Article

Inside-Out: Beyond the Internal/External Distinction in Legal Scholarship

A distinction commonly drawn in legal scholarship deserves scrutiny. To begin to see why, consider this question: What do the following claims by legal scholars have in common?

(1) That economic explanations of tort law which interpret it as an instrument for achieving social goals, such as compensating victims or deterring unreasonably risky conduct (or both), are defective in part because such explanations do not “work through” the concepts that judges invoke in their opinions when deciding cases.

(2) That historical accounts of judicial decisionmaking that explain case outcomes by reference to judges’ ideologies or economic self-interest do not threaten normative interpretations of legal practice because such explanations do not offer courts any concrete guidance as to how to decide future cases.

(3) That political scientists’ criticisms to the effect that legal scholars commonly make fallacious inferences miss their mark because they fail to understand that legal scholars have rhetorical goals not shared by political scientists.

(4) That the arguments of philosophers and neuroscientists about the existence or nonexistence of free will are irrelevant to “all the participants in the legal system” because the criminal law assumes that people are morally responsible for their actions irrespective of whether those actions were causally determined.

(5) That when lawyers write history they ignore evidence and distort the facts in order to rationalize and legitimize legal practice.

There are really two answers to this question. The first is that legal scholars have framed all of these claims around a distinction between “internal” and “external” forms of explanation, criticism, or argument. In each case, the suggestion made is that a particular kind of analysis is flawed or misguided because it is not of the right sort. It is “external” when what is required is an “internal” analysis, or vice versa. That all of the above claims have been conceptualized in this way may at first seem odd since they otherwise seem to be about quite unrelated issues.

From a broader perspective, however, each of these arguments can be seen as making judgments about what counts as a “legal” argument or critique and what does not. That is the second thing these claims have in common. They are all efforts to draw the boundaries of law. Each seeks to distinguish, for one reason or another, the aims and methods of law from those of other academic disciplines. In particular, they seek to distinguish law from those disciplines whose methods are aimed at better understanding the natural or social world, whether in the humanities, sciences, or social sciences. Here I do not mean “law” in the sense of those rules or principles that are (or properly ought to be) enforced by the state, but rather “law” in the sense of those materials, methods, and values that influence the form and content of those rules and principles. In other words, each makes a claim about the nature and boundaries of what John Chipman Gray called the “sources” of law.

Such an effort does not alone warrant criticism. To the contrary, questions about which materials and values judges and other legal decision makers ought to rely on are foundational ones. So, too, are questions about which materials, methods, and values they actually do rely on, whether they should do so or not. Moreover, the internal/external distinction captures well a powerful intuition. Some forms of scholarship—say, traditional doctrinal analyses—do seem to be in some sense launched from within the legal enterprise, whereas others—such as empirical studies of judicial behavior—appear to offer descriptions or critiques from outside it.

In part for these reasons, the internal/external distinction has now become so entrenched in the consciousness of legal scholars that recently a pair of prominent scholars has felt compelled to call out various judges and legal theorists for having committed what they call the “inside/outside fallacy.” According to Professors Eric Posner and Adrian Vermeule, these theorists suffer from “methodological schizophrenia” because they adopt “internal” and “external” perspectives simultaneously. The point of their article, they explain, is not that one type of scholarship is better than the other, but rather simply that the two perspectives are fundamentally incompatible and so must be kept separate. Although they recognize that the inconsistency they identify could be framed in other terms, they dub it the “inside/outside fallacy” on the ground that the internal/external distinction has been “traditionally a central issue for legal theory.”

And that is true. Or at least it is true if by “traditionally” one means “for the past few decades.” For most of the twentieth century, legal theorists did not conceptualize philosophical, historical, or sociological investigations as “outside” of, or “external” to, law. Although a version of the distinction has a long history in the philosophy of the human sciences (or what are today called the “social sciences”), its introduction into legal theory is relatively recent. H.L.A. Hart famously invoked the distinction between two “points of view”—one “internal,” the other “external”—in his 1961 jurisprudential classic, The Concept of Law. But the distinction did not really take hold as a way of interpreting theoretical claims about law until after Ronald Dworkin made use of a similar distinction in his 1986 opus, Law’s Empire.

The distinction’s recent popularity thus invites at least three sorts of questions. The first is historical or explanatory: Why has the distinction come to play such a leading role in legal theory? Why does it today seem so natural? The second is conceptual or analytic: Is there really just one internal/external distinction or are there several going under the same name? And if there are multiple distinctions, what are they and how do they relate to each other? The third set of questions is evaluative or normative: What functions does the distinction serve? And are those functions useful ones for legal theory, practice, or education?

The aim of this Article is to offer some answers to these questions. It argues, in brief, that the distinction has taken hold as a result of both intellectual and institutional changes in the legal academy in the last few decades of the twentieth century. These changes created a need for, and a method of, reconciling increasingly popular forms of interdisciplinary scholarship with more traditional legal scholarship. The internal/external distinction has largely met that need, which contributes to its popularity. But it has done so in part by trading on a crucial ambiguity—between a substantive distinction, on the one hand, and various methodological distinctions, on the other. That ambiguity first appeared in Hart’s work and has clung to the distinction ever since. Thus, distinguishing among, and clarifying the meaning of, the different versions of the distinction is a worthwhile endeavor in itself.

Moreover, even when its meaning is clear, today the various methodological versions of the distinction do more harm than good. When used as a methodological criterion, the distinction rarely serves as a useful conceptual tool to clarify issues or open up avenues of inquiry. Instead, it operates mainly as a rhetorical weapon whose function is to insulate particular substantive views from arguments deemed to be threatening to it. Its tendency has thus been to cabin scholarly debate about the nature and purposes of law, rather than to widen it, and to dampen original thinking about such questions, rather than to stimulate or provoke it.

The burden of this Article is to support these broad claims. It will do so in four Parts. Part I will seek to identify and distinguish among the three main versions of the distinction as they first appeared in the jurisprudential writings of H.L.A. Hart and Ronald Dworkin. It will also describe the historical context in which Hart and Dworkin wrote in an effort to explain why the distinction may have appeared when it did. The primary aim of Part I, however, is to show the role that the “internal point of view” played in each theorist’s philosophy of law and, more specifically, to show how the methodological versions of it enabled both thinkers to obscure or evade difficult questions. In other words, the point is to show that the trouble with the internal/external distinction began at its inception.

Parts II through V will then analyze and evaluate each of these three versions of the distinction in more detail. In each case, the particular version of the distinction that Hart and Dworkin drew has been applied in other contexts by scholars who have then modified it in subtle ways. For this reason, it is possible to see Hart’s and Dworkin’s particular distinctions as instances of more general dichotomies. I will give each of these more general versions a new name in order to distinguish them more clearly from each other. Part II will take up the Genuine versus Instrumental Rule Follower distinction, which is a substantive one about how legal actors use (or fail to use) rules in guiding their conduct. This is the least objectionable use of the distinction, though even here I will suggest that framing the distinction as one of “point of view” or “perspective” is misleading and results in confusion. Part III will then consider the first methodological version of the distinction. I call this the Participant Perspective versus Non-Participant Perspectives distinction, which distinguishes between a variety of different ways of understanding or explaining social (and hence legal) phenomena. I will argue that only a commitment to very controversial epistemological or metaphysical views could justify adopting either the Participant Perspective or any of the Non-Participant Perspectives to the exclusion of other methods and that, instead, the defense of any method of social inquiry should lie in the adequacy of the explanations it offers.

Reasonable Expectations of Anonymity

New technologies and methods of data analysis are being used by the government to monitor the public in ways that were unimaginable a decade ago. Law enforcement agencies ranging from municipal police forces to the Department of Homeland Security are using tools such as genetic databanks, biometric scanners, roadside cameras, and cell phone metadata analysis to gather detailed information about the lives of individuals who are not suspected of any wrongdoing. The meaningful question in this area is no longer what information the government can obtain about us, but rather what information is beyond its reach.

The reason for this is that the Supreme Court has concluded that the Fourth Amendment’s protections do not apply to any information that has been exposed to the public or third parties. This includes information about our public movements, Internet usage, cell phone calls, and so on. Such information is per se fair game for police collection by any means.

This Article argues that the Court’s conclusion derives from a mistaken conflation of privacy and anonymity, and that understanding the difference between these concepts reveals strong substantive and formal reasons for interpreting the Fourth Amendment to protect not only reasonable expectations of privacy, but also “reasonable expectations of anonymity.” Further, it demonstrates that the incorporation of this new analytic concept into Fourth Amendment jurisprudence yields significant value: first, by identifying otherwise-unrecognizable ways in which new techniques of big data implicate the Constitution, and second, by delivering on the unfulfilled promise of the Supreme Court’s teaching that “the Fourth Amendment protects people, not places.” A more detailed roadmap of this argument follows.

The argument begins, in Part I, with an analysis of the Fourth Amendment right to be free from unreasonable “searches”—a term that the Supreme Court has, ever since Katz v. United States, interpreted to mean violations of reasonable expectations of privacy. The key contribution of this Part is clarifying what the Court means by “privacy” in the Fourth Amendment context, which has been the subject of much confusion in the literature. A close analysis of the case law reveals that the Court has adopted what can be termed an “epistemic,” rather than a normative, conception of privacy. The clarification of this point provides the foundation for a discussion of two doctrines that significantly limit the scope of the Fourth Amendment’s protections: the public exposure and third party doctrines, under which the Supreme Court has concluded that the Fourth Amendment’s protections do not apply to any information that has been exposed to the public or third parties.

The question that motivates this Article is whether the Supreme Court has erred in reaching this conclusion. The dominant view in the privacy scholarship is that the Court has failed to account for the ways in which privacy can exist in degrees. While this critique is correct as far as it goes, this Article demonstrates that it only identifies part of the problem.

The even deeper problem, identified in Part II, is that courts—along with most scholars—have incorrectly assumed that there is only one way of protecting a piece of personal information from public access: the one we call “privacy.” In doing so, they have overlooked a distinct and equally important way of doing so: through anonymity. This oversight derives from the fact that anonymity and privacy have been mistakenly conflated.

An example helps introduce the key distinction that has gone unrecognized. Imagine, for instance, that a person’s medical file contains a piece of paper with the results from his blood test, but his doctor removes the paper and places it in a blank file. If we subsequently obtained access to this person’s medical file, without the test results, we would describe the situation using the concept of privacy: We would say “the privacy of the person is protected,” or “the associated information is private.” If, on the other hand, we obtained access to the test results, without the medical file, we would describe the situation using the concept of anonymity: We would say “the anonymity of the test results is protected,” or “the associated person is anonymous.”

What this example illustrates is two basic points about anonymity and privacy that have been misunderstood. The first is a point about their substantive difference. Although both anonymity and privacy prevent others from gaining access to a piece of personal information, they do so in opposite ways: Privacy involves hiding the information, whereas anonymity involves hiding what makes it personal. The second point is about their formal relationship. Anonymity and privacy have the same causal origin and thus are flip sides of each other: They describe opposite sides of a single underlying event.

This account of the nature of anonymity, when combined with the insight that Katz and its progeny adopt a purely epistemic conception of privacy, has significant legal implications. As identified in the final Section of Part II, it reveals strong substantive and formal reasons for reading the Fourth Amendment to protect not only reasonable expectations of privacy, but also “reasonable expectations of anonymity.”

It is perhaps worth highlighting here that this is not a normative argument about what our constitutional law should be, but rather a legal argument about the best way to interpret the Fourth Amendment precedents that we have. Thus, I do not question whether Katz and its progeny provide the best interpretation of the text of the Fourth Amendment, but rather make a claim about the best reading of this case law, accepting that it provides a controlling reading of the text. Further, and relatedly, I do not question the premise that the Fourth Amendment does not prohibit the government from collecting personal information that has been knowingly exposed to the public, but rather show that this premise does not support the conclusions reached by courts in many of the public exposure cases—that the logic of the public exposure doctrine imposes limits that have not been recognized. This is not to say, however, that my argument is at odds with those of scholars who argue for more radical revisions of Fourth Amendment jurisprudence on normative grounds. Rather, a normative approach might reach the same conclusions on many issues, as will become clear in Part III.

The practical payoff of incorporating the concept of “reasonable expectations of anonymity” into Fourth Amendment jurisprudence is the focus of Part III, which identifies two general dimensions in which it yields significant insights. The first dimension is analytic, where thinking in terms of anonymity identifies otherwise-unrecognizable ways in which many new techniques of big data implicate the Fourth Amendment. This is demonstrated by reference to the question of whether two new techniques of data aggregation and analysis can constitute Fourth Amendment searches. One is a form of genetic identification known as “familial searching,” in which a criminal DNA database is used to identify persons who do not meet the legal criteria for inclusion, but happen to be related to people who do. The other is the use of tools such as biometric-equipped video cameras, GPS, and the metadata from cell phone calls to conduct long-term locational tracking of people’s movements in public.

Both of these techniques have faced significant criticism in the privacy scholarship, and there is language in judicial opinions questioning their legitimacy, but neither the literature nor the judicial opinions have offered a strong legal argument for how they can constitute Fourth Amendment searches. The reason for this is that the constitutional problem cannot be sufficiently explained in terms of privacy.

What is needed is the concept of reasonable expectations of anonymity, which not only reveals the Fourth Amendment interests that are violated by these specific techniques, but also provides a meaningful standard that can be used more generally to determine when data aggregation implicates the Fourth Amendment and when it does not. In these ways, the concept helps solve difficult puzzles left open by the concurring opinions in United States v. Jones.

In addition to providing the analytic power necessary to understand the unconstitutionality of many new techniques of big data, the incorporation of anonymity into Fourth Amendment jurisprudence will help deliver on the unfulfilled promise of the Supreme Court’s teaching that the Fourth Amendment is meant to protect “people, not places.” There are two central ways in which it does so, as the final Section of Part III demonstrates.

The first is by revealing that the structural features of the world that are capable of protecting Fourth Amendment interests are far more complex and expansive than the Supreme Court has recognized. Although the Court has moved beyond a property-based conception of Fourth Amendment interests, the only structural features of the world that the Court has recognized as protecting these interests are those that protect the “privacy” side of secrecy: Homes, car trunks, envelopes, and other containers all hide facts about a person whose identity might be known. Yet the structures that are capable of maintaining the secrecy of “personal information” are not limited to those that hide the piece of information. Rather, as this Article makes clear, they can also include structures that hide what makes that information personal or, in other words, structures that make it anonymous. For example, the size of a city, the layout of its streets, and the presence of crowds can all contribute to making someone’s public actions anonymous. By uncovering the legal significance of these structures, attention to anonymity opens up new types of public spaces to the Fourth Amendment’s protections.

The second and related way in which attention to anonymity can help deliver on the promise of the Fourth Amendment is by expanding the sources of law and norms that can provide the basis for its protections. Although property law is often cited as the quintessential enabling source of law for reasonable expectations of privacy, reasonable expectations of anonymity may be created by sources of law ranging from whistle-blowing statutes and agency law to copyright and the First Amendment, all of which protect anonymity rights. In the First Amendment context, for example, the Supreme Court has held that “an author’s decision to remain anonymous . . . is an aspect of . . . freedom of speech.” Thus, an anonymity-based understanding of Fourth Amendment claims could ground them in new legal and normative foundations, including other constitutionally protected liberties.

Further, these two lessons—along with the other insights of this Article—are not only applicable to the Fourth Amendment. Rather, as suggested in the Conclusion, they are relevant to the many other sources of law that provide legal protection to reasonable expectations of privacy. Across all of these domains, attention to the distinct concept of anonymity can reveal important and viable interests in the secrecy of personal information that have gone unrecognized, clarify new ways in which these interests are being threatened, and provide insights into how they can be better protected by our courts and our law.

The Original Source of the Cause of Action in Federal Courts: The Example of the Alien Tort Statute

Federal common law causes of action—actions created neither by Congress nor by state law—have long generated debate among judges and scholars. In Erie Railroad Co. v. Tompkins, the Supreme Court famously rejected “federal general common law.”  Nonetheless, the Court has cautiously embraced several specific enclaves of federal common law over the ensuing decades.  The question of federal judicial power to recognize federal common law causes of action arises in a range of contexts in the field of federal courts. For instance, may federal courts recognize an implied cause of action for the violation of a federal statute that does not itself create a cause of action? Relatedly, may federal courts recognize an implied cause of action for the violation of the Constitution when neither the Constitution nor a federal statute specifically creates one? Although courts and scholars continue to debate these questions, they have not reached a consensus on how to resolve them.

Recently, the power of federal courts to recognize federal common law causes of action has emerged as a key question under the Alien Tort Statute (“ATS”).  Congress enacted the ATS in 1789 as part of the First Judiciary Act. The ATS grants federal courts subject matter juris-diction over claims by aliens for torts in violation of the law of nations, but creates no cause of action itself.  In the last decade, the Supreme Court has twice interpreted the ATS and, in the process, has suggested that, although the statute is purely jurisdictional, federal courts have limited power to recognize a small handful of federal common law causes of action when exercising this jurisdiction.

Over time, judges and scholars have reached different conclusions in different contexts about the power of federal courts to recognize federal common law. From the Founding through the nineteenth century, the Supreme Court did not recognize any “federal common law”—that is, “federal rules of decision whose content cannot be traced directly by traditional methods of interpretation to federal statutory or constitution-al commands.”  To be sure, in certain cases, exemplified by the Supreme Court’s decision in Swift v. Tyson, early federal courts applied general law—a transnational source of law that included the law merchant, the law maritime, and the law of state-state relations. General law, however, did not preempt contrary state law or create causes of action. Moreover, general law was not federal common law. Unlike modern federal common law, general law neither supported federal question jurisdiction nor preempted contrary state law. The Supreme Court stopped applying general law as such in 1938 when it held in Erie that “[t]here is no federal general common law.”  Nonetheless, following Erie, the Court recognized several distinct “enclaves” of federal common law. In recent decades, the Court has been reluctant to recognize new enclaves because of concerns that judicial creation of federal common law is in tension with Erie, and with principles of separation of powers and federalism more generally.

Against this background, the Supreme Court interpreted the ATS for the first time in 2004. In Sosa v. Alvarez-Machain, the Court concluded that “the ATS is a jurisdictional statute creating no new causes of action.”  Nonetheless, the Court believed that “[t]he jurisdictional grant is best read as having been enacted on the understanding that the com-mon law would provide a cause of action for the modest number of international law violations with a potential for personal liability at the time.”  The Court rested this belief on the assumption that the First Congress would have understood “the ambient law of the era” to pro-vide the causes of action that federal courts would adjudicate in exercising their ATS jurisdiction.  In other words, the Court “assume[d] that the First Congress understood that the district courts would recognize private causes of action,” derived from ambient law, “for certain torts in violation of the law of nations.”  On the basis of this assumption, the Court suggested that federal courts today may “recognize private claims under federal common law” for a narrow range of international law violations.  Commentators have generally embraced Sosa’s vision of ambient law and federal judicial power at the Founding with little independent historical analysis or verification.

In fact, the claim that early federal courts relied on “the common law” in the abstract to supply causes of action in civil suits rests on a false historical premise. Ambient or general law neither supplied nor was understood by the Founders to supply the cause of action in civil cases (including ATS cases) within the jurisdiction of early federal courts. Rather, Congress enacted specific statutes that prescribed the civil causes of action available in federal courts, as well as related matters. Although the full import of these statutes is largely overlooked to-day, they provide important context for understanding the kind of judicial power that federal courts exercised within their limited subject matter jurisdiction. Members of the First Congress considered and debated many aspects of federal judicial power over civil disputes—including whether litigants would enjoy the right to a jury trial,  how expansively federal courts would exercise equity jurisdiction,  how expensive and otherwise inconvenient federal litigation would be,  and how federal courts would order executions on their judgments.  In addition to these questions, but integrally related to them, Congress considered and pro-vided the source of the causes of action available in federal court. The resolution of all these questions depended in large part on the forms of proceeding that federal courts generally would use in civil cases. Congress addressed these questions by enacting a series of early federal statutes that specified the forms and modes of proceeding that federal courts were to apply.

To understand these statutes, one must understand the status of the common law in the United States prior to their enactment. Before the Constitution was adopted, state courts generally relied on common law forms of proceeding to adjudicate cases before them. During British rule, the colonies had applied common law as British law. After independence, the individual states chose to adopt the common law as state law. Each of the original thirteen states took action to receive the com-mon law—including its forms and modes of proceeding—by statute, constitutional provision, or judicial decision. The resulting state law forms of proceeding defined the remedies that were available to plain-tiffs for particular wrongs, and how state courts would determine a plaintiff’s right to a particular remedy. In other words, the traditional forms of proceeding adopted by the states defined the causes of action available to plaintiffs and the procedures to be used for adjudicating them. Over time, individual states molded these forms of proceeding in response to local circumstances, resulting in variations among state causes of action.

Accordingly, when Congress exercised its power to create lower federal courts in 1789,  there was no single body of “common law” that applied throughout the United States. Congress made no attempt to fol-low the states’ lead by adopting its own version of the common law as a whole for the nation, in part because any such attempt would have exceeded enumerated federal powers as then understood. Nor did Congress adopt uniform forms of proceeding for use in federal court, apparently because it was unable (or unwilling) to do so.  Rather, in the Process Acts of 1789 and 1792, Congress instructed inferior federal courts adjudicating common law suits to borrow the forms and modes of proceeding then in use by the states in which they sat. In this legislation, Congress balanced the need to create an effective federal judiciary with a desire to heed anti-Federalist concerns about consolidated national power at the expense of the states.  Members of Congress argued that the interests of the people would be “more secure under the legal paths of their ancestors, under their modes of trial, and known methods of decision.”  Accordingly, the First Congress established a “species of continuity” with diverse state practices by adopting the forms of proceeding of each state as the governing forms of proceed-ing for federal courts located in that state.  In cases in equity and admiralty, the First Congress directed federal courts to use the traditional forms of proceeding that applied in such cases. In doing so, Congress did not leave federal courts free to derive the causes of action they would employ from “ambient law.” Rather, Congress specifically adopted several preexisting, well-developed bodies of law for use in federal court.

This original source of the cause of action in federal courts has been largely forgotten by today’s lawyers and judges, not only because it is no longer relevant to their work, but also because modern legal sensibilities no longer identify “process” as the source of a “cause of action.”  When Congress adopted the Process Acts of 1789 and 1792, however, legal and equitable forms of proceeding defined the specific causes of action available to litigants. In light of this background, the First Congress contemplated that federal courts would hear only those causes of action already available under existing legal and equitable forms of proceeding. At the time, lawyers, judges, and other public officials understood that these forms of proceeding—not ambient law—defined the causes of action available to litigants. Once established, this connection would have been sufficiently obvious to members of the First Congress and the judiciary that it warranted little, if any, discussion.

This background has important implications for interpreting the ATS. The Supreme Court has self-consciously sought to identify and implement the First Congress’s understanding of the ATS. The Court has proceeded, however, on the false premise that the First Congress assumed that federal courts would adopt causes of action in ATS cases by looking to “the ‘brooding omnipresence’ of the common law then thought discoverable by reason.”  The Process Acts demonstrate that the First Congress made no such assumption. Instead, the Process Acts instructed federal courts adjudicating any of the legal claims over which they had subject matter jurisdiction—including ATS claims—to apply the forms of proceedings used by the courts of the state in which they sat. Neither early congressional legislation nor early federal judicial practice supports the Supreme Court’s suggestion that courts today should employ novel—and artificially narrow—federal common law causes of action in ATS cases. To the contrary, long-standing historical practice suggests that state law may continue to define the causes of action available when federal courts exercise jurisdiction under the ATS—not under the now-defunct Process Acts, but under Erie and the Rules of Decision Act.

This Article will proceed as follows. Part I will describe how the Supreme Court has recently interpreted the ATS to authorize the creation of limited federal common law causes of action. The Court’s approach is based on the mistaken historical premise that the ambient law of the era—rather than the Process Acts—would have supplied the causes of action available to early federal courts exercising jurisdiction under the ATS.

Part II will describe how the Process Acts of 1789 and 1792 adopted state forms of proceeding in cases at law and traditional forms of proceeding in equity and admiralty as the causes of action available in federal court. The Process Acts marked a victory for opponents of expansive federal judicial power, especially insofar as the Acts required federal courts to follow state forms of proceeding in common law cases.

Part III will describe how early federal courts understood their authority to entertain legal and equitable causes of action. In a range of contexts across jurisdictional grants, federal courts adjudicated only those causes of action authorized by the Process Acts of 1789 and 1792, absent contrary instructions from Congress in other statutes.

Part IV will describe some of the implications of this history for the source of the cause of action in ATS cases. Although this Article will not attempt to work out all of the implications of the history it presents, this Part will use the ATS to illustrate how a proper understanding of the original source of the cause of action in federal court can both in-form and transform debates over federal judicial power.

©Virginia Law Review Association. All rights reserved.