Thursday was another busy day at the Court, seeing the release of six (fine, five-plus) decisions:

  • In Andy Warhol Foundation v. Goldsmith (No. 21-869), a 7-2 Court held, in a lengthy decision expounding on the “purpose and character” prong of the fair-use test, that the licensing of an iconic Andy Warhol portrait of Prince to a magazine was not protected by the fair-use defense to a copyright-infringement brought by the photographer who took the “source” photograph;
  • In Twitter v. Taamneh (No. 21-1496), the Court unanimously held that Twitter did not “aid and abet” terrorism by failing to take stronger steps to remove content posted by ISIS. Based on that holding, the Court vacated and remanded in Gonzalez v. Google (No. 21-1333), an essentially identical case involving Google;
  • In Amgen v. Sanofi (No. 21-757), a unanimous Court held that two of Amgen’s patents for antibodies were invalid, because Amgen’s patent filings didn’t provide enough information “to enable any person skilled in the art” to make the patented antibodies;

Let’s start with Andy Warhol Foundation v. Goldsmith (No. 21-869), where Justices Sotomayor and Kagan faced off over the meaning of “fair use” in copyright law and, in a real sense, what the case was actually about.

Everyone more or less agreed about the facts. Back in 1981, a rising-star music photographer, Lynn Goldsmith, took (or, in fairness, created) several photographs of a rising-star musician, Prince. The studio photographs depicted Prince as a “vulnerable, uncomfortable” person on the edge of stardom. A few years later, Goldsmith licensed one of the photographs (to which she held the copyright) to Vanity Fair to serve as an “artist reference for an illustration” to be created by another artist, which would run alongside a feature story on Prince. The other artist Vanity Fair hired was Andy Warhol, already a legend of “pop art,” known in particular for his stylized, silkscreen portraits of celebrities like Marilyn Monroe. Through a painstaking process, Warhol took Goldsmith’s photograph and transformed it from what you see on the left, to what you see on the right (each image used fairly here, we daresay, for the purpose of “news reporting,” see 17 U.S.C. § 107):

Warhol’s image, “Purple Prince,” ran alongside a feature on the singer, with requisite credit to Goldsmith for the “source photograph.” But Warhol didn’t stop there. He made 14 additional silkscreens (and two drawings) derived from Goldsmith’s original photograph, which came to be known as the “Prince Series.” Those 16 works are now in the hands of galleries and private collectors, as well as the Andy Warhol Museum in Pittsburgh. The Andy Warhol Foundation (AWF), which took title to Warhol’s copyrights after his death, continues to license images of the works.

After Prince’s death in 2016, Vanity Fair’s parent company, Condé Nast, reached out to AWF for permission to reuse Purple Prince for a commemorative magazine. AWF said sure, but did you know there are 15 other images from the series you can choose from? Condé Nast selected one of the other images, “Orange Prince,” to run on the cover of its commemorative magazine:

Condé Nast paid AWF $10,000 for the license. But this time, Goldsmith received neither a fee nor source credit. Goldsmith didn’t even know about Orange Prince (or the other works in the Prince Series) until she saw the Condé Nast magazine cover. She promptly contacted AWF to inform it of the apparent infringement of her copyright. AWF responded by filing a declaratory-judgment action against Goldsmith, alleging noninfringement and, alternatively, fair use. Goldsmith counterclaimed for infringement and the matter was submitted for summary judgment in the Southern District of New York.

The District Court granted summary judgment to AWF. Applying the four “fair use” factors, which are derived from common law and codified in the Copyright Act at 17 U.S.C. § 107, the District Court found that all four favored AWF. In truth, though (and not without some precedential justification), it relied almost entirely on the first factor: “the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes.” In the District Court’s view, Warhol’s use of Goldsmith’s photograph in the Prince Series was “transformative,” in that his images “transformed Prince from a vulnerable, uncomfortable person to an iconic, larger-than-life figure,” such that “each Prince Series work is immediately recognizable as a ‘Warhol’ rather than as a photograph of Prince.” Because the District Court found Warhol’s use to be “transformative,” it spent little time analyzing the other three fair-use factors. On appeal, the Second Circuit criticized this overreliance on the first factor and also found that, properly considered, the “purpose and character of” Warhol’s use was actually not so different from Goldsmith’s that the first factor favored AWF. To the Second Circuit, it couldn’t be enough that the secondary work merely added a “new aesthetic or new expression to its source material.” Instead, it concluded that the proper question was “whether the secondary work’s use of its source material is in the service of a fundamentally different and new artistic purposeand character.” Here, Goldsmith and AWF used the source material (Goldsmith’s photograph) for essentially the same purpose: to illustrate, depict, and accompany a magazine article on the musician Prince.

With that lengthy, but largely uncontested exposition out of the way, we get to the contested stuff. Because the majority and dissenting opinions in the Supreme Court seemed at odds not only on the law—what does “purpose and character” mean in the first fair-use factor—but also on the more fundamental question of what use was at issue: Warhol’s original use of Goldsmith’s photograph to create his transformative Prince Series or AWF’s later licensing of one of those images to Condé Nast without payment or credit to Goldsmith. As a result, we have two lengthy and compelling opinions that, while frequently taking jabs at each other in footnotes, seem largely to talk past one another.

To the majority—led by Justice Sotomayor, and joined in full by all but the Chief and Justice Kagan—the key question addressed by the first fair-use factor is not whether Warhol’s use of Goldsmith’s photograph was “transformative” in the sense that it conveyed a different artistic “meaning or message” than the original, but rather whether the specific “use” complained of—AWF’s licensing of Orange Prince to Condé Nast—had a “further purpose or different character” relative to the original. Justice Sotomayor emphasized that this is a question of degree, with the degree of difference being weighed against other considerations like commercialism. It was not enough that Warhol’s works added new expression and artistic meaning to Goldsmith’s photograph; the fact remained that the purpose of both AWF’s use of Orange Prince and Goldsmith’s use of her original photographs was the commercial purpose of depicting the subject of a magazine article. “Although new expression may be relevant to whether a copying use has a sufficiently distinct purpose or character, it is not, without more, dispositive of the first factor.” Therefore, “[e]ven though Orange Prince adds new expression to Goldsmith’s photograph,” the majority “agree[d] with the Court of Appeals that, in the context of the challenged use, the first fair use factor still favors Goldsmith.”

The dissent, meanwhile, focused almost exclusively on Warhol’s original creation of the Prince Series, rather than AWF’s licensing of Orange Prince. In a characteristically lively opinion joined by the Chief, Justice Kagan illustrated (“with pictures!”) how, in art (as in love), “nothing comes from nothing, nothing ever could.” The history of art is one long and continuously unfolding story of artists appropriating and improving on the works of their predecessors. This iterative process, Kagan insisted, was essential to the very purpose of the Constitution’s Copyright Clause: to promote the progress of science and useful arts. And, in cases like this one, where one artist’s copying is truly transformative—when it results “in the creation of new information, new aesthetics, new insights”—then it necessarily has a different enough purpose and character from the original to satisfy the first fair-use factor. Kagan acknowledged that the first factor is but one of four, and that in fact the fourth factor—“the effect of the use upon the potential market for or value of the copyrighted work”—is often the most important. But the first factor is the only one that focuses on the copying work itself, and “what the copier’s use of the original work accomplishes.” And it is the most significant to achieving the purposes of copyright law, for “when a transformation of the original work has occurred, the user of the work has made the kind of creative contribution that copyright law has as its object.” To her, it was clear as day that Warhol’s transformative use of Goldsmith’s photograph made that kind of contribution, and therefore the first factor weighed heavily in AWF’s favor. The majority’s recasting of the first factor as solely concerned with the profane purpose of a use, rather than its sacred character disserves both the fair-use doctrine and copyright law generally. “It will thwart the expression of new ideas and the attainment of new knowledge. It will make our world poorer.”

The disagreement between the principal opinions extended to the meaning and import of the Court’s fair-use precedents, in particular Campbell v. Acuff-Rose Music (1994), which held that the rap group 2 Live Crew’s song “Pretty Woman” made fair use of Roy Orbison’s song “Oh, Pretty Woman.” Campbell has long been seen as the Supreme Court’s most important fair-use decision, and it has long been assumed (at least by lower courts) that it endorsed a rule similar to that espoused by Justice Kagan: that when a copying work is transformative of the original, the first factor is satisfied. But in her majority opinion, Justice Sotomayor argued that Campbell required more. It wasn’t sufficient that 2 Live Crew’s song was different from Orbison’s; key to the Campbell holding was that the second song was a parody of the first. And parody (as opposed to general satire) requires a borrowing of the original in the same way that a book review or newspaper article discussing an original work might. So in Campbell too, the majority insisted, the first factor depended on the purpose of the use: it wasn’t just to make money or make a statement about women generally, but to parody the specific work that was copied. Justice Kagan, of course, rejected that view, urging readers to “go take a look at the decision.”

Justice Gorsuch offered a slightly different explanation for the majority’s seeming shift away from Campbell. In a useful concurrence joined by Justice Jackson, he pointed out that “the question before us is a narrow one of statutory interpretation.” While the fair-use factors were developed through judicial decisions, they are presently codified in the Copyright Act, which requires courts to consider “the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit or educational purposes.” While the dissent and AWF were preoccupied with the purpose the creator (Warhol) had in mind and with the character of his resulting work, the majority properly focused on “the purpose and character of the challenged use.” Here, Goldsmith had (at least by the time the case came to the Supreme Court) limited her challenge to AWF’s licensing of Orange Prince to Condé Nast. That use was of exactly the same type that Goldsmith herself made of her Prince photographs. In Campbell, the owner of Orbison’s song challenged 2 Live Crew’s right to use the original work in the first instance. Here, no one challenges Warhol’s, or any other artist’s, right to copy an original in the process of creating a transformative new work. Indeed, as Gorsuch pointed out, “[t]he Court today does not even decide whether the Foundation’s image of Prince infringes on Ms. Goldsmith’s copyright.” It decides only that, if there is infringement, the first fair-use factor favors Goldsmith, not AWF, though only with respect to this particular use. If AWF wanted to display the Prince Series in a museum or even license them for a book commemorating Twentieth Century art, the result may well be different.

And so, at the end of the day, the Court sided with Goldsmith. But just what that means for AWF and other artists going forward remains to be seen. Certainly, anyone who’s copied a photograph should think twice before licensing it for the same purpose as the original. But it seems that the owners of other Warhols that hang on walls are safe for now. Taking the majority at its word, this is just a decision stating that AWF should have paid Goldsmith a licensing fee before it accepted a similar fee from Condé Nast for Orange Prince. That said, it cannot be denied that the decision casts a new light on the first fair-use factor, potentially exposing once safely shadowed types of copying and derivative use to new legal scrutiny.

Next up are Twitter v. Taamneh (No. 21-1496) and Gonzalez v. Google (No. 21-1333). While the cases drew attention for other issues (including internet service providers’ immunity under Section 230 of the Telecommunications Act), the Court resolved both by clarifying the elements of a claim under the Justice Against Sponsors of Terrorism Act (JASTA). Among other things, that statute imposes secondary liability on anyone “who aids and abets, by knowingly providing substantial assistance, or who conspires with the person who committed such an act of international terrorism.” In Twitter, a unanimous Court held that the plaintiffs failed to state a JASTA aiding-and-abetting claim against Twitter. And in Google, a unanimous Court remanded to the Ninth Circuit to consider whether the same should be true of the plaintiffs’ suit against the search giant.

In the early hours of January 1, 2017, an armed terrorist entered a nightclub in Istanbul, Turkey, firing more than 120 rounds and killing or injured more than 100 people. The next day, ISIS released a statement claiming responsibility for the attack. The family of one of the victims who was killed sued Twitter, Google, and Facebook under JASTA (codified at 18 U.S.C. § 2333(d)(2)), alleging that ISIS used the defendants’ platforms for recruiting, fundraising, and spreading propaganda. The family also alleged that defendants knowingly allowed ISIS to use their platforms and to benefit from the platforms’ “recommendation” algorithms, and that defendants profited from advertisements placed on ISIS’s social media activity. After the District Court dismissed the complaint for failure to state a claim, the Ninth Circuit reversed in part, holding that the defendants could be held secondarily liable for the nightclub attack. At the same time, the Ninth Circuit affirmed the dismissal of certain claims against Google seeking to hold it directly liable, both because Google was immune from those claims under Section 230 and because plaintiffs had failed to allege proximate cause. In Twitter, the defendants then sought and obtained certiorari as to whether the family stated a JASTA secondary-liability claim, while in Google, the family obtained certiorari as to whether Section 230 applied.

In Twitter, a unanimous Court reversed in an opinion written by Justice Thomas. His opinion focused on two statutory interpretation questions: (1) What does JASTA mean by “aid and abet”? And (2) what act or person must the defendant have “aided and abetted” for liability to attach?

Starting with the first, Congress did not define “aid and abet” in Section 2333(d)(2). But the statute did explicitly point to a 1983 D.C. Circuit decision, Halberstam v. Welch, as providing the “proper legal framework” for civil aiding and abetting under JASTA. Taking that cue, Thomas closely analyzed Halberstam, as well as the common law of aiding and abetting, and concluded that JASTA’s use of the term “aid and abet” denotes “a conscious, voluntary, and culpable participation in another’s wrongdoing.”

On the second question, the plaintiffs argued that to be held secondarily liable under JASTA, the defendants need only have generally aided and abetted “the person” who committed the act of terrorism. Hence, if Twitter aided ISIS generally, the family need not show that the defendants specifically aided the Reina nightclub attack. The defendants, by contrast, argued that a plaintiff must show a “strict nexus” between the assistance and the “act of international terrorism,” that is, the Reina attack in particular. Thomas rejected both these rigid formulations. Instead, he concluded that the concept of aiding and abetting (including in JASTA) necessarily requires that the defendant aid a wrongful act. But at the same time, a defendant need not have “known all particulars of the primary actor’s plan.” In some circumstances, then, a defendant’s role in an illicit enterprise “can be so systemic” that the defendant aids and abets each of the enterprise’s wrongful acts, even without a close nexus between the aid provided and the specific act complaint of.

Turning to the facts at hand, Thomas concluded that the family’s allegations fell short of showing that the defendants gave the kind of knowing and substantial assistance to ISIS that would amount to them aiding and abetting the Reina attack. The only affirmative conduct the defendants undertook was creating their social media platforms (used by billions worldwide) and setting up algorithms to manage how content is displayed on them. The plaintiffs made no allegations, for example, that the defendants encouraged or gave special treatment to ISIS. While groups like ISIS can use the defendants’ platforms for malicious ends, “[t]he mere creation of those platforms . . . is not culpable.” And defendants’ “recommendation” algorithms are part of the platform’s basic infrastructure. Such “passive assistance” does not qualify as “active abetting.” Because the plaintiffs did not allege “a strong showing of assistance and scienter,” they failed to state a claim for secondary liability under JASTA. A contrary holding “would necessarily hold defendants liable as having aided and abetted each and every ISIS terrorist act committed anywhere in the world” and would extend liability to providers of cell phones, email, and more.

Finally, Thomas addressed allegations that Google, specifically, approved ISIS videos on YouTube as part of a revenue-sharing system, thus splitting advertising revenue with ISIS. Thomas found those allegations deficient, because they “allege[d] nothing about the amount of money that Google supposedly shared with ISIS, the number of accounts approved for revenue sharing, or the content of the videos that were approved.” Without more, the allegations insufficiently charged Google with providing substantial assistance to the nightclub attack.

Justice Jackson wrote a brief concurrence, in which she emphasized the fact-specific nature of the Court’s opinion. “Other cases presenting different allegations and different records may lead to different conclusions,” she cautioned. The general principles relied upon by the majority’s opinion thus “do not necessarily translate to other contexts.”

Turning now to Google, as we said, the Ninth Circuit had dismissed plaintiffs’ direct-liability claims against the company based on Section 230 of the Telecommunications Act. That provision immunizes providers of “interactive computer services” (i.e., websites) from civil liability based on content posted by their users. The Court had granted certiorari to review that holding. But it found it unnecessary to delve into that issue given the plaintiffs’ concession that their ultimate claims against Google were “materially identical to those at issue in Twitter.” As a result, the Court found it “sufficient to acknowledge that much (if not all) of plaintiffs’ complaint seems to fail under either our decision in Twitter or the Ninth Circuit’s unchallenged holdings below.” The Court accordingly vacated the judgment and remanded so that the Ninth Circuit could consider the complaint in light of Twitter.

Our fourth decision for today, Amgen v. Sanofi (No. 21-757), involves the “enabling” requirement of the Patent Act. Two drugmakers, Amgen and Sanofi, both came up with artificial antibodies that help the body reduce bad cholesterol and thereby prevent heart disease and strokes. These antibodies target a naturally occurring protein called PCSK9 that interferes with the receptors in the body that remove bad cholesterol from the bloodstream. By latching on to the amino acid “sweet spot,” these PCSK9-blockers keep the receptors available to do their cholesterol-mitigating work. Amgen and Sanofi both obtained patents for these antibodies in 2011.

But Amgen wanted more. In 2014, it also sought to patent “the entire genus” of antibodies that ward PCSK9 off by binding to its sweet spot, identifying 26 amino acid sequences that did the trick. Amgen first claimed it had a “roadmap” to discover other such antibodies, but the map was pretty do-it-yourself: In substance, Amgen advised making new antibodies, testing them to see if they stuck to PCSK9, checking if the latch was in the sweet spot, and then making sure that the antibody kept PCSK9 away from bad-cholesterol receptors. Or, Amgen suggested in a second patent application, you might start with a proven antibody, swap out some amino acids for others with similar properties, and then see if the antibody still works. Remarkably, the U.S. Patent Office approved both of Amgen’s applications. Unremarkably, Amgen then sued Sanofi for infringing its patents.

That didn’t end well for Amgen. Both the district court and the Federal Circuit sided with Sanofi, which argued that Amgen’s patent application was inadequate because it failed to enable other scientists to create the universe of antibodies subject to Amgen’s patent. The Supreme Court agreed in a unanimous opinion authored by Justice Gorsuch.

As Gorsuch explained, patents embody a bargain between rewarding past inventions and fostering future innovation. That’s why Article I, Section 8 of the Constitution gives Congress the power to “promote the Progress of Science and useful Arts” through patents, and why Congress has exercised that authority since the Patent Act of 1790 to grant a limited monopoly to inventors who demonstrate the originality of their contributions. To ensure the public receives its share of the benefit, the inventor has an “enablement” obligation of disclosing with specificity how the invention works, so that other skilled artisans can replicate and build upon the invention after the patent term expires. But when the patent seeks to cover an entire realm of technical advancement without specifying an enabling technology of commensurate breadth, a patent monopoly is unlawful. Thus, when Samuel Morse invented the telegraph, he was able to patent his unique method of combining multiple circuits to maintain strong electromagnetic forces over long distances, but not the idea of telegraphic communication itself.

Through similar examples ranging from Edison’s lightbulb to Perkins’ glue, Gorsuch elucidated the requirement that a patent’s specification must enable the full scope of the invention so that others skilled in that art can make and use the entire class of patented processes based on their general qualities. Although that specification need not be absolutely exhaustive—some degree of ad hoc experimentation and tailoring is permitted—that allowance must be reasonable in light of the patent’s scope.

Under that standard, Amgen’s patents for the entire genus of PCSK9-blocking antibodies could not stand. Although Amgen had adequately explained how to make the 26 antibodies it identified, it could not monopolize the entire class of antibodies that reduce cholesterol by clinging to PCSK9’s sweet spot merely by granting “two research assignments”—or, even less generously, “a hunting license”—to future scientists. Trial-and-error and conservative substitution might indeed prove valuable to future discovery of other antibodies in this class, but that process was no more than the scientific method at work. Because Amgen’s patents “claim[ed] a lot, but enabled only a little,” it could not block Sanofi or other members of the researching public from attacking the sweet spot of PCSK9.

Now let’s take a break from technology and IP law with Ohio Adjutant General’s Department v. Federal Labor Relations Authority (No. 21-1454). There, the Court considered whether an entity could be a federal “agency” for purposes of the Federal Service Labor-Management Statute (FSLMRS) even though it wasn’t among the exhaustive list of “agencies” identified in the statute. In a 7-2 decision emphasizing function over form, the Court held that the petitioners—the Ohio National Guard and related parties—were “agencies” because they “function[] as an agency.”

The underlying case involved a dispute about the labor rights granted under the FLSMRS and the extent to which those rights may be enforced by the Federal Labor Relations Authority (FLRA). The FLSMRS guarantees rights to federal “employees,” and the FLRA has jurisdiction over federal “agencies” to enforce those rights. These seemingly straightforward definitions were complicated by the fact that this case involved the labor rights of a “rare bird” of federal employment: the “dual-status technician.” Dual-status technicians are federal employees (e.g., they receive federal civil-service pay) but are “employ[ed] and administer[ed]” by a State National Guard, here Ohio’s. During negotiations over a collective bargaining agreement, the Guard apparently reneged on a deal with the union representing the dual-status technicians and stopped deducting union dues from their paychecks. The union complained to the FLRA, which ultimately found that the Guard had violated the FLSMRS.

But did the FLRA have jurisdiction over this dispute in the first place? The parties agreed that the dual-status technicians at issue in the case were “employees” under the FLSMRS. But they disagreed about whether the Guard was an “agency” subject to the FLRA’s jurisdiction. Unsurprisingly, the FLRA upheld its own authority, finding that the Guard was an agency. And in a challenge to the FLRA’s decision, the Sixth Circuit agreed. The Court then granted cert, but only on the jurisdictional question of whether the Guard was an “agency.”

Writing for a majority of seven (all but Justices Alito and Gorsuch), Justice Thomas affirmed. He concluded that the Guard was an “agency” for purposes of the FLRA because it “acts as a federal ‘agency’ when [it] hire[s] and supervise[s] dual-status technicians.”

The Guard resisted this functional analysis, pointing out that FLSMRS defines an “agency” as an “Executive Agency.” “Executive Agency,” in turn, means an “Executive department,” and the statute then exhaustively lists the 15 Cabinet-level Departments. Pointing to these statutory definitions, the Guard argued it could not be an “agency” because it was not one of these 15 enumerated departments.

Thomas rejected this approach for three reasons. First, he observed that the dual-status technicians’ work for the Guard was much like the work such technicians do for the Department of the Army or Air Force. But those entities aren’t specifically identified as “agencies” either. They nonetheless are subject to the FMLA’s jurisdiction, because they are “components” of the Department of Defense, which is a listed “agency.” The Guard should be treated no differently. Second, the Guard only had the authority to employ dual-status technicians based on a “designation” of authority from the Secretaries of the Army and Air Force. This meant that dual-status technicians were ultimately employees of these covered agencies. As Thomas observed, “it would be passing strange” if these individuals had rights under the FLSMRS, but their supervisor had no duty to safeguard them. Third, the Court drew on the body of law that governed federal labor before the enactment of the FLSMRS, pursuant to a savings clause that permitted later courts to rely on these earlier rulings. Because one such earlier case “rejected arguments virtually identical to those petitioners advance here,” the Court “presume[d]” that Congress intended the FLSMRS to have the same jurisdictional reach.

Justice Alito dissented, joined by Justice Gorsuch. He focused narrowly on the statutory text: The FLSMRS provided an exhaustive list of “agencies,” and the petitioners were not on that list. Alito then critiqued the majority’s three main points: First, in his view, it simply does not follow that because an entity is a “component” of an agency or is designated to exercise an agency’s authority, it is itself an agency. Second, he found it “perfectly ordinary” that dual-status technicians might qualify as employees under the FLSMRS, but that the FMLA might not have the authority to protect their rights. Finally, he found the majority’s reliance on precedent unconvincing, because the FLSMRS’s savings clause expressly provided that courts could “disregard” prior precedent, meaning that caselaw provides no real guidance to the meaning of the statute. For these reasons, Alito and Gorsuch would have reversed.

Finally, we come to Polselli v. Internal Revenue Service (No. 21-1599), which addresses what kind of notice the IRS must give when it issues summonses to investigate and collect from delinquent taxpayers. By statute, the IRS can serve a summons to collect any tax or to figure out how much a taxpayer owes. Normally, if the IRS issues a summons implicating a third party, it has to provide that third party with notice, which gives the third party the chance to try to quash the summons. But the IRS need not provide such notice if it issues the summons “in aid of the collection of” past-due tax. Resolving a circuit split, the Court unanimously read this “in aid of” language broadly, rejecting a taxpayer’s argument that notice is excused only if the delinquent taxpayer holds an interest in the object of the summons.

In 2017, the IRS determined that Remo Polselli owed more than $2 million in back taxes. It launched an investigation to find any assets that Polselli might have that could be used to pay the debt. An IRS investigator issued summonses to several banks seeking information about accounts that were held in the name of Polselli’s wife and others, which the IRS believed Polselli himself controlled. The IRS didn’t provide those account holders with notice of the summonses, but they found out anyway and moved to quash them. The District Court dismissed their efforts and a divided Sixth Circuit affirmed, disagreeing with decisions from the Ninth Circuit that read this statutory exception to apply only if the delinquent taxpayer had an interest in the object of the summons.

Writing for a unanimous Court, Chief Justice Roberts affirmed the Sixth Circuit’s dismissal. He started (and more or less ended) with the text of the statute. By default, when the IRS issues a summons under 26 U.S.C. § 7609, it must provide notice. But the statute contains an exception if three requirements are met: (1) the summons is “in aid of . . . collection”; (2) the collection must be of “an assessment made or judgment rendered”; and (3) the summons must aid in collection of assets or judgments “against the person with respect to whose liability the summons is issued.” Contrary to the Ninth Circuit, the statute says nothing about a requirement that the taxpayer have a “legal interest” in the object of the summons. Other related provisions of the Internal Revenue Code, however, do turn on whether a taxpayer has a “proprietary interest” in an account. The absence of a similar requirement in this statute thus seemingly reflects a deliberate choice by Congress.

The Chief then addressed three counterarguments. First, he rejected the petitioners’ argument that unless the taxpayer has a legal interest in the object of the summons, the summons can’t “aid . . . the collection” of a tax debt. That interpretation reads “aid” too narrowly: The IRS issued the summonses in this case because it suspected Polselli was using these other accounts to shelter assets that he literally controlled, even though he did not have a formal “legal interest” in the accounts. Figuring out whether that was so would “aid” the IRS’s efforts to collect. Second, the Court rejected the argument that reading the statute in this way would render another notice exception—one for a “fiduciary or transferee” of the taxpayer—superfluous. Not so, because that exception had additional requirements not found in this exception, so they cover different circumstances. Finally, the Chief brushed aside the petitioner’s argument that the notice provision must be read narrowly because Congress had heightened notice requirements in response to prior cases in which the Court had found notice wasn’t required. That (assumed) congressional goal was no reason to disregard the plain text of the statute Congress wrote.

With those points out of the way, the Chief closed by emphasizing that the Court’s decision was not an attempt to exhaustively catalog all the situations when a summons could be said to be issued “in aid of collection.” Instead, the parties below had focused exclusively on the Ninth Circuit’s “legal interest” requirement. The Court’s decision simply rejected that idea as contrary to the statute, leaving to another day whether particular summonses were issued in aid of collection.

Justice Jackson, joined by Justice Gorsuch, concurred. She emphasized that notice to third parties is the default rule of the statute, so the IRS can only decline to give notice in specific enumerated circumstances. Those circumstances reflect Congress’s effort to balance the privacy interests of individuals with the IRS’s goal of collecting validly assessed taxes. At the same time, Jackson noted that some summonses might be too attenuated from the IRS’s collection activities to count as summons “in aid of collection.” She thus called on both the IRS and the lower courts to “be ever vigilant” to ensure that the IRS’s summons authority is not abused.