The Supreme Court has roundly rejected prior restraint, but U.S. intelligence agencies nonetheless prohibit literally millions of former public servants from speaking or writing about government policy without first obtaining the government’s approval. On Thursday, the Court will consider a petition, filed by the ACLU and the Knight Institute, asking the Court to revisit Snepp v. United States, the forty-year-old case atop which the intelligence agencies’ far-reaching system of prior restraint has been built. Snepp is a glaring anomaly in relation to the rest of the Court’s First Amendment jurisprudence, and the contemporary prepublication review system is a free-speech train wreck.

As we and our colleagues have written here before, the prepublication review system originated in a set of contractual obligations imposed on a small number of public servants with access to the nation’s most sensitive secrets. But what was once a narrow regime has grown into a byzantine network of regulations and policies that restrict the speech of millions of people.

We owe this system in large part to the Supreme Court’s 1980 decision in Snepp, which affirmed the imposition of a constructive trust over the proceeds of a book that a former CIA officer had published without first submitting it for review—an act the agency characterized as a breach of fiduciary duty even though it conceded that the book did not contain classified information. The intelligence agencies interpreted Snepp as a green light for a dramatic expansion of their prepublication review regimes, and lower courts have interpreted it to mean that the agencies’ prepublication review regimes are not subject to the stringent First Amendment review that would ordinarily be applied to prior restraints.

It’s amazing to us that the intelligence agencies have been able to squeeze so much juice from Snepp’s meager fruit. The narrow question the Supreme Court addressed in Snepp was whether the agency was justified in imposing a constructive trust. In addressing this question, the Court mentioned the First Amendment only twice, including once in the Court’s summary of decisions below. The only First Amendment analysis appears in a footnote, and it consists mainly of the bare conclusion that prepublication review is “reasonabl[e]” in light of the government’s “vital interest” in protecting “both the secrecy of information important to our national security and the appearance of confidentiality so essential to the effective operation of our foreign intelligence service.” The Court did not address what materials agencies may constitutionally require former employees to submit for review, on what bases an agency may constitutionally withhold permission to publish, or how quickly agency review must be completed.

That Snepp is the foundation of the modern prepublication review system is, again, simply amazing. But it is especially so because of the highly irregular way in which the case was decided.

Here is that procedural history: The district court enjoined Snepp from further violating his secrecy agreement and imposed a constructive trust on his book-proceeds, as the government requested. On appeal, the Fourth Circuit upheld the injunction but rejected the trust, reasoning that Snepp hadn’t published classified information and that accordingly the agency was entitled to recover only actual damages. The government was satisfied with this result, but Snepp petitioned the Supreme Court to hear the case, in part maintaining that his contract was an unenforceable prior restraint. After Snepp filed his petition, the government filed a conditional cross-petition, asking the Court to grant its petition if the Court granted Snepp’s.

And at this point, the case took a bizarre turn. The Court simply decided the case on the basis of the petition and conditional cross-petition, without inviting briefing on the merits and without hearing oral argument. The Court’s per curiam opinion was issued over a lengthy dissent from Justice Stevens, who complained not just about the Court’s broader reasoning but about its dismissive treatment of First Amendment issues that warranted more serious consideration.

This unusual procedural history perhaps goes some way in explaining why Snepp’s First Amendment footnote is so difficult to square with the Court’s First Amendment jurisprudence. The Snepp opinion doesn’t discuss the constitutional presumption that prior restraints are unenforceable. It doesn’t attempt to reconcile its reasoning with the rule that any system of prior restraint must include “narrow, objective, and definite standards to guide the licensing authority” (from Shuttlesworth v. City of Birmingham), and must include “procedural safeguards designed to obviate the dangers of a censorship system” (from Freedman v. Maryland). It doesn’t discuss the interest of former public servants in speaking about the operations of government, or the interest of the public in hearing that speech (interests the Court highlighted in cases like United States v. National Treasury Employees Union). Snepp’s terse First Amendment footnote seems essentially oblivious to fundamental First Amendment principles.

The result, forty years later, is that intelligence agencies’ prepublication review regimes look nothing like the licensing schemes the Court has permitted in other contexts. The prepublication review regimes lack narrow, objective, and definite substantive standards to guide government officials and cabin their discretion when reviewing materials. They also lack narrow, objective, and definite standards as to what former employees must submit for review—essentially requiring the submission of virtually anything former employees write about the government or national security, regardless of whether they rely on information they learned in the course of their employment. Finally, the regimes also lack the procedural safeguards required to mitigate the risk of illegitimate censorship—including perhaps most critically any firm timeline for government review, meaning the government can take months or more to review authors’ work and leave authors with no recourse other than expensive as-applied lawsuits. The prepublication review system, as Jack Goldsmith and Oona Hathaway have observed, is “racked with pathologies.”

Writing in dissent in Snepp, Justice Stevens remarked that the majority “seem[ed] unaware” that its cursory First Amendment analysis had potentially “fashioned” a “drastic new remedy” to “enforce a species of prior restraint on a citizen’s right to criticize his government.” 444 U.S. at 526 (Stevens, J., dissenting). He was right. The Court now has the opportunity to consider the First Amendment interests implicated by prepublication review for, truly, the first time. It should take it.