A law can exist and not exist at the same time. One morning, it binds the U.S. government and reassures the public; by afternoon, it has been secretly rescinded without altering the official public record. On paper, it still stands. In practice, it is gone. That is the odd physics of America’s executive orders. Even the most visible directives—such as Executive Order 14086, the foundation of U.S.–EU data transfers—may appear fully in force while having already been set aside in secret.

Executive orders (EOs) are signed in public ceremonies, published in the Federal Register, and widely treated as reliable markers of policy. Yet they rest on a fragile premise: very little prevents a President from rescinding or modifying them in secret. An EO is not a distinct legal form but a tool for communication—one of many the President may use, alongside memoranda, classified national security directives, or even oral commands. A President could therefore rescind an order’s effect through a quiet classified instruction while leaving its text untouched. That is the essence of a “Schrödinger’s Executive Order”—a rule that remains on the books but vanishes in practice.

The first public glimpse of this issue came in 2007, when Senator Sheldon Whitehouse revealed a Justice Department Office of Legal Counsel opinion on the Senate floor stating that “[t]here is no constitutional requirement for a President to issue a new Executive order whenever he wishes to depart from the terms of a previous Executive order. Rather than violate an Executive order, the President has instead modified or waived it.” In other words, the mere act of departure is treated as an amendment.

That logic is especially consequential in intelligence law, where much of the governing framework is found in executive orders rather than statutes. Orders like 12333 and 14086 may appear intact in the Federal Register, but in practice they could be rescinded in secret, suspended when convenient, or waived on a case-by-case basis. The result: outside observers (including, potentially, Congress) have no way of knowing whether the published rules are the rules actually in force.

This phenomenon unsettled some members of Congress. Senator Russ Feingold called it “secretly changing the law,” warning that “since the Executive Order stays on the books, Congress and the public are misled about what the real law is.” In 2008, he and Senator Whitehouse introduced the Executive Order Integrity Act, which would have required the President to provide public notice in the Federal Register within thirty days of any revocation, modification, waiver, or suspension. The bill never advanced out of committee.

The President’s Duties

To analyze the President’s duties, we should start with the Federal Register Act. It requires publication of presidential proclamations and executive orders in the Federal Register—but with a major caveat. Directives “not having general applicability and legal effect or effective only against Federal agencies or persons in their capacity as officers, agents, or employees thereof” need not be published at all. The law was intended to ensure that binding presidential directives would be accessible to the public, yet its exception is broad and seems to be applicable to intelligence law executive orders that only bind intelligence agencies.

At first glance, the Freedom of Information Act (FOIA) seems like it might offer a route to challenge this through agencies. But the statute contains broad exemptions—including for internal agency practices and for matters “specifically authorized under criteria established by an Executive order to be kept secret in the interest of national defense or foreign policy.” Those carve-outs leave significant uncertainty about how far FOIA reaches into executive-branch policymaking.

This sets up the legal scaffolding for secret rescissions of EOs. In practice, this gap has long been cushioned by congressional oversight and trust. Historically, congressional intelligence committees often received notice through routine hearings, staff-level calls, or direct communication between senior officials and committee leadership. These exchanges rested on norms of reciprocity and confidence built over years of working relationships. Yet those guardrails are only as strong as the people who honor them. It is increasingly clear that the machinery of interbranch trust can atrophy quickly. When regularized consultation erodes, the risk that executive-law changes occur entirely behind closed doors becomes a far more pressing prospect.

The Broader Landscape of “Secret Law”

The possibility that EOs can be waived or rescinded in secret is part of a larger phenomenon often described as “secret law.” National security governance has long relied on rules that bind government actors but remain hidden from the public.

One form appears in classified statutory annexes. Since the 1970s, Congress has increasingly used classified annexes to intelligence authorization bills. The classified schedule of authorizations is often incorporated by reference. Robert S. Litt, then-General Counsel of the Office of the Director of National Intelligence, took the position that only this schedule of authorizations has the force of law, while the “remainder of the annex is report language explaining the positions of the committee on a variety of issues, and has no more force than any other committee report.”

Another form comes from the White House. Presidential directives—whether labeled National Security Decision Directives, Presidential Policy Directives, or some other variant—can impose obligations across the executive branch. They are not always published, and oversight committees often lack even an authoritative list of what is in effect.

A third form lies within the executive branch’s own legal machinery. Opinions of the Justice Department’s Office of Legal Counsel are treated as “controlling” across agencies. Yet, OLC opinions concerning intelligence law and national security are rarely released, and only persistent FOIA litigation has begun to pry some of them into public view.

Tacit modification of executive orders occupies a different corner of this landscape. Here, the text of the law is public, but practice quietly diverges.

Case Study: Executive Order 14086

The Schrödinger’s Executive Order phenomenon would not only affect internal agency procedures—it could affect international commerce and foreign policy. Consider EO 14086, issued in 2022 to implement U.S. commitments under the EU-U.S. Data Privacy Framework. The order introduced new safeguards on U.S. signals-intelligence activities and established a redress mechanism for EU individuals, satisfying conditions set by the European Court of Justice after it invalidated the prior “Privacy Shield” arrangement.

When the European Commission adopted its adequacy decision of the EU-U.S. Data Privacy Framework in 2023, it explicitly relied on EO 14086’s requirements. In EU law, those commitments now form part of the legal foundation that permits data transfers between the European Union and the United States. This has led some observers to suggest that the Executive Order’s limits have been “internationalized”—that they can no longer be unilaterally altered because they underpin a bilateral regulatory regime.

Yet the order remains, at bottom, a unilateral presidential directive. It was not ratified by the Senate, nor formalized as a treaty. Thus, the only constraint is political: a known revocation would likely trigger the collapse of the European Commission’s adequacy finding and disrupt trans-Atlantic commerce.

This makes EO 14086 a paradigmatic Schrödinger’s Executive Order: simultaneously domestic and transnational, relied upon as permanent yet vulnerable to invisible revocation. Its status reveals how executive directives can become global legal infrastructure without ever ceasing to be presidential instructions.

How Secret Modification Might Work in Practice

If a President can tacitly amend or waive an executive order, what might that look like in practice? Several hypothetical models illustrate the possibilities.

One possibility is government-wide disregard. A President could treat a standing intelligence order (such as EO 12333) as completely irrelevant and secretly direct the intelligence bureaucracy to disregard it, rendering an effective rescission. Because the Federal Register Act exempts purely internal directives from publication, the order could remain on the books while being quietly ignored.

A second model is agency-wide discretion. Instead of discarding the order altogether, the President might authorize an agency like the CIA to treat its provisions as optional. On paper, the EO remains, but operationally it becomes a guideline subject to waiver whenever the agency deems necessary.

A third, more targeted approach is a mission-specific waiver. For a particular covert program, EO constraints could be lifted. This model avoids the problem raised by Judge Ellen Huvelle in 2013, when she ordered the release of Presidential Policy Directive 6, because of its “widespread dissemination[] within the Executive Branch” which rendered the presidential communication privilege inapplicable. A waiver tied to a single mission, by contrast, would be closely held and unlikely to trigger litigation or FOIA disputes.

Finally, authority to waive Executive Orders could be delegated to a single official or a set of officials. The President might empower the CIA Director or another senior officer to set aside EO requirements at their discretion. This keeps control tight, limiting the risk of broad distribution that could undermine secrecy.

What Congress Can Do

The failure of the Feingold–Whitehouse bill does not mean the subject is closed. If Congress were to revisit the issue, there are several options on the table.

One option would be to follow the Feingold–Whitehouse model and require notice of changes to published orders. This path would likely run into constitutional issues surrounding the President’s foreign affairs powers. Relatedly, Congress could add this provision to a classified annex. In fact, it is possible that such a requirement already exists in a classified annex of an appropriation bill and that is why the issue has not been raised in over a decade. While this would not meet the constitutional requirements of bicameralism and presentment, in practice, agencies tend to follow such directives.

Another possibility would be for Congress to incorporate intelligence-law executive orders by reference directly into legislation. By doing so, provisions that begin as presidential directives would acquire the force of statutory law, and a President could no longer rescind or waive them unilaterally. Congress has taken this path before, elevating EO-based policies such as sanctions regimes and whistleblower protections into the U.S. Code. In principle, the same technique could be applied to intelligence-law orders, ensuring that rules now subject to tacit modification would instead rest on a legislative foundation.

Nevertheless, Congress’s ability to act here is constrained by constitutional limits. Any attempt to lock in or oversee intelligence-law executive orders would run up against the President’s broad Commander-in-Chief and foreign affairs powers, as well as the increasingly relevant unitary executive theory. Notwithstanding Justice Jackson’s Youngstown framework, which imagines a sliding scale of presidential power against congressional opposition, the courts have historically granted the President wide latitude in intelligence and foreign affairs.

And even if Congress were to press forward, most of these disputes are unlikely to be justiciable. Challenges to the modification or disregard of an executive order would likely be dismissed for lack of standing or political question concerns. That leaves Congress with the tools it has always had: political pressure, control of appropriations, and oversight.

Schrödinger’s Executive Orders may exist, or they may not. However, in an era of fraying interbranch trust and decaying norms, it is easy to see how they could. Intelligence law has always depended on trust, restraint, and professional judgment as much as on formal limits. When those norms erode, what once looked like a theoretical vulnerability begins to feel like a practical one.

Tal Feldman is a 2L at Yale Law School.