the “they believed” part. And if that means I’m not really an originalist, I can live with that. I’ve been content in the past to call myself a kind of originalist, but in the end, the reason I adhere to the framers’ views is not because they were the framers, but because their views were right.

That brings to mind a question I’d like to ask Prof. Rappaport and other originalists: do you believe that the authors of a legal text can actually be mistaken as to its meaning? To be more precise, if we are bound to follow a text as understood by the authors of that text, it would seem to follow that the authors are the definitive sources with regard to its meaning. Thus if a statute is passed in that, say, imposes a tax on the importation of fish, by a legislature whose members think that dolphins are fish, and later on, scientists discover that dolphins are not, in fact, fish—does the statute apply to dolphins or not? Do you rely on something like Prof. Rubenfeld’s distinction between “Application Understandings” and “No Application Understandings”? Or do you distinguish between the interpretation of a written text on one hand—which is subject to the standard analytical tools—and the nature of an obligation to obey a written text on the other, which includes other considerations? I mean this as a serious question, because as I’ve said, I’m not a student of originalism per se, and it’s possible these questions have already been answered without my knowing it.

I want to address Prof. Lawson’s hypothetical segregation case, because I think it helps address some of the originalism critiques I commented on earlier. Lawson asks whether my argument proves too much: under my theory, would a racial segregation law have been constitutional if it had been enacted before the Bill of Rights? My answer, as Prof. Lawson predicts, is no, precisely because I “pack[] [my] analysis of constitutionally legitimate federal action into the very concept of law.” This means that the “Fifth Amendment’s due process clause is essentially a truism—though no less important or significant for having that status.”

I suspect some readers will find this more troubling than Prof. Lawson does, for two reasons. First, it seems implausible. How could the concept of lawfulness, to which the founders cheerfully subscribed, have already prohibited practices that we regard as horrifying but which they took for granted? And second, doesn’t such a reading render much of the rest of the Constitution redundant?

As I write in my HJLPP article, I’m not so bothered by the second question. Yes, the Due Process of Law Clause is often redundant—it’s an extremely old provision, intended to prohibit arbitrary government action of all sorts; subsequent generations have added more specific prohibitions—for example, state constitutional clauses forbidding special legislation—which prohibit what can also be characterized as arbitrary action, but which later generations thought it would be better to specify as particularly bad. But while doing so, they have rightly seen no reason to eliminate the older, more general protection of the Due Process Clause.

Due Process of Law is therefore kind of a palimpsest or catch-all provision that often overlaps with other protections. This should not disturb us. Keep in mind that the framers of the Constitution thought the entire Bill of Rights was unnecessary and redundant, and expected federal courts to strike down lawless government acts even without any Due Process Clause! So even if my interpretation of the Due Process of Law restriction does make it echo other constitutional protections, that doesn’t work against my reading—as Michael Kent Curtis puts it, lawyers say everything twice.

But it still seems weird to suggest that the Due Process of Law Clause could have forbidden something like Prof. Lawson’s hypothetical segregation law—or that the very concept of law itself could have forbidden it even before the Due Process Clause was added to the Constitution, right?

Actually, no. This theory was well known even to the founding generation. English courts, as I observed earlier, had half a millennium of experience applying Magna Carta’s law of the land clause to forbid state actions on the grounds that, when considered by moral reason, they were inconsistent with the concept of lawfulness. Prof. Lawson’s hypothetical even sounds a bit like Somerset v. Stewart. If my approach is correct—focusing not on the meaning that some particular group ascribed to a word, but rather on a conceptual analysis of what that word objectively means—then it’s entirely possible for even the authors of a legal provision to be unaware of that provision’s full meaning, or of the consequences of the words to which they subscribe.

This, too, should not trouble us; it happens all the time. Statutory interpretation is often about whether a provision that applies in general terms to all Xes should also apply to Ys which are similar to Xes but aren’t quite Xes. The intention of the authors is probative, but not dispositive, in this inquiry, since it’s just a characteristic of human knowledge that we are always groping toward the future, not always knowing the full implications of the things we say and believe. We pass a law, for instance, that applies to all “motor vehicles,” not realizing that its wording could encompass airplanes, and thus—to twist Prof. Lawson’s hypothetical—that it would also applies to the Space Shuttle, even though the Space Shuttle had never been thought of at the time. Legal concepts—all concepts—are open-ended; they include all specific instances within the defined class. It should therefore come as no surprise that constitutional language can legitimately apply in ways that authors did not expect—as in the Kyllo and Jones cases I mentioned earlier. That’s just how language works. So, too, an argument that (say) slavery was already unconstitutional prior to ratification of the Thirteenth Amendment makes a lot more sense than many modern, historicist or positivist scholars are willing to acknowledge. Again to quote Justice Scalia, what we’re enforcing is not a specific 18th century law, but an open-ended, 18th century guarantee against arbitrary government action of all sorts.

Of course the generality of a text means there’s good reason to reiterate certain specifics. So although we could say—and courts have said—that the Due Process of Law Clause itself forbids uncompensated takings of property, or takings not for public use, it’s probably wise to reassert those requirements in separate, more specific clauses like the Takings Clause. It would be perverse if this attempt to provide double security were used as a basis for robbing the Due Process Clause of its force. Again, remember that the entire Bill of Rights was itself seen as a redundancy, added, in Jefferson’s words, as an additional “brace.”

But what about “this Constitution”? Prof. Lawson points out that only “this Constitution” is the supreme law of the land, which suggests that the document should be read as its authors read it at the time. I’ve made a similar argument before, about the proper role of stare decisis. But I don’t think that this provision works against my argument here—in fact, it strengthens it. I’m trying to address the meaning of this Constitution by asking what the words mean—as opposed to the subjective understandings of the authors of the language. It’s like trying to understand the meaning and internal logic of Heart of Darkness as a work of art, as opposed to seeing it as an artifact of the life of Joseph Conrad. By abstracting from the Constitution’s language to understand its principles and philosophy, we are truer to this Constitution than if we limit our inquiry to what the authors thought they were saying. We thereby focus on this “artificial perfection of reason,” instead of “any man’s natural reason.” We thereby focus on this Constitution, instead of any particular person’s Constitution. This—or so I’ve always thought—is the sort of objective approach that originalism promised us.

Timothy Sandefur’s essay on whether his conception of substantive due process is consistent with the Constitution’s original meaning contains in my view the kernel of a defensible position about substantive due process. First, though, I address what I regard as two false starts toward a defense of substantive due process.

The first is contained in Sandefur’s opening essay, with its claim about the meaning of the term “law.” Sandefur claims that “ultra vires government action is not “law” for purposes of the Due Process of Law requirement,” but the historical evidence is hardly uniform on this point. As I pointed out in my response essay, this is not the sense in which either the Constitution itself or Marbury v. Madison use the term “law,” to use just two examples — both regard ultra vires enactments as “law,” even if they exceed congressional power under the Constitution. If, on the other hand, the Fifth Amendment were understood to prohibit ultra vires federal legislation, why did Chief Justice Marshall fail to invoke it in Marbury? Moreover, this view turns the Due Process Clause into surplusage since ultra vires actions also are declared as beyond the scope of federal power by the Tenth Amendment. On Sandefur’s view, the Fifth Amendment does the same work as the Tenth. It may be that lawyers occasionally draft with “belt and suspenders,” but surely any view that turns important parts of the Constitution into redundancies should viewed with some skepticism. In any event, even if Sandefur were correct that the Due Process Clause was understood to forbid ultra vires legislation, that does not establish that the Due Process Clause was itself a source of unenumerated limitations on legislative power beyond those stated elsewhere in the Constitution. As I indicated in my opening essay, there are plenty of framing-era sources that describe due process in solely procedural terms. Ryan Williams is correct that later in the nineteenth century, there is more evidence of a substantive understanding of due process, but even he refrains from claiming that the evidence is uniform on this point. There is plenty of evidence from the Fourteenth Amendment’s framing era that due process was understood in procedural and not substantive terms. The historical and textual evidence goes no further than to establish ambiguity.

The second false start is the effort to pack substantive due process into the asserted constitutional obligation against “arbitrary” government action. Sandefur embraced this view in his opening essay, and it is embraced as well by Gary Lawson, who recognizes a “general ban on arbitrary federal legislative action” that he claims “is actually well grounded in the fiduciary obligations imposed on Congress by its status as a public agent who received delegated power,” in particular, “in the requirement that laws for the implementation of federal power be ‘necessary and proper’ for executing those powers, with ‘proper’ serving as a standard term to describe such fiduciary obligations.” He concludes: “One of the most basic obligations of a fiduciary is to behave non-arbitrarily … .” Professor Lawson recognizes that his view creates surplusage — in his view, the Due Process Clause adds nothing already not required by the Necessary and Proper Clause (and, I could add, the Tenth Amendment as well). Indeed, he writes: “If the very concept of law is as value-laden as Sandefur would have it, then the original, unamended Constitution already forbade arbitrary federal action from June 21, 1788 onwards.” On this view, the Due Process Clause at best “codified and clarified this basic requirement of (procedural and substantive) legality but did not create it.” Again, I am more reluctant to read the Due Process Clause as a redundancy. But put all this aside. The concept of “arbitrary” governmental action will not bear all the weight that Lawson and Sandefur place on it. Consider Professor Lawson’s exemplar of substantive due process — “a hypothetical federal statute requiring interstate railroads to maintain racially segregated cars.” Whatever one wants to say about racial segregation, it was surely not “arbitrary.” A Jim-Crow-era legislature’s decision to require segregation by race as oppose to, say, hair color, was anything but arbitrary. It was part of a quite considered view of relative position of the races; one that can even be reconciled with the “fiduciary obligations” that Professor Lawson supposes are implicit in the concept of “law.” I am reminded of Charles Black’s reference to “the entirely sincere protestations of many southerners that segregation is ‘better’ for the Negroes, is not intended to hurt them. But I think a little probing would demonstrate that what is meant is that it is better for the Negroes to accept a position of inferiority, at least for the indefinite future.” Indeed, one can find this sort of “benevolent” view of a racial caste system in Jefferson’s Notes on the State of Virginia. One cannot condemn racial segregation merely by calling it “arbitrary,” at least if one is concerned with accuracy. Due process may prohibit systems of racial caste — at least I think it does — but not because these systems are “arbitrary.”

A more promising route to substantive due process begins with Professor Lawson’s question whether it matters that the meaning of due process may have changed between the ratification of the Fifth Amendment in 1791 and the Fourteenth Amendment’s ratification in 1868. He asks whether constitutional amendments are “subject as well to the interpretative conventions applicable to ‘this Constitution’—meaning the Constitution of 1788—including the interpretative convention that makes the hypothetical reasonable observer of 1788 the referent for meaning? If so, there is at least a plausible case that the original public meaning of 1788 determines the constitutional meaning even of amendments ratified in the distant future.”

Yet by 1868, in a line of cases stretching from at least McCulloch v. Maryland to Dred Scott, the Supreme Court had made clear the expectations of Professor Lawson’s “hypothetical reasonable observer of 1788″ were not the sole or even the most important interpretive conventions to be employed in constitutional interpretation. A “hypothetical reasonable observer” in 1868, in other words, would not think that constitutional amendments would be interpreted exclusively based on the original understanding; McCulloch would have put him on notice that “it is a constitution we are expounding,” one that was “intended to endure for ages to come, and consequently, to be adapted to the various crises of human affairs,” and which did not “deprive the legislature of the capacity to avail itself of experience, to exercise its reason, and to accommodate its legislation to circumstances.”

It had become clear by 1868, if not before then, that the Constitution was not chained to the original expectations as to what powers the legislature could exercise. This recognition, in turn, suggests a more durable basis for a doctrine of substantive due process than simply labeling everything one finds distasteful or wrongheaded as “arbitrary.” The eighteenth-century understanding of due process may have been primarily, if not exclusively, procedural, but it had evolved in a legal system where the legislature exercised unfettered power over substantive law. The new Constitution’s Supremacy Clause, however, subordinated legislative power to the Constitution itself. As I suggested in my opening essay, in a republic, “due process,” when it comes to the wisdom of government policy, is ordinarily provided by the political process, but it is likely the case that we do not regard every issue as properly resolved by majoritarian institutions.

As Ryan Williams observes, most of the supporting historical evidence for the concept of substantive due process involved legislation benefiting discrete classes or impairing what were regarded as vested rights. Accordingly, he, echoed by Professor Michael Rappaport, doubts that the original meaning of the Due Process Clause can support a general doctrine of substantive due process. But due process was sometimes articulated in an even broader way – Dred Scott invalidated any federal prohibition on slavery in the territories without characterizing the law it struck down as either class legislation or an impairment of vested rights (it was neither). By 1868, due process had come to be a malleable concept without precise contours. More important, Sandefur is quite correct to distinguish between original intended applications and original meaning. The Due Process Clause was drafted at a higher level of generality than simply a prohibition on class legislation or the impairment of vested rights.

When it comes to the Fourteenth Amendment’s Equal Protection Clause, for example, what binds us is the general principle embodied in the constitutional text — equal protection — and not the specific expectation that the vast majority of framers surely had that the Fourteenth Amendment would not outlaw segregation. After all, the Fourteenth Amendment did not codify a special rule about segregation, but rather a more general principle about equality. Similarly, when it comes to due process, what we are bound by is not the framers’ views on vested rights or special legislation — which they did not bother to codify in the Constitution’s text — but instead the more abstract concept that they did place in the text — due process of law. We must decide when the political process supplies all the process that is “due” when it comes to legislation affecting life, liberty, or property, and when it does not.

To be sure, our Constitution creates a republic in which policy is ordinarily made by the legislature. That is why the political process ordinarily supplies all the process that is “due” when it comes to what are ultimately policy disputes as to the wisdom or propriety of legislation. But our conception of “due process” does not mean that legislative process is the place to resolve every debate, or that majorities can impose just about any disability they wish on a minority. Americans have always worried about the tyranny of the majority, from the framing era to the present. Defining the limits of majoritarian power is the task of substantive due process. It is not easy, and it must be done with appropriate respect for our republican form of government, and with appropriate regard for the institutional limitations on the judiciary. It cannot be done, however, simply by slapping the label “arbitrary” on whatever one regards as bad policy. That is tyranny of a different sort.

I’m bewildered by Prof. Rosenthal’s references to “humility” and “prudence” when advocating for a legal position that would leave legislatures virtually unconstrained to violate the rights of individuals. I would have thought that someone advocating humility and restraint would want a strong judiciary to protect people against the brazen, often imprudent, passion-driven actions of legislatures. That, after all, was exactly the reasoning that led the founding fathers to create a strong, independent judiciary: they believed government should be humble and prudent, and should not see people as the raw materials of legislative “experimentation,” and the only preventative to this was an actively engaged judiciary.

In a democracy, writes Hamilton in Federalist 71, the people commonly intend the public good, but they do not “always reason right about the means of promoting it.” They

sometimes err; and the wonder is that they so seldom err as they do, beset, as they continually are, by the wiles of…sycophants… [and] the artifices of men who possess their confidence more than they deserve it…. When occasions present themselves, in which the interests of the people are at variance with their inclinations, it is the duty of the persons whom they have appointed to be the guardians of those interests, to withstand the temporary delusion, in order to give them time and opportunity for more cool and sedate reflection. Instances might be cited in which a conduct of this kind has saved the people from very fatal consequences of their own mistakes, and has procured lasting monuments of their gratitude to the men who had courage and magnanimity enough to serve them at the peril of their displeasure.

Humility and prudence mean respecting the rights of others and acting only in carefully considered, fully conscious, fair, non-arbitrary ways. Judicial enforcement of constitutional limits serves these ends. One who truly believes in humility and prudence should therefore welcome an engaged judiciary that will help ensure that legislatures act, not arbitrarily, but in humble and prudent ways. Observe that it’s in our era of judicial deference that government has most clearly abandoned these virtues, and imposed increasingly intrusive, ambitious, burdensome schemes—whether aimed at “social justice” or at a “morally upright society”—all built on violations of our liberty and erosions of our constitutional rights.

Take Kelo v. New London. If that case is anything, it’s a monument to judicial deference—and to government’s lack of humility and prudence when not constrained by meaningful judicial enforcement of constitutional limits. Judicial deference is the opposite of humility and prudence—it is the open door to legislative ambition and folly. (I’ll never forget the attorney for the City borrowing a famous cliché during his oral argument before the Court—the city, he said, must make the trains run on time. This, of course, was famously, and falsely, said of Mussolini.) And Kelo is hardly the only example.

As I’ve argued elsewhere, the dangers of deference far outweigh the dangers of an overly active judiciary. First, deference blinds courts to the dangers of rent-seeking and to what Cass Sunstein has called “naked preferences”: the self-interested use of government power by the politically influential. Wrongfully upholding laws is more dangerous than wrongfully striking them down, since the hardship caused by an error falls much more heavily on an oppressed individual than on a temporarily stymied legislature, and courts are poorly suited to determine what sorts of things are best relegated to the democratic process and what things are not.

In the end, the answer was given by Justice Sutherland, when he responded to Justice Brandeis’ famous dissent in New State Ice Co. v. Liebmann. Many people know of Brandeis’ dissent because he argued that states serve as “laboratories of democracy” that can “experiment” with new kinds of legislative schemes. Not many remember Sutherland’s answer: states have no legitimate power to “experiment” on human subjects against their will:

Unreasonable or arbitrary interference or restrictions [on liberty] cannot be saved…merely by calling them experimental…. [I]t would be strange and unwarranted doctrine to hold that [states] may [“experiment”] by enactments which transcend the limitations imposed upon them by the federal Constitution. The principle is imbedded in our constitutional system that there are certain essentials of liberty with which the state is not entitled to dispense in the interest of experiments.

In a majority-rules society like ours, legislative experiments are the primary threat to humility and prudence—and judicial enforcement of individual rights are the guarantors of those values.

Before I begin, I would like to say what a great honor it is to have the opportunity to comment on Mr. Sandefur’s thought-provoking arguments and to be included in such a distinguished group of commentators. I would also like to thank Mr. Sandefur for his most recent response, which I hope will provide an opportunity to clarify our differences regarding how we view the due process concept and our underlying methodological approaches.

As an initial matter, I must confess to being a bit uncertain regarding what precisely Mr. Sandefur has in mind when he says that he is “interested in what the Constitution’s language means” rather than “in what any particular people at any particular time thought it meant.” To the extent this statement is meant to convey merely that meaning is not necessarily reducible to the subjective intentions or understandings of a particular, identifiable speaker or group of speakers, I don’t necessarily disagree. But to the extent Mr. Sandefur means to suggest that there is some objective “meaning” of constitutional language that can be wholly divorced from subjective understandings of some group of speakers and listeners, then I think the view is fundamentally mistaken. Human language has meaning only in context and that context must have reference to the way in which some historically situated interpretive community uses and understands the particular language at issue.

To take an example suggested by Professor Lawrence Solum, the word “deer” in Middle English was used to refer generically to all animals of any kind rather than as a reference to the particular animal contemporary English speakers use the word “deer” to identify. Thus, in order to know what was being referred to in a very old English document that uses the term “deer,” it will be important to know when that document was created. I therefore think that some resort to history will often be unavoidable, if only for purposes of identifying the interpretive community we care about and determining how members of that community generally used and understood the language in question.

To the extent Mr. Sandefur is claiming that there was a widely shared public understanding among Americans in 1791 that understood either “due process of law” or “law of the land” to encompass an abstract protection against all forms of arbitrary government decisionmaking, I believe the evidence for such a proposition is lacking for the reasons explained in my original response. In short, the English understanding of the “law of the land” concept that Americans of the Founding generation inherited viewed this provision solely as a restriction on the executive and the courts, not the legislature. Though it was certainly possible to formulate readings of the “law of the land” concept that would allow it to be extended to legislatures (as subsequent experience demonstrated), such readings were relatively rare until the middle decades of the nineteenth century.

Mr. Sandefur is certainly correct that the incipient nature of American judicial systems and case reporting in the late eighteenth and early nineteenth centuries limit the information we have about how members of the public actually understood “due process of law” and “law of the land.” And such limitations in the historical record must certainly be kept in mind in considering the degree of confidence with which we can make claims about historical understandings. But these limitations cut both ways and raise just as many problems for those claiming “due process of law” and “law of the land” were widely understood as substantive restraints on legislative power. Much turns, therefore, on which side bears the burden of proof in such arguments and the standard of proof to be applied to such claims. In my view, the late eighteenth-century English understandings of “due process of law” and “law of the land” provide a useful baseline for judging claims about how Americans of the Founding generation would most likely have understood those concepts. Those seeking to establish that members of the Founding generation shared an understanding that diverged significantly from the more traditional English understanding should thus bear the burden of proof in establishing that proposition. I do not believe that the existing evidence of American understandings in the late eighteenth and very early nineteenth centuries can support such a claim.

Finally, in response to Mr. Sandefur’s intriguing final question, I do not think it is particularly controversial that authors of legal texts can sometimes make mistakes regarding meaning. This is why the scrivener’s error doctrine exists. More generally, I think most modern originalists would agree that the enactors’ expectations about how a particular text will be applied in a given factual context are not necessarily determinative of that text’s meaning. In an important article, Professors Mark Greenberg and Harry Litman illustrated this point with a hypothetical statute requiring all persons with “contagious diseases” to be quarantined before entering the United States. As Greenberg and Litman observe, if at the time that provision was adopted the public mistakenly believed that a noncommunicable disease such as psoriasis was contagious, that mistaken belief would not necessarily control how subsequent interpreters should apply the statute after the true medical facts were discovered. This is so because the term “contagious” had an objective meaning that was independent of the erroneous expectation regarding how the statute would apply to a particular disease.

By contrast, Mr. Sandefur’s example of a statute taxing the importation of “fish,” as applied to dolphins, provides an excellent example of the importance of context and history in determining the actual original linguistic meaning (as opposed to expected application) of legal language. If the hypothetical statute Mr. Sandefur has in mind were adopted against a societal background where the basic principles of modern evolutionary taxonomy are widely accepted, then the application of the statute to dolphins would probably involve a linguistic mistake, even if members of the enacting public were confused about the phylogenetic relationship between fish and dolphins. By contrast, if at the time the statute was enacted, most individuals accepted a pre-modern taxonomic system that focused primarily on outward morphology rather than genetic relationships, then it would probably be correct to classify dolphins as “fish” in accordance with the widely shared notion of what the term “fish” referred to. In other words, we need to know whether a reasonable reader of the time would have understood the term “fish” as referring to some underlying genetic or evolutionary relationship among similar organisms or to the outwardly observable characteristics (e.g., fins, tail, living in the water) that animals referred to by the term generally exhibit. There is no innate quality of “fishness” to which we can refer to answer this question that is independent of how individuals in a particular time and place actually use and understand words.

I contend that the word “law” refers to something that has certain characteristics. Something lacking those characteristics, even if it goes through the ordinary process of promulgation, is not a “law.” Consequently, to penalize a person for disregarding it would be to deprive that person of life, liberty, or property without due process of law. But Prof. Rosenthal sees “law” in purely formal terms: “The only requisites for the existence of ‘law,’ in Constitution’s view, are procedural,” so that anything whatsoever, so long as it is promulgated by both houses and signed by the President, is, ipso facto, law.

I admit that we have some linguistic difficulties here (nothing new to lawyers!). But notwithstanding our habit of calling all legislative actions, whether legitimate or not, by the word “law,” invalid enactments are not laws, and it is technically inaccurate to refer to an ultra vires legislative act as a “law,” except in what Lon Fuller called “the Pickwickian sense in which a void contract can be said to be one kind of contract.” And this isn’t just an abstraction: it’s part of the Constitution itself, as Alexander Hamilton observed. In Federalist 33 he explained that

acts of the [federal government] which are not pursuant to its constitutional powers…will [not] become the supreme law of the land. These will be merely acts of usurpation, and will deserve to be treated as such. Hence we perceive that the clause which declares the supremacy of the laws of the Union…only declares a truth, which flows immediately and necessarily from the institution of a federal government. It will not, I presume, have escaped observation, that it expressly confines this supremacy to laws made pursuant to the Constitution; which I mention merely as an instance of caution in the convention; since that limitation would have been to be understood, though it had not been expressed.

That’s about as clear as it gets, but Hamilton wasn’t alone. Justice Chase, in his justly famous Calder opinion, explained that a statute that exceeds constitutional boundaries (either explicit or implicit boundaries) is only “[a]n ACT of the Legislature (for I cannot call it a law).” Similar phrases can be found in the works of many others. This observation, that an invalid enactment really doesn’t deserve to be called a “law” at all, is one of the oldest in our legal tradition, tracing to Aristotle, Cicero, Aquinas, and others, who explained that law differs from arbitrary decrees precisely because law contains elements such like regularity and generality, and that an enactment or command that lacks such elements is not actually a law at all. More recently, Lon Fuller and H.L.A. Hart, whatever their other differences, agreed that the concept of law just means that not everything promulgated according to procedures is ipso facto a law. Ordinary people, I think, understand this—they, too, tend to stumble over terminology if you press them on whether an invalid law is really a law in the first place.

Of course, the point isn’t to count up all the legal thinkers who have observed the technical imprecision of using the word “law” for invalid legislative enactments. It’s to indicate that the fact that “in common parlance” we use the word “law” even to refer to invalid acts, is really trivial. This is just because our language lacks a handy term for such things, so we resort to “invalid law” and such locutions. It would have been remarkable indeed if the Constitution’s authors had suddenly devised some ingenious new terminology to resolve this longstanding problem. It is enough that they were conscious of it, addressed it, and wrote a Constitution with it in mind.

More to the point, it’s easy to show that Prof. Rosenthal is wrong in saying that anything that goes through the process of promulgation is ipso facto a “law.” Suppose Congress were to pass a statute containing a self-contradiction: one which simply declares, “This is not a law,” or a statute that simultaneously requires and forbids the same act. Would this be a law?

Chief Justice Vaughan in said no in 1677, and it seems obvious that to call such a thing a law would be to mutilate that word into semantic absurdity. A self-contradiction is nothing; it is null; it is no more a law than it is a pigeon or a sneeze, and a court is therefore justified in disregarding it, even if passed with full procedural formalities. Yet nothing in the Constitution expressly forbids self-contradictory statutes. This prohibition is implicit–embedded in the logic of law itself. As the framers were aware, Francis Bacon made the same point when explaining that a legislature cannot create an unrepealable law (perpetua lex est nullam legem). Nothing in the Constitution expressly forbids unrepealable statutes. Yet such things are prohibited by what Bacon called their “impertinency,” or by what Hamilton called “the nature and reason of the thing.” There are limits on legislative authority imposed by the logic of law itself. That’s why even Blackstone, who believed that government possessed “supreme, irresistible, absolute authority” still admitted it could not do something that is “naturally impossible.” To make an arbitrary, ipse dixit, unauthorized use of force into a “law” is naturally impossible.

If one concedes this much, then one must admit that mere promulgation is not sufficient to make something a “law.” Instead, the legal status of a promulgated rule must be determined at least in part by its content. That is, by its substance: we must ask what a law is, and determine whether something containing the substantive provisions in question meets that definition. That just is substantive due process of law.

Finally, assume Congress passes a bill that reads as follows, and that the President signs it: “SHweio%#gkl Gw24 Vas#$3a2g shall be punishable by death.”

1) Is this a law?

2) How do you know?

3) Assuming your answer to 1) is “no,” how would you distinguish your answer from the account of substantive due process that I’ve given?

My answer is obvious: it’s not a law because it’s unintelligible and therefore not a general application of the state’s coercive power in the service of a public purpose. It lacks the substantive criteria by which it would become a law. Applying it to my client would be fundamentally unfair—would violate certain manifestations of fairness such as the “void for vagueness” doctrine—and would therefore deprive my client of due process of law. But I’m interested in seeing answers from the other side of the table.

A few quick thoughts on the substantive due process debate thus far:

1. Much of the discussion to this point seems to be focusing on what can best be called matters of institutional politics: What sorts of issues are judges fit to decide? That is a conceptually separate question from the correct understanding of the meaning of the Constitution. For all I know, it might very well be unwise for judges to try to decide real-world cases in accordance with the Constitution, on the ground that they are likely to do it badly and cause great harm in the process, so that the “best” theory of constitutional adjudication may well relegate the actual Constitution to a secondary, or even minimal, role in the adjudicative process. That is certainly an apt description of modern practice. I have no professional scholarly opinion on that point, which is within the domain of political and moral theory rather than law and therefore beyond any expertise that I can plausibly (or implausibly) assert. I am interested only in figuring out what the Constitution actually means; what anybody does with that meaning once they have it is not my concern.
2. Timothy Sandefur’s response essays point out that one cannot really discuss the meaning of any part of the Constitution – to the extent that the meaning of the Constitution is really the issue (see above) –without first clarifying how that meaning is determined. That is a whole different conversation from the one that we were charged with having. I will say only that Sandefur’s attempt to decouple constitutional meaning from anyone’s intentions, like his more specific argument about substantive due process, is partly right and partly wrong. Communicative instruments such as the Constitution have meaning only in light of intentions – that is what makes them communicative instruments rather than the equivalent of goat entrails. The question is: Whose intentions? Real people’s intentions? In that case, you need to specify which real people count. Or imaginary people’s intentions? For reasons too complicated to explain in this dialogue, I think the answer is the latter: the Constitution’s meaning is determined by the hypothetical, legally constructed intentions of a hypothetical, legally constructed reasonable observer of 1788. Of course, as Professor Rosenthal aptly points out, there is a long tradition in the case law, and in American politics more broadly, that makes crystal clear that my hypothetical observer’s legally constructed views are “not the sole or even the most important interpretive conventions to be employed in constitutional interpretation.” But I am not trying to give a sociologically or historically accurate account of what various people have said (or are likely in the future to say) about the Constitution. I am just trying to figure out what the Constitution actually means. (And to answer Sandefur’s question: Yes, the historically real authors of a provision can be mistaken about its meaning in any circumstance in which meaning is determined by the intentions of someone else.)
3. Professor Rosenthal also correctly points out that a non-arbitrariness principle, even if applicable to most actions by the federal government, will not necessarily yield all of the results that Sandefur (or I) think are prescribed by the Constitution, such as the unconstitutionality of a federal statute forbidding segregated rail cars. That is quite correct. The obligations of a fiduciary include non-arbitrariness, which I emphasized because Sandefur emphasized it, but also include a range of other obligations. Most notably, eighteenth-century fiduciaries were under a presumptive obligation to treat all principals fairly, which roughly corresponds to what today we call “equal protection.” So yes, I do believe that something resembling “federal equal protection” is validly part of the federal Constitution – not because of some weird reverse incorporation of the Fourteenth Amendment, but because the principle was part of the original Constitution as of June 21, 1788.
4. I would still like to see Sandefur say more than he has thus far about the federal/state distinction. From a constitutional standpoint, does he think that state and local laws stand on the same footing as federal laws? For example, I assume that he believes (as do I) that the individual mandate to have health insurance in the Patient Protection and Affordable Care Act is unconstitutional. Does he also believe (as I do not) that the equivalent provision in Massachusetts is unconstitutional?

Timothy Sandefur believes that my view of due process “would leave legislatures virtually unconstrained to violate the rights of individuals.” This statement, of course, assumes its conclusion. It is true only if “the rights of individuals” include a right to judicial review to determine if legislation is what Sandefur regards as “arbitrary” (more on that later). In any event, what Sandefur overlooks, of course, is that legislators who must account for their actions at the next election are never “virtually unconstrained.”

Consider the example Sandefur deploys of “virtually unconstrained” legislative power — the City of New London’s decision to condemn property for redevelopment at issue in Kelo v. City of New London. One can disagree with the decision — most of us can find any number of legislative actions with which we disagree — without concluding that it was “virtually unconstrained.” The most obvious constraint on New London was that it was obligated to pay property owners “just compensation” within the meaning of the Fifth Amendment. All taxpayers in New London (and those to whom taxes are passed on, such as renters), in turn, had to be taxed to fund the payment of that just compensation. All property owners in New London (and, for that matter, all who resided in that property), understood as well that they were also subject to the power of eminent domain, and could be subject to forced relocation by virtue of future condemnations. Thus, if the redevelopment was improvident, the voters would have ample reason to hold the elected officials of New London accountable at the next election. In a republic, of course, this is ordinarily the “process” that is “due” when it comes to legislative decisions that some regard as improvident. And, in a republic, when it comes to debate over the wisdom or propriety of a policy judgment, the majority is ordinarily entitled to prevail, as long as the minority has a chance to convert more to its side at the next election.

There is surely more than a little irony in Sandefur’s reliance on Federalist 71 to support his view. After all, the Constitution that was defended by Hamilton, Madison and Jay in the Federalist Papers contained no due process clause. The authors of the Federalist were not arguing for a judicial veto as a means of preventing abuses of legislative power; they were arguing instead that the system of political checks and balances that the Constitution had created would be sufficient to constrain government power. To be sure, Madison was later persuaded to draft what became the Fifth Amendment’s Due Process Clause, but he never expressed anything resembling Sandefur’s view of substantive due process. Hamilton, for his part, described due process solely in procedural terms.

On Sandefur’s view, however, the next election is somehow not a meaningful constraint on legislative power — all legislative decisions must also be subject to a judicial veto. I am happy to agree that there are some occasions in which policy is not appropriately made by majoritarian institutions, but Sandefur seems to believe that policy never ultimately rests in majoritarian hands — it is always subject to judicial veto. This is not the “due process” of a republic. Sandefur claims that this veto only applies to “arbitrary” decisions, but he does not really mean to limit judicial review to “arbitrariness,” as I have demonstrated in my earlier essays. Sandefur seems to think that a great many decisions that are not “arbitrary,” at least as we use that term in common parlance, nevertheless violate the Due Process Clause. His invocation of Kelo provides yet another example. Whatever one thinks of New London’s decision to condemn property for redevelopment, it surely was not “arbitrary.” New London made a considered judgment about how to revitalize a neighborhood that it regarded as underdeveloped. What Sandefur appears to mean by “arbitrary” — especially in light of his professed aversion to any form of judicial deference to politically accountable branches of government — is that the Due Process Clause permits the courts to set aside any decision with which they disagree. One conception of tyranny is the existence of unaccountable power. If so, it seems that Sandefur’s view of due process creates a judicial tyranny.

The fact that something like substantive due process was embraced for the first time by the Supreme Court in Dred Scott v. Sandford should give us pause. After all, are we not told that those who forget the mistakes of the past are condemned to repeat them?

Scientists who study evolution are sometimes challenged to produce the “missing link.” Those who have the patience to answer point out that there is no such thing as the missing link, and that to seek one is to misunderstand the basic idea of evolution. Indeed, as the old chestnut goes, every time you do find a link, it only opens up two more “gaps”! Evolution isn’t a bunch of links—it’s a process, a natural phenomenon, and we can comprehend and illustrate it by reference to specific fossils, but it is not to be found in any specific fossil. It is, so to speak, the line that connects the dots, and you just can’t draw a line by accumulating dots. To see a line—to understand evolution—we must engage in a process of abstraction, so that we grasp a general principle, which we can then test against the facts we have.

It seems this same “missing link” problem is often present in originalism discourse. Any particular piece of evidence from the founding era—for example, James Madison’s speech when introducing the Bill of Rights in the First Congress—is only one little fossil, and in isolation it tells us only so much. But if we take it and other “fossils”—say, James Wilson’s remarks at the Philadelphia ratification convention, Hamilton’s Federalist 84, and the correspondence between Madison and Jefferson about a Bill of Rights—then we can draw a line: we can put together a conceptual-level picture, a theoretical construct, of the theory of rights that the Constitution’s authors had in mind. Lord Coke, in a passage I quoted earlier, called this process grasping the “artificial perfection of reason” that is the law. It allows us to better understand, say, the Ninth Amendment’s reference to “other rights,” and, indeed, to understand more broadly what Randy Barnett has recently called the “presumption of liberty.” Yet some critics attacked Barnett, as they do me, by saying that he was only making an abstract philosophical argument; these critics demanded something like a “missing link”—some document that will say “Here’s exactly what we meant by the Ninth Amendment.” Any time we do offer up a fossil of originalism to show how it fits into our approach, we’re told it only opens two new gaps.

I think Prof. Rosenthal does something very like this. For example, he asks why Chief Justice Marshall didn’t cite the Fifth Amendment in Marbury v. Madison. Marshall is, of course, notorious for the paucity of citations in his decisions—Fletcher, Dartmouth College, and many of his other great rulings cite virtually no authority for anything; instead, they proceed irresistibly through logical argument from premises to conclusion, and are not the less ironclad for that.

But anyone who reads section two of Marbury—where Marshall addresses whether the law affords a remedy for Marbury’s injury—will find that it’s based profoundly on the theory I’ve advanced here. Marshall says that, except in special cases, an injured person is entitled to a remedy when he has been harmed by a lawless, unauthorized, arbitrary government action. If Marshall had believed, as Prof. Rosenthal does, that whatever government chooses to do according to rules of promulgation is automatically law, then it’s hard to imagine how he could have concluded either that Mr. Marbury was wrongfully deprived of his commission or that the Judiciary Act was unconstitutional. Searching for the “missing link” would make us miss the crucial point about one of the most important of all constitutional precedents: the principle that the government is not above the law—that unauthorized government acts are not law—and that, if the Court had jurisdiction, this would have entitled Marbury to the writ he sought. To regard Marbury as not fitting the model I’ve presented because it doesn’t cite the Fifth Amendment seems to me like arguing that “Oh Captain! My Captain!” isn’t about Lincoln because Whitman doesn’t use his name, or that Fanny Hill isn’t pornographic because it contains no profanity. It is, in short, formalism: it is “mechanical jurisprudence,” and, I think, it misconceives the process of legal reasoning.

Obviously I take the founders and their writings very seriously, and I’ve tried hard to provide good fossil specimens from their work to buttress my interpretation of the Constitution’s promise that government will act in a lawful, non-arbitrary manner. But in the end, we cannot look for a “missing link”: we must instead use the evidence to comprehend the broader picture suggested to us by these documents—a picture that is written in abstract principles, applicable to all men and all times.

Prof. Rosenthal’s mention of the Dred Scott case reminds me that I should address this recurring myth: that substantive due process originated in Dred Scott. As near as I can tell, this idea originated in John Hart Ely’s Democracy and Distrust, and was taken up by Robert Bork in The Tempting of America. I can’t quite decide if it’s an ad hominem or the genetic fallacy, but either way, the claim, while not entirely false, contains only enough truth to be misleading. Dred Scott was not the origin of substantive due process, and while that case contains many faults, its brief, single-sentence reference to substantive due process (on page 450 of the decision, if you want to look it up) is not one of them. I explain more fully in The Right to Earn a Living, but I’ll address it briefly here, and then ask a question of my positivist friends.

First, some preliminaries. Dred Scott is a complicated and bizarre decision. It’s one of the longest in Supreme Court history, and it was written in bizarre circumstances: Chief Justice Taney apparently refused to allow his colleagues even to see the decision before he issued it—which is why Justice Benjamin Curtis resigned from the Court, the only Justice ever to do so over a matter of principle.

And it’s a labyrinthine opinion, one that goes off in directions that were unnecessary to the resolution of the case. Taney appears to have thought his decision could resolve the entire slavery question, and in particular, that he could deal a death blow to the anti-slavery constitutional theory advanced by people like Charles Sumner, Lysander Spooner, Joel Tiffany, and others. That hope was obviously disappointed, but not merely by the Civil War. Very soon after the decision was issued, New York’s highest court refused to follow it, issuing a decision in Lemmon v. People (20 N.Y. 562 (1860)) that rightly threw down the gauntlet at the High Court, only to be superseded by events.

Briefly put, here’s what the Dred Scott decision says:

A) For the government to deprive a person of slave property under a statute that exceeds Congress’ lawmaking authority would be a deprivation of property without due process of law.

B) Congress has no constitutional authority to ban slavery in the western territories, which is what it attempted to do, resulting in depriving Mr. Sandford of his slave.

∴) Therefore taking away Mr. Sandford’s slave property deprives him of property without due process of law.

As a logical matter, this is a valid argument. But it’s unsound, because the minor premise is false: the Constitution quite clearly gave Congress power to ban slavery in the territories, and Taney’s elaborate attempts to deny this—which take up the bulk of the opinion—are denials of plain fact. There are other things wrong with the opinion, but as I said, it goes in so many directions that I can’t discuss them all.

The major premise, however—that ultra vires legislative action cannot count as a law, and therefore violates the due process of law guarantee—is entirely sound, and was well established and non-controversial by 1857. We’ve already seen that Daniel Webster’s invocation of this idea, which became the standard text in nineteenth century substantive due process decisions, came almost half a century earlier, and state courts had been applying the idea long before Dred Scott. Even Prof. Williams, who thinks substantive due process originated much later than I believe, acknowledges that it was not a novelty in Dred Scott, and that whatever other well-deserved criticisms were leveled at the decision when it was issued, nobody at the time seems to have regarded that aspect of the case as problematic.

Thus while substantive due process did play a role in Dred Scott, it is highly misleading to characterize it as a “substantive due process case” (let alone as the source of that theory). It’s like saying that the problem with Korematsu was that it was “a supremacy clause case,” or that the problem with Plessy v. Ferguson was that it was “an equal protection case.” It just won’t do to smear substantive due process with the stain of Dred Scott. Substantive due process is an analytical test—a legal theory, like “standing.” Judges can be wrong about standing and can even abuse it to reach wrong decisions. But that’s an argument for being careful about the judges we choose—not an argument that the law itself should be discarded.

But now that I’ve explained (one reason) why I think Dred Scott was wrong, I have a question for my positivist friends. Why do you think Dred Scott was wrong? How could it possibly be wrong, according to your theory? If you believe, as Prof. Rosenthal does, that “[t]he only requisites for the existence of ‘law’…are procedural,” and that law is simply “whatever rules have been enacted consistent with the procedural requirements for lawmaking, whether we consider [the results] justified or not,” then how is it possible for Dred Scott to have been wrongly decided? It was, after all, promulgated by the duly constituted Supreme Court, in accordance with all of the procedural requirements. How could it not have been law?

I see no sense in saying that mere promulgation is sufficient for the legislature but not for the courts (certainly nothing in the Constitution draws the line here). If promulgation by constituted authorities makes something law, then Dred Scott must have been the law—that is, it must have been what it purported to be: a correct interpretation of the Constitution. And yet…Dred Scott says that ultra vires Congressional action doesn’t satisfy the due process clause, which you hold to be incorrect. What, then, is the status of this decision, which was duly promulgated and must therefore be law—and yet which says that mere promulgation is not enough to make something law?

I sense here the same Cretan Paradox that I address in footnote 222 of my article. In Lawrence v. Texas, the Court said that “Bowers [v. Hardwick] was not correct when it was decided.” Yet Bowers was duly promulgated by the constituted authorities. It therefore must have been the law when it was decided. Thus Lawrence must be in error to say that Bowers was wrong. Yet Lawrence was duly promulgated by the constituted authorities, and it, too, therefore must be law. (And you can’t get out of this by saying that Court decisions can be law but still mistaken interpretations of the law, since they purport to be nothing other than correct interpretations of the law, and disavow any lawful effect if they are shown to be incorrect interpretations.)

Finally, the straw man. I’m obviously not advocating “judicial tyranny.” I believe in no sort of tyranny, but because in our system, “the invasion of private rights is chiefly to be apprehended, not from acts of Government contrary to the sense of its constituents, but from acts in which the Government is the mere instrument of the major number of the constituents,” I believe in a strong, engaged judiciary as a bulwark against legislative violation of individual rights and lawful order—problems to which today’s vogue for judicial deference leaves us vulnerable. On that point, let’s pause on what Prof. Rosenthal calls the “irony” of my citation to the Federalist. He considers that ironic because “the Constitution…contained no due process clause” when it was written. This leads him to conclude that Hamilton and his colleagues “were not arguing for a judicial veto as a means of preventing abuses of legislative power; they were arguing instead that the system of political checks and balances that the Constitution had created would be sufficient to constrain government power.”

But Hamilton wrote that when Congress passes statutes that are “not pursuant to its constitutional powers,” those statutes “will [not] become the supreme law of the land. These will be merely acts of usurpation, and will deserve to be treated as such…. [T]hat limitation would have been to be understood, though it had not been expressed.”

“Treated”…by whom? Prof. Rosenthal mentions “checks and balances”—what check and balance is Hamilton hinting at here? If we can’t guess, Hamilton gives us the answer: “the courts,” he says, have “the right[]…to pronounce legislative acts void, because contrary to the Constitution.” Courts, he explained, “were designed to be an intermediate body between the people and the legislature,” entrusted with the power “to keep the latter within the limits assigned to their authority. The interpretation of the laws is the proper and peculiar province of the courts. A constitution is, in fact, and must be regarded by the judges, as a fundamental law. It therefore belongs to them to ascertain its meaning, as well as the meaning of any particular act proceeding from the legislative body.”

This was, indeed, said even before the Fifth Amendment was ratified. But that Amendment, like the rest of the Bill of Rights, was regarded by everyone concerned—Antifederalists and Federalists alike—not as establishing new provisions, but as reiterating what was already implicit in the Constitution: they are limitations that “would have been to be understood, though they had not been expressed,” and were added as a redundant precaution. The fact that the Bill of Rights and the limitation in the Supremacy Clause were considered redundant is no reason to ignore them. Certainly nobody in 1788 believed—as Prof. Rosenthal seems to—that before the Fifth Amendment was ratified, the federal government was allowed to deprive people of life, liberty, or property without due process of law!

Probably the most basic myth about substantive due process is that it is anti-democratic. It is not. Courts are part of our lawful democratic process—they are one of what Prof. Rosenthal calls the “political checks and balances” in our system. The people, knowing their legislators—and even they themselves—make mistakes, or are sometimes driven by passion instead of reason, or otherwise violate individual rights, chose to create in the Constitution a system of separated powers, including a judiciary, which is entrusted with the job of ensuring that the democratic process doesn’t violate the fundamental law of the land, and become destructive of individual rights, both enumerated and unenumerated. To regard the courts as standing outside our lawful democratic process is to misunderstand that process. The judges are chosen indirectly by the people, and their powers are subjected to meaningful checks by the President and Congress. The courts and their power of judicial review stand within, not outside our system of lawful democratic order. Ours is not a simple majoritarian system combined with some anti-democratic “veto.” It’s a structure by which the people give law to themselves—by which they limit their own democratic power within a system of law—law that courts (and other branches) are supposed to enforce. As Jefferson put it, “An elective despotism was not the government we fought for; but one which should not only be founded on free principles, but in which…no [branch] could transcend their legal limits, without being effectually checked and restrained by the others.”

Gary Lawson is one of the foremost contemporary scholarly advocates of originalism. Anything he writes on that subject must be taken with the utmost seriousness. Accordingly, I wish to treat with his approach to constitutional interpretation because it illustrates one of the principle quandaries faced by today’s constitutional originalism.

Professor Lawson writes that “Sandefur’s attempt to decouple constitutional meaning from anyone’s intentions … is partly right and partly wrong.” He elaborates: “Communicative instruments such as the Constitution have meaning only in light of intentions … . [T]he Constitution’s meaning is determined by the hypothetical, legally constructed intentions of a hypothetical, legally constructed reasonable observer of 1788.” Professor Lawson goes on to agree with me that “a non-arbitrariness principle, even if applicable to most actions by the federal government, will not necessarily yield all of the results that Sandefur (or I) think are prescribed by the Constitution, such as the unconstitutionality of a federal statute forbidding segregated rail cars.” Nevertheless, he thinks that the original meaning of the Constitution would forbid a statute mandating racial segregation because “eighteenth-century fiduciaries were under a presumptive obligation to treat all principals fairly, which roughly corresponds to what today we call ‘equal protection.’” In his scholarship, Professor Lawson has assembled support for the claim that in the eighteenth century, governments were widely viewed as exercising a kind of fiduciary power, to which obligations of reasonableness and evenhandedness applied.

Based on this understanding of the original meaning of the Necessary and Proper Clause, as well as the Fifth Amendment’s Due Process Clause, Professor Lawson concludes that racial segregation violated the 1788 Constitution, as well as the 1791 addition of the Fifth Amendment. Readers with even a passing familiarity with the history of segregation, of course, will note that something extraordinary is going on. There is, of course, good reason that Professor Lawson invokes “the hypothetical, legally constructed intentions of a hypothetical, legally constructed observer” rather than the actual intentions of any actual observers circa 1788 or 1791. That is because pretty much no one actually alive then thought that the Constitution contained anything that prohibited racial segregation. As Michael Klarman has demonstrated, even after the Constitution added an express requirement of “equal protection” in the Fourteenth Amendment, it was widely agreed that segregation was permitted in both state and federal legislation.

Professor Lawson’s “hypothetical, legally constructed observer” would have been considered quite the oddball in 1788 if he announced that the Necessary and Proper Clause, or the Due Process Clause, forbade segregation in the District of Columbia, where it was instead widely practiced well into the twentieth century. Thus, although Professor Lawson tells us that the meaning of a legal text cannot be determined without reference to the intentions of those who are properly regarded as its authors, he offers an account of original meaning that matches the actual intentions of few if any of those actually involved in the framing and ratification of the 1788 Constitution or the amendments that followed in its wake.

Presumably, what Professor Lawson means is that we need not worry about the actual intentions of 18th-century observers when it comes to racial segregation (or pretty much any other concrete application of the Constitution) because when the Constitution’s text is written at a very high level of generality – such as the Necessary and Proper and Due Process Clauses – it embodies quite abstract principles, such as a requirement of fair treatment. Even putting aside the strength of the historical evidence supporting this conclusion, Professor Lawson’s interpretive methodology raises many questions. For present purposes, I will confine myself to three of them.

First, if we are to ignore the concrete intentions of the framers as to how the Constitution should be applied and treat its broadest language as not rule-like but rather as imposing abstract, standard-like principles that can lead us, as in the case of racial segregation, to quite different concrete applications than were contemplated in the framing era, does this brand of originalism differ from living constitutionalism? If the Necessary and Proper or the Due Process Clause, understood in terms of their original meaning, give us no guidance other than a guarantee of “fairness,” then this brand of originalism can be used to justify virtually any result that the modern judge regards as “fair.” If a judge thinks that taxes are too high and exemptions for the wealthy are too generous, thereby depriving some taxpayers of “liberty” or “property” unjustifiably while awarding equally undeserved lenient treatment to capital gains, why not invalidate the Internal Revenue Code as “unfair”? Or, if the judge doesn’t like the siting decisions that the federal government makes for oil refineries, why not do the same? As I argue in a paper forthcoming in the Indiana Law Journal, the central problem faced by the prevailing brand of originalism, which rejects original intended applications as controlling, is that in practice it operates no differently than nonoriginalism. If the fiduciary obligations of those who exercise governmental political power include “fairness,” a judge can decide that pretty much anything is unconstitutional, no?

Second, why is Professor Lawson so uncertain about whether his conception of the fiduciary obligations imposed on government apply to the states? If it is really true that one of the basic obligations imposed on any exercise of governmental power is that it be “fair,” for example, is this not an aspect of “due process of law” applicable to the states by virtue of the Fourteenth Amendment?

Third, even if we are to accept Professor Lawson’s account of the framing-era normative conceptions about the manner in which governmental power should be exercised, what is the evidence that original meaning included a conception that these limitations would be judicially enforceable? After all, as Ryan Williams has demonstrated, there was not a robust tradition of judicial review of the substance of legislation at the time of the Fifth Amendment’s adoption. Did the framers really mean to give the judiciary the power to apply highly abstract standards? After all, we know that the framers intended to create a republic. If they also created a life-tenured and politically unaccountable judiciary that could set pretty much any legislation at naught by branding it “arbitrary” or “unfair,” perhaps what they created fell far short of a republic — something more like a judicially supervised nanny state. And, if we can only preserve our republican form of government by insisting on a degree of judicial deference to legislative policy judgments, is there really that much difference between Professor Lawson’s originalism and my plea for judicial prudence and humility?

One basis for modern opposition to “substantive due process” is that the Due Process of Law guarantee only applies to the courts or the executive, not to the legislature. The superficial plausibility of this argument fades upon examination, though.

As Prof. Nicholas Rosenkranz has recently emphasized, we should take care to properly identify the subjects and objects we’re talking about. The executive is charged with executing the law enacted by the legislature, whereas the legislature, strictly speaking, doesn’t act at all. It merely prescribes the rule, and the executive justifies his actions (if at all) by reference to the enacted rule. So, in the example I gave in my opening article, I imagined Congress passing a law establishing a national religion; a person refuses to go to church, on the grounds that this “law” is unconstitutional, and is therefore void and of no effect. But at this point, his rights have not yet been violated—he has not yet been deprived of life, liberty, or property, and cannot assert a violation of the Due Process Clause or anything else. It is only when the sheriff—i.e., the executive officer—arrives to arrest him that the issue is joined. Now the dissenter argues that he’s being deprived of liberty without due process of law because the sheriff is acting without authority. The sheriff, meanwhile, points to the statute as the source of his authority. The question now is whether that statute qualifies as “law” for purposes of the Due Process Clause. There’s a colloquial sloppiness in saying at this point that “the legislature has violated the Due Process Clause.” In fact, it’s the sheriff who has deprived the man of his liberty, without legitimate authority, because the statute is not “law” for purposes of the Due Process of Law Clause. It’s the legislature and the executive in concert that have violated the Constitution.

Keep in mind that the due process of law guarantee originated in Magna Carta at a time when there was no significant understanding of separation of powers, and the wording of the clause in that document indicates no such distinction. Not only did the king exercise what we would consider judicial and legislative powers, but Parliament was more like a court than a legislature until relatively recently. Like Magna Carta’s law of the land clause, the Fifth or Fourteenth Amendments’ due process of law clauses make no distinction between executive, legislative, or judicial: they simply declare that no person shall be deprived of rights without due process of law—not that no person shall be deprived by the judiciary or the executive without due process of law.

What would it mean to say that the due process clause doesn’t restrict the legislature? In the example I’ve given, the sheriff who arrests a dissenter under an unconstitutional law can point to the unconstitutional law as the source of his authority, and—then what? His action is now authorized because the legislature enacted a statute, so that the arrest does not violate the due process of law clause? If the due process of law clause does not apply to the legislature, then we really would have the situation that Justice Bronson characterized as “you shall not do the wrong, unless you choose to do it”! (I discuss Bronson argument further on pp. 335-342 of my article.)

I don’t think the framers fell for this argument, either. In the crisis over the Sedition Act, some argued that Act didn’t violate the Constitution because the First Amendment was modeled on the British common law and thus allowed the legislature to punish speech post hoc. Madison answered that the American Constitution was fundamentally different from that of Britain. “In the British government, the danger of encroachments on the rights of the people is understood to be confined to the executive magistrate,” whereas Parliament was held to be sovereign. Thus the “ramparts for protecting the rights of the people,” including Magna Carta and the British Bill of Rights, were aimed at the executive, rather than the legislature. But in the United States,

[t]he people, not the government, possess the absolute sovereignty. The legislature, no less than the executive, is under limitations of power. Encroachments are regarded as possible from the one as well as from the other. Hence, in the United States, the great an essential rights of the people are secured against legislative as well as executive ambition. They are secured, not by laws paramount to prerogative, but by constitutions paramount to laws. This security of the freedom of the press requires that it should be exempt, not only from previous restraint of the executive, as in Great Britain, but from legislative restraint also….

What about Alexander Hamilton? It seems every time I read about the founders and due process, I encounter the assertion that Hamilton believed the due process clause did not apply to the legislature, and it nearly always includes this quote:

The words due process of law have a precise technical import, and are only applicable to the process and proceedings of the courts of justice; they can never be referred to an act of the legislature.

But when you consult Hamilton’s actual speech, and read it in context, it’s obvious that Hamilton is making exactly the opposite point for which he is being quoted! The question being debated was whether the legislature could exclude certain former Loyalists from voting. Hamilton opposed this, arguing that it would constitute a “legislative disqualification,” and, therefore, a violation of the due process of law guarantee. To disqualify a person from voting, or impose any other punishment on him, the government must first accord him due process of law—and an act of the legislature is not enough. That was the point Hamilton was making: “Some gentlemen hold that the law of the land will include an act of the Legislature,” he says—that is, some people argue that the law of the land guarantee is satisfied by the legislature passing a statute—but:

Lord Coke, that great luminary of the law, in his comment upon a similar clause in Magna Charta, interprets the law of the land to mean presentment and indictment…. But if there were any doubt upon the Constitution, the bill of rights enacted in this very session removes it. It is there declared that no man shall be disenfranchised or deprived of any right but by due process of law or the judgment of his peers. The words “due process” have a precise technical import, and are only applicable to the process and proceedings of courts of justice; they can never be referred to an act of the Legislature. Are we willing, then, to endure the inconsistency of passing a bill of rights and committing a direct violation of it in the same session? In short, are we ready to destroy its foundations at the moment they are laid?

In other words, for the legislature to disqualify a person or impose a deprivation on a person would be a direct violation of the due process of law clause. I’m no fan of Hamilton, but I feel sorry for a man who has been so brazenly misquoted.

Hamilton’s “legislative deprivation” theme was a common one in pre–Civil War substantive due process cases, such as Wynehamer or Billings v. Hall. And the question in these cases came down to this: can the government deprive a person of life, liberty, or property without legitimate justification, simply by passing a statute for that purpose? This just is my hypothetical of the sheriff. And it just is Lochner v. New York. And it just is Lawrence v. Texas.

These cases all involved “legislative deprivations.” In other words, this just is substantive due process theory. Hamilton’s words indicate that some acts of the legislature are not “law” for purposes of the due process of law guarantee—that is, arbitrary, ipse dixit, self-interested, irrational acts do not satisfy the due process of law guarantee even if the legislature votes in favor of them. Legislative “process” is not enough.

Our conversation is starting to turn to the question of how we access legal meaning—that is, into the epistemological question about how language and ideas work. This is too big a question to be more than hinted at here (I would urge readers to consult the works of Tara Smith, David Brink, and the like). But I will do my best to explain briefly why I believe what Prof. Williams does not: that “that there is some objective ‘meaning’ of constitutional language that can be wholly divorced from subjective understandings of some group of speakers and listeners.” Or something like that, anyway.

We use words to denote concepts, which are mental abstractions from the perceived facts of reality. These concepts are mental constructions, which stand for real qualities in the world. It’s the very fact that those qualities are real that explains how it’s possible that we may be mistaken about them. The word “fish,” for example, stands for a concept—not for any particular fish, but for all creatures with those fish-defining qualities. These qualities are facts in reality: the concept does not merely exist because we subscribe to it; it’s not a social construct. A fish really is not a dolphin, and even if everyone in our society chose to say that fish and dolphins are the same, they would be wrong. Note that I am not claiming that there is some intrinsic, Platonic form of “fishness.” But fishness does exist, in one sense: it’s a concept—a mental abstraction about real shared qualities between things. And it is objective—meaning that it is not invested with truth value simply by our decision to believe it; there is a fact of the matter about what a fish is, and we can submit that concept to scientific tests to see if it is correct.

Thus there is an objective meaning to words which can be divorced from the subjective understandings of persons, which is why it is possible for an author to be mistaken about the meaning of a written text. As Tara Smith writes,

words are shorthand for countless particular things of the relevant kind (“man” for all men; “shopping” for all shopping, and so on). The written law is also shorthand, drafted in deliberately broad terms in order to govern an array of cases that are similar in principle but different in particulars. The Constitution is written in particularly broad terms…. To objectively interpret the concepts used in law is to understand kinds…. While judgment is needed to identify which referents properly fall under the terms of a given law, then, that judgment is bounded by the words of the law. Stable law…requires not an immutable list of specific concretes, nor that we blindly obey the flawed or incomplete understanding of concepts and their distinguishing criteria held by particular men of an earlier era, but constancy in the kinds of concretes that the law refers to.

This much is sufficient for my purposes, because I am in the position of the hypothetical son of the dying inventor in Lon Fuller’s Morality of Law. The son, deciding to carry on and complete his father’s unfinished last invention, figures out how the device is supposed to work not by consulting his father’s intentions, but by analyzing the device in terms of its purpose and clues about how the machine is to operate. His concern, and mine, is with what and how—what is the Constitution and how does it work? In this sense, the task is only descriptive, and the objective theory of concepts serves us well.

But there is a second layer of complexity: why? What obliges us to complete the machine, or to follow the promise embodied in the law? And if, having committed ourselves to that promise, it turns out to mean something other than what we at first thought, are we still bound by it? This is what Lawrence Solum calls the distinction between semantic and normative originalism. And here I must answer that I simply don’t know. I don’t think the law knows, either. We are familiar with the doctrine of objectivity in contract law, under which a person cannot escape his obligation by claiming afterwards that he was only kidding. And if I agree to a promise containing non-specific terms—especially such a broad word as “liberty”—then I can’t say I’m not obliged to accept the results just because it turns out a way I didn’t expect. If I agree to submit a claim to arbitration, or even to a coin toss, I can’t refuse to accept the decision that goes against me. It likewise seems wrong to remove protections for a person’s freedom on the basis that he’s doing something people didn’t anticipate him doing when they used the word “liberty” in the Constitution. The concept of liberty includes that indeterminacy. Since a court decision typically purports to be a description of the content of the law, not the formation of an agreement, it seems reasonable to hold people to their promises even when the court, upon examining a promise, finds that it requires something the author did not expect. And although there are exceptions even in the law of contract whereby a mistaken belief as to the meaning of the words in the agreement might excuse performance, those situations generally involve an existing false belief about the meaning of the terms, as opposed to a simple lack of belief about how the term will be applied, or even an agreement to use a term, knowing that it is indeterminate, which is more often the case in disputes about the application of constitutional language.

It’s beyond the scope of this discussion to go further, in part because I’m convinced that the framers were familiar with the theory that we today call “substantive due process,” or that it is logically entailed by what they did believe. I have to second Prof. Lawson’s characteristically frank conclusion: “I am interested only in figuring out what the Constitution actually means; what anybody does with that meaning once they have it is not my concern” (a point that recalls one of my favorite law review articles, his marvelous The Rise and Rise of the Administrative State. One needs only read the first sentence to see why it’s dear to my heart). My focus is on what the Due Process of Law Clause means, which I believe can be differentiated from what particular people thought it meant at any one time.

Prof. Lawson has asked me to elaborate about the relationship of state and federal governments. What difference, if any, is there between the limit on states imposed by the Due Process of Law guarantee in the Fourteenth Amendment, and the limit imposed by the similar guarantee in the Fifth? And, to put it in a concrete case, supposing I think the individual mandate in Obamacare is unconstitutional (as indeed I do; I am currently litigating one of the many challenges to the mandate)—do I also think a state mandate is unconstitutional?

First, some background. I think it’s actually not as clear as Prof. Lawson thinks whether or not states “trace their basic legal authority to the federal Constitution.” As I explain more fully in my article, Privileges, Immunities, and Substantive Due Process, antebellum legal scholars were divided over this issue due to certain ambiguities in American legal history. Many argued that the states were basically sovereign entities, united by a sort of treaty. The most extreme of these was John C. Calhoun, but even relative moderates, like Pennsylvania’s Chief Justice Jeremiah S. Black, argued that except where explicitly barred by the federal Constitution, the states were free to act in whatsoever manner they pleased—that states enjoyed the sovereignty that William Blackstone called “supreme, irresistible, absolute authority.”

Pitted against them were the Lockeans, most notably John Quincy Adams and his admirers, like Charles Sumner, William Seward, and Abraham Lincoln. They argued that the principles of the American constitutional order—derived from the Declaration of Independence, the Anglo-American common law, and other sources—limited the authority of states even in the absence of explicit constitutional limits. This view was generally associated with anti-slavery forces, either radicals like Frederick Douglass, who argued that slavery was already unconstitutional even before the ratification of the Thirteenth Amendment, or moderates who thought that while the Constitution protected slavery in states where it already existed, it barred, or allowed Congress to bar, its expansion. They all agreed that no state could be legitimate and claim “supreme, irresistible, absolute authority.”

These thinkers rejected the states-rights theory in favor of the theory of “paramount national citizenship,” which held that people were citizens first of the nation, and then only derivatively of their home states. They relied heavily on the original Privileges and Immunities Clause in Article IV—look, for example, at Adams’ objection to the admission of Missouri, or Joel Tiffany’s treatise on the Constitution. Rufus King and others had explained in 1787 that the states had never been sovereigns, lacked “the peculiar features of sovereignty,” and “as political beings, they were dumb” and “deaf”—and had derived their sovereignty from their membership in the federal union. Lincoln reiterated that point in 1861 when he argued that states derived their sovereignty from the federal union.

After the clash of visions about federal and state power had erupted into Civil War, victorious Republicans hoped to settle the question forever by adding the Fourteenth Amendment, to constitutionalize what they had always thought true: that Americans are citizens of the nation first, and of states only derivatively, and that their natural and common law rights are protected against state interference by the federal government. That’s why the Fourteenth Amendment starts out by defining citizenship nationally. (States don’t even have the constitutional authority to determine who their own citizens are! That’s a pretty pathetic “sovereignty”!) Then the rights of all Americans are secured against state infringement by an additional Privileges or Immunities Clause, a second Due Process of Law Clause, and a brand new Equal Protection Clause.

Of course, this effort to write the theory of “paramount national citizenship” into the Constitution was crippled in the cradle by the Slaughterhouse Cases, which ruled—without any basis, and without any inquiry into original meaning or anything at all except the Court’s mere say-so—that the Fourteenth Amendment wasn’t meant to “radically change[] the whole theory of the relations of the State and Federal governments to each other and of both these governments to the people.” In fact, that’s precisely what it was meant to do. (It comes as no surprise that Jeremiah Black, one of the great spokesmen for the states-rights theory of the Constitution, was the attorney for Louisiana in Slaughterhouse, and that he exulted over his role in demolishing one of the keystones of Reconstruction.)

Before I get too far afield, the point is that, while states don’t derive their “basic legal authority” from the same source as the federal government, citizens of the United States do derive their basic political protection from, and owe their basic political allegiance to, the union, not the state. And that is why states must remain within the boundaries of Due Process of Law just like the federal government. States are not departments of the federal government, and they are not governments of enumerated powers, but they are bound within federal constitutional limits, including provisions that require them to respect natural rights, common law rights, and the rights articulated in the federal Constitution.

This is why the “incorporation” doctrine is best seen as an aspect of the Due Process of Law guarantee. Properly understood, incorporation is a generalized limit against arbitrary state action—not a straight transfer of the Bill of Rights guarantees to the states. The federal government may not take property without paying for it, for instance, because of the specific command of the Fifth Amendment. But in my view—and in the Chicago B. & Q. Court’s view—the due process clause forbids states from taking property without paying for it only because, and to the extent that, doing so would be an unauthorized, arbitrary assertion of power. Thus the Fourteenth Amendment’s Due Process Clause places a more generalized restriction on states, one that still leaves them an indefinite range of powers. It puts a ceiling on their ambitions by depriving them of the ability to choose their citizens, protecting the privileges or immunities of those citizens, and then barring states from acting in ways that violate the principles of lawfulness—principles which, of course, include the aforementioned rights, but also include the anti-arbitrariness principles I’ve mentioned. These overlap, but are not identical. For example, I think one could make a good argument that laws banning gay marriage do not deprive persons of natural rights, traditional common law rights, or rights articulated in the Constitution, but are arbitrary exercises of ipse dixit power (and more obviously violate the Equal Protection Clause). This position hardly serves my libertarian predilections, but I think it’s the best reading of the post-Civil War Constitution’s federalist structure.

While the individual mandate in Obamacare is unconstitutional because it exceeds Congress’ power under the Commerce and Necessary and Proper Clauses—and is thus outside the enumeration of powers—the Massachusetts individual mandate could only be unconstitutional if it deprives a person of a natural right, of a traditional common law right, or of a right articulated in the Constitution, or if it is an exercise of arbitrary, ipse dixit power. I don’t know enough about that mandate to reach a conclusion, but I think the right approach is suggested by Loan Association v. Topeka. That case differs from the mandate cases in that it involved a tax that went directly to a private industry, rather than a statute compelling consumers to buy something from a private firm, but the principles remain sound:

It is undoubtedly the duty of the legislature which imposes or authorizes municipalities to impose a tax to see that it is not to be used for purposes of private interest instead of a public use…. [I]n deciding whether, in the given case, the object for which the taxes are assessed falls upon the one side or the other of this line, [courts] must be governed mainly by the course and usage of the government, the objects for which taxes have been customarily and by long course of legislation levied, what objects or purposes have been considered necessary to the support and for the proper use of the government, whether state or municipal. Whatever lawfully pertains to this and is sanctioned by time and the acquiescence of the people may well be held to belong to the public use, and proper for the maintenance of good government, though this may not be the only criterion of rightful taxation.

But in the case before us, in which the towns are authorized to contribute aid by way of taxation to any class of manufacturers, there is no difficulty in holding that this is not such a public purpose as we have been considering. If it be said that a benefit results to the local public of a town by establishing manufactures, the same may be said of any other business or pursuit which employs capital or labor. The merchant, the mechanic, the innkeeper, the banker, the builder, the steamboat owner are equally promoters of the public good, and equally deserving the aid of the citizens by forced contributions. No line can be drawn in favor of the manufacturer which would not open the coffers of the public treasury to the importunities of two-thirds of the businessmen of the city or town.

One major problem with the individual mandate, of course—as highlighted in our brief in the Florida case—is that it is not a tax, but a device concocted to evade the politically unpopular route of taxation. Taxing people to pay for government-operated hospitals would probably be constitutional, but would have polled badly. So, by forcing consumers to buy from a private firm, and forcing private firms to serve consumers, Congress evaded the traditional democratic process for providing government services. RomneyCare likewise essentially co-opts the insurance industry to provide a subsidized public service instead of taking the honest, but politically unpopular, route of seizing the insurance industry through eminent domain or establishing an alternative socialized medicine scheme. I would think those who consider “due process” to be satisfied so long as the legislature abides by traditional rules of promulgation would be unnerved by such a novel evasion of political accountability.

His recent posts provoke three thoughts on Timothy Sandefur’s argument for a robust doctrine of substantive due process.

1. Text: Sandefur marshals historical evidence that on some occasions in the framing era, the term “law” was used to connote not only enactments that had satisfied the procedural requirements for lawmaking, but instead to enactments which also conformed to express and implied limits on the scope of legislative authority. Yet he refrains from claiming that the term “law” was ALWAYS used in this fashion. Sometimes, the term was used merely to refer to enactments that had satisfied the procedural prerequisites for lawmaking, as in Article I of the Constitution, which provides that if a bill is passed by both houses of Congress and signed by the president, or if the president’s veto is overridden, “it shall become Law.” At best, Sandefur has identified an ambiguity in the text. Of course, an ambiguous text cannot tell us how to resolve the ambiguity. Textual argument is no better than a draw.

2. History: In an effort to address this ambiguity, Sandefur resorts to history. He is tripped up by the same problem. Sandefur marshals the evidence that supports his view, and there is some. There were a number of legal thinkers in the framing era who embraced Sandefur’s view of the meaning of “law,” as well as “due process of law.” Ryan Williams is quite right that the evidence supporting Sandefur is not as strong at the time of the Fifth Amendment’s framing as at the framing of the Fourteenth Amendment, and he is right as well that most (but, I think, not all) of this evidence involves a less robust conception of substantive due process than Sandefur’s (primarily relating to vested rights and some kinds of class legislation). But, there is certainly some historical evidence supporting Sandefur’s view. But, there is also inconsistent evidence. Sandefur likes Justice Chase’s opinion in Calder v. Bull, but Justice Iredell’s view was quite different. Sandefur cites the opinion in Wynehamer v. People, but the opinion of the Rhode Island Supreme Court in State v. Keeran took a directly opposing view. I could go on, but the point should be plain. This historical evidence is messy and in conflict. Sandefur makes no effort to demonstrate the contrary. Historical evidence supplies no reliable basis for decision.

3. Structure and prudence: In light of what I regard as the conflicting evidence from text and history, I have turned to structural and prudential argument. While I am willing to embrace a narrow conception of substantive due process, I think there are powerful structural and prudential reasons for the judiciary to defer to the legislature on policy issues where judicial expertise is lacking and political accountability is likely to be a more effective means of redress for unwise decisions than granting an unaccountable judiciary an effective veto over matters about which it is likely to know little. As an example of the dangers of an overbroad doctrine of substantive due process, I have referred to Dred Scott, the first occasion on which the Supreme Court embraced something like the doctrine that Sandefur advocates.

In his post on Dred Scott, Sandefur, after a fair amount of huffing and puffing, ultimately admits that the Court held that the Due Process Clause prevents Congress from forbidding slavery in the territories. His complaint about Dred Scott, it seems, is that he regards the decision as wrong. Why? Because the Constitution “quite clearly gave Congress power to ban slavery in the territories … .” He cites the Territorial Clause, which provides: “The Congress shall have power to dispose of and make all needful Rules and Regulations respecting the Territory or other Property belonging to the United States; and nothing in this Constitution shall be so construed as to Prejudice any Claims of the United States, or of any particular State.” Does this language “quite clearly” grant Congress the power to ban slavery in the territories? A majority of the Supreme Court did not see matters as so clear. Moreover, even if the Territorial Clause conferred this power on Congress, why is that power not limited by the Fifth Amendment’s prohibition, at least as Sandefur understands it, on “arbitrary” legislation? Why is it not similarly limited if we accept Professor Lawson’s view that the Constitution, properly understood, bans “unfair” legislation? In Dred Scott, the Court wrote: “[A]n Act of Congress which deprives a citizen of the United States of his liberty or property, merely because he came or brought his property into a particular Territory of the United States, and who had committed no offense against the laws, could hardly be dignified with the name of due process of law.” It seems that the Court thought that a ban on slavery was “arbitrary” or “unfair,” to the extent we can even understand what Sandefur or Lawson mean by these terms.

All adjudication involves a risk of error. The more imprecise the “rule” to be applied, the greater the risk of error. Sandefur’s rule is pretty imprecise — it prohibits “arbitrary” legislation. Despite my fancy education and credentials, I do not really understand what he means by “arbitrary.” After all, Sandefur seems to believe that a lot of carefully considered decisions to which he objects on grounds that have little to do with arbitrariness — such as racial segregation or the use of eminent domain to promote private redevelopment — violate his rule. Professor Lawson’s far more candid view is that a great deal of legislation that is not arbitrary nevertheless runs afoul of what he regards as a constitutional obligation to treat citizens “fairly.” In the abstract, these pronouncements are unobjectionable, but because they are so broad and abstract, any effort to apply them in litigation would confer enormous powers on courts and inject an enormous risk of error. If you let courts strike down anything that they regard as “unfair,” you should not be surprised if Dred Scott — or its contemporary equivalent — is the result.

If we are to have a doctrine of substantive due process, it should be one that injects only an acceptable risk of error into constitutional adjudication, and one that is sensitive to the limits of judicial expertise, and to the unsuitability of the adversarial process for making what are essentially policy judgments. Is that too much to ask?

There’s a new article about substantive due process in the Yale Law Journal, this one written by Nathan Chapman and Michael McConnell. Like Prof. Williams, they are primarily concerned with history, and they give us some good history. But there are three basic flaws in their approach.

First and most importantly, Chapman and McConnell approach the subject with a strawman understanding of modern substantive due process doctrine, which they characterize as being “about judicial creation of fundamental rights outside the reach of legislative amendment.” This is not what substantive due process is, and by looking for it in the legal precedent, the authors are in the position of the man who tried to play Solitaire with a deck of Tarot cards. Second, they fail to grapple with the dramatic shift in the understanding of political legitimacy—of sovereignty—that took place with the birth and re-birth of the United States in 1776 and 1865. Third, by looking so closely at the cases, Chapman and McConnell lose track of the broader picture that is drawn for us by the cases themselves: they see each particular tree, but don’t think there’s a forest.

Chapman and McConnell begin, rightly, with the law of the land clause in Magna Carta, describing how Lord Coke and other judges saw it as protecting the subject against arbitrary or unauthorized government acts. But they think Coke was only arguing procedure: “[l]egislative acts violated due process not because they were unreasonable or in violation of higher law, but because they exercised judicial power or abrogated common law protections…. Coke was not arguing that the government could not [pass the restriction in question]; rather, he was asserting that the king could do so only in coordination with parliament…. [O]nly the legislature can make law that deprives persons of established rights.”

But this is not right. First, judicial enforcement of separation of powers in Coke’s day was a species of “creating” rights that are “outside the reach of legislative amendment”: even if we assume the law of the land clause only imposes procedural boundaries, those boundaries ARE beyond legislative alteration. Is it procedural or substantive? Well, as we’ve seen, the distinction between the two is not at all clear. Nor does it make sense to say simultaneously that the due process requirement was not about a higher law, but that it required the legislature to abide by common law protections. Those common law protections, especially according to Coke, were manifestations of abstract principles of reasonableness: the “artificial reason of the law.” A man could not be a judge in his own case, for example. This was considered a fundamental principle of common law precisely because it would be unreasonable and arbitrary to allow a man to be a judge in his own case. This is “higher law” reasoning, and, like the other rights protected by substantive due process, it can alternately be characterized as “procedural” or “substantive.” Chapman and McConnell describe this principle as one of “customary law”—but that’s just the point: it was a principle of reason, and for that reason embedded in the common law, and therefore valid even against Parliament’s authority. And if no man can be a judge in his own case, then that also IS a species of “creating rights” that are “outside the reach of legislative amendment”: at the very least, a right not to be subject to a law that makes a man a judge in his own case! (One thinks here of the Fugitive Slave Law.) This “procedural” protection is just as substantive as anything.

The point is that Chapman and McConnell are opposing what is basically a strawman caricature of substantive due process. That theory was not—and does not purport to be, today, except in the mischaracterizations of the theory’s opponents—about simply inventing rights. It’s a general guarantee against arbitrary, lawless government action. And that guarantee requires a court in a particular case to decide whether the government acted arbitrarily or lawlessly—an analysis that necessarily incorporates normative principles. Put that way, it’s easy to see that “substantive due process” was very much a part of seventeenth and eighteenth century jurisprudence.

Take another example: Chapman and McConnell dispute that Daniel Webster’s argument in Dartmouth College was an articulation of substantive due process theory. Instead, they argue, Webster was only arguing for a separation of powers principle: that the legislature is may not enact “quasi-judicial measures” that are “tantamount to a judicial sentence.” It’s certainly true that Webster argued this—the same “legislative deprivation” argument that we’ve seen Hamilton discussing. But Webster was contending that legislative deprivation was one type of lawless, unauthorized, arbitrary government action. And the reasons he gave for categorizing it as such show that he recognized a broader principle—that government may not impose arbitrary or lawless deprivations on people. To contend that Webster’s argument was not substantive due process theory is tantamount to saying that Areopagitica was only about the freedom to publish books about divorce, because the censorship case that inspired John Milton to write it involved a book about divorce. That would be silly, because Milton was arguing for a broad principle that would apply to all expression. Similarly, Webster was arguing that Dartmouth College had suffered a legislative deprivation which was one kind of arbitrary action that violates the due process guarantee. To argue otherwise is affirming a disjunct.

But there’s an even more important point that Chapman and McConnell overlook: the change in the understanding of sovereignty that took place with the American Revolution and was vindicated in the Fourteenth Amendment. Lord Coke’s articulation of the law of the land clause came at a time when debate over Parliament’s sovereignty was accelerating toward the English Civil War. That war was the first of a series of wars that climaxed with the American Revolution, which decided that, whatever British legal thinkers might say about Parliament’s sovereignty, in America, legislatures were not supreme. This added another layer of “higher law” reasoning to the protection of individual rights.

William Blackstone and others had argued that Parliament enjoyed “supreme, irresistible, absolute” authority. But the Declaration of Independence committed the United States to a different principle: there could be no legitimate “absolute” government, because legislative supremacy was limited by the natural rights of individuals. As James Madison put it, “the Sovereignty of the Society as vested in & exerciseable by the majority, may do any thing that could be rightfully done, by the unanimous concurrence of the members; the reserved rights of individuals (of Conscience for example), in becoming parties to the original compact, being beyond the legitimate reach of Sovereignty, wherever vested or however viewed.” Pro-slavery forces in the nineteenth century clung to the Blackstonian argument that legislative power was supreme and absolute, and that therefore the legislature could (in Chapman and McConnell’s words) “make law that deprives persons of established rights so long as it abided by procedural requirements. But others (whom I’ve called Lockeans) argued that legislatures were not supreme—their legitimate authority was restricted by the natural rights articulated, among other places, in the Declaration of Independence. The most articulate spokesman for this understanding was John Quincy Adams:

[State] sovereignty is not, and never was, a sovereignty as defined by Blackstone and the English lawyers, identical with unlimited power; that sovereignty, thus defined, is in direct contradiction to the Declaration of Independence, and incompatible with the nature of our institutions…. [T]he States, united, and the States, separate, are both sovereign, but creatures of the people and possess none but delegated powers…. [The doctrine that legislatures have power to enact laws that violate individual rights would mean that] [t]he Declaration of Independence will become a philosophical dream, and uncontrolled, despotic sovereignties will trample with impunity, through a long career of after ages, at interminable or exterminating war with one another, upon the indefeasible and unalienable rights of man.

In other words, the legislature can not “make law that deprives persons of established rights,” as the British Parliament could. At the end of the Civil War, the victorious Republicans sought to constitutionalize their rejection of the Blackstonian conception of legislative power. The Fourteenth Amendment—drafted largely by Adams’ protégés—was meant to put an end to the previous half-century of debates over this issue, and to make clear that our natural rights put boundaries around legislative power. When the legislature steps beyond those boundaries, its acts are unauthorized and are therefore not law for purposes of the Due Process of Law Clause.

Sadly, this revolution in the understanding of legitimate sovereignty is not discussed in Chapman and McConnell’s article. They are concerned primarily with legal decisions, and although I agree for the most part with what history they provide, it’s crucial to step back and see these cases in the context of the broader clash of political philosophy. If we focus too narrowly only on a specific case, we can be misled into seeing it only standing for that specific controversy. Thus we might, for example, be misled into reading Dartmouth College as being only about legislative deprivations, rather than as one permutation of the anti-arbitrariness principle. This truly would be missing the forest because no particular tree looks like one.

Most of all, we handicap our understanding both of the theory and the history itself if we approach the case law with a flawed picture of substantive due process in mind—namely, that it’s about “creating rights.” Debates over rights do arise necessarily from any substantive due process analysis, because deciding whether a law is a law for purposes of the Due Process Clause requires such a determination—we saw as much in my earlier comment on Gideon v. Wainwright. At the very least, when a plaintiff asserts that he’s been deprived of a right without due process of law, the court must first determine whether he actually had the right he claims was lost—and at that point, it can be accused of “creating” rights. Yet this is done in virtually any case, and it hardly shows that substantive due process is illegitimate. If we approach the history with a flawed image of substantive due process in mind, then it can come as no surprise when we fail to find it in the old decisions. But if we take a step back and look at the principles involved, we see that any such mismatch is only an artifact of our own misunderstanding, not of a nefarious change in the law.

The anti-arbitrariness principle we today call “substantive due process” is centuries old; it has been the subject of much disagreement and debate; it has manifested itself in different ways in different factual circumstances; and it took on a crucial new importance with the discovery of the limits on legislative power imposed by individual rights. If the post-Fourteenth Amendment era represents a kind of “Cambrian explosion” of substantive due process, we must not be misled by it in the way that some people are misled by the “Cambrian explosion” itself. The era saw a flourishing of substantive due process decisions for many reasons, including the increasing number of states; growing literacy and access to courts; intellectual fashions like the temperance movement that led to previously unheard-of legal controversies; profound changes in the way sovereignty was understood—and, of course, the clash over slavery itself. Throughout all these changes, the commitment to lawful, non-arbitrary rule remained in place. But this “explosion” is not evidence of some act of arbitrary special creation. Substantive due process has a strong pedigree in the principles of lawfulness and the political philosophy that informed the framers.

I’ve been defending a kind of reasoning, not any particular court decision. Thus it’s a bit beside the point to argue that the Court decided, say, Lawrence v. Texas, or Kelo v. New London, or even Dred Scott, rightly or wrongly. The point is that the Due Process of Law Clause requires courts to analyze whether the infringement on a person’s life, liberty, or property has been done in a lawful or non-lawful manner—and in making that decision, the court cannot avoid the ultimately value-laden analysis of whether the government’s act includes those elements (generality, rationality, public-orientation, and so on) that distinguish law from a mere act of force.

Ultimately, even many of those who say that the Due Process Clause only promises procedural protections are still doing just this analysis. Even those who think that the Clause only protects your right to a trial, for example, would probably concede that a coin-toss is not a “trial.” Why not? Because it’s arbitrary: it lacks the elements of fairness that make something a trial. And look at Prof. Rosenthal’s comments on Lawrence and Kelo. The statute in Lawrence, he argues, was not arbitrary, and neither was the condemnation in Kelo. I disagree, of course, but this is a substantive due process argument; the point is that the Due Process of Law guarantee requires us to have just that debate.

The reason is that the basic seed of everything we call “the rule of law” is the recognition of a gap between Law and the ruler’s Will. As in the famous Euthyphro dilemma, when such a gap is recognized, there must be criteria for determining which of the ruler’s actions do, and which do not, qualify as law. People can then deliberate about those criteria, and attempt to apply them in principled ways. And as that regime becomes more complicated, those criteria will interact in a matrix of increasing complexity. Some of these “rules of recognition” can be called procedural (a bill must receive a majority vote in both houses) and some can be called substantive (no bill establishing a religion is valid), but either way, in determining whether the government act qualifies as a “law” will require at least some determination of what law really means. And as Lon Fuller so beautifully explained, that is a question that can’t be answered without resort to normative considerations—many of which will be implicit, unspoken but inherent restrictions on the lawmaker’s power. That, in turn, means the judges, whoever they are, must make some value-laden decisions about what sorts of government acts do and do not deserve the appellation “law.”

Sadly, in modern times, it’s popular to declare that law simply is the will of the ruler, and that any more nuanced understanding of the subject is just an intellectually dishonest game, and one that clashes with democratic values. It’s sad that such simplistic cynicism is taken as a serious intellectual position, since as a matter of logic, it cannot be the case that the will of the ruler is the law. Take my still-unanswered question about whether an unintelligible statute is law. Where law is identified with the will of the ruler, the ruler becomes a kind of solipsist; power becomes an undifferentiated, impenetrable mass, making discussions over legitimacy and lawfulness literally impossible.

Consider some of the great literary depictions of lawful versus arbitrary rule. In the classic Twilight Zone episode “It’s a Good Life,” the villagers are governed entirely by Anthony’s will. Their society may be called many things, but it cannot rightly be called a lawful rule. Neither can the hellish village of Shirley Jackson’s “The Lottery,” even though it is governed by a regular, orderly procedure—a procedure that is utterly arbitrary and unreasoned, and exists solely because it’s always been that way. In both of these stories, law is a mere command, ultimately the ruler’s senseless ipse dixit. When the ruler governs by mere will, he may call his decrees “law,” but they are not law in fact—they are not general principles relating to the public good. By contrast, we see in The Oresteia the beginnings of lawful rule in the marriage of the passions with reason—not the domination or command of the passions, but a marriage, presided over by Athena and conducted through the medium of Persuasion. This is crucial, because law contains an element of reciprocity and of at least rough equality. Law is not an order; it is a promise. It is not mere will; it is reason. As Aristotle said, a lawful government “ha[s] a regard to the common interest [is] constituted in accordance with strict principles of justice,” whereas “those which regard only the interest of the rulers are all defective and perverted forms, for they are despotic, whereas a state is a community of freemen.”

Our Constitution was written light of this difference between lawful and non-lawful rule. It’s an attempt by the people to give law to themselves; an effort by the people to bind themselves while exercising their own power at the same time. That’s a paradox, but the framers attempted to resolve it with a multilayered process whereby the people indirectly choose judges and entrust them with the duty of enforcing the Constitution even when the people themselves might not want it enforced. Calling them “unelected judges subverting the will of the people” is really a demagogic way of mischaracterizing what judges do in such cases. Yes, they make wrong decisions—sometimes even corrupt ones. But that doesn’t mean the theory of substantive due process, an idea at the heart of so much that is great in the common law tradition, deserves to be treated with contempt and ridicule.