That the judge’s task is to find the law, not to make it, was once a commonplace of our legal culture. Today, decades after Erie, the idea of a common law discovered by judges is commonly dismissed—as a “fallacy,” an “illusion,” a “brooding omnipresence in the sky.” That dismissive view is wrong. Expecting judges to find unwritten law is no childish fiction of the benighted past, but a real and plausible option for a modern legal system.
This Article seeks to restore the respectability of finding law, in part by responding to two criticisms made by Erie and its progeny. The first, “positive” criticism is that law has to come from somewhere: judges can’t discover norms that no one ever made. But this claim blinks reality. We routinely identify and apply social norms that no one deliberately made, including norms of fashion, etiquette, or natural language. Law is no different. Judges might declare a customary law the same way copy editors and dictionary authors declare standard English—with a certain kind of reliability, but with no power to revise at will.
The second, “realist” criticism is that law leaves too many questions open: when judges can’t find the law, they have to make it instead. But uncertain cases force judges to make decisions, not to make law. Different societies can give different roles to precedent (and to judges). And judicial decisions can have many different kinds of legal force—as law of the circuit, law of the case, and so on—without altering the underlying law on which they’re based.
This Article claims only that it’s plausible for a legal system to have its judges find law. It doesn’t try to identify legal systems that actually do this in practice. Yet too many discussions of judge-made law, including the famous passages in Erie, rest on the false premise that judge-made law is inevitable—that judges simply can’t do otherwise. In fact, judges can do otherwise: they can act as the law’s servants rather than its masters. The fact that they can forces us to confront the question of whether they should—and, indeed, whether the Erie doctrine itself can outlive its mistaken premises. Finding law is no fallacy or illusion; the brooding omnipresence broods on.
Judges ought to remember, that their Office is Jus dicere, and not Jus dare; To Interpret Law, and not to Make Law, or Give Law.
This Article defends the view that unwritten law can be found, rather than made. Suffice it to say that this view is not in vogue. To modern scholars, law is always made by somebody: written law is made by legislators, and unwritten law is made by judges. The notion “that the common law had a positive source independent of judicial decisions” is said to have “no modern adherents.” Maybe Blackstone thought judges were not “to pronounce a new law, but to maintain and expound the old one”; today “[i]t would be only a slight exaggeration to say that there are no more Blackstonians.” Some still do assign the courts a duty “to say what the law is, not to prescribe what it shall be.” Yet the late Justice Scalia, who wrote those words, also took unwritten law to be “law developed by the judges,” and he viewed “playing common-law judge” as akin to “playing king—devising, out of the brilliance of one’s own mind, those laws that ought to govern mankind.”
Since Erie Railroad Co. v. Tompkins, many judges and academics have treated this modern approach as the only conceivable one. To them, Erie not only overruled Swift v. Tyson, but “overruled a particular way of looking at law.” Erie agreed with Justice Holmes that law “does not exist without some definite authority behind it”—and that courts rendering decisions, much like legislatures enacting statutes, establish new rules of law in a “voice adopted by the State as its own.” So Erie left no room for a common law to stand apart from courts or legislatures, or to be found instead of made. As one scholar put it, “Erie’s real significance is that it represents the Supreme Court’s formal declaration that this view of the common law . . . is dead, a victim of positivism and realism.”
If that’s what Erie declared, then Erie is wrong. A system of positive law, with fallible people as judges, can still expect those judges to find unwritten law and not to make it. That kind of system is a real possibility, not what John Austin called a “childish fiction” of the benighted past. Whether any legal systems actually work this way is an empirical question, which this Article doesn’t address; the claim here is that it’s plausible for a legal system so to arrange things.
In making that claim, this Article adopts a different strategy than other recent responses to the modern view. Some scholars contrast judge-made law to natural law or to moral principles, rules “out there” that judges might discover. Others distinguish it from bodies of “community custom,” like those of merchants or sailors, that judges might seek to preserve. This Article doesn’t contest those theories, but it doesn’t rely on them either. It sets a higher bar by focusing on positive law—law that’s “in some important sense a social fact or set of social facts”—and on law that binds society as a whole, not just a single tight-knit community. Finding this kind of law is impossible, the modern view argues, because there’s nothing out there to find: hence the derision of a “‘transcendental body of law outside of any particular State’” as a “fallacy,” an “illusion,” a “brooding omnipresence in the sky.” Even if law could rest on custom, societies are too diverse to share customs across the board; so their unwritten law has to come from some other source, like judicial decisions. If this law isn’t made by legislatures, then it has to be made by judges—for who could believe, to use Austin’s phrase, in a “miraculous something made by nobody”?
What’s strange about this argument, though, is that we follow somethings-made-by-nobody all the time. People routinely conform their conduct to familiar norms of fashion, etiquette, or natural language. These norms are addressed to society as a whole, and they’re generally perceived as binding, without anyone in authority having formally enacted them or laid them down. Just like legal norms, these social norms can sometimes be contested, changeable, controversial, political, or morally fraught. Yet in any given society, and at any given time, they can also have determinate content, offer broad guidance for the future, and stand apart from the style manuals or Miss Manners columns in which they’re expressed. If it’s possible for ordinary people to “find etiquette,” then it doesn’t seem strange that judges, lawyers, or private citizens might “find law.” In that case, the slogan that unwritten law is “whatever judges say it is” might be true only in the sense that standard English is whatever English teachers, dictionary authors, and copy editors say it is—with no copy editor or society of copy editors, however influential in practice, having any right to revise it at will. Positive law depends on social facts, but the social facts are “out there” for diligent jurists to find.
That leaves a second, realist strain of modern arguments against finding law: that courts are inherently lawmaking institutions. Even if there’s some unwritten law for judges to find, there’ll never be quite enough. To fill the gaps, judges have to make new law: “the process of adjudication necessarily entails articulating rules to elaborate and clarify” the law already on the books. And if making law is inevitable, then Erie got it right: whether a state’s rules “shall be declared by its Legislature in a statute or by its highest court in a decision” is only a matter of detail.
This Article again sets a higher bar by assuming, for argument’s sake, that some legal questions lack a right answer. Judges in unclear cases do have to make decisions. But we shouldn’t assume that, in making decisions, they’re also making law. A judge’s decision can have different force in different cases or different legal systems. And a decision can be influential, or even binding, in future cases without ever altering the underlying law. Instead, a system might instruct its courts to treat a past decision as if it stated the law, taking that decision as the law of the case, the law of the circuit, the dictate of precedent, and so on, without taking it to supplant whatever law was there before. The losing party in a traffic case might be estopped from later asserting that his light was green, but there’s still a fact of the matter; the light was the color it was, whatever future judges might have to assume. In the same way, doctrines like precedent or preclusion can serve as temporary stand-ins for the actual law, whatever it might be—settling certain questions among judges, without necessarily settling them right. The law is one thing, the decisions of courts another.
To be clear, this Article won’t defend any number of other views often associated—sometimes pejoratively—with a “declaratory theory of law.” (Say, that judges’ decisions all follow mechanically from precise legal rules, that judges are never influenced by policy or politics, that common law doctrines have all existed since time immemorial, etc.) Some of these are indeed matters of “childish fiction,” and in any case they’re irrelevant to the central point. It’s both possible and sensible to task judges with finding the law, though they retain their human failings even after donning robes.
If that’s true, it has real consequences for the American legal system. In other countries, the roles of state and federal courts might have little to do with legal theory; a federal system can allocate authority how it likes. But Erie’s account of American law depends quite heavily on the Court’s most abstract and theoretical claims. If law can be found as well as made, then Erie’s strongest pillar collapses, and the “Erie doctrine” itself—as to both state law and “federal common law”—may collapse as well.
Judges can only find law if there’s something there to find. The old vision of a body of unwritten law, already in place and ready to hand, is now widely seen as “a delusion” or a “fairy tale,” stemming perhaps from a “self-deceiving refusal to face the reality of legal decision making.” That “judges make the common law” is said to be something that “[a]ll lawyers know” and that “[e]very beginning law student is taught.” Open declarations that “judges had made up the common law” appear in the Federal Reporter with little comment and no dissent. Despite confirmation-hearing claims that judges can apply the law as it stands, many modern lawyers think otherwise. They assume that a state’s unwritten law is “fundamentally like the written law of each state,” though “made by a different branch of the state government: written law is made by legislatures and unwritten law is made by appellate courts.”
On the modern account, unwritten law is nothing but case law: a special kind of written law, found in judicial opinions rather than statutes. An opinion might not read like an enactment, and it might need some interpreting before it yields a general rule, but in the end the precedent is the source of authority. Vacate or reverse the judgment, and the legal rule goes away. As Kermit Roosevelt puts it, “the positive source of the common law is just the judicial decisions in which it is embodied.”
This vision of unwritten law is often associated with positivist views of law in general. On Abbe Gluck’s account, “the idea of a body of ‘natural,’ general, or universal legal principles”—something that judges might find rather than make—has given way “to a more positivistic understanding of law as something specific,” namely “a policy choice linked to a particular jurisdiction.” If that policy choice was first recorded in a judicial decision, then it stands to reason that the people who made the choice were judges, and that the unwritten law is whatever the judges say it is. For scholars like these, “[p]ositivism has thoroughly eroded” the notion “of a general law existing independently of any territorial sovereign”; the “positivist view” is simply “that judges ‘make’ new law.”
Accordingly, a few recent defenses of finding law have begun by questioning positivism. Allan Beever, for example, describes past exponents of the declaratory theory as arguing that the common law “include[s] the natural law.” Judges don’t make the natural law, so they might not make the common law either. Gerald Postema charts a different path, describing the classical conception of the common law as a “process of practical reasoning.” Because this process looks past the propositions stated in judicial opinions to the reasoning that underlies them, he sees his view as “incompatible with both orthodox natural law thought and with orthodox legal positivism.”
These responses, right or wrong, rest on stronger assumptions than necessary. Not all positive law is necessarily posited—“set, or prescribed, . . . laid down by humans to humans,” in the form of explicit “statutes [or] court decisions.” Positive law might just be like other normative systems, such as grammar, etiquette, or fashion, which are solidly rooted in social facts without having been formally adopted by anyone. No one disparages natural language as a “miraculous something made by nobody.” And to borrow H.L.A. Hart’s phrase, it’d be “merely dogmatic”—indeed, rather absurd—to say that nothing can be a rule of grammar “unless and until it has been ordered by someone to be so.”
Other defenses of finding law focus on preserving popular custom. For example, if merchants traditionally allow each other three “days of grace” before payment, the law might take the custom of that community into account when resolving disputes. Legal officials might then look to “the usual or ordinary understandings of parties to a commercial transaction,” or what the Uniform Commercial Code calls the “usages of trade,” as something “out there” to find.
These defenses, too, are more limited than they need to be. A society can have a customary rule without many members of that society needing to take part in the custom. Plenty of nonlegal social conventions are highly obscure in practice, like the forms of address for various dignitaries or the “popular names” of minor constellations. But they’re still bona fide social conventions, matters of widespread agreement among those widely believed to know such things. In the same way, the doctrine of anticipatory breach or the rule against perpetuities might be known primarily to an elite group of legal experts, yet still be part of a customary law that belongs to society as a whole. Judges bound to apply that law might then be expected to apply these prevailing standards without alteration: to find the rules, and not to make them.
Finding these prevailing rules may not be easy. There are difficult questions of judgment in extracting a particular customary rule from a diverse society, in tracking changes in a custom over time, and so on. But in contexts other than law, few would call these tasks impossible or incoherent. People who can’t explain how prevailing norms are grounded on complex social facts can still tell you whether a given outfit would be out of place at an important business meeting, or whether an ordinary English phrase would be ruled out-of-bounds in English class. When it comes to fashion, etiquette, or grammar, we routinely distinguish everyday practice from the prevailing standard, usually without thinking.
And while these kinds of social norms may seem fuzzy or indistinct, at least compared to the extraordinary technical detail of modern legal systems, there’s nothing mysterious about law resembling or resting on social norms like these. On what Mitchell Berman calls the standard positivist picture, societies can produce any number of “independent artificial normative systems,” of which law is a wholly “non-exceptional” example. We regularly distinguish what’s customary from what just happens (the “done thing” from what’s frequently done), or separate “hard” customary obligations from “soft” ones (say, bad grammar from bad writing). We can equally well distinguish legally binding norms from the informal social customs of a legal elite, drawing a line between intestate succession and judges’ wearing black robes. Whatever faculty helps us sort out one custom from another can help us sort out law from custom. And just as we sometimes find it useful to designate particular people to apply our customary standards—say, the grader of a college entrance exam, or the amateur referee of a pickup basketball game—we might appoint particular officials to apply standards that are already taken as legally obligatory. A rule of customary law doesn’t have to wait for a judge’s ruling to make it so, any more than a rule in a pickup game waits around to be born when a referee first applies it.
Theories of positive law don’t always agree. But it’s far from clear that anything in the nature of law would prevent us from finding it. If anything, there’s good reason to think that unwritten law is regularly found in practice. Indeed, without it, courts couldn’t make many familiar types of decisions, including the federal “Erie guess” about the content of state law. That should make us appropriately skeptical of Erie-based arguments against finding law: if judge-found law didn’t exist, Erie would have us invent it.
We live in a world chock-full of social rules. How we dress, how we act, how we write: all these are governed, not merely by practice or rote habit, but by shared standards of assessment and criticism. These shared standards are more than just coincidences of judgment or taste. Many of them are social objects—norms that we identify, accept, and apply together—and they profoundly shape our everyday lives, without any authoritative procedure to enact them or give them force.
In the legal world, though, the idea of an authoritative procedure dies hard. The way we identify rules of grammar or etiquette seems too wishy-washy to be useful for identifying rules of law. To Mark Greenberg, for example, “[t]elling lawyers to look to customs . . . doesn’t take us very far,” at least not without answers to some very basic questions: “Which customs matter? Do customs of ordinary people count for anything? Of ordinary lawyers or only certain elite lawyers?” For that matter, how do unwritten rules ever change, unless someone changes them? And how can judges declare this changing practice, without deciding the answers themselves? These questions are sharp ones, and they’re made sharper by a legal system that demands hard-and-fast answers—under pressure from talented advocates—to what might otherwise be questions of degree.
At the same time, though, the questions are hardly unanswerable. Most people might agree that the law can incorporate ordinary popular customs: say, providing by statute that “a single hand of five-card draw poker” will be used to break ties in local elections. If the parties later dispute whether three of a kind beats two pair, they won’t be disagreeing about what the statute means, but what the rules of poker actually are—rules that no one ever laid down or enacted. As a matter of procedure, these disputes might end up being answered by various judges or election officials. But it’s still the external social practice, and not some official decision, that makes the answers right or wrong. (The statute says “poker,” not “poker, as it subsequently may be defined by judges and election officials.”)
If customary rules like these can exist and be given legal force, then the same might be true of other customary rules specific to the legal system. A statute or procedural rule might invoke the traditional set of remedies in equity, the traditional grounds for awarding new trials, or the traditional rules for the issuance of writs, in the same way that it might invoke a traditional ranking of poker hands. Or it might incorporate an entire body of traditional rules wholesale—such as by declaring “the common law of England” to be “in full force” in Virginia.
Whether traditional rules can have legal force on their own, or whether they first need the blessing of a statute or sovereign pronouncement, is discussed below. But despite the many questions surrounding customary law—what it is, how it arises, who it belongs to, how it changes, and so on—it seems entirely plausible that a legal system might include customary rules like these. Or, at least, this seems no less plausible than the idea that poker players can reliably identify and apply the rules of poker. In other contexts, for other systems of customary norms, we manage to solve these problems every day. And if these problems are surmountable in other areas of life, then maybe they should trouble us less with respect to law.
To find unwritten law, we have to know where to look. “Written” law might be unclear, but it’s no mystery: the statute says what’s in the statute book, and it’s law because the legislator said so. But “unwritten” law has an air of obscurity. If there’s nothing authoritative for us to read, then what’s the law, and where does it come from?
The mystery is made worse by Blackstone’s ecstatic descriptions of the common law—as “the universal rule of the whole kingdom,” discerned by judges as “depositaries of the laws” and “living oracles,” who “do not pretend to make a new law, but to vindicate the old one from misrepresentation.” There’s a strong temptation to view these as fairy tales, or even as deceit: a con job by judges, who never really set their “own private judgment” aside when declaring “the known laws and customs of the land.” Hence the critics’ mockery of the “transcendental body of law,” the “mysterious something made by nobody,” and so on.
But the mystery dissolves once we remember how good we are at identifying unwritten rules. Language, fashion, etiquette, and other customary systems are all unwritten in this way. A rule of standard English might not be “the universal rule of the whole kingdom,” but it’s close: with surprising consistency, we spell words correctly, compose full sentences, follow shared norms of grammar and word order, and so on. What’s more, we do this without any authoritative list of social rules analogous to the Statutes at Large. A social norm can be taught and transmitted through writings (textbooks, fashion magazines, Dear Abby columns, and so on), without being founded on these writings; no grammar book establishes rules of English in “the way that statute books establish rules of law.”
What makes these rules “unwritten” isn’t whether they can be expressed in words—they can—but why the words matter. A statute would still be “written” if it were “a string of ones and zeros in ASCII format,” or “a set of interpretive dance steps,” or even “if we all just memorized it, taught it to our children, and then burned the National Archives.” It would “still contain particular terms, adopted on a particular occasion, that carry legal significance by virtue of their adoption.” Rules of grammar or etiquette, by contrast, don’t have to be adopted in any particular way; they’re not enacted through some procedure with validating effect, as prescribed by yet another set of rules (what Hart called “secondary” rules). Instead, grammar and etiquette rules typically rest on a general practice that different people can describe differently, with any standard formulations (like “i before e, except after c”) useful only insofar as they get the practice right. The authorities in these fields are epistemic ones, “living oracles” and “depositaries” of practice like Ann Landers, your third-grade teacher, or Strunk & White. Their job is to report the “customs of the land” without “pretend[ing] to make a new [rule],” just as almanacs report the tide schedule without pretending to command the tides.
Unwritten law can work the same way. Brian Simpson saw “[f]ormulations of the common law” as resembling “grammarians’ rules, which both describe linguistic practices and attempt to systematize and order them.” The rules discussed by Blackstone—say, that “there shall be four superior courts of record, the chancery, the king’s bench, the common pleas, and the exchequer”; that “wills shall be construed more favourably, and deeds more strictly”; that “money lent upon bond is recoverable by action of debt”; or that the crime of burglary “must be by night”—are all written down in his book, but they aren’t founded on any particular writings. They’re just things that competent lawyers were supposed to know. Even today, without any statute to tell us so, we know that duress is a defense to certain crimes, that the defendant has the burden of proving it, and so on; a judicial decision might illustrate the rule, but the rule long predates the decision.
Because these rules are unwritten, they don’t always need to be expressed in particular terms, so long as there’s agreement on their content. Six torts professors might give six different explanations of res ipsa loquitur; if they’re similar enough to “secure general agreement,” that’s all that matters. Some unwritten rules do have a classic formulation, like John Chipman Gray’s version of the rule against perpetuities (that “[n]o interest is good unless it must vest, if at all, not later than twenty-one years after some life in being at the creation of the interest”). What makes the formulation classic isn’t some special legal power vested in John Chipman Gray, but the successful reception and use of his formulation by subsequent generations of law students—just as the standard lyrics to “Jingle Bells, Batman Smells” are determined by playground practice, and not their attribution to some long-lost author. The formulation is meant to summarize an existing practice, and it’s only good law to the extent that it’s a good summary.
So we might well identify unwritten legal rules largely as Blackstone did: by their “long and immemorial usage, and by their universal reception.” The adjectives have to be taken with a grain of salt. Neither grammar rules nor Gray’s formulation have really been used always and everywhere. Yet they may still reflect the practice here and now. What’s important is that the relevant norm is drawn from current practice; it isn’t dependent on some special authorizing event for its “original institution and authority,” the way that “acts of parliament are.”
Custom is said to arise from current practice and also to create rules for the future. How do we get from the ‘is’ to the ‘ought’? If milkmen were to regularly “adulterate the milk supplied to their customers,” Dickinson asked, would that create a custom of milkmen—or even a customary law? Maybe, as Jeremy Bentham wrote, “a law is to be extracted [from the cases] by every man who can fancy that he is able: by each man, perhaps a different law.”
To solve this problem, customary law has long been said to demand two things: that there be a widespread practice, and that the practice be followed from a sense of obligation (opinio juris). This definition strikes some as circular, even “mysterious”: “the legal obligation is created by a . . . belief in the existence of the legal obligation.” So critics have offered various other explanations for these practices—for example, that the actors are motivated by self-interest instead.
For practices of language or etiquette, though, opinio juris makes a lot of sense—and seems in accord with recent research on social norms. A speaker who conjugates her verbs doesn’t have to be motivated by a love of conjugation or of rule-following; she might have perfectly ordinary and self-interested reasons, such as wishing to seem well-educated or to be understood. Still, she’s acting in full compliance with the rule, and if questioned she might cite the rule to explain or justify what she did. As Hart explains, we don’t need to delve into psychology to explain why “chess-players will move the bishop diagonally,” so long as the rules matter in guiding, explaining, and justifying what they do. Opinio juris reflects whether they take a putative social rule as a rule, distinguishing “the adult chess-player’s move from the action of the baby who merely pushed the piece into the right place.” (As Leslie Green notes, the most powerful rules are the ones we obey without even thinking: “Few men wake up in the morning, mentally rehearse the gender-rules about dress, and then put on trousers instead of a skirt in a deliberate attempt to conform to that norm.”)
So practice-plus-obligation fits our norms rather well. What counts as standard English may vary over time, but it’s not just a corpus-linguistics catalog of whatever words people happen to use. Instead, it’s a normative practice—a practice of following a particular set of social rules, complete with do’s and don’t’s, accepted standards of behavior and shared grounds for criticism. The longstanding battle between “descriptivists” and “prescriptivists” overlooks the fact that we always act in both roles at once: we can describe our system only in terms of the prescriptive norms in current use. Whatever our statistics on word usage (“Fifty-two percent of respondents approve of the singular ‘they’”), there might be no magic threshold that suddenly counts as general acceptance, just as there’s no magic number of sand grains that suddenly count as a heap. Still, we understand that some usages are generally permitted in our language while others are not. Someone grading the Advanced Placement English Language and Composition exam has to decide what counts as a “lapse in diction or syntax,” something that’s unintelligible absent a shared understanding of standard English as ruling some usages in and others out.
All this can also be true of law. Just as English-speakers don’t have to like formal English, and just as store clerks don’t have to believe that “the customer is always right,” law-abiding citizens don’t have to be motivated by their admiration for various rules of law. They only need, when the point is raised, to understand the customary rules as rules, and not just as things that lots of people do.
To identify a rule from social practice, we need to know whose practice matters. Sometimes customary law draws on the practice of a regulated community, like that of merchants, ranchers, or nation-states—distinct communities whose informal norms might be absorbed into the formal law. To critics of this model, including the early twentieth-century scholar John Dickinson, such a “customary theory of law breaks down in a complex society.” For the legal system as a whole, “practically no legally pertinent customs are universal, but nearly all are partial, fluid, conflicting.” Matt Adler similarly contends that the modern world is too diverse, its social norms too specific to individual groups, for there to be shared norms that define the law. To Adler, there’s no single community that defines dress norms for all of Manhattan, and perhaps no single community that defines legal norms for the entire United States.
Social practice is very diverse. But for nonlegal norms, there’s often one practice that dominates the others, for better or for worse—and we often have no trouble identifying it. No matter how varied the fashions on the Manhattan subway, in a lineup most people could pick out those dressed “professionally.” And no matter how polyglot the city of New York, we can still distinguish (in John Fisher’s words) between the “formal, official language in which business is carried on” and “the various casual dialects of familiar exchange.” People who rarely use standard American English can still identify it as the standard—or at least can identify the elite practitioners whose opinions set the standard. Nothing about this process need be democratic, or even all that fair, for it to be a distinctive feature of a society.
The same thing happens in law. What’s customary for ordinary people, the custom in pays (“in the country”), coexists with a more specific custom in foro (“in the court”)—in Simpson’s terms, a “body of traditional ideas received within a caste of experts.” It’d be silly to treat the rule against perpetuities as a popular custom, like eating with knife and fork; but it makes perfect sense to place it among the internal customary practices of a legal elite. As Joseph Beale saw, unwritten law reflects the “body of principles which is accepted by the legal profession”—shaped by the “teachers of law,” the “expressed opinion of writers,” and “the argument of practicing lawyers.” The relevant customs are those of jurists, or of the expert legal class.
Because these are our jurists, there’s also a clear sense in which custom in foro belongs to society as a whole. We often rely on what Hilary Putnam called a “division of linguistic labor”: people with no idea how elm trees differ from beech trees can still talk about them as separate kinds of trees, trusting that expert botanists will know the difference. We equally rely on lawyers or grammar snoots to know their field’s rules and to tell us what they are. (And these groups, too, sometimes contract out to a smaller class of specialists—say, admiralty or tax lawyers, whose views on admiralty or tax law are taken as those of the legal profession as a whole.) What makes standard English “standard” is precisely this sort of incorporation-by-reference: it’s enforced as everyone’s “good grammar,” not just as the dialect of one cloistered group. The same process can turn the legal customs of elites into the legal customs of society at large. So long as we can pick out the experts, and the experts can pick out the rules, the rules still belong to us all.
(Again, this is true whether or not the process is all that democratic or fair. The rules belong to “us,” not in the sense that we ourselves would choose them—they might be terrible—but in the sense that they happen to be the rules of the society in which we live. That society, to borrow Hart’s phrasing, “might be deplorably sheeplike; the sheep might end in the slaughter-house. But there is little reason for thinking that [such rules] could not exist,” or alternatively for “denying [them] the title of a legal system.”)
For some legal rules, the “us” extends quite far. Judges once claimed to find, not just the common law of New York or South Carolina, but the common law: something that transcended particular jurisdictions while deciding real cases within them. That kind of law might seem hard to base on social facts, and easy to mock as a “brooding omnipresence”—until we remember that customs aren’t confined by political boundaries, and that they can be shared across borders and across cultures, too. Whatever the linguistic differences from Manhattan to Myrtle Beach, English teachers across the map generally aim at enforcing shared standards, making it strange to deny (per Justice Holmes) “that there is this outside thing to be found.” And just as American and British English are variations of a transatlantic language (indeed, a global one), the common law applied in New York might be mostly the same as that applied in Charleston or London, with all three jurisdictions drawing from a single transatlantic well. In Beale’s day, courts often used the term “common law” without indicating which kind they meant; but this loose talk is no more unusual than saying that Americans speak “English,” without mentioning the many other languages spoken here or the many kinds of English spoken elsewhere in the world. If it’s the local custom of the jurisdiction to apply a more general customary rule—supplemented, perhaps, by other local variations or local usages—then there’s nothing mysterious about this general custom being “outside of any particular State but obligatory within it.”
Finding custom is one thing; accounting for change is another. “If judges never make any laws,” an American realist once asked, “how could the body of rules known as the common law ever have arisen, or have undergone the changes which it has?” On the one hand, it’d be silly to claim that the common law “exist[s] . . . from eternity,” as Austin mockingly suggested—or, to quote a 1910 doggerel, “[t]hat there prevailed in Babylon / [t]he law of motor cars.” On the other hand, accepting changes in the common law might mean accepting that judges change it, amending the law every time they overrule a prior case.
This dilemma resembles what international lawyers call the “chronological paradox”: how could a custom possibly get started? If custom is practice-plus-obligation, the very first person to engage in a practice couldn’t have been required to do it by custom. Yet if she didn’t act from customary obligation, then her practice couldn’t have established a custom to obligate the second person, and so on. One response to the paradox is to deny that any custom exists; the other is to have the courts impose it themselves, as “a binding norm going forward” that’s “socially and morally desirable.” For any common law rule, the Supreme Court reasoned in 1907, “there must . . . be a first statement,” which will be “found in the decisions of the courts,” and which “presents the principle as certainly as the last.”
As applied to nonlegal norms, though, the “paradox” loses its bite. The first person to wear a skinny tie in 1950s America didn’t respond to what was already the fashion; nevertheless, at some point, some other people did. The same goes for changes in accepted spelling—say, from Chaucer’s time to Shakespeare’s to today. Maybe the first American to drop the ‘u’ from ‘behaviour’ misunderstood the custom; or maybe it was just a printer’s error, or a deliberate attempt at subversion on the part of Noah Webster. Eventually, the new spelling caught on, and then became de rigueur. Likewise, the “early adopters” of a standard might have been mistaken about its popularity, or perhaps they deliberately set out to enforce a new rule; but none of that matters to the strength of the custom today. Neither an English teacher nor a person getting dressed in the morning needs to resolve any philosophical problems before they can follow the current practice.
Because it’s founded on practice, a custom can change without anyone needing authority to change it. No one has to issue a decree on tie width for the fashion to evolve over time. It just does, as people haphazardly revise their beliefs and actions. If Vogue editor Anna Wintour declared that skinny ties were “in,” she’d at best be trying to hasten this change in practice—to “make ‘fetch’ happen,” so to speak. Given her position, she might be successful. But a causal power to influence others isn’t the same as a norm-conferred authority to legislate. Wintour might persuade others, or she might not; but there’s no preexisting social rule authorizing the editor of Vogue to establish new rules of tie width at will.
At this point, the legal analogy should be clear. Unwritten law certainly changes over time, as a function of changes in how lawyers and officials understand the law. Courts are especially well-equipped to bring about these changes, even when their opinions aren’t binding—as was the case for many prominent district court opinions. But this, again, is a causal and not a legal power. George W. Bush and Will Ferrell between them managed to get “strategery” into the Oxford English Dictionary, but no one lists this among the powers vested in the President by Article II.
To criticize these changes for their lack of authority, as Dickinson did, is to recite the chronological paradox again. Unlike written law, custom doesn’t need to be enacted. Once enough of the right people drop the ‘u’ from ‘behaviour,’ or treat parol contracts as requiring consideration, a good dictionary editor or legal treatise writer is obliged to recognize the new standard. Tracing the exact development of the change over time would be burdensome and pointless; it even bored Blackstone, who found nothing “more difficult than to ascertain the precise beginning and first spring of an antient and long-established custom.”
In fact, worrying too much about the origins of a particular rule is a symptom of ignoring the distinctions between written and unwritten law. Rules of written law trace their validity to an initial enactment, made by particular people in a particular way; as Simpson writes, a statute “is both the only reason and a conclusive reason for saying that this is the law.” This may be the normal way of thinking about American constitutional law, which gives pride of place to written sources. But unwritten law, like a natural language, derives its content from usage today, not from whatever happened a long time ago. It simply doesn’t matter which obscure case, “decided say in 1540,” was the first to hold “that parol contracts require consideration”; the rule doesn’t “derive its status as law today from this antique decision,” and the old decision wouldn’t even be “good authority for the rule” in a modern brief. Nor does it matter whether today’s law of choses-in-action started with some shenanigans pulled by Lord Mansfield in the eighteenth century. The existing doctrine might have been “judge-made” in some causal sense, but it’d be deeply misleading to view that doctrine as “the product of a series of acts of legislation” by unremembered judges. Its current validity rests on current acceptance; Lord Mansfield has nothing to do with it.
Living with past judicial shenanigans doesn’t mean giving carte blanche to future ones. At any given time, officials are obliged to conform to the law as it stands. Past changes in law don’t offer legal ground for new departures, any more than past changes in spelling license each English teacher to invent some more. Dickinson criticized “the paradoxical conservatism . . . of the historical jurist, which holds that because law has continually changed in the past it is somehow impossible to change it in the present.” But changing the law plainly isn’t impossible; it might just be forbidden by other rules of law. A legal system can coherently give legislative power only to certain officials (like legislators), while expecting other officials (like policemen or judges) merely to apply that law, whatever it is and wherever it came from. The fact of change over time doesn’t mean that there’s no legal standard to apply right now; and if there weren’t any standard to apply right now, we couldn’t talk coherently about change over time.
Analogizing law to social rules is only the first step: we still have to decide which social rules really are rules of law. As Mark Greenberg argues, the same facts can be used to support many different and conflicting norms. So which facts produce legal norms, and which norms do they produce? Whatever norms we pick might seem a deliberate choice of the judges, not the result of some dispassionate process of discovery. And if law is something wielded by real officials, rather than just brooding omnipresently in the sky, then perhaps (the modern argument goes) it’s the officials who get to make the law.
In fact, ordinary custom can do plenty of work on its own, without first needing an official pronouncement to transform it into law. Facts about social practice might be consistent with multiple legal rules, but that’s true of all customary rules; we should be wary of any argument that’d make it impossible to identify custom at all. Similarly, ordinary customary rules can share all the hallmarks of legal rules—establishing complex hierarchies of rules, empowering officials, addressing topics of great moral weight, and so on. If we can reliably identify distinct bodies of custom, then we can also identify a distinct body of legal rules, even before the judges get their hands on it. And it seems unlikely that only written sources (like statutes or judicial decrees) can give rules their legal status, for then we’d have to ask what makes those written sources count. A system that relies on unwritten rules for recognizing particular pieces of paper as “statutes” or “constitutions” can just as easily recognize particular customary rules as “rules of common law.” Maybe something in the nature of law requires custom to be blessed by judges before it can serve as part of the law; but that claim is highly controversial, and there may not be much reason to believe that it’s true.
Whatever our social practices might be, Mark Greenberg argues, in theory they could reflect any number of social rules. Just as dots can be connected by an infinite number of curves, a given set of cases might establish any number of “lines” of case law, if we use “a non-standard or ‘bent’ model” to read them. For example, how can we be sure that the doctrine of quasi-contract doesn’t lapse on December 31, 2049? That kind of “bent” rule seems like “a non-starter”; but how can we rule it out, if all of the cases on quasi-contract are from before 2050? Everything the legal system has done to date—including every discussion of quasi-contract in a law school classroom, and so on—occurred well before 2050, and so is equally consistent with both rules.
To Greenberg, if all we have is social practice, then it’s not clear what keeps these “bent” rules out. A bunch of data points on a page won’t spontaneously fit themselves to a line; we have to do that ourselves, and that means making choices. Something has to limit which aspects of a practice are or aren’t legally relevant (say, “occurring before 2050”); but those limits can’t be supplied by the law, as ‘the law’ is precisely what we’re trying to discover. We might just take those limits as brute facts, but that doesn’t seem right either; there ought to be intelligible reasons why people’s actions and beliefs have the legal consequences they do. Yet if we look only to facts about practice, it’s not clear what those reasons are.
This is a powerful argument, but maybe it’s too powerful. As Greenberg recognizes, it isn’t “limited to the law”; it follows the same structure as other well-known problems of language and induction, and it applies just as much to other social rules. (Do chess bishops move diagonally only before 2050? If not, why not?) Without reviewing other possible answers to these puzzles, it’s enough to note that, if the problems have the same structure, then they ought to be handled with the same solutions. If practice generates real social norms for chess, then maybe it can do the same for quasi-contract.
In fact, it may be crucial to Greenberg’s theory that the general problem be solvable. He would rule out “bent” interpretations with “value facts,” which privilege certain ways of inferring obligations from practice. A statute’s legal content might rest in part on ordinary linguistic usage, because that’s evidence of what Congress intended or how the provision is reasonably understood, and those things matter for reasons of democracy or fairness. A new ordinance that “cars must drive on the right” might affect our obligations by changing how other people are likely to act, which matters for other things we value (like preventing accidents). But to conclude that there is such a thing as ordinary linguistic usage—or that we have a social norm of driving on the right, that others reading this statute are likely to keep doing so, and so on—we have to extract general rules from past practice. Simply to apply the value facts correctly, we need to rule out in advance certain “bent” interpretations of our linguistic or social practices (say, “use ‘right’ and ‘left’ in the traditional way only until this afternoon, then flip”). And whatever method Greenberg uses to rule those out, anyone else can use too. Customary law, like other kinds of custom, can be rooted in practice even if we can’t spell out the precise relation between the two. It can be intelligible enough for our purposes, without needing to be intelligible all the way down.
Language, etiquette, and fashion are all systems of social norms. Is law different in kind? Drawing analogies to these other systems might threaten to blend them—leaving us unable, say, to tell whether “curb your dog” or “don’t wear white after Labor Day” are rules of law or not. Something has to keep law and custom apart, and maybe that something is the courts.
In our society, law is marked by its involvement with formal proceedings and authoritative pronouncements, along with a vast and complex structure of powers, immunities, and officials. No one is arrested by the fashion police (yet), but law speaks with the sovereign’s voice and is backed by the use of force. That’s why Greenberg discounts analogies to “rules of practices (including organizations, games, and so on),” because “familiar practices, such as etiquette, . . . have no equivalent to legal officials, let alone to the acceptance by officials of a rule of recognition.” Even Blackstone thought the “customs or maxims” that “shall form a part of the common law” were distinguished by their being “known,” and their “validity . . . determined,” by “the judges in the several courts of justice.” Is what differentiates law, then, the fact that officials make it?
As it turns out, legal and nonlegal norms are remarkably similar. Almost every important feature of legal norms—including secondary rules, reliance on officials, or the morally significant use of force—can also be found outside the law. What truly differentiates legal norms from social norms might just be what differentiates different kinds of social norms from each other: their use in particular fields and their application to particular problems.
No less than legal systems, nonlegal groups often adopt secondary rules—rules about rules—in addition to their ordinary rules of primary conduct. Informal clubs or student associations sometimes follow customary rules of practice, and some of their rules incorporate other authorities by reference: for example, the traditional pride of place given to Robert’s Rules of Order. Natural languages, too, can have these secondary rules. It might be that a customary rule of the French language authorizes the Académie française to define or exclude new words (like “le email”); recognizing the Académie’s authority would then be part of what’s required to be a competent French-speaker. There’s little distance between the informal social recognition that “what Robert’s Rules prescribe is proper procedure,” or “what the Académie approves is proper French,” and a recognition by legal officials that “what the Queen-in-Parliament enacts is law.”
These kinds of authoritative resolutions are especially common in law, which gives legal disputes a kind of crispness that disputes over grammar or fashion usually lack. (How would a “grammar court,” hearing a “grammar suit,” even begin to weigh the expert testimony of disputing authorities?) But treating that crispness as unique to law is largely a sleight-of-hand, because it presupposes a thoroughgoing consensus about other social rules, including those which specify the relevant legal sources. If our grammar or fashion norms were as determinate as all that, we’d have no trouble resolving grammar or fashion disputes. Even the technical, artificial languages used in computer programming—which might leave no room for doubt about proper syntax or interpretation—rest ultimately on a consensus within various programming communities about which features of the language are canonical, and which are idiosyncratic “forks” or departures from the norm. Law is no more or less subject to disagreement than anything else. And, indeed, when the sociolegal consensus breaks down—as happened during the Dorr Rebellion, when two rival governments claimed power in Rhode Island—we might well feel more confident in our judgments of contemporary fashion than in many judgments of contemporary law.
Reliance on officials is also very widespread, even in customary domains. Once we identify shared norms, we regularly task some persons or institutions with applying them. A teacher hired by the College Board to grade AP English essays is expected to ignore her own preferences or pet peeves and to grade the exams according to common standards. Even the unorganized group of friends who choose a referee for a pickup basketball game are empowering a neutral official, by means of a social rule, to render authoritative resolutions of disputes.
So while official constructions of law do carry an authority often lacking in private applications of social rules, this is at most a difference in degree and not in kind. Hart famously compared judges to the official scorers of games, who have authority to apply the rules and to make binding determinations. This in-game authority entails that “the score is what the scorer says it is”; but it doesn’t eliminate every other rule, collapsing the game into one of “scorer’s discretion.” A player without such authority can still privately “assess the progress of the game,” using the same rules that the scorer would “find . . . established as a tradition and accepted as the standard for [his] conduct.”
In the same way, the fact that customary legal rules have to be applied by judges doesn’t entail that the judges get to make the rules. Instead, they might be expected to apply rules already known and established by custom. A customary legal rule might eventually fade away if it weren’t accepted by most of the judges, most of the time; but that doesn’t make the judges who use the rule its author, any more than a basketball referee is the author of the three-point line. (A $5 million award in Maine recently turned on the absence of an Oxford comma in the statute; certainly judges didn’t create that rule, even if they could decide whether or not to apply it.) Richard Ekins points out that a court need have “no more authority than any other subject of the law to interpret” what the law requires; its job might simply be to resolve disputes, with interpretation only becoming relevant when the disputes turn on contested understandings of the law. The rest of the time, the court is engaging in essentially the same activity as the lawyer who renders opinions in her office, or the Monday-morning referee who watches the replay and questions the call.
Nor is it a significant distinction—for these purposes, at least—that legal norms are frequently controversial, comprehensive in scope, of great moral weight, or characterized by the use of force. Most of the time, there’s not much morally at stake in linguistic practice; people can talk however they want. But some language use might be highly controversial, as in the case of gender pronouns or the singular “they.” Law’s broad field of application isn’t unique either; for example, virtually every human activity might fall within the domain of some field of etiquette. And other social norms can also give rise to urgent moral concerns: think of the code duello, or “honor killings,” or the bloody unwritten rules of Jim Crow. People have always used violence to impose and enforce social norms, whether or not those norms had the name of law. The fact that legal norms sometimes give the state a monopoly on violence may be a point in their favor, but it doesn’t make them fundamentally different from other sorts of social rules.
That shouldn’t surprise us, because on the standard positivist picture, legal rules simply are social rules, or are indirectly derived therefrom. We can usually tell social rules apart without any kind of formal apparatus; understanding a custom means understanding what the custom is not. We can tell fashion from etiquette simply by knowing their respective domains, notwithstanding the occasional edge case where the same conduct breaches both (like wearing a long white dress to a wedding); and we can tell bad grammar from bad writing simply by knowing the conventions applicable to each. We’re already pretty good at distinguishing Shakespearean English from modern English, or kickball from dodgeball, or the ordinary rules of etiquette from the nonlegal conventions specific to law or politics (judicial robes, faithless electors, State of the Union addresses, pre-Roosevelt limits on presidential terms).
Occasional edge cases notwithstanding, then, it needn’t be any more difficult as a theoretical matter to distinguish ordinary social rules from rules of law: say, to distinguish ordinary gossip from common law defamation, or aggressive driving from a moving violation. To delineate those differences, we might look to treatises or court decisions, but that doesn’t mean the judges or the treatise writers made them up. The custom in foro and the custom in pays are different customs, and we distinguish them the same way we distinguish all the others.
That might sound a little hand-wavy, but it’s borne out by our ordinary experience. Divisions between normative systems aren’t always cut-and-dried, but they do real work, and they can do real work for law. Berman, for example, suggests that legal systems are marked out by their connections to politics, serving as “political communities’ normative Swiss-Army knives.” If so, they’d inherit all the messiness involved in distinguishing political communities from many other overlapping social structures—including those defined by class, culture, profession, religion, region, or ethnicity. Yet the field of political science hasn’t collapsed for lack of borders, and neither has the law. People don’t need a crisp set of necessary and sufficient conditions to know a legal rule when they see it: the degree of legal obligation attached to a custom in foro may be something the custom itself tells us.
Whether unwritten rules are really rules of law might depend on what counts as law in the first place. If your theory grounds law in official action, as many positivist theories do, then you might insist (with Dickinson) that a custom has legal force “only when and so far as the courts have determined to accept it as such.” If courts get to choose which customs are legally binding, then they might be said to choose the law—making it hard to say that this law is found rather than made.
But are those theories correct? Some claims about law, like that it’s always the command of a sovereign or a prediction of official action, are now thought far less persuasive than they seemed in Justice Brandeis’s day. And other claims, such as that only legal institutions can distinguish legal rules from nonlegal ones, remain unproven at best.
Erie itself, citing Justice Holmes, relied on an Austinian command theory: law is the command of the sovereign, and it speaks with the sovereign’s voice. So a common law rule needs “some definite authority behind it,” such as a declaration of the sovereign’s court. Since Erie, though, Austin’s theory has come under withering scrutiny. As Brian Bix describes it, the theory is now “almost friendless, and is today probably best known from Hart’s use of it as a foil.” On Hart’s account, the custom comes first: the written rule is special only because an unwritten rule makes it so. To paraphrase an argument by Stefan Sciaraffa, a judge’s decree that a custom has legal force would still “be a dead letter absent a custom among the system’s legal officials of conforming” to the judge’s decrees. That custom continues to matter even after the decree issues, because whether the decree “is live or a dead letter comes in degrees.”
Indeed, in a regime in which the very existence, identity, and jurisdiction of the courts was determined by rules of customary law—say, that “there shall be four superior courts of record, the chancery, the king’s bench, the common pleas, and the exchequer”—it’d seem rather circular to base these rules’ legal force on their having been announced by the courts they constitute. The very idea of appellate precedent (with distinctions between holdings and dicta, binding and persuasive authority, and so on) itself trades on judges’ fidelity to unwritten rules, despite an awareness that their superior courts simply “can’t reverse everything.” In the end, we can’t do without customary rules.
Elsewhere Justice Holmes suggested a different theory, that law depends on official action because it’s just a prediction of how officials will act: “The prophecies of what the courts will do in fact, and nothing more pretentious, are what I mean by the law.” That theory is popular among cynics as well as practicing lawyers, who get paid to advise clients on what judges might do. But among legal experts, it’s also “nearly friendless,” with some authors denying that the realists ever really believed it. A judge consulting a law book isn’t trying to predict her own actions, any more than chess players merely predict that they’ll move bishops diagonally. For our purposes, moreover, the prediction theory proves too much, and thereby fails to say anything interesting about unwritten law in particular. We want to know whether courts might treat unwritten standards the way they treat statutes—or, conversely, whether unwritten law is necessarily judge-made in the way that some kinds of law are not. If law is always and everywhere a prediction about judges, then all these sources are evidence together, and there’s nothing any less law-like about an unwritten rule. (Nor is it clear why judges are the officials who matter; what your clients really want to predict is the behavior of policemen, judicial marshals, and the 101st Airborne. On the other hand, if we bring judges and statutes back into the picture—say, because marshals usually listen to judges, and judges usually listen to statutes—then it would also matter whether judges usually listen to customary law.)
A third theory, suggested by Joseph Raz, is that courts are responsible for identifying customary legal rules because of the judiciary’s fundamental role in the legal system. On Raz’s account, a legal rule is “part of the system only if it is recognized by legal institutions”—in particular, by “primary law-applying” institutions like courts. Courts do have an enormous impact on legal practices, and law-applying institutions do need some way of separating legal norms from nonlegal ones when the occasion arises. But do they need an institutional decision to draw this line? Perhaps these institutions could instead respond to what’s already the case—to the fact that a particular custom might already be recognized as law, even before its first discussion by a court, just as a particular grammar rule is operative even before its first invocation by a grader of AP exams. Raz notes that there are some customs on which courts are already disposed to act, as to which the courts are “merely recognizing and enforcing” existing standards and not making their own. And courts are hardly the only law-appliers in town. If all sorts of officials and private citizens are routinely engaged in applying the law, and if they already consider particular customary rules to be an ordinary part of the law that they apply, then why must they wait for a court to speak first? A system in which the courts routinely enforce different rules than everyone else would certainly be unstable, just as a sporting event will be unstable if the scorer has gone rogue. But the practical need for a faithful scorer doesn’t mean that rules only go into effect once a scorer has applied them.
To Hart, it seemed clear that courts might “apply custom, as they apply statute, as something which is already law and because it is law”; to exclude this possibility was “merely dogmatic.” If a society can recognize as law whatever’s written in a particular book, whatever’s “carved on some public monument,” and so on, it can also recognize as law whatever’s identified by a certain kind of customary practice.And one of the things this customary law might do would be to incorporate other bodies of customary rules by reference. Some jurisdictions might have a local custom or usage of adopting the general common law—essentially, a reception statute without the statute—while other jurisdictions might not. If adopted in this way, the general law would really be “law in [the] jurisdiction,” but “only because the jurisdiction’s officials or inhabitants said so.” The situation would be much like that of the law of nations, which was commonly said to be to be part of the common law—a part that the local jurisdiction couldn’t actually amend or change, though one it could always abrogate or choose not to apply. A system founded on social practice could thus incorporate rules external to its own society; these incorporated rules would be given sovereign authority, without needing a sovereign author.
There may be good responses to this account, and reasons why custom, local or general, has to wait for the courts to be made law. But without having those reasons in hand, the burden of proof seems to rest with those who disagree—who claim that customary law can only be created by courts, and that it can’t be found by judges.
One last reason to believe that judges can find law is that they actually do it quite often. (As in the old saw about believing in infant baptism: “Hell yes, I’ve seen it done!”) This Article doesn’t claim that any particular society has charged its judges only to find the law, and never to make it. But at least some legal norms have been found throughout history, a practice that still continues today.
The common law is often identified with case law—that is, with judicial legislation structured by rules of stare decisis. But English law used unwritten rules for centuries before such doctrines took their modern form. In the early eighteenth century, Sir Matthew Hale described the traditional doctrine as treating judicial opinions as mere evidence of the law:
It is true, the Decisions of Courts of Justice, tho’ by Virtue of the Laws of this Realm they do bind, as a Law between the Parties thereto, as to the particular Case in Question, ‘till revers’d by Error or Attaint, yet they do not make a Law properly so called, (for that only the King and Parliament can do); yet they have a great Weight and Authority in Expounding, Declaring, and Publishing what the Law of this Kingdom is, especially when such Decisions hold a Consonancy and Congruity with Resolutions and Decisions of former Times; and tho’ such Decisions are less than a Law, yet they are a greater Evidence thereof than the Opinion of any private Persons, as such, whatsoever.
Hale’s account was no pious fiction. “Legal historians widely agree,” Postema writes, “that before the eighteenth century there was no firm doctrine of stare decisis in English common law.” In other words, “if historians are correct, English common law functioned well enough for over 500 years without the one thing that, according to current orthodoxy, held the practice together as a form of law.”
Instead, according to David Ibbetson, the English relied on customs in foro, customs not “of the English people as a whole but of the lawyers.” The customary law thus created was not “just judge-made: it was the product of the whole legal culture focused first on Westminster Hall and later on the Inns of Court, where lawyers lived, discussed, taught and learned together.” Even today, as Peter Tiersma notes, “a remarkable amount of orality has survived in the English common law.”
These English practices were carried across the pond. In the early nineteenth century, as Judge Fletcher famously described, federal and state courts managed “to develop a uniform body of law” on marine insurance, seeing themselves as “engaged in the joint endeavor of deciding cases under a general common law.” Justice Brandeis himself found it difficult to resist the practice; shortly after Erie, he decided a case in which “[m]ost of the issues . . . involve questions of common law and hence are within the scope of [Erie],” yet he saw no claim “that the local law is any different from the general law on the subject.”
In our own day, federal courts constrained by Erie frequently act as if they’re finding rules of unwritten law. This might happen in cases involving accretion and avulsion of littoral property; waiver on appeal; uncodified criminal defenses, such as duress or necessity; the doctrine that interest goes with the principal; and so on. Many rules of so-called federal common law are, in substance, just the old general-law doctrines in disguise. Caleb Nelson documents how courts applying common law concepts in federal statutes will look to general principles of torts or agency—“how most states do things,” as opposed to what any one state has said. This looks a great deal like courts finding law rather than making it.
Most importantly, federal courts are told to find law in the one place where we might expect the contrary: the decision of state-law issues under Erie. In George Rutherglen’s quip, Justice Holmes’s “predictive theory of law has been everywhere discredited as a theory of adjudication—except in its application to state law under the Erie doctrine.” Federal courts deciding state-law issues must perform the “Erie guess,” predicting how the issues might be decided in the state court of last resort. As Justice Holmes put it almost three decades before Erie, the “fiction” of Swift and of general common law “had to be abandoned” once everyone realized “that decisions of state courts of last resort make law for the State.” If court decisions make law, rather than find law, then federal courts have to apply the law made by those courts, and they should try to reach the same ruling that the state courts would reach.
What’s strange, though, is that the Erie guess presumes that federal courts are indeed capable of finding law—so long as it’s the law of a state. When a federal court makes an Erie guess, it’s guessing what state courts would do by their own lights, not deciding what those courts should do. To Justice Holmes, a federal court has power “only to declare the law of the State,” based on existing sources, and “not an authority to make it” out of whole cloth. But if a federal court can do that, then presumably it’s also capable of declaring law of other sorts, without any “authority to make it”—of making guesses about how a given issue might be decided by courts throughout the United States (or the whole common-law world), based on the existing legal sources and according to prevailing views. That is, it could go about finding, and not making, the preexisting law.
Suppose that judges sometimes find law. Could that be all they do? Or do they sometimes need to make it instead?
This Article assumes, for argument’s sake, that the “right-answer thesis” is false: not every legal question has to have a uniquely correct right answer, and certainly not one that’s easy to find. When such questions arise, the legal system needs a way of settling them, like handing them over to a judge. If the system has rules of precedent, then that decision will be taken as a standard for the future—making new law, some would say, in order to fill the gap. If legal uncertainty will always be with us, the argument goes, then so will judicial lawmaking.
Both courts and scholars have accordingly treated judge-made law as inevitable. The Supreme Court borrowed from Austin the view “that judges do in fact do something more than discover law; they make it interstitially by filling in with judicial interpretation the vague, indefinite, or generic . . . terms that alone are but the empty crevices of the law.” Hart, too, concluded that in “legally unregulated cases” in which the law dictates “no decision either way,” a judge “must exercise his discretion and make law for the case”—employing “law-creating powers” to choose “between the competing interests in the way which best satisfies us.” Some realists described this as a crucial “discovery about the way our courts work,” one that “helped to bring about the Erie decision—the realization that the judicial process is not a mechanical process of ‘finding’ or ‘discovering’ an already existing law, but quite often the creative job of making new law.” Even Justice Scalia conceded that judges must make law, with the caveat that they should act as if they don’t: making law “as judges make it, which is to say as though they were ‘finding’ it—discerning what the law is, rather than decreeing what it is today changed to, or what it will tomorrow be.”
Forbidding judges to make law is actually more plausible than it seems. Even when judges can’t help breaking new ground in their decisions, they’re still just making decisions; they don’t have to be making law. The legal force of their decisions rests on other doctrines in the legal system, and a decision can be legally influential, or even binding, without changing the law on which it’s based. A legal rule might be “the law of the case” or “the law of the circuit” without being “the law”; it stands in for the actual law without supplanting or altering it. So a system might provide that the decisions of courts, even those of last resort, only establish stand-in obligations like these—leaving future judges and officials free to pursue the real law, and leaving that law to be found rather than made.
Requiring judges to find the law is consistent with hiring fallible human beings to be judges. Plenty of people, including other officials and private parties, have to make decisions under legal uncertainty—sometimes facing precisely the same kinds of problems as the uncertain judge. We can expect them to follow the law, as a normative matter, at the same time that we expect them as an empirical matter to fail repeatedly in doing so. But what we expect of them still matters, and it makes a real difference to the law.
The basic argument that decisions make law is rather simple. To Hart, if courts can “make authoritative determinations of the fact that a rule has been broken, these cannot avoid being taken as authoritative determinations of what the rules are.” If future courts have to decide similar cases in a similar way, then “these judgments will become a ‘source’ of law,” resembling “the exercise of delegated rule-making power by an administrative body.”
But not all legal systems treat precedent this way. Some civil law systems have treated it as persuasive authority only. Some past common law systems waited for a line of cases, not just a single decision, before declaring a matter settled. And even modern-day systems needn’t treat a court’s judgment as equivalent to a statute—something that, in Simpson’s phrase, “is both the only reason and a conclusive reason for saying that this is the law.” Precedent can make a powerful difference without having to make law, and a system can be committed to judicial precedent without being committed to judicial lawmaking.
A court’s judgment can be “a ‘source’ of law” in more than one way. It might be res judicata, the law of the case, the law of the circuit, stare decisis, and so on. Each of these doctrines treats a prior judgment as having a certain amount of legal force in the future. But each does so by treating the judicial decision as if it were law, and not by substituting that decision for the underlying legal standards on which it’s based.
A judgment is certainly a source of law in a particular case. The final judgment in a civil action, unlike an interlocutory order, can only be corrected by limited means. An appellate judgment won’t be second-guessed on remand, under the mandate rule—or even on a subsequent appeal, under the law-of-the-case doctrine. But this doesn’t mean that either kind of judgment actually sets out the law. It only sets out what a court must assume is the law for purposes of a particular decision. The law-of-the-case rule has an exception if the prior decision is “clearly erroneous, and would work a substantial injustice”; that error can only be judged in light of the actual legal standards, not those determined by the prior court.
A judgment can also be a source of law for the parties. It can bar certain claims or arguments through preclusion doctrines without affecting the legal system as a whole. When a plaintiff loses on a particular issue—say, whether the light was green or red, or whether the defendant had a duty of care—she might be collaterally estopped from challenging that determination in future cases, sometimes even as to third parties. But the decision doesn’t change the fact of the matter: the light was actually green or red, or the defendant did or didn’t owe a duty, no matter what the court said about it and no matter what future courts must assume.
A judgment can be a source of law for other courts as well. The Fourth Circuit’s search-and-seizure holdings bind district courts in Maryland, but not in Delaware, even though the same Fourth Amendment applies in each state. The force of those holdings might vary by legal issue: the Federal Circuit applies its own law to patent issues, but “the law of the regional circuit” to others. In this context, no one would confuse “the law of the Fourth Circuit” with “the law”; otherwise it’d be impossible, say, for the Federal Circuit to read the same (nonpatent) statute according to the rival interpretations of different circuits in different cases. Judges of the Federal Circuit, like district judges in Maryland, are occasionally required to assume that Fourth Circuit precedent is correct—just as they’re occasionally required to assume the irrelevance of a waived argument, the resolution of a precluded issue, or the truth of a well-pleaded complaint. They may have to say it, and even act on it; but they don’t have to believe it, and it doesn’t have to be right.
The same theory can be applied to courts of last resort. As a matter of legal theory, there’s no reason why the holdings of a court like the Supreme Court of the United States must necessarily be taken to represent “the law,” as opposed to “the law of the Supreme Court,” binding on other courts within the reach of its appellate jurisdiction. When the Court construes state law, for example, its decision isn’t always binding on the courts of that state. Even within the federal system, the Court acknowledges a distinction between the true law and its precedents every time it describes a past decision as “wrong the day it was decided.”
Societies have good reasons for occasionally forcing their courts to assume false things. Acting as if a certain proposition were law helps achieve a variety of social goals: it avoids relitigation, provides stability, and so on. But a legal practice of occasionally ignoring the real answers doesn’t suggest that there are no real answers, either as to facts or as to law—or that the answers the court gave somehow became the real ones, as opposed to our being required to act as if they were. (Indeed, we often refuse to apply issue preclusion to certain “unmixed questions of law,” precisely to avoid forever binding the parties to something other than the actual law.)
So Justice Scalia was too hasty when he claimed that “the requirement that future courts adhere” to a decision thereby “causes that decision to be a legal rule.” Rules of precedent might make a past decision of obvious legal interest; they might sometimes require actors to treat the decision as if it were the law. But precedent alone doesn’t require “that the decisions of Courts constitute laws”—something that the Supreme Court, prior to Erie, thought would “hardly be contended.”
For some, this distinction between precedent and law may seem gossamer-thin. It’s one thing to say that precedent doesn’t displace law when we know what the law is. But if there is no law on a subject, or we really can’t tell what it is, how could it possibly matter whether the precedent makes “real” law or just a stand-in? Why even try to differentiate the two?
Distinguishing “real” law from “as-if” law might seem quite difficult. Does an administrative agency make new law when it issues a regulation, changing people’s legal rights and obligations? Or is it merely issuing an instruction, with which some other legal rule (like the organic statute) requires as-if compliance? Nondelegation worries aside, either view seems fraught with dangers. On the one hand, if the agency makes new law, then it’s hard to deny that private parties make new law when they sign a contract, that senior partners make new law when they give reasonable ethics instructions to junior associates, or that flight attendants make new law when FAA regulations require passenger compliance with lighted signs and crewmember instructions. On the other hand, maybe none of these things are real lawmaking, as opposed to the exercise of a preexisting legal power, or some other action to which some other rule lends as-if force. But how far up the chain does that argument go? Are we sure that Congress makes any laws, as opposed to producing mere pieces of paper that Article I requires us to respect? If we don’t want to go that far, are we sure where to place judicial decisions on the scale?
In fact, we can distinguish judges from these other actors without needing a general theory of who makes law and how. The proper way to classify a judicial holding might depend on its goal: say, to conform to a preexisting set of legal entitlements (as determined by external standards), or alternatively to lay down new standards for the future. To put it another way, we might ask whether the holding has a “mind-to-world” or “world-to-mind” direction of fit. A private contract might be unwise, or even unlawful, but it usually won’t be incorrect: it’s not trying to match some legal norm already in the world, but rather to do something new. The same is true of a newly adopted rule of court: the new rule’s purpose might be to recodify an old one, and to that extent it might be correct or incorrect about what the old rule did—that is, in the same colloquial sense in which a new tax cut might be right or wrong about economics. But although either the tax cut or the rule of court might be the consequence of human error, and thus fail to achieve its intended purposes, neither can be wrong about what the current rule now is. By contrast, other judgments about law can quite easily be mistaken: say, a judicial marshal’s interpretation of a court order. In some legal systems, it might equally make sense to describe a court’s judgment in a given case as incorrect on the law, even if it turns out to have a variety of useful consequences for the future. As Nelson argues, “[a]ll modern lawyers would understand” a claim that “[t]he Constitution plainly establishes Rule X, but the Supreme Court has interpreted it to establish Rule Y instead,” even if “the Court is not going to overrule that interpretation.” So long as the court’s holding is supposed to comport with an external standard, it can make sense to describe that standard, and not the holding, as the law.
When people argue that judicial lawmaking is inevitable, they tend to focus on cases in which these external standards are absent (or at least really hard to find). But if we’re going to treat judges as necessarily having authority to make real law, they can’t have that authority only in the gaps where “law runs out,” for the simple reason that no one knows where those gaps are. Legal questions don’t come neatly separated into two piles, “obvious” and “obscure.” Even if the set of actually indeterminate cases were fixed over time, the set of cases we believe to be indeterminate is not. Sometimes we discover legal reasons we didn’t know of before: new evidence of constitutional or legislative history, important authorities that the parties forgot to mention in their briefs, and so on. At that point, a court might have legal reason to change its mind, wholly independent of any considerations of good public policy. In doing so, the court wouldn’t be repealing an enacted rule, so much as determining that its prior decision was wrong the day it was decided. Yet if a precedent truly makes law—serving as “both the only reason and a conclusive reason for saying that this is the law”—then the law the precedent has made is complete and self-sufficient; it’s hard to see why the new information should matter.
Alternatively, if we imagine courts to be making law whenever the issues are actually indeterminate, and to be making mere precedent whenever a right answer actually exists, then we’ll regularly be in a state of ignorance as to whether any given precedential rule is really a rule of law or not. (Who knows what new evidence the future might bring?) Yet the entire point of treating precedent as law, on this gap-filling account, is to avoid the strangeness of drawing an important distinction between two things—law and precedent—that we can’t tell apart in practice. Limiting lawmaking to actually obscure cases does precisely the same thing, one level up.
The problems caused by conflating law and precedent are particularly severe when dealing with other systems of law. Suppose that our choice-of-law doctrines require us to decide a particular case according to the law of Japan. The Supreme Court plainly has no power to “make” Japanese law, any more than it can “make” state law in an Erie guess. The conflicts rule may be part of our legal system, but the Japanese legal rule is not; as Green writes, American officials “can neither change it nor repeal it, and [the] best explanation for its existence and content makes no reference to [American] society or its political system.” The actual law of Japan is determined by Japanese social practices, and the most the Court can do is to state the American system’s best guess of what Japanese law might currently be. If the Court missed out on some relevant aspect of Japanese social practice, the best explanation is simply that it misstated Japanese law, not that it made a new rule of American law relating to Japan (which might be criticized for policy reasons, but not for legal error). Declaring that the Supreme Court must be making new law, or that American law can’t incorporate Japanese law by reference, would seem to be an inappropriate limitation on what sorts of law American social practice might support.
Treating a court’s application of a legal standard as necessarily generating a new standard—e.g., maintaining that the Fourth Amendment really does require different things in Maryland and Delaware—would also create all kinds of jurisprudential headaches. For one thing, why would state judges in Maryland be allowed to disagree? For another, why would judges be the only ones with this authority? A district attorney’s decision to drop a set of cases might set an internal precedent for her office, but it needn’t change the elements of the charged offense. Claiming that the charging decision, by eliminating the threat of criminal punishment for certain conduct, necessarily alters “the law of the 14th Judicial District of North Carolina” ignores the fact that the same conduct in the same jurisdiction will still be treated as unlawful in myriad other ways (postconviction review, official impeachment standards, negligence per se in civil actions, and so on). Disaggregating these issues into an innumerable set of separate and independent legal questions—a “bundle of sticks,” some governed by the district attorney and some not—would assume away all of the systematic features that make the law a coherent system of norms. It would also bring on all the disadvantages of the predictive theory, as the rules of internal precedent in prosecutors’ offices are subject to disaggregation too. The only thing determining this defendant’s punishment would be whether this assistant D.A. is likely to bring a case—but of course the assistant D.A. isn’t trying to predict her own behavior. Precisely the same could be said of any attempt to treat an appellate decision as really establishing “the law of the Fourth Circuit,” as we’d still need to ask why the Fourth Circuit gets to make law across its entire jurisdiction in a way that the Fourth Amendment apparently can’t.
Finally, even if we thought that the Fourth Circuit necessarily makes law for district judges, it’s not clear that its authority has to bind anyone else. A legal system has a certain unity to it: the sale of a car between private persons will bind the U.S. government, altering its powers with respect to Fourth Amendment searches (“effects”), Fifth Amendment takings (“property”), and so on. That kind of legal unity makes sense if we take precedents as stand-in law: everyone else can recognize, in a unified way, that the decisions of the Fourth Circuit create stand-in legal reasons for the district judges in Maryland while leaving the actual law intact. But it’s not necessarily true if we take precedents as statements of the law in general. There’s a live and longstanding controversy between departmentalist views of the judicial power, in which a court’s judgment only determines the law for the parties or for lower courts, and judicial-supremacy views, in which a precedential holding is binding on everyone in the jurisdiction. If a holding states the law, then every legal actor is bound by it (even if only within the Fourth Circuit). But if executive officials, legislators, or state judges can adopt their own views, then the law-precedent distinction makes a real difference. At the very least, departmentalism appears to be a plausible option that a legal system could choose—meaning that it’s no less plausible for a legal system to distinguish sharply between precedent and law.
There are plenty of reasons why a society might give appellate courts a certain influence over legal questions, without granting them any power to change the law. In a hierarchical court system, having lower courts follow higher courts might help reduce uncertainty, process cases quickly, and kick important questions upstairs—all without reverting to a system of “scorer’s discretion.” Any doctrine of precedent that’s binding on courts will matter to parties who find themselves before courts; lawyers will tell their clients about the precedents, and clients will change their behavior accordingly. But so would the Holmesian ‘bad man,’ who doesn’t care about legal rules anyway. Those who do care about legal rules can’t make a priori assumptions about the scope of judicial lawmaking; the force of precedent is itself something that has to be determined by law.
Judges are human beings, not mechanical rule-followers. It may be theoretically coherent to distinguish judicial decisions from the law on which they’re based, but does that distinction exist in real life? When judges are given only vague rules to follow, they can’t help but make choices of their own—many of which end up transforming the rules that were in place. So can judges truly be expected not to make law?
The answer is “yes,” but it turns on what one means by “expect.” We generally expect people to obey the law, but we also create police departments and prisons on the assumption that many people won’t. For the same reasons, no one should be so foolish as to predict that judges will always get the law right, or even that they’ll always try in good faith. But predictions aren’t the point. We can make good faith a norm for judges’ conduct, even if we know that they’ll often fall short. Put another way, we can expect more of judges than we expect from them.
Finding law sounds most unrealistic as an account of the process of judicial decision. Real judges in real cases don’t flip through their books until they find the answer, then stop. But that objection confuses legal reasoning with judicial psychology, or with the phenomenology of judicial decision-making. Real-life judges also don’t proceed in the manner of a standard judicial opinion: they don’t sit down and start by pondering subject-matter jurisdiction, then the facts and prior proceedings, then the standard of review, and so on. Judges might proceed by hunches or guesswork, subject to a thousand biases and nonlegal influences; but their opinions have to give legal arguments in favor of the judgment, subject to ordinary standards of legal justification. As Hart noted, the role of legal reasoning isn’t found in the “methods of discovery,” but rather in the “standards of appraisal” that judges “respect in justifying decisions, however reached.” As in Green’s example of the person who gets dressed in the morning without consciously adverting to social norms, the judge’s “[r]ule-following behavior is . . . displayed ex post actu, when rules are produced in justifications, used in communicating decisions to others, and [in] explaining what was done or . . . defending it against criticism, actual or anticipated.”
So the fact that judges, especially in unclear cases, don’t always feel like they’re “finding” law doesn’t mean that they’re making law instead. Hard cases often involve not a shortage of legal reasons, but a surplus; the cases are hard because too many legal reasons are in play at the same time. Confronted by all of these at once, uncertain judges aren’t forced into the position of choosing the best rule in the abstract, but of deciding how this menagerie of legal instructions is best satisfied on these facts. To Hart, “when the explicit law is silent, judges do not just push away their law books and start to legislate without further guidance from the law”; instead, courts proceed “by analogy,” accounting for the decision “in accordance with principles or underpinning reasons recognized as already having a footing in the existing law.” This process helps explain why such terms as “‘choice’, ‘discretion’, and ‘judicial legislation’ fail to do justice to the phenomenology of considered decision,” including “its felt involuntary or even inevitable character.”
This process also ought to make us suspicious of the idea that the judge, in making a hard decision, is forced into the position of making new law. When a court addresses a legal standard to a particular set of facts, we usually call that activity “applying” law, not “making” it. Judges who disagree about the proper decision in a particular case—say, whether the plaintiff’s fear was “reasonable”—are disagreeing about the law’s proper application, and not necessarily about its content. We can all agree on what “reasonable” means without agreeing, in a particular case, on how the reasons come out. Some applications of the rule might be less persuasive than others, but so what? “[T]here is a great deal of ruin in a nation,” and a great deal of legal error in a nation’s courts; still, we generally manage to get by.
After all, judges are hardly the only ones who have to make hard legal decisions. Prosecutors, policemen, and private citizens do so routinely. We regularly speak of juries as “finding facts,” even in contexts where the truth is somewhat in the eye of the beholder (whether the plaintiff’s work environment was truly hostile, the pain-and-suffering damages truly adequate, and so on). Relatively few people would call this a “childish fiction”; so why should it be odd to speak of judges in similar circumstances as “finding law”? Juries also make complex and largely unreviewable decisions in murky cases, but saying that the jury “makes facts” would sound paranoid or bizarre. And if a jury did exercise active choice over the facts—say, by flipping a coin or favoring a sympathetic party—it’d plainly be acting ultra vires. A court faced with a similarly hard question might come to a satisfying answer or an unsatisfying one, but the mere fact that it’s acting under uncertainty doesn’t mean that it’s “making law.”
Unlike juries, of course, judges are supposed to explain their reasoning. But even in uncertain cases, they can do so while attempting to adhere to the existing standards. Berman distinguishes two categories of judicial doctrine: (1) the “operative propositions” that the relevant sources of law actually establish, and (2) the judge-created “decision rules”—the menagerie of tiers of scrutiny, n-factor tests, and so on—which “direct courts how to decide whether [an] operative proposition is satisfied.” Even without a power to lay down new rules, a court might certainly come up with such n-factor tests on its own, as a way of formalizing its thought process in answering difficult questions—the way someone having trouble choosing a house might draw up a list of relevant factors, like “move-in ready” or “easy commute.” Future courts might then adopt the same list of factors as a proxy for the underlying legal considerations. Perhaps, in the fullness of time, the customary law might change, and adhering to the list might itself become a customary rule. But none of this suggests any judicial power to alter the operative propositions, or to transform their legal requirements at will. As things currently stand, the proxy is not “a conclusive reason for saying that this is the law”; whatever its policy merits, it might still be a bad proxy, one that fails to track the operative propositions.
This understanding of decision rules isn’t just a word game: it rules out otherwise-attractive judicial tactics and common attempts at evasion. Say that the most accurate standard would involve seven factors, but it’s much easier for district courts to administer a standard that has only three. A court with authority to make new law could simply abrogate the old test and lay down a simpler one. But a court without that authority can’t sacrifice the best legal standard for easy administration, any more than it can introduce administrability concerns or cost-benefit analysis into a statute that ignores them. Doing so is equivalent to saying that there should have been a different legal rule, one responsive to fewer or more specific considerations than it was. The fact that judges frequently do rely on judicially authored n-factor tests doesn’t show that they’re making new law; what matters is how the tests are constructed, used, and understood.
The United States might have a legal system in which judges are allowed to impose new legal standards, or it might not. Again, this Article doesn’t seek to describe our existing institutions, or even to suggest which choice is a better idea, all things considered. Its goal is merely to show that a world in which judges are charged to find the law, and not to make it, is hardly a strange hypothetical. It might be a world suspiciously resembling our own.
If law can be found, and not just made, what are we to make of Erie? No other decision has so exalted a place in American legal scholarship—regarded as “a sea change in how judges view law,” or even “a change in the nature of law itself.” Before Erie, judges were said to be “the living oracles of a preexisting natural law”; afterward, they apparently became “lawmakers in a relativistic legal world,” in which “the common law was nothing more than [their] decisions.” Yet among many legal philosophers, Erie’s Austinian legal theory is nowadays thought to be wrong—obviously, laughably, “friendless[ly]” wrong. So if Erie was wrong about the nature of law, then what else was it wrong about?
The answer may well be: nearly everything. What’s become known as the “Erie doctrine” is founded on a mistake. Though Erie’s flawed jurisprudence might be distinguished from its doctrinal legacy, the structure of the American legal system means that the two will always be intimately linked. On the state level, by denying that courts could find law, Erie undermined the states’ efforts to incorporate separate bodies of general law as their own. And on the federal level, Erie helped to birth an entirely ahistorical category of “federal common law,” which federal courts are encouraged to make rather than find. The result was a wholesale revision of numerous areas of doctrine, based largely on an error about legal theory. Whether one praises or assails the decision as a matter of policy, as an intellectual matter Erie was worse than a crime—it was a blunder.
Had circumstances been otherwise, Erie’s errors might have been confined to the abstract and theoretical realm. According to Jack Goldsmith and Steven Walt, Erie didn’t need Austin anyway; a “general theory about the nature of law” should have “no implications for the allocation of authority between the state and federal governments.” It’s true that general theories don’t always have to affect concrete doctrine. But whether they do depends in part on what the doctrine says. In the particular circumstances of American federalism, Erie’s theory made an enormous doctrinal difference.
Goldsmith and Walt’s contrary argument is straightforward. At its core, Erie was a case about who defers to whom for what; it held that federal courts must defer to a state court’s understanding of general law rather than applying that law for themselves. A view that judges make or find law could be compatible with either Erie or its opposite. Maybe state judges make state unwritten law, which federal judges must then apply; or maybe federal judges make national unwritten law, which state judges must then apply. On the other hand, maybe judges can find law instead, in which case we’d still need to decide whether either a federal or a state court should defer to what the other finds. As Walt puts it, either theory of judicial authority is consistent with either approach to judicial federalism, so positivism is at most a “superfluous premise in Erie’s rationale and its constitutional holding”: general jurisprudence “says nothing about which roles are appropriate for federal courts,” and Erie can’t rest on whether judges make law or find it.
The problem with Goldsmith and Walt’s argument isn’t that it’s wrong, but that it’s too general. By looking only to which legal regimes are possible, it ignores the question of whether any existing constitutional rules, federal or state, do rest Erie’s holding on a particular legal theory.
Sometimes the details of a legal system force us to deal in high theory. If a case should turn on the law of postwar Germany—say, involving retroactive civil liability for wartime acts—we might have to decide, in court, whether horrific Nazi laws were really laws at all. Usually that kind of question is reserved for legal philosophers. But when judges have to assess foreign law, the Federal Rules leave them to their own devices, telling them to consider “any relevant material or source.” So an American judge might have to decide the postwar status of Nazi law under the right theory of jurisprudence, whatever that is. That may be unfortunate, but it’s also their job; judges regularly decide whether a given chemical is really a carcinogen, or a given piece of evidence is really probative, so why not whether a given norm is really law?
There’s good reason to think that this is what happened in Erie: a garden-variety legal question happened to turn on a deep question of jurisprudence. To determine if Tompkins had trespassed on the railroad’s right-of-way, Justice Brandeis had to determine the content of Pennsylvania law—and, if Pennsylvania had incorporated the general common law by reference, whether recent state-court decisions had displaced that law, or whether they merely served as evidence thereof. Brandeis reasoned that “the law to be applied” was in either case “the law of the State,” and that this law might just as well be “declared by its Legislature in a statute or by its highest court in a decision”; the choice between them was “not a matter of federal concern.” Yet if the choice were really up to the state, some states might not have wanted their high courts to announce the law. Maybe their legal systems still treated state judicial decisions as evidence of state law, and not as law themselves. Whether that choice is a coherent one, deserving of the federal courts’ respect, depends at least in part on whether courts are capable of finding law. If not, then a federal court charged by the Rules of Decision Act to apply “[t]he laws of the several States” is obliged to treat state decisions as making law rather than finding it, even if the state courts pretend otherwise. In a system with a particular approach to judicial federalism, these theoretical questions make a real difference.
As it happens, Erie’s theoretical error may have led it to take a seriously incorrect view of states’ unwritten law. On some historical accounts, the early American states inherited a tradition in which courts were charged to find law rather than make it. Whether by reception statute or by local custom, they had adopted the general common law as their own, instructing state officials and state courts to follow a legal tradition extending beyond their own borders. Swift v. Tyson expected the federal courts to follow this cross-reference faithfully, except where “local statutes or local usages” otherwise prescribed: so long as the state claimed to be adhering to the general rule, the federal courts would adhere to it too. Toward the end of the nineteenth century, federal courts began to act more aggressively in enforcing general law, even when state courts had openly recognized local customs or usages authorizing departures from the general rule. But Erie hardly devoted itself to investigating this history or enforcing the limits of Swift. Instead, it insisted that the states had adopted precisely the same views toward general law as happened to be held by Justices Holmes and Brandeis.
As Michael Steven Green described in “Erie’s Suppressed Premise”—an article that others might have titled “One More Reason Why Erie is Wrong”—the Erie Court declined to investigate any specific state legal systems to see if their courts were empowered to make law. Instead, Justice Brandeis simply assumed that finding general law was impossible, so the courts must be making it instead. (The “fallacy” of a “transcendental body of law,” and so on.) The underlying theory was provided by Justice Holmes in the Taxicab Case: if a state constitution chooses to delegate legislative powers to the state courts, providing that “the decisions of the highest Court should establish the law,” then the federal Constitution shouldn’t interfere. And, Holmes assumed, this is the right way to read any state constitution with a court in it: “[W]hen the constitution of a State establishes a Supreme Court it by implication does make that declaration as clearly as if it had said it in express words.”
This is an extraordinary claim, especially when one remembers that many state constitutions—including, say, the Massachusetts Constitution of 1780—were enacted well before the modern acceptance of judge-made law. Why would Holmes assume that every state constitution must be read this way? Because, he wrote, each state supreme court necessarily “says with an authority that no one denies, . . . that thus the law is and shall be. Whether it be said to make or to declare the law, it deals with the law of the State with equal authority however its function may be described.”
If judges can find the law instead of making it, this last sentence is plainly false. A court charged only to find state law might occasionally be mistaken, but its decision and the law are different things. Except as the decision is used by lower courts or similar institutions, or as a practical guide for the parties, its authority is limited to its correctness. By contrast, if a state court necessarily makes state law, then it attains Bishop Hoadly’s “absolute Authority to interpret,” and state law becomes something that it can’t be wrong about: whatever it decides is the law. (To paraphrase President Nixon, when a state supreme court does it, that means it is not illegal.) These two authorities are nothing alike.
If Holmes was wrong to confuse the two—and this Article argues that he was—then the whole logic unravels. The mere existence of a court system isn’t enough to infer a delegation of lawmaking authority to state courts. And without a delegation of lawmaking authority to state courts, the federal courts have no clear obligation, whether under the Constitution or the Rules of Decision Act, to defer to them on what the state itself considers to be matters of general law.
One might try to resuscitate Erie with something like the following argument. When applying Vermont law, federal courts usually assume that Vermont’s courts know what their law is. Federal courts will even defer to state-court constructions of Vermont statutes, although the state courts don’t get to “make” statute law (and although Vermont’s other branches of government might, as an internal matter, assert a departmentalist independence from the state judiciary). So why shouldn’t federal courts defer to state-court constructions of unwritten law too, regardless of whether Vermont’s courts are considered, as an internal matter, to make or find the law? After all, the Vermont Supreme Court usually won’t tell anyone not to defer to it, as the deference question rarely arises at home. So the federal courts’ obligation to follow state law, and “the necessity that some court must have final authority to settle the meaning of state law,” suggest that whatever the Vermont Supreme Court says, goes.
This argument might be challenged on its premises: perhaps state statutory-construction decisions weren’t always taken as conclusive—at least not until “a series of decisions” could “settle the rule” and make it “the settled law of the state.” But the real problem with the argument is that it assumes its conclusion, namely that the state makes no real distinction between local and general law. Suppose that Vermont really wants to incorporate some other rules by reference; say, that a statute discussing “the law of the place of contracting” makes a case turn on the laws of Japan. Maybe Vermont’s courts are owed deference as to what the statute means—as if it had really said “law*” or “contracting*,” with the asterisks denoting implicit limitations to be named later. But if Vermont’s courts read the words the same way everyone else does, then there’s no issue of local law to be discussed. Vermont’s courts would enjoy deference on whether Vermont really cross-references the law of Japan, but not necessarily on what the laws of Japan actually are—a question that other courts can investigate on their own, and as to which the “state tribunals are called upon to perform the like functions as ourselves.”
It’s certainly possible for states to incorporate other rules by reference—and even to do so blindly, making no attempt to review the details of what’s incorporated. And they might incorporate, not only the laws of another jurisdiction, but other bodies of rules as well—like the American Law Institute’s Restatements, or even a past edition of Robert’s Rules of Order. Had Vermont reprinted the text of Robert’s Rules in its session laws, maybe we’d believe its courts’ claim to be construing the idiosyncratic language of a local statute. Or if its courts claim to be relying on some procedural usage specific to Vermont, modifying the otherwise-applicable provisions of Robert’s Rules, maybe we’d have to believe them—for “it is a principle, that the general common law may be, and in many instances is, controlled by special custom.” But if the statute book merely says “Robert’s Rules of Order” with no claim of any asterisk, then garden-variety disputes about some provision buried in Robert’s Rules don’t obviously turn on any questions of local law.
At the very least, this is a plausible way for a state to organize its legal system. As a matter of historical fact, we know that states have sometimes treated portions of their law as true incorporations-by-reference in this way, seeking to achieve uniformity across borders rather than within them. That makes it hard to argue that definitive settlement of general-law questions by a single state court is truly a “necessity”—or that federal courts must, as Holmes and Brandeis suggested, treat state court decisions as conclusive on every question that arises from the state’s law.
Today, eight decades on, Justice Holmes’s vision of judge-made law is widely echoed. Perhaps our state constitutions now do confer, or have been authoritatively construed to confer, lawmaking powers on state courts. This Article takes no position on that question, which is ultimately one of empirics rather than theory. Still, that possibility leaves Erie on shaky ground. Some states might dissent from the modern view (Green suggests Georgia); others might do so in particular areas (say, the interpretation of the Uniform Commercial Code), seeking to conform their legal rules to those of a broader tradition. In either case, simply citing state court decisions as the sole source of authority—as the “Erie doctrine” is thought to require—might well get the state’s law wrong. If Erie was wrong on the theory, then it’s probably just wrong about state law.
When it comes to federal law, the situation is even worse. The federal judicial power was parceled out in 1788, long before Justice Holmes could get his mitts on it. If federal courts back then were supposed to find the law—especially in areas where the states were incompetent to legislate—then Erie’s preference for judge-made law shouldn’t stand in the way. Recognizing the theoretical flaws in the Erie doctrine might well mean tossing its whole invented edifice of “federal common law,” in favor of federal courts finding actual common law rules.
On the day it decided Erie, the Court also announced that an interstate water dispute involved questions “of ‘federal common law’ upon which neither the statutes nor the decisions of either State can be conclusive.” Today, “federal common law” is said to preempt contrary state law, to authorize federal-question jurisdiction, and to allow for judicial legislation in the pursuit of “sound policy.” All of this may turn out to be a modern invention, and one in deep tension with the Constitution’s actual text. But if so, the invention was foreordained by Erie’s insistence that law always has to be made by somebody—and if not by Congress or by a state, then why not by the federal courts?
One alternative, of course, is that federal courts might sometimes find law instead. When a case arises in which no state’s law controls, and to which no federal statute or treaty provides an answer, the federal courts might be required to apply preexisting sources of law that the Constitution left intact: in the Court’s words, to look to “known and settled principles of national and municipal jurisprudence,” such as “the common law,” “the law of equity,” or “the law of nations.” The United States is a common law jurisdiction, whose courts hear cases “in Law and Equity,” or of “admiralty and maritime Jurisdiction.” In Judge Fletcher’s famous formulation, the general common law—together with the rules of equity and admiralty—might be law “for the United States,” though “not of the United States”; law appropriate for use when necessary by federal courts, albeit without the preemptive status of “federal law.” Whenever the Constitution vests judicial power in the federal courts but leaves them without a rule of decision, this kind of law might still remain available for courts to apply. So the fact that courts can find law might entail that, under our system, they must do so—at least in particular enclaves such as admiralty, federal procedure, or customary international law.
Again, this Article focuses on what’s possible, and not what the American legal system (or any other system) actually requires. Whether Article III conferred a lawmaking power two centuries ago, or whether that power has somehow been wrested from Congress over time, are questions beyond this Article’s scope. But once we admit that finding law is a real possibility, we may well have to change our attitude toward the courts’ existing practice. As noted above, federal courts rarely use their oft-proclaimed power to shape new rules of federal common law; typically, they “adopt” the rules that were already in place. Yet they also feel competent to revise those rules when it strikes them as necessary. Whether the federal courts possess such a power may well depend on whether our legal system really authorizes them to make law, or whether it expects them merely to find it.
Accepting that judges can find unwritten law does more than correct a jurisprudential or historical error. Like its antithesis in Erie, it has the power to transform modern attitudes toward law.
Almost a century later, it’s instructive to go back and see what Holmes actually envisioned for judicial legislation. In Southern Pacific Co. v. Jensen, a page away from his famous line about the “brooding omnipresence,” one finds the following:
A common-law judge could not say “I think the doctrine of consideration a bit of historical nonsense and shall not enforce it in my court.” No more could a judge exercising the limited jurisdiction of admiralty, say, “I think well of the common-law rules of master and servant and propose to introduce them here en bloc.”
Holmes’s restraint here is surprising. For this is precisely how many judges, especially state judges, view their role today—as having been hired to examine common law rules and then to retain them or toss them aside. As New York’s Chief Judge Judith Kaye wrote in 1995, “Time and again, state courts have openly and explicitly balanced considerations of social welfare and have fashioned new causes of action where common sense justice required”—though wisely tempered by the need to “exercise that responsibility with care.” At its extreme, this picture presents each state judge as a swashbuckling Lord Mansfield, bravely reforming the law of the land (with due consideration and humility, of course).
This model is not without its advantages; there are reasons why Mansfield is still remembered today. But there are also drawbacks: limiting democracy, frustrating expectations, presenting legislators with a moving target, and depriving them of their ordinary ability—so crucial to striking bargains on other issues—to choose to leave well enough alone.
As the model of finding law has receded, drawbacks like these have been shunted aside. After all, one reason why judges adopt the swashbuckling role is a belief that there’s no real alternative, that this is just what a common law judge does. As Chief Judge Kaye put it, “[f]or state judges, schooled in the common law, to refuse to make the necessary policy choices when properly called upon to do so would result in a rigidity and paralysis that the common-law process was meant to prevent.” Any residual “anxiety about ‘legislating from the bench,’” or worries that certain changes might be “best left to the Legislature,” are dismissed as naïveté or weakness of will—to be rebutted with the “inevitability” of gap filling, or the discovery that “all judges are activists.” In other words, it’s much easier to breeze past the ordinary critiques of judicial lawmaking if the alternative is philosophically defunct.
This conceptual error also leads to revisionist readings of the history. Larry Kramer, for example, has explained how eighteenth-century common lawyers approached their subject very differently than many lawyers do today. But if “judge-made law is unavoidable,” as he argues, “simply because there is no clear line between ‘making’ and ‘applying’ law,” then it was just as unavoidable in the past. This leads to a sort of historical sleight-of-hand, with evidence of past courts’ finding law trotted out to support modern courts’ making it. Kramer argues that no one should “object when state courts take the initiative in fashioning common law,” for this practice—the modern practice, mind—necessarily “was, is, and always has been allowed in all these states.” And in the federal courts too; for if making law is just what courts do, then Article III must necessarily confer a “traditional judicial power to make common law.” Thus the Supreme Court’s famous, and famously self-serving, complaint that “the judicial hand would stiffen in mortmain if it had no part in the work of creation.”
This change in attitude toward the common law seems to be rooted in a giant intellectual mistake. According to Kramer, the removal of limitations on judicial lawmaking “results not from doctrinal changes, but from changes in our beliefs about the nature of law and the lawmaking process.” It’s only because “[w]e have come to see that even the fundamental principles of the common law were ‘made’ by judges” that “the ‘natural’ limits of pre-modern common law disappear, and the potential for making common law becomes as broad as we are willing to let judges go.” Surely the judicial process could have used some demystification; surely the history of the common law, under Lord Mansfield as well as others, is replete with examples of judges playing fast and loose with unwritten law. But the real motive force here seems to be a simple error about the nature of law: that it’s a “fallacy” or “illusion” to suppose “that there is this outside thing to be found.” And such errors, once made, don’t restrict themselves to unwritten law: cavalier judicial attitudes toward the common law have seeped into statutory and constitutional arguments as well.
Again, nothing in this Article addresses the actual norms of actual legal systems—whether in Blackstone’s England, New York State, or the United States as a whole. Maybe today’s legal norms really do empower judges, federal or state, to trade in their black robes for superheroes’ capes, or to play “junior-varsity Congress” with unwritten law. But in light of Holmes’s own reticence, it’s important to remember that this is not the only possible approach; that history and legal theory do offer alternatives; that different polities can choose, through their own constitutional systems, the powers they want their judges to enjoy. To make this choice, we need to restore, at least at the level of possibility, the consensus that such a choice exists. Unwritten law can be found, as well as made; the brooding omnipresence broods on.
Stephen E. Sachs: Professor of Law, Duke University School of Law. For advice and comments, I am grateful to Lawrence Alexander, Matthew Adler, William Baude, Allan Beever, Mitchell Berman, Samuel L. Bray, Nathan Chapman, Ryan D. Doerfler, Allan Erbsen, Matthew X. Etchemendy, Richard H. Fallon, Jr., Abbe R. Gluck, Neal Goldfarb, Christopher R. Green, Michael Steven Green, Mark D. Greenberg, James Grimmelmann, Alexandra D. Lahav, Margaret H. Lemos, Michael J.Z. Mannheimer, Ralf Michaels, Ketan Ramakrishnan, Michael Rappaport, Richard M. Re, Kermit Roosevelt, Chaim Saiman, Amanda Schwoerke, Adam Steinman, James Y. Stern, Brian Z. Tamanaha, and Todd J. Zywicki; to participants in workshops at Duke University School of Law, the Federalist Society’s Annual Faculty Conference, George Mason University’s Antonin Scalia Law School, George Washington University Law School, Oxford University’s Jurisprudence Discussion Group, UCLA School of Law, University of Pennsylvania Law School, Washington University School of Law, and William and Mary Law School; and to the students in my Harvard Law School reading group on “The Common Law in the Federal System.” I am also grateful to Adam F. Griffin for excellent research assistance.
From and after September 1, 2019, this Article may be reproduced, excerpted, or redistributed in any format, for educational purposes and at or below cost, so long as any excerpt identifies the author, provides appropriate citation, and includes the disclaimer “Originally published in the California Law Review—Volume 107 (2019)” as well as this copyright provision.
. Am. Trucking Ass’ns v. Smith, 496 U.S. 167, 201 (1990) (Scalia, J., concurring in the judgment); see also Rogers v. Tennessee, 532 U.S. 451, 472 (2001) (Scalia, J., dissenting) (quoting Bacon’s aphorism).
. Antonin Scalia, Common-Law Courts in a Civil-Law System: The Role of United States Federal Courts in Interpreting the Constitution and Laws, in A Matter of Interpretation: Federal Courts and the Law 4, 7 (Amy Gutmann ed., 1997).
. Larry Kramer, The Lawmaking Power of the Federal Courts, 12 Pace L. Rev. 263, 283 (1992); see also Jesner v. Arab Bank, 138 S. Ct. 1386, 1409 (2018) (Alito, J., concurring in part and concurring in the judgment) (asserting that, after Erie, “the general common law was no more”); Kermit Roosevelt III, Resolving Renvoi: The Bewitchment of Our Intelligence by Means of Language, 80 Notre Dame L. Rev. 1821, 1840 (2005) (attributing to Erie the “devastating objection” that “there is no such thing”).
. See, e.g., Beever, supra note 3, at 425–26; Gerald J. Postema, Philosophy of the Common Law, in The Oxford Handbook of Jurisprudence and Philosophy of Law 588, 592–93 (Jules L. Coleman et al. eds., 2004); cf. Brian Zamulinski, Rehabilitating the Declaratory Theory of the Common Law, 2 J.L. & Cts. 171, 171 (2014) (discussing judges’ “ability to apprehend moral truths”).
. See, e.g., Caleb Nelson, A Critical Guide to Erie Railroad Co. v. Tompkins, 54 Wm. & Mary L. Rev. 921, 932–33 (2013); accord Caleb Nelson, The Legitimacy of (Some) Federal Common Law, 101 Va. L. Rev. 1, 10–13 (2015); Ernest A. Young, A General Defense of Erie Railroad Co. v. Tompkins, 10 J.L. Econ. & Pol’y 17, 31 (2013).
. Lea Brilmayer, Untethered Norms After Erie Railroad Co. v. Tompkins: Positivism, International Law, and the Return of the “Brooding Omnipresence,” 54 Wm. & Mary L. Rev. 725, 734 (2013) (criticizing this view).
. See The Nomination of Elena Kagan To Be an Associate Justice of the Supreme Court of the United States: Hearing Before the S. Comm. on the Judiciary, 111th Cong. 103 (2010) (statement of Elena Kagan) (stating that “it’s law all the way down”).
. Nelson, Critical Guide, supra note 15, at 930 (criticizing this view); cf. Morris R. Cohen, The Process of Judicial Legislation, 48 Am. L. Rev. 161, 168 (1914) (“[J]urists and legislators actively participate in different ways in the process of law making.”). But see Michael W. McConnell, Tradition and Constitutionalism Before the Constitution, 1998 U. Ill. L. Rev. 173, 197 (critiquing the “modernist misconception . . . that the common law is simply legislation by judges,” rather than “aris[ing] from custom”).
. Cf. H.L.A. Hart, The Concept of Law 134 (3d ed. 2012) (noting that “there is no single method of determining the rule for which a given authoritative precedent is an authority,” although “[t]he head-note is usually correct enough”).
. Cf. U.S. Bancorp Mortg. Co. v. Bonner Mall P’ship, 513 U.S. 18, 26 (1994) (citing the benefits of “[j]udicial precedents” as a reason not to vacate a judgment (internal quotation marks omitted)).
. Roosevelt, supra note 2, at 1076; see also Kramer, supra note 11, at 281 (describing “the modern understanding of common law as a form of positive law made by judges”); cf. Nelson, Legitimacy, supra note 15, at 14 (describing views that treat precedent as the source of common law).
. Fallon & Meltzer, supra note 4, at 1760; accord Richard S. Kay, Construction, Originalist Interpretation and the Complete Constitution, 19 U. Pa. J. Const. L. Online 1, 10 n.47 (2017) (describing Justice Brandeis’s position in Erie as “the now prevalent positivist view”).
. Id. (arguing that a judge can’t “unilaterally and finally fix the scope or meaning of a rule through his or her decision, regardless of how carefully crafted the language of the opinion is,” because “the quality and force of the reasoning, not the public utterance of it . . . lends authority to a court’s rationale”).
. Lawrence Lessig, Erie-Effects of Volume 110: An Essay on Context in Interpretive Theory, 110 Harv. L. Rev. 1785, 1791 (1997); see Duxbury, supra note 14, at 340; cf. Stephen E. Sachs, From St. Ives to Cyberspace: The Modern Distortion of the Medieval ‘Law Merchant,’ 21 Am. U. Int’l L. Rev. 685 (2006) (cautioning against hasty attributions of mercantile law to mercantile practice).
. U.C.C. § 1-303(c) (Am. Law Inst. & Unif. Law Comm’n 2017) (“A ‘usage of trade’ is any practice or method of dealing having such regularity of observance in a place, vocation, or trade as to justify an expectation that it will be observed with respect to the transaction in question.”).
. See, e.g., CIGNA Corp. v. Amara, 563 U.S. 421, 439 (2011) (interpreting a statutory entitlement to “appropriate equitable relief” to include “those categories of relief that, traditionally speaking[,] . . . were typically available in equity” (citations, emphasis, and internal quotation marks omitted)).
. See Judiciary Act of 1789, ch. 20, § 14, 1 Stat. 73, 81–82 (empowering courts of the United States to issue “all other writs not specially provided for by statute, which may be necessary for the exercise of their respective jurisdictions, and agreeable to the principles and usages of law”) (codified as amended at 28 U.S.C. § 1651 (2012)).
. An Ordinance to Enable the Present Magistrates and Officers to Continue the Administration of Justice, ch. 5, § 6 (Va. 1776), in A Collection of All Such Public Acts of the General Assembly, and Ordinances of the Conventions of Virginia, Passed Since the Year 1768, as Are Now in Force 37, 37 (Richmond, Thomas Nicholson & William Prentis 1785) (codified as amended at Va. Code Ann. § 1-200) (emphasis omitted).
. Cf. Mark Forsyth, The Elements of Eloquence: Secrets of the Perfect Turn of Phrase 45–46 (2014) (noting that size adjectives are placed before color, such that “great green dragons” will scan but “green great dragons” won’t).
. See, e.g., Restatement (Third) of the Foreign Relations Law of the United States § 102(2) (“Customary international law results from a general and consistent practice of states followed by them from a sense of legal obligation.”).
. See, e.g., Cristina Bicchieri, Norms in the Wild: How to Diagnose, Measure, and Change Social Norms 35 (2016) (defining a social norm as “a rule of behavior such that individuals prefer to conform to it on condition that they believe that (a) most people in their reference network conform to it (empirical expectation), and (b) that most people in their reference network believe they ought to conform to it (normative expectation)” (emphasis omitted)).
. See generally Bryan A. Garner, Making Peace in the Language Wars, 7 Green Bag 2d 227 (2004); David Foster Wallace, Tense Present: Democracy, English, and the Wars over Usage, Harper’s, Apr. 2001, at 39.
. AP English Language and Composition: Sample Scoring Guidelines for the Synthesis Essay, Coll. Bd. (2006), http://apcentral.collegeboard.com/apc/public/repository/samplescoringguidelin_51461.pdf [https://perma.cc/SV4U-887H].
. Cf. Matthew Kramer, Of Final Things: Morality as One of the Ultimate Determinants of Legal Validity, 24 Law & Phil. 47, 59 (2005) (noting that “lexicographers and grammatical experts,” to be “minimally successful in the fulfillment of their assigned functions,” must “accurately represent the workings of the language” and “reflect the norms that are actually presupposed in the communications of the relevant language’s users”).
. See, e.g., Lessig, supra note 52, at 1791; cf. Emily Kadens, Custom’s Past, in Custom’s Future: International Law in a Changing World 11, 21–30 (Curtis A. Bradley ed., 2016) (describing this process).
. Simpson, supra note 47, at 362; see also Jeremy Bentham, A Comment on the Commentaries and a Fragment on Government 182–84 (J.H. Burns & H.L.A. Hart eds., 1977) (1823); accord Postema, supra note 13, at 592.
. 1 Joseph H. Beale, A Treatise on the Conflict of Laws § 4.7, at 40 (1935); cf. Matthew X. Etchemendy, Legal Realism and Legal Reality 14–19 (June 10, 2018), http://max-etchemendy.com/wp-content/uploads/2018/06/Legal-Realism-and-Legal-Reality-6-10-18.pdf [https://perma.cc/9CRT-GN4L] (discussing Beale’s view of unwritten law).
. Harry R. Blythe, A Theory, 22 Green Bag 193, 193 (1910); cf. 1 Beale, supra note 113, § 4.7, at 39 (“[B]y a process of backward projection, it is argued that unless the courts changed the law the law must have been the same in 1200 that it is today.”).
. See David Micklethwait, Noah Webster and the American Dictionary 148–49 (2000); cf. Johnson, The Error of Our Ways, Economist, Feb. 2, 2019, at 81 (describing how “a napron” later became “an apron”).
. See John Gardner & Timothy Macklem, Book Review, Notre Dame Phil. Revs. (Dec. 8, 2011), https://ndpr.nd.edu/news/legality [https://perma.cc/AH6W-XTEV] (reviewing Scott J. Shapiro, Legality (2011)) (discussing the “chicken and egg problem”).
. See Dan Priel, Not All Law Is an Artifact: Jurisprudence Meets the Common Law, in Law as an Artifact 239, 247 (Luka Burazin et al. eds., 2018) (“Dress norms are often customary, but some people (‘fashion gurus’) may acquire a quasi-authoritative role that may shape norms in a conscious way.”).
. See 1 Beale, supra note 113, § 4.13, at 49; accord 1 James Wilson, Of the Common Law, in Works of James Wilson 423, 425 (James DeWitt Andrews ed., Chicago, Callaghan & Co. 1896) (describing it as part of “the very nature of a system of common law” that it “insensibly” adapts “to the situation and circumstances of the people, by whom it is appointed”).
. See, e.g., Zubulake v. UBS Warburg, 217 F.R.D. 309 (S.D.N.Y. 2003) (establishing influential cost-shifting factors for electronic discovery); Zippo Mfg. Co. v. Zippo Dot Com, Inc., 952 F. Supp. 1119 (W.D. Pa. 1997) (proposing an oft-cited test for Internet jurisdiction).
. Strategery, Oxford English Dictionary, http://www.oed.com/view/Entry/33737630 (last visited Jan. 4, 2018); see Saturday Night Live (NBC television broadcast Oct. 7, 2000), https://www.nbc.com/saturday-night-live/video/cold-opening-gore–bush-first-debate/n11360; Strategery, Wikipedia (Dec. 26, 2017), https://en.wikipedia.org/w/index.php?title=Strategery&oldid=817180017 [https://perma.cc/4P35-3MPX].
. See David A. Strauss, Common Law Constitutional Interpretation, 63 U. Chi. L. Rev. 877, 885 (1996) (noting, to his dismay, that the “common law method” is “not [one] we usually associate with a written constitution”); see also Stephen E. Sachs, Originalism as a Theory of Legal Change, 38 Harv. J.L. & Pub. Pol’y 817, 820 (2015) (“If the law was X at the Founding but is supposed to be Y today, the natural follow-up question is what happened in between—and why whatever happened (an amendment, a statute, a shift in custom or usage) was legally capable of making that change.”).
. Cf. Livingston v. Jefferson, 15 F. Cas. 660, 664 (C.C.D. Va. 1811) (No. 8411) (Marshall, Circuit Justice) (refusing to depart from an “innovation on the old rule” because it had since become the standard “for a long course of time”).
. Cf. 1 Blackstone, Commentaries, supra note 3, at *70–71 (noting that “a positive law” might be “fixed and established by custom, which custom is evidenced by judicial decisions; and therefore [could] never be departed from by any modern judge without a breach of his oath and the law”); Sachs, supra note 138, at 844 (“Our law requires us, at one and the same time, to overlook past violations and to commit to being rule-governed in the future; to go, and sin no more.”).
. Mark Greenberg, How Facts Make Law, in Exploring Law’s Empire: The Jurisprudence of Ronald Dworkin 225, 249–51 (Scott Hershovitz ed., 2008); see also Mark D. Greenberg & Harry Litman, The Meaning of Original Meaning, 86 Geo. L.J. 569, 614–17 (1998).
. E.g., David Lewis, New Work for a Theory of Universals, 61 Australasian J. Phil. 343, 375–76 (1983); David Plunkett, A Positivist Route for Explaining How Facts Make Law, 18 Legal Theory 139 (2012).
. Cf. James Grimmelmann, All Smart Contracts Are Ambiguous 1 Pa. L. & Innovation J. (forthcoming 2019) (manuscript at 10–14), https://ssrn.com/id=3315703 (discussing the Python programming language).
. Cf. Leslie Green, Legal Positivism, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., Fall 2009), https://plato.stanford.edu/archives/fall2009/entries/legal-positivism [https://perma.cc/4VQ8-4DFT] (drawing this distinction in the context of mathematics or foreign law).
. See Deborah Cameron, Gender and the English Language, in The Handbook of English Linguistics 724, 736–38 (Bas Aarts & April McMahon eds., 2006); John E. McIntyre, Everyone Goes Their Own Way, Balt. Sun (Jan. 6, 2015), http://www.baltimoresun.com/news/language-blog/bal-everyone-goes-their-own-way-20150106-story.html.
. Brian Bix, John Austin, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., Summer 2017) (citation omitted), https://plato.stanford.edu/archives/sum2017/entries/austin-john [https://perma.cc/2P4Q-RE45].
. Leslie Green, Law and Obligations, in Oxford Handbook of Jurisprudence and Philosophy of Law 514, 517 (Jules Coleman & Scott Shapiro eds., 2002); accord Fallon & Meltzer, supra note 4, at 1763 (“The fallacies of a ‘predictive theory’ of law, which Holmes is often taken to have asserted, are well known.”).
. See, e.g., Brian Leiter, Legal Realism and Legal Positivism Reconsidered, in Naturalizing Jurisprudence: Essays on American Legal Realism and Naturalism in Legal Philosophy 59, 60, 71 (2007) (arguing that the Realists “had nothing explicit to say about the concept of law”). But see Green, supra note 95, at 11–12 (arguing that some realists did believe this); cf. John Chipman Gray, The Nature and Sources of the Law 124–25 (Roland Gray ed., 2d ed. 1921) (arguing that “all the Law is judge-made law,” for “it is only words that the legislature utters; it is for the courts to say what those words mean”).
. See Leslie Green, Introduction to Hart, supra note 37, at xvii (“It would be like being told God doesn’t exist, only to find out that the interlocutor doesn’t believe in the existence of dogs either.”).
. See Livingston v. Jefferson, 15 F. Cas. 660, 665 (C.C.D. Va. 1811) (No. 8411) (Marshall, Circuit Justice) (“This common law has been adopted by the legislature of Virginia. Had it not been adopted, I should have thought it in force.”).
. See Black & White Taxicab & Transfer Co. v. Brown & Yellow Taxicab & Transfer Co., 276 U.S. 518, 534 (1928) (Holmes, J., dissenting) (“Louisiana is a living proof that it need not be adopted at all.”).
. David Ibbetson, Custom in Medieval Law, in The Nature of Customary Law: Legal, Historical and Philosophical Perspectives 151, 165 (Amanda Perreau-Saussine & James Bernard Murphy eds., 2007); see also Oran Doyle, Conventional Constitutional Law, 38 Dublin Univ. L.J. 311, 315 (2015).
. Tiersma, supra note 36, at 1188; see also Chaim Saiman, The Domain of Private Law: An Anglo vs. American Comparison 22–23 (Jan. 4, 2019) (unpublished manuscript on file with author) (describing the role of general law in the Commonwealth).
. See, e.g., McKenna v. Ortho Pharm. Corp., 622 F.2d 657, 661 (3d Cir. 1980) (making “a prediction of how the state’s highest court would decide were it confronted with the problem”); accord Bradford R. Clark, Ascertaining the Laws of the Several States: Positivism and Judicial Federalism After Erie, 145 U. Pa L. Rev. 1459, 1461 (1997).
. Cf. Young, supra note 15, at 32 (“That distinction [between following others and acting by one’s best lights] exists even in contemporary practice prescribed by Erie itself, as federal courts must try to follow state law in diversity cases while enjoying greater autonomy in enclaves of federal common law.”).
. Hart, supra note 37, at 272; see also id. at 272 (describing “the judge as filling the gaps by exercising a limited law-creating discretion”); accord Kramer, supra note 11, at 269 (“Deciding individual cases thus generates some common law because the process of adjudication necessarily entails articulating rules to elaborate and clarify the meaning and operation of statutory texts.”); Postema, supra note 13, at 588 (discussing the view that, “[i]n novel cases, where law arguably is silent, judges fill the silence with new binding precedents,” and in this way “massage precedents and in the process make new law”).
. Charles E. Clark, Federal Procedural Reform and States’ Rights; to a More Perfect Union, 40 Tex. L. Rev. 211, 223–24 (1961). But see generally Tamanaha, supra note 31 (questioning how revelatory this really was).
. See Chris Green, TerBeek, Theories, Movements, and Direction of Fit, Originalism Blog (May 12, 2017), http://originalismblog.typepad.com/the-originalism-blog/2017/05/terbeek-theories-movements-and-direction-of-fit.html [https://perma.cc/95DV-W9NR].
. Compare Comment, James Durling, The Intercircuit Exclusionary Rule, 128 Yale L.J. 231, 232, 239 (2018) (noting that courts’ attempts to apply sister-circuit Fourth Amendment law conflict with the traditional principle “that federal law is unitary”), with id. at 240 (describing federal law’s uniformity as “a myth,” but only “[f]or practical purposes” like forum-shopping).
. See James Y. Stern, Property, Exclusivity, and Jurisdiction, 100 Va. L. Rev. 111, 162–63 (2014) (on the “bundle of sticks”); Jeremy Waldron, “Transcendental Nonsense” and System in the Law, 100 Colum. L. Rev. 16, 25–26 (2000) (on systematicity).
. Cf. Abraham Lincoln, First Inaugural Address (Mar. 4, 1861), http://avalon.law.yale.edu/19th_century/lincoln1.asp [https://perma.cc/DGP6-FGW8] (“[I]f the policy of the Government upon vital questions affecting the whole people is to be irrevocably fixed by decisions of the Supreme Court, . . . the people will have ceased to be their own rulers, having . . . practically resigned their Government into the hands of that eminent tribunal.”).
. Cf. John Finnis, The Fairy Tale’s Moral, 115 Law Q. Rev. 170, 172–73 (1999) (describing the declaratory theory less as a “descriptive assertion” than as “an important element in judicial duty”).
. H.L.A. Hart, Problems of the Philosophy of Law, in Essays in Jurisprudence and Philosophy 88, 105–06 (1983); cf. Gottlob Frege, Introduction to The Foundations of Arithmetic, at x (J.L. Austin trans., 2d rev. ed. 1980) (urging theorists “always to separate sharply the psychological from the logical”).
. See Timothy A.O. Endicott, Legal Interpretation, in The Routledge Companion to Philosophy of Law 109, 109 (Andrei Marmor ed., 2012) (“If, on the best interpretation, the law requires you to do what is reasonable, you will need a technique other than interpretation in order to identify the reasons at stake.”).
. See Tenn. Valley Auth. v. Hill, 437 U.S. 153, 184 (1978) (reading the Endangered Species Act as requiring certain protections, “whatever the cost”); see also Asher Steinberg, Supreme Court 2016 Statutory Term in Review: Purposively Interpreting Statutes that Don’t Say Much in Midland Funding, LLC v. Johnson, The Narrowest Grounds (Aug. 26, 2017), http://narrowestgrounds.blogspot.com/2017/08/supreme-court-2016-statutory-term-in_26.html [https://perma.cc/E5EL-DQHH]. But cf. Tara Leigh Grove, Tiers of Scrutiny in a Hierarchical Judiciary, 14 Geo. J.L. & Pub. Pol’y 475, 476–77 (2016) (defending such tests as a matter of the supervisory power over lower courts).
. Compare, e.g., Henry J. Friendly, In Praise of Erie—and of the New Federal Common Law, 39 N.Y.U. L. Rev. 383, 421 (1964) (describing Erie as “forging a new centripetal tool incalculably useful to our federal system”), and Young, supra note 15, at 113 (arguing that Erie “fit[s] beautifully with the federalism doctrine that would emerge after the New Deal”), with Michael S. Greve, The Upside-Down Constitution 222 (2012) (arguing that Erie “mobilizes plaintiffs’ opportunistic state forum choices to enhance state power over interstate commerce”), and Suzanna Sherry, Wrong, Out of Step, and Pernicious: Erie as the Worst Decision of All Time, 39 Pepp. L. Rev. 129 (2011) (taking a somewhat negative view); cf. Nelson, Critical Guide, supra note 15, at 965–70 (arguing that Erie may turn out to increase legal uncertainty, rather than diminish it).
. Cf. Kleinwort Benson Ltd. v. Lincoln City Council  2 AC 349 (HL) 376 (Lord Goff of Chieveley) (appeal taken from Commercial Ct.) (UK) (discussing whether a payment made under a settled view of the law, later unsettled by judicial decision, was really a “mistake of law”).
. See Coon v. Med. Ctr., Inc., 300 Ga. 722, 730 (2017) (adhering to the view that “there is one common law that can be properly discerned by wise judges, not multiple common laws by which judges make law for their various jurisdictions”).
. 28 U.S.C. § 1652 (2012) (“The laws of the several states, except where the Constitution or treaties of the United States or Acts of Congress otherwise require or provide, shall be regarded as rules of decision in civil actions in the courts of the United States, in cases where they apply.”).
. See 41 U.S. (16 Pet.) 1, 18–19 (1842); see also Bridwell & Whitten, supra note 293, at 99 (“General commercial law disputes were treated independently by the federal courts because they were cases in which the states themselves purported to adhere to an extraterritorial body of customary principle.”).
. See, e.g., Edward A. Purcell, Jr., Brandeis and the Progressive Constitution: Erie, the Judicial Power, and the Politics of the Federal Courts in Twentieth-Century America 51–52 (2000); Bridwell & Whitten, supra note 293, at 3 (discussing practice after 1860); Young, supra note 15, at 37–38.
. 28 U.S.C. § 1652 (2012) (requiring that “[t]he laws of the several States . . . be regarded as rules of decision in civil actions in the courts of the United States, in cases where they apply”); see also Nelson, Critical Guide, supra note 15, at 951–56 (discussing Brandeis’s misconstruction of the Act); Young, supra note 15, at 26–27 (same).
. See Wainwright v. Goode, 464 U.S. 78, 84 (1983) (“[T]he views of the State’s highest court with respect to state law are binding on the federal courts.”); see also Murdock v. City of Memphis, 87 U.S. (20 Wall.) 590, 626 (1875) (“The State courts are the appropriate tribunals . . . for the decision of questions arising under their local law, whether statutory or otherwise.”); Young, supra note 15, at 105–07.
. See Murdock v. City of Memphis, 87 U.S. (20 Wall.) 590, 626 (1875) (“The State courts are the appropriate tribunals . . . for the decision of questions arising under their local law, whether statutory or otherwise.”); see also Swift v. Tyson, 41 U.S. (16 Pet.) 1 (1842) (discussing “the positive statutes of the state, and the construction thereof adopted by the local tribunals”).
. See, e.g., Alaska Organic Act, ch. 53, § 7, 23 Stat. 24, 25 (1884) (incorporating wholesale as Alaska territorial law “the general laws of the State of Oregon now in force”); 15 Cong. Rec. 529 (1884) (statement of Sen. Benjamin Harrison) (noting that “the committee had [not] made any careful study of the laws,” but had “supposed that the Oregon code was in a more mature or satisfactory shape”).
. V.I. Code Ann. tit. 1, § 4 (establishing “the rules of the common law, as expressed in the restatements of the law approved by the American Law Institute,” as “rules of decision”), recognized as abrogated by statute, Virgin Islands v. Connor, 60 V.I. 597, 602–03 (2014).
. See Stewart Jay, Origins of Federal Common Law: Part One, 133 U. Pa. L. Rev. 1003, 1010 (1985) (“There was no coherent concept in the early nineteenth century of ‘federal common law’ as we now make use of that expression.”); Stewart Jay, Origins of Federal Common Law: Part Two, 133 U. Pa. L. Rev. 1231, 1274 (1985) (“Nothing like the theory of jurisdiction just articulated was generally accepted until far into the nineteenth century.”).
. Compare, e.g., E. River S.S. Corp. v. Transamerica, 476 U.S. 858, 865 (1986) (endorsing the federal courts’ power to apply “newly created rules” of admiralty), with Am. Ins. Co. v. Canter, 26 U.S. (1 Pet.) 511, 545–46 (1828) (“A case in admiralty does not, in fact, arise under the Constitution or laws of the United States. These cases are as old as navigation itself; and the law admiralty and maritime, as it has existed for ages, is applied by our Courts to the cases as they arise.”).
. Compare, e.g., Shady Grove Orthopedic Assocs., P.A. v. Allstate Ins. Co., 559 U.S. 393, 416 (2010) (arguing that when no federal authority provides a rule of decision, “state law must govern because there can be no other law” (citations and internal quotation marks omitted)), with Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 49–50 (1825) (denying that “the practice of the Federal Courts, and the conduct of their officers, can be indirectly regulated by the State legislatures”), and Rule, 2 U.S. (2 Dall.) 411, 413–14 (1792) (“The Court considers the practice of the courts of King’s Bench and Chancery in England, as affording outlines for the practice of this court . . . .”).
. Compare, e.g., Sosa v. Alvarez-Machain, 542 U.S. 692, 729 (2004) (endorsing, in the field of customary international law, the occasional “judicial recognition of new substantive rules” as federal common law), with N.Y. Life Ins. Co. v. Hendren, 92 U.S. 286, 286–87 (1876) (refusing to exercise federal-question review of an issue under the law of nations, which must be decided on “principles of general law alone”); see also Ernest A. Young, Sorting Out the Debate Over Customary International Law, 42 Va. J. Int’l L. 365 (2002) (describing customary international law as part of the general law).
. 244 U.S. 205, 221 (1917) (Holmes, J., dissenting); cf. Prentis v. Atl. Coast Line Co., 211 U.S. 210, 226 (1908) (Holmes, J.) (“[T]he making of a rule for the future . . . is an act legislative, not judicial, in kind . . . .”).
. See, e.g., William N. Eskridge, Jr., Dynamic Statutory Interpretation, 135 U. Pa. L. Rev. 1479, 1479 (1987) (arguing that statutes “should—like the Constitution and the common law—be interpreted ‘dynamically,’ that is, in light of their present societal, political, and legal context”); accord Guido Calabresi, A Common Law for the Age of Statutes (1982).