Faculty Bibliography

In this Article, I examine a phenomenon unfolding within the United States’s military legal establishment: an effort by a segment of military lawyers to define how the law of armed conflict (LOAC) applies to the wars they anticipate fighting in the future. I refer to this effort as LSCO lawyering: the development, advancement, and institutionalization of a vision of LOAC tailored to large-scale combat operations (LSCOs), understood here as multi-domain warfare against a peer adversary such as China. Drawing on doctrinal materials, planning documents, and conversations with current and former armed-forces legal advisers from the United States and allied or partner forces, I trace how LSCO lawyering reflects a diagnosis of interpretive and institutional crisis—a perceived mismatch between prevailing legal expectations and the operational realities of high-intensity war—and a response that aims not to dismantle LOAC but to reassert it on terms viewed as credible and feasible under conditions of existential conflict. At the heart of this project lies a reconfiguration of what I call American international humanitarian law (IHL): the United States’s distinctive assemblage of legal interpretations, operational practices, and normative commitments that shape its approach to the conduct of hostilities. While LSCO lawyering is often framed internally as a modest clarification of existing law, I suggest that it functions as a far-reaching attempt to reshape the interpretive ecosystem within which LOAC is applied by privileging internal coherence, institutional discretion, and operational speed over policy overlays, external scrutiny, and extensive civilian-protection norms. The LSCO lawyering project does not reject the law, but it does aim to narrow its aperture to ensure that legal interpretation does not require, in the view of its proponents, normatively undue or operationally unsustainable limits on commanders preparing to fight—violent, vicious, and fast—in a potentially existential war, which would entail extraordinarily high consequences for civilian death and destruction. The emergence of LSCO lawyering sheds light on deeper conditions within the law of armed conflict’s normative and interpretive architecture. It brings to the surface long-standing tensions—between operational feasibility and civilian protection, between internal judgment and external review, between doctrinal minimalism and progressive development—that have shaped the field for decades. As a project grounded in anticipatory planning, LSCO lawyering highlights the degree to which LOAC interpretation is shaped not only by treaty text or customary practice but also by institutional culture, professional memory, and perceived strategic necessity. In that sense, it offers a revealing case through which to examine the evolving contours of LOAC as a legal, operational, and epistemic system—one whose boundaries are still being contested and whose authority remains under active construction.

U.S. states traditionally play a minor role in establishing national security policies, which generally fall within the federal government’s remit. But the return of great power competition with China and Russia and the accompanying proliferation of threats have spurred states to act on national security concerns. With unprecedented speed, breadth, and frequency, U.S. states have taken it upon themselves to address perceived security concerns with TikTok, purchases of real estate by foreign actors, and foreign-made drones, as well as commercial dealings with Russian firms. Drawing on their police powers, they have enacted security-related laws that sometimes parallel and sometimes go beyond the federal government’s actions. We term this phenomenon “entrepreneurial federalism” and explain its unique features. The increasing frequency and breadth of states’ national security–focused actions have set U.S. states and the federal government on a collision course. Private parties have launched a range of legal challenges to state laws, arguing that courts should hold that those laws are preempted based on existing federal statutes or on broader doctrines that disable states from acting in foreign relations. Courts may be tempted to do so, especially because China and Russia are near-peer threats that require careful federal management. But if the courts adopt broad preemption doctrines in this space, they may inadvertently foreclose two constructive phenomena that can arise from acts of entrepreneurial federalism: useful supplementation by the states of federal efforts to address national security threats and the productive friction that states can introduce into policymaking to improve the quality of U.S. national security policies. Even when there are good reasons for courts to hold that state actions that implicate the U.S. relationship with China or Russia are preempted, judicial decisions that reach that result too readily — or that use a broader form of preemption than necessary — may unintentionally impose longer-term costs on U.S. national security. This Article documents the rise of states’ national security actions, distinguishes them from earlier academic models of federalism, and proposes ways that the courts, Congress, the Executive, and the states can foster a positive role for states while minimizing the downsides that could flow from state actions in the national security space.

Where a country maintains a higher-legal (‘constitutional’) script of institutional forms, procedures and directives for the state’s exercise of lawmaking powers, we may demand or expect of that script that it be designed and applied with an overriding view to securing conditions of humanity, dignity and freedom for all the people of the country. The constitutional directives may, accordingly, include guarantees respecting named basic interests and needs of the people. Looking here into appraisals in South Africa of that country’s subsisting Constitution (and judicial applications of it), I find debates extending to a question of what sort of work a higher-legal constitution should anyway be doing in that country. As a nation-state among nation-states, South Africa today presents the aspect of a regime of civil government, authorised and responsible to oversee by law the ordering and conduct of affairs across its bounded territory. We tend to assume that any such nation-scale regime of government by law will have to rest on its population’s convergence on a script of procedures and directives for the state’s exercise of lawmaking powers: a constitution. This constitutional-existential assumption might not necessarily hold at every moment for every country. Here I advance it provisionally, with a view to charting implications from it that for some readers may bolster doubts — or dampen hopes — that the constitutional-existential proposition is fully fit to the South African case. Impending upon such doubts or hopes is a widely held sense of a collision here of culturally entrenched predispositions — ‘African,’ say, with ‘Western’ — towards conceptions of being-in-the-world and persons-in-society. I ask here about the resulting prospects for South African constitutionalism. I do so partly with a view to tempering intuitions about how to judge the doings of the Constitution’s drafters and its first-generation judicial custodians.

Functionalists, especially those inspired by American Legal Realism, downplay the importance of abstract concepts and their interrelations in legal reasoning. Hence, they stress shallow, narrow, and isolated concepts, in order to be close to the facts in a transparent way. In this paper, we address an often-overlooked function of law: managing legal concepts themselves. That is, one aspect of law is partially self-referential. Various devices involved in legal reasoning are designed to make the system of legal concepts work better, from fine-tuning the results of the use of concepts to tinkering with the concepts themselves. This requires us to look at law as a system, but one that is both inward-looking and operating on itself, and, at the same time, concerned with how law operates in the real world. Law also uses various devices to manage the system of concepts, and these devices in turn serve various functions. These functions include maintenance, modulation, and modification of concepts, depending on the actual or potential misfiring of the concepts in question. As we will suggest, this offers deeper insight into the role that legal fictions, equity, and presumptions play in the law.

A controversial case on Elon Musk's $56 billion pay has laid bare our poor under-standing of ratification in corporate law and possibly in other legal domains. What exactly is ratification and what is its internal logic? In this Article, I recon-struct the concept of ratification within a general theory of shared power and show why the prevailing vocabulary in corporate law is misleading. I argue that what courts call shareholder ratification is not ratification at all, but a different form of shared legal power that I call interposition, a term of art borrowed from Roman law. Ratification is delegation that survives a misstep: someone cures a defective act by their representative. Interposition, by contrast, coordinates parties with equal authority but asymmetrical roles. Because no party has superior authority, no one can unilaterally cure defects: decision-making power is structurally shared among multiple parties. Three features distinguish the two models — whether representation is structurally necessary or merely contingent, whether the repre-sented may give binding instructions, and whether the represented may cure the representative's defective acts — and these features are not independent design choices but interlocking corollaries of the basic distinction between equal and une-qual authority. Neither model is inherently superior, but within a chosen model internal coherence has independent value: the features are plausibly economic complements, and mixing them degrades predictability, administrability, and le-gitimacy. Applied to Delaware corporate law, the framework reveals that the shareholder vote in conflicted transactions is not a curative ratification but a step in a multi-party interposition among board, independent directors, shareholders, and the Court of Chancery. On this view, the Court of Chancery's refusal to treat a post-trial shareholder vote on Musk's pay as a curative ratification is not an affront to "shareholder will," as many observers lamented, but a faithful applica-tion of the interpositional architecture of Delaware corporate law.

The cases of Galette v. N.J. Transit Corp. and N.J. Transit Corp. v. Colt reduce to a simple question: who is the legal person before the Court, against whom the plaintiffs request relief? The common-law immunity States retained at the Founding was a personal immunity, one belonging to specific legal persons, namely the sovereign States. Each State may decide the internal structure of its government; yet each also has power to create separate and distinct legal persons that are not sovereign, appearing before courts as creatures of States rather than States themselves. That is what New Jersey did. It created a new public corporation—petitioner New Jersey Transit Corporation—whose appearances in court are not appearances of the State, whose judgment debts are not debts of the State, and whose adversely decided issues are not issues decided against the State. And the Corporation in turn created its own subsidiary corporations—including petitioner NJ Transit Bus Operations, Inc.—as legal persons distinct even from itself, still more plainly the sort of “lesser entities” that lack the State’s sovereignty and immunity. The only reason this case might seem difficult is that this straightforward historical rule has since been swallowed up by judicial confusion. The Court should restore the original test: whether the requested relief acts against a separate legal person (which is amenable to federal process) or against a sovereign State (which is not). Because the State of New Jersey chose to create the Corporation as a distinct legal person from itself, with neither bound by judgments against the other, the Corporation cannot claim the State’s immunity.

For many years, liberals have favored broad standing and conservatives narrow standing. Yet that pattern has disappeared and is now reversing. We studied the Supreme Court justices' votes on standing during the fifteen-year period from October Term 2010 through October Term 2024. Our results show a stark change in voting patterns around October Term 2020, when the Court obtained a conservative supermajority. We label this development "the standing realignment," discuss several interrelated explanations, and suggest implications.

This letter brief, submitted in King v. Bon Charge, No. 1:25-cv-105-SB (D. Del. filed Jan. 24, 2025), argues that the Court should apply Rule 4(k)(2) to exercise jurisdiction over a foreign defendant sued on a federal claim. In particular, it argues (1) that the Court should adopt other circuits’ practice of expecting a defendant who resists Rule 4(k)(2) jurisdiction to name a state where jurisdiction might be available, (2) that Rule 4(k)(2) incorporates actual Fifth Amendment standards, not the state-mirroring standards that many thought might apply before Fuld v. PLO, and (3) that the original Fifth Amendment did not restrict Congress’s power to vest the courts with personal jurisdiction over defendants found abroad.

The Healthy Technology Act of 2025, a bill being considered by Congress, would permit artificial intelligence (AI) systems to prescribe medications without human sign-off. While allowing AI to practice independently may still be a stretch, large language models are already being used to support diagnosis and treatment.1 These applications challenge the current regulatory framework for clinical AI, which has focused on software as a medical device (SaMD). The US Food and Drug Administration (FDA) has cleared more than 1000 AI tools using this framework, mostly for narrow, well-defined tasks. However, generative models can be applied across domains, straining the SaMD framework. In this Viewpoint, we explore the application of a licensure paradigm to clinical AI systems, arguing that this offers a more reliable regulatory framework.

The convergence of Artificial Intelligence (AI), drug discovery, and genomics represents one of the most transformative shifts in biomedical science. Traditional drug development is costly, slow, and fraught with uncertainty, yet AI-driven methodologies now promise to accelerate discovery, improve precision, and reduce attrition rates across the pharmaceutical pipeline. This paper provides an in-depth exploration of how AI enables predictive modeling of molecular interactions, target identification, and genomic-driven precision medicine. Machine learning algorithms, generative models, and systems biology frameworks are redefining the speed and scope of therapeutic innovation—allowing researchers to model diseases at the molecular level and design novel compounds with unprecedented accuracy. The study further examines the integration of genomics, transcriptomics, and proteomics data with AI, enabling stratified medicine and individualized therapy design. Ethical, regulatory. and data governance dimensions are critically assessed, including issues of data privacy, algorithmic bias, and explainability in medical AI systems. By analyzing current case studies and future trends, this paper reveals how AI not only accelerates drug discovery but also redefines the paradigm of innovation itself—ushering in an era of intelligent, personalized, and predictive medicine.

Amid plans to mark the nation’s semiquincentennial, many are asking whether or not the people really do rule, and whether the law is still king.

In the Supreme Court’s tariffs case, the conservative Justices will weigh two conflicting impulses regarding Presidential authority.

Should platforms be liable for harms suffered by users? A platform enables interactions between firms and users. Harmful firms impose larger costs on users than safe firms. If firms have deep pockets and are fully liable for harms, platform liability is unnecessary. If firms have limited liability, holding platforms liable for residual harm increases platforms' incentives to raise interaction prices and invest in auditing to deter, detect, and block harmful firms. The social desirability and optimal level of platform liability depend on whether interactions require user consent, the degree to which users internalize harms, and the observability of platform effort.

Nudges are behavioral interventions that influence decision-making by subtly altering the choice environment without restricting freedom. Rooted in behavioral economics, they have shown promise in health care by improving adherence to guidelines and promoting preventive behaviors. Learning Health Systems offer infrastructure for implementing nudges at scale through tools like electronic health records and decision support systems. In cardiovascular-kidney-metabolic care, nudges targeting both patients and providers, ranging from defaults and reminders to gamification, have improved prescribing, vaccination rates, and physical activity. Frameworks such as EAST and MINDSPACE guide effective design, emphasizing timing, audience, and framing. Looking ahead, artificial intelligence–powered nudges promise personalized, adaptive interventions that respond to real-time behavior and performance, enhancing scalability and sustainability. By aligning behavioral science with health technology, nudges can help optimize care delivery, reduce variation, and improve outcomes across complex health systems.

This Viewpoint reviews employer protection initiatives for family caregivers in Illinois and other states.

His posts and rants are omnipresent, ugly, and unhinged. Don’t look to history to make it make sense.

Congress wrote statutes with the apparent assumption that whoever held the office of the Presidency would use the powers they granted in good faith.

In the last one hundred years, the United States Constitution has remained essentially the same. About halfway through that century, however, constitutional theory reimagined it. In the time before, everyone agreed that the U.S. Constitution was undemocratic. That was the point of it, for its defenders; and it was the cause of America’s crises and thus the predicate of reform, for its critics. Nor was this fact especially surprising: the point of constitutionalism, ancient or modern, was generally to limit democracy or not to have it. After a brief transition, in the middle of the twentieth century, almost everyone came to agree that the U.S. Constitution — indeed constitutionalism generally — is democratic, actually and potentially. Nothing about the document had changed; indeed, it had changed more through formal amendment before this development, compared to since. This turn is best understood as the pseudo-democratization of American constitutional theory. The purpose of this chapter is descriptive and narrow. It is to establish the fact of pseudo-democratization. The core of the chapter is an enumeration of the forms of the new (pseudo)democratic constitutional theory.

A privately-informed entrepreneur may withhold material information from prospective investors who may sue the firm ex post for (alleged) non-disclosure. Absent liability, the entrepreneur has an excessive incentive to withhold bad news and pursue socially-wasteful projects. Liability deters inefficient non-disclosure and prevents capital misallocation. Any damage award received by investors is partially offset by a reduction in equity value. Depending on the likelihood of court error and litigation cost, the socially-optimal damage award may be either zero or the minimum necessary for full deterrence. The private incentive to waive liability may be socially excessive or insufficient. Positive and normative implications are discussed.

As perhaps the most conservative Supreme Court of the past century, the Roberts Court regularly touts its commitment to traditional ideals, including the rule of law. Yet Trump v. United States seems ready to place presidents entirely or largely above the law by granting them absolute or near‑absolute immunity from criminal prosecutions for official conduct. And Trump is not a one-off: from judicial recusal to qualified immunity to due process decisions, the Court frequently seems to display indifference to values of accountability and impartiality that have long been thought integral to the rule of law. What is going on? We call this puzzle “the Roberts Court Paradox.” This Essay explains how the Court arrived at this untenable position and what has led it astray. Justice Antonin Scalia’s famous essay The Rule of Law as a Law of Rules provides our starting point. It argues that judges who wish to remain true to the rule of law must aspire to decide cases by formulating bright-line rules instead of nuanced rules, standards, or balancing tests. At the core of this argument are two ideas, both facially plausible. First, bright-line rules provide individuals with fair notice of what the law is and what sanctions they face. Second, they promise to prevent appellate judges from reading their own values into the law and thereby usurping the powers of other branches. We show that the Roberts Court is where it is in large part because of this equation of the rule of law with bright-line rules. Unfortunately, Justice Scalia’s argument withers under scrutiny, and it is especially weak as applied to constitutional interpretation undertaken by the Supreme Court. This is not only because it entirely fails to consider the centrality of accountability and equality before the law, which are both foundational to the rule of law. It is also because Justice Scalia’s “notice” and “limited power” arguments are far from compelling. While the value of notice is indeed critical in areas such as criminal law, it is much less important in constitutional law, which is primarily about power-conferring and power-constraining rules, not about sanctioning individuals for violating duty-imposing rules. And while judicial usurpation of power is indeed a rule-of-law issue, it is pure dogma to suppose that sticking to bright‑line rules is the only approach that can keep judges in their lane. More importantly, Supreme Court justices who create bright-line rules that preclude accountability for other branches of government are plainly abandoning the core value of limited powers, not protecting it. In sum: the Roberts Court, following Justice Scalia’s lead, has allowed its fondness for bright lines to overwhelm its duty to acknowledge the foundational place of accountability in the Anglo‑American legal tradition.

Article V recognizes two ways that the People might speak: through regular legislatures and through popular conventions. Over time, the convention method has become moribund, mirroring the decline of Article V as a whole. Americans now rely on courts to solve constitutional problems that were supposed to be solved with voting and politics. But we could regain many of the benefits of conventions through a “rolling” convention, in which states begin the work of constitutional change. Through a new amendment, we could create another avenue for change that flips the order of Article V, letting three fourths of states propose amendments that two thirds of Congress could ratify. Flipping Article V preserves the constitutional threshold for amendment, the chance for national deliberation, and the relative power of the states. But it breaks some of the modern barriers to amendments—and helps return constitutional politics to the political process.

A proporcionalidade, aceita como princípio geral do direito constitucional por muitos países, exige que as interferências do governo nas liberdades individuais sejam justificadas, que interferências mais intensas tenham justificativas robustas e que as punições reflitam a gravidade relativa da infração. A proporcionalidade, enquanto doutrina desenvolvida pelas Cortes, como no Canadá, forneceu um quadro metodológico estável, promovendo decisões estruturadas e transparentes, até mesmo sobre valores constitucionais fortemente contestados. Outros benefícios incluem o potencial de aproximar o direito constitucional da justiça constitucional, fornecer um discurso comum sobre direitos para todos os ramos do governo e ajudar a identificar os tipos de falhas no processo democrático que justificam um escrutínio judicial mais rigoroso. Nos Estados Unidos, debates anteriores acerca da “ponderação” (balancing) não levaram em conta a recente experiência comparada à doutrina estruturada da proporcionalidade e dos seus benefícios. Muitas áreas do direito constitucional estadunidense incluem alguns elementos que, em outros contextos, são chamados de análise de proporcionalidade. Defendo aqui um uso mais amplo dos princípios e da doutrina, bem como argumento que a análise judicial de proporcionalidade não é a resposta para todas as questões relativas a direitos constitucionais. A liberdade de expressão pode se beneficiar de presunções categóricas, mas, em sua aplicação e formulação, a proporcionalidade pode ser relevante. A Quarta Emenda, que garante um “direito” contra “buscas e apreensões irrazoáveis”, está repleta de regras categóricas que protegem a conduta policial do escrutínio judicial. Uma análise mais casuística da “irrazoabilidade” ou desproporcionalidade da conduta policial protegeria melhor os direitos e o Estado de Direito. Demandas por igualdade baseadas em “impacto desproporcional” poderiam ser mais bem avaliadas por meio de padrões de análise mais proporcionais; a revisão de sentenças prisionais, à luz da Oitava Emenda, iria se beneficiar de um uso maior dos princípios da proporcionalidade. Assim, reconhecer as vantagens – e os limites – da proporcionalidade permitiria ao direito constitucional dos Estados Unidos da América (EUA), ao mesmo tempo, proteger direitos e favorecer uma autogovernança democrática eficaz.

At present, American constitutional law gives short shrift to the democratic value of alignment (congruence between governmental outputs and popular preferences). But it doesn't have to be this way. In this symposium contribution, I outline three ways in which constitutional law could incorporate alignment. First, alignment resembles federalism in that it's a principle implied by the Constitution's text, structure, and history. So doctrines analogous to those that implement federalism could be crafted to operationalize alignment. Second, comparative constitutional law recognizes democratic malfunctions that involve misalignment as well as innovative judicial remedies for these problems. Likewise, American constitutional law could appreciate the full arrays of misaligning threats and potential judicial responses to them. Lastly, one of the key concepts of modern originalism is the construction zone, in which disputes must be resolved on grounds other than the constitutional text's original meaning. Alignment could be a factor that courts consider in the construction zone, pushing them to further, not to frustrate, this value.

The use of AI, or more generally data-driven algorithms, has become ubiquitous in today’s society. Yet, in many cases and especially when stakes are high, humans still make final decisions. The critical question, therefore, is whether AI helps humans make better decisions compared to a human-alone or AI-alone system. We introduce a methodological framework to answer this question empirically with minimal assumptions. We measure a decision maker’s ability to make correct decisions using standard classification metrics based on the baseline potential outcome. We consider a single-blinded and unconfounded treatment assignment, in which the provision of AI-generated recommendations is assumed to be randomized across cases, conditional on observed covariates, with final decisions made by humans. Under this study design, we show how to compare the performance of three alternative decision-making systems—human-alone, human-with-AI, and AI-alone. Importantly, the AI-alone system encompasses any individualized treatment assignment, including those not used in the original study. We also show when AI recommendations should be provided to a human-decision maker, and when one should follow such recommendations. We apply the proposed methodology to our own randomized controlled trial evaluating a pretrial risk assessment instrument. We find that the risk assessment recommendations do not improve the classification accuracy of a judge’s decision to impose cash bail. Furthermore, replacing a human judge with algorithms—the risk assessment score and a large language model in particular—yields worse classification performance.

A resounding win for the university in court still leaves the Trump Administration with plenty of ways to force schools into submission.