The dissent is confident about the ability of federal courts to “improve foreign relations” and make judgments that are “ ‘consonant with U. S. foreign policy interests.’ ” Post, at 16 (opinion of Sotomayor, J.). The Constitution’s allocation of power, however, requires greater judicial humility.

 While few plaintiffs have successfully litigated ATS suits to final judgment, many more have sued and obtained large settlements. See C. Ewell, O. Hathaway, & E. Nohle, Has the Alien Tort Statute Made a Difference?: A Historical, Empirical, and Normative Assessment, 107 Cornell L. Rev. 1205, 1251–1252, 1278 (2022) (discussing ATS settlements, some of which resulted in “multimillion-dollar payments to plaintiffs”). Shutting off the possibility of additional ATS liability thus also provides clarity to defendants.

 The dissent recruits the Blackstone three as support for judicial power to create ATS actions. If the Blackstone three, the dissent asks, then why not more? Post, at 12. The Blackstone three do not give the dissent what it is looking for, because Sosa did not invoke them as evidence that the ATS originally granted courts the authority to create causes of action. Under the prevailing jurisprudence of the time, Sosa said, such torts were understood to be “found or discovered” by courts rather than “made or created.” 542 U. S., at 725. Offering examples, Sosa stated that offenses against ambassadors “appea[r] to have been” at top of mind, violations of safe conduct “were probably understood to be actionable,” and “individual actions arising out of prize captures and piracy may well have also been contemplated.” Id., at 720. Those are the only ATS actions this Court has ever specifically mentioned, and to the extent there has been reliance on the availability of those three actions, we see no need to revisit them. But while we say “this far and no further,” the dissent would go very far indeed. In the name of honoring the expectations of the First Congress, it would leverage the Blackstone three to justify power that would have confounded the First Congress.

 We have noted that the TVPA “contemplates liability” against those “who do not personally execute the torture”—in particular, those who “giv[e] an order to torture.” Mohamad v. Palestinian Authority, 566 U. S. 449, 458 (2012); see also post, at 25 (opinion of Sotomayor, J.) (relying on this assertion). Such liability, however, is much more limited than aiding-and-abetting liability.

SUPREME COURT OF THE UNITED STATES

_________________

No. 24–856

_________________

CISCO SYSTEMS, INC., et al., PETITIONERS v.
DOE I, et al.

on writ of certiorari to the united states court of appeals for the ninth circuit

[June 23, 2026]

Justice Sotomayor, with whom Justice Kagan and Justice Jackson join as to Parts I, II, III, and V, dissenting.

Petitioner Cisco Systems, Inc., allegedly knowingly designed and intentionally built a mass-surveillance system for the Chinese Communist Party to use to identify, track, arrest, and torture thousands of religious minorities. As all agree, if respondents’ allegations were to be proved true, then that would mean Cisco violated universally recognized norms of international law. Respondents thus sued Cisco in a federal action under the Alien Tort Statute (ATS), which vests jurisdiction in district courts over “any civil action by an alien for a tort only, committed in violation of the law of nations,” 28 U. S. C. §1350. In Sosa v. Alvarez-Machain, 542 U. S. 692 (2004), this Court held that the ATS permits federal courts to find implied causes of action authorizing plaintiffs to sue defendants who violated international law.

The Court nonetheless closes the courthouse doors not just to respondents, but to virtually every future litigant seeking redress for a violation of international law under the ATS. It thus overrules Sosa, without even acknowledging that it is doing so. Today’s decision marks yet another low point in this Court’s esteem for its precedents.

Further, the Court also errs by ignoring the plain meaning of the Torture Victim Protection Act of 1991 (TVPA) and slamming the door completely shut to claims by U. S. citizens against those who aid and abet torture.

I respectfully dissent from both of the Court’s holdings.

I

A

The following factual allegations are drawn from respondents’ amended complaint and are assumed to be true because this appeal arises from the grant of a motion to dismiss. Ashcroft v. al-Kidd, 563 U. S. 731, 734 (2011).

Respondents are members of Falun Gong, a Chinese spiritual movement. Falun Gong members believe in the three core tenets of “Truthfulness, Compassion, and Tolerance.” App. 2. Many of its members have vocally opposed the Chinese Communist Party (CCP).

In 1999, the CCP designated Falun Gong organizations as illegal and called for violent political crackdowns against them. During these “violent struggle” campaigns (or “douzheng”), CCP and Chinese Government officials subjected Falun Gong targets to brutal human-rights abuses, including torture, arbitrary detention, and reeducation through labor. Id., at 12; see 73 F. 4th 700, 712 (CA9 2023) (detailing “beatings with steel rods[,] . . . shocking with electronic batons, sleep deprivation, . . . violent force-feeding,” and more). The United States State Department estimates that hundreds of thousands of Falun Gong believers have been persecuted, and media sources report that thousands have been tortured to death.

To target Falun Gong members, the CCP determined that it needed to develop a sophisticated, nationwide internet-surveillance tool to track Falun Gong activity online. It thus devised the anti-Falun Gong “ ‘Golden Shield’ ”: a “ ‘vast and multi-tiered’ ” bespoke “ ‘surveillance system’ ” that the CCP contemplated would obtain and organize all manner of data on the Falun Gong members across China, including address information, internet history, and financial and family information. Id., at 710. At the time the CCP came up with the idea, however, Chinese engineers lacked the technical expertise to create such technology. As a result, the CCP turned to American technology companies and invited them to submit proposals to design its Golden Shield system.

Petitioner Cisco answered the CCP’s call with an aggressive campaign to secure the Golden Shield contract. Cisco’s CEO (and other senior executives) personally met with China’s President and other officials to discuss and “explicit[ly] support” “the Golden Shield’s douzheng objectives and goals.” App. 62. Internal files reveal Cisco’s “pledge to satisfy the repressive anti-Falun Gong purposes of the” program, while characterizing Falun Gong and its members as “ ‘viruses,’ ” “ ‘despicable,’ ” and an “ ‘evil cul[t]’ ”—all mirroring CCP propaganda denigrating Falun Gong members as “subhuman.” Id., at 20–21. Cisco’s public Chinese marketing materials also described the company’s “promis[e] to tailor the apparatus” to meet the CCP’s objectives, including “the douzheng of ” Falun Gong. Id., at 22–23; see 73 F. 4th, at 710 (describing Cisco brochures offered at Chinese trade shows marketing its services “as useful to the ‘douzheng’ of Falun Gong”).

Cisco’s push succeeded. In 2001, the CCP selected Cisco to submit designs for the Golden Shield and later awarded it contracts to develop the program. Cisco, using engineers working from the United States, then created a massive surveillance system that “ ‘analyzed patterns of Falun Gong Internet activity to enable the intelligent identification of individual Falun Gong Internet users,’ ” while providing “ ‘real time monitoring’ ” of “ ‘Falun Gong Internet traffic patterns’ ” and sharing the results with the CCP to “facilitate the . . . forced conversion through torture’ ” of Falun Gong members. Id., at 711. Over time, Cisco engineers in San Jose, California continued to provide support while also upgrading and expanding the system. For example, Cisco later created a video-surveillance system that, combined with facial-recognition technology, has become a “ ‘primary means’ of identifying Falun Gong practitioners through non-internet activities, such as protests or religious practice.” Ibid.

Armed with the vast suite of technological tools Cisco provided, the CCP identified, arrested, and tortured thousands of Falun Gong members. Id., at 712.

B

Respondents in this case are among those victims or are family members of those who have disappeared and are suspected or confirmed dead. They allege that the CCP used information collected and stored by the Golden Shield program in forced-conversion sessions to which they were subjected. For example, on top of prolonged isolation and physical torture, Chinese officials leveraged information about some of respondents’ family members (and employed threats against those family members) to coerce respondents into renouncing their religious beliefs.

In 2011, respondents filed this action against Cisco and two of its executives, suing under the ATS and alleging that Cisco aided and abetted seven international-law violations: torture; cruel, inhuman, or degrading treatment; forced labor; prolonged and arbitrary detention; crimes against humanity; extrajudicial killing; and forced disappearance. One of the respondents, who is a U. S. citizen, also brought a claim alleging that two individual defendants aided and abetted his torture in violation of the TVPA.

After the District Court dismissed the action, the Ninth Circuit reversed, holding that aiding-and-abetting liability was available under both the ATS and the TVPA. Id., at 709. This Court then agreed to review the Ninth Circuit’s judgment. 607 U. S. 1120 (2026).

II

A straightforward application of this Court’s settled case law should have allowed respondents’ ATS claims against Cisco to proceed.

A

The ATS, enacted in 1789, permits foreign nationals harmed by violations of international law to obtain compensation for their injuries in U. S. court. See 28 U. S. C. §1350 (“The district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States”). The statute is limited: As this Court explained in Sosa v. Alvarez-Machain, 542 U. S. 692, the ATS “is in terms only jurisdictional,” meaning that it gives courts authority to entertain suits but does not itself create any causes of action that permit plaintiffs to sue someone. Id., at 712. Even so, consulting the historical context surrounding the ATS’s passage, the Court determined that Congress “intended” for the ATS “to have practical effect the moment it became law.” Id., at 724. Consistent with that understanding, Sosa held that, although federal courts should be “restrained” in doing so, they may find implied private causes of action under the ATS “for violations of any international law norm” comparable to “the historical paradigms familiar when [the ATS] was enacted.” Id., at 725, 732. Those paradigms include three international-law offenses that were well established when the ATS was enacted: “violation of safe conducts, infringement of the rights of ambassadors, and piracy.” Id., at 715. As all agree, the ATS allows implied private rights of action for at least these three offenses.

To invoke an additional implied cause of action under the ATS, Sosa explained, the plaintiff must identify a norm of international law that has sufficiently “definite content” that has gained “acceptance among civilized nations.” Id., at 732. As part of that inquiry, courts must also make a “judgment about the practical consequences of making that cause available to litigants in the federal courts.” Id., at 732–733. Applying this test, the Court rejected the arbitrary-detention claim that had been raised there. The plaintiff had not alleged a sufficiently definite norm of international law. Recognizing the plaintiff ’s proposed cause of action would also have “breathtaking” consequences by permitting “a cause of action in federal court for any arrest, anywhere in the world,” so long as it is “unauthorized by the law of the jurisdiction in which it took place.” Id., at 736. The Court thus declined to allow a new cause of action under the ATS. Ibid.

“In the years since, this Court has read Sosa to announce a two-step test for recognizing the availability of a cause of action under the ATS.” Nestlé USA, Inc. v. Doe, 593 U. S. 628, 648 (2021) (Sotomayor, J., concurring in part and concurring in judgment). At step one, courts ask “ ‘whether a plaintiff can demonstrate that the alleged violation is “of a norm that is specific, universal, and obligatory.” ’ ” Ibid. If the plaintiff makes that required showing, then the court must determine “ ‘whether allowing [a] case to proceed under the ATS is a proper exercise of judicial discretion.’ ” Ibid. That discretionary inquiry must account for “the potential implications for the foreign relations of the United States of recognizing such causes [of action],” as courts must be “particularly wary of impinging on the discretion of the Legislative and Executive Branches in managing foreign affairs.” Sosa, 542 U. S., at 727.

B

This case succeeds at each of Sosa’s steps. At step one, there is no dispute. Respondents allege that Cisco aided and abetted the violation of seven international-law norms: torture; cruel, inhuman, or degrading treatment; forced labor; prolonged and arbitrary detention; crimes against humanity; extrajudicial killing; and forced disappearance. See 73 F. 4th, at 713. Before this Court, Cisco does not make any arguments at all at step one. See Brief for Petitioners 17. It does not dispute that all seven norms are specific, universal, and obligatory, as required by Sosa. Nor does it dispute that international law recognizes aiding-and-abetting liability, either in general or as a substantive part of each of these seven norms.

Cisco, joined by the United States as amicus curiae, instead exclusively focuses on Sosa’s second step, at which courts ask “ ‘if there are sound reasons to think Congress might doubt the efficacy or necessity of a damages remedy.’ ” Jesner v. Arab Bank, PLC, 584 U. S. 241, 264 (2018). Both fail to demonstrate that it would be improper to allow this case to proceed.

First, Cisco and the United States argue that an aiding-and-abetting theory of liability can never satisfy Sosa’s second step because such liability will always endanger foreign-policy concerns. Those concerns, however, are best addressed on a case-by-case basis, rather than a categorical one. See Sosa, 542 U. S., at 733, n. 21 (discussing possibility of “case-specific deference to the political branches”).

This case disproves the idea that recognizing aiding-and-abetting liability will necessarily impair U. S. foreign-policy interests. The political branches have already consistently condemned China’s treatment of Falun Gong members. The Solicitor General confirmed in this very case that “the United States [has] long condemned China’s treatment of Falun Gong practitioners,” and has sanctioned Chinese officials for “gross violations of human rights,” including “particularly severe violations of religious freedom.” Brief for United States as Amicus Curiae on Pet. for Cert. 9. Similarly, in 2020, the Trump administration released a report on the Government’s strategic approach to China that highlighted how the Government had released “statements calling on the [Chinese] government to respect the rights of ” religious minorities including “Falun Gong adherents, all of whom face repression and persecution in China”; the report also announced that the “United States will continue to take a principled stand against the use of our technology to support China’s military and its technology-enabled authoritarianism.” In addition, the State Department has “denounce[d]” China’s “particularly abhorrent behavior” toward Falun Gong. Dept. of State, Annual Report on International Religious Freedom 2000, 106th Cong., 2d Sess., xxix, xxxii (Joint Comm. Print 2000). So has Congress. See, e.g., H. Res. 605, 111th Cong., p. 5 (2010) (condemning the Chinese Government’s “campaign to persecute, intimidate, imprison, and torture Falun Gong practitioners”); see also Brief for Members of Congress et al. as Amici Curiae 22–26 (listing other congressional resolutions and actions condemning China’s persecution of Falun Gong). Given that both political branches have so publicly and directly condemned China’s persecution of Falun Gong members, it strains credulity to say that allowing a private suit against an American company to proceed would meaningfully change the state of relations between the United States and China.

In addition, any potential foreign-policy consequences are further reduced because neither China nor any Chinese instrumentality is a party to the case itself. Rather, the suit is against an American company and focuses primarily on its conduct in the United States. China also could have appeared in this case if it opposed it. It could have filed a brief in this case saying so, just as it did in a previous ATS case that involved China. See Doe v. Qi, 349 F. Supp. 2d 1258, 1264, 1296–1301 (ND Cal. 2004) (noting that, in an ATS suit against Chinese local-government officials, China, “through the United States Department of State, submitted a letter to this Court urging this Court not to assert jurisdiction over the instant cases”). Other foreign states have filed similar materials in other ATS cases. See, e.g., Jesner, 584 U. S., at 271 (citing brief filed by the Hashemite Kingdom of Jordan in ATS suit against a Jordanian bank). That China has not done so here is additional evidence that this case is unlikely to aggravate U. S.–China relations.

Second, Cisco and the United States argue that this Court’s decision in Central Bank of Denver, N. A. v. First Interstate Bank of Denver, N. A., 511 U. S. 164 (1994), forecloses aiding-and-abetting liability here, reading that case to permit aiding-and-abetting liability only if a statute expressly so states. Central Bank has minimal bearing here. It involved the Securities Exchange Act of 1934, the text of which limits liability to those who “ ‘directly or indirectly’ ” engage in certain kinds of conduct. Id., at 171. The Court there recognized that Congress “has taken a statute-by-statute approach to civil aiding and abetting.” Id., at 182. The Court reasoned that the text of that statute did not extend liability to aiders and abettors because they do not necessarily engage in the underlying activity they facilitate. Id., at 176–177. That conclusion was buttressed by the fact that it was “uncertain” whether state common law supplied “a deeply rooted background of aiding and abetting tort liability.” Id., at 181–182, 184. Here, by contrast, the ATS does not limit liability to those who engage in certain acts. Instead, it authorizes courts to entertain suits involving a “tort . . . committed in violation of the law of nations.” §1350. As noted above, Cisco does not dispute that liability for aiding and abetting the seven underlying violations is universally recognized in international law. Central Bank’s holding about a textually distinct statute rooted in domestic law thus says little about what the ATS recognizes under international law.

Because Cisco does not dispute that Sosa’s first step is satisfied and has not identified a sound reason for this suit not to proceed at Sosa’s second step, the Court should affirm the Ninth Circuit’s judgment.

III

The majority, however, is in a rush. Not wanting to dally on niceties like binding precedent, or to confine itself to deciding the case before it, the majority casts aside Sosa’s two-step framework altogether. It disclaims any future authority by any court to find additional causes of action in all cases to come, even though it declines to disturb the three implied causes of action previously recognized under the ATS. Ante, at 1. In short, it overturns Sosa. In doing so, the Court brushes past the stare decisis factors that this Court considers when deciding whether to overturn precedent. It also slams the door in the faces of victims of horrific mistreatment without giving any reason to think that Congress, whom the Court purports to respect, would have wanted to do so. The majority errs at each turn.

A

To begin, the majority fails to demonstrate that Sosa is wrong, never mind that there is the special justification needed to overturn it.

Start with the common ground: There is no dispute that the ATS contains judicially implied causes of action. Specifically, it contains implied causes of action for “ ‘violation of safe conducts, infringement of the rights of ambassadors, and piracy.’ ” Ante, at 4 (quoting Sosa, 542 U. S., at 715). Those causes of action appear nowhere in the ATS’s text. They instead come from authoritative accounts of international law, most notably the work of 18th century English jurist William Blackstone. See 4 W. Blackstone, Commentaries on the Laws of England 68 (1769). (Hence the moniker “Blackstone three.”) As Sosa explained, suits alleging violations of safe conducts, infringements on the rights of ambassadors, and acts of piracy all may proceed under the ATS because Congress likely had these three offenses “in mind” when it passed the ATS, even if it did not say so expressly. 542 U. S., at 724–725.

If the ATS allows these three implied causes of action drawn from international law, then it is hard to see why it does not allow others similarly drawn from that body of law. Indeed, this is a key part of Sosa’s reasoning and its central holding. See Bucklew v. Precythe, 587 U. S. 119, 136 (2019) (“[J]ust as binding as [a case’s] holding is the reasoning underlying it”). When Congress passed the ATS, “the accepted conception” of the common law was that federal courts, when adjudicating individual cases, were “discover[ing]” a pre-existing and “ ‘transcendental body of law outside of any particular State’ ” and applying it. Sosa, 542 U. S., at 725. Working under that assumption, Congress omitted any specific causes of action from the ATS, see §1350, and left it to courts to draw upon the body of international law, which included the Blackstone three, and apply relevant principles in other, specific cases. Id., at 724–725, 730.

Ample historical evidence corroborates Sosa’s account of the ATS. The First Congress passed the ATS after the Continental Congress (operating under the Articles of Confederation) “was hamstrung by its inability to ‘cause infractions of treaties, or of the law of nations to be punished.’ ” Id., at 716 (quoting J. Madison, Journal of the Constitutional Convention 60 (E. Scott ed. 1893)). For example, on at least two occasions prior to the Constitution’s ratification, foreign diplomats were assaulted on American soil. 542 U. S., at 716–717. The Continental Congress called on States to pass legislation authorizing their courts to hear cases involving violations of international law, but just one answered the call. Id., at 716. Understandably offended, at least one foreign ambassador lodged a protest that was discussed during the Constitutional Convention. Id., at 717. As Sosa explained, given the “anxieties of the preconstitutional period,” it is unlikely that Congress intended for the ATS to serve only “as a jurisdictional convenience to be placed on the shelf for use by a future Congress or state legislature that might, someday, authorize the creation of causes of action.” Id., at 719. Rather, Congress intended for its handiwork to have immediate effect, without further express creation of causes of action.

Contemporaneous legal sources discussing international incidents that transpired shortly after Congress passed the ATS further support Sosa’s holding. In 1792, Secretary of State Thomas Jefferson and Attorney General Edmund Randolph each issued an opinion addressing incidents in which Americans entered Spanish Florida and French Haiti, abducted slaves, and returned them to the United States. See T. Jefferson, Opinion on Offenses Against the Law of Nations (Dec. 3, 1792), in 24 The Papers of Thomas Jefferson 693–696 (J. Catanzariti ed. 1990) (Jefferson Opinion); E. Randolph, Opinion on Offenses Against the Law of Nations (Dec. 5, 1792), in 24 id., at 702–703 (Randolph Opinion). Both agreed that federal courts could hear civil suits relating to these incidents under the ATS. Jefferson wrote that the ATS would cover any suit by “an alien . . . for a tort only,” Jefferson Opinion 694 (emphasis deleted), while Randolph agreed that “damages may be recovered in the courts of the United States, under the jurisdictions established by the judicial law”—that is, under the ATS, Randolph Opinion 702. Similarly, in 1795, Attorney General William Bradford issued an opinion explaining that “Americans who had taken part in the French plunder of a British slave colony in Sierra Leone,” and thus arguably breached a declaration of neutrality, “ ‘no doubt’ ” could be held civilly liable “ ‘in the courts of the United States; jurisdiction being expressly given to these courts in all cases where an alien sues for a tort only, in violation of the laws of nations, or a treaty of the United States.’ ” Sosa, 542 U. S., at 721 (quoting 1 Op. Atty. Gen. 57, 59).

Taken together, the opinions of these early American statesmen confirm that the ATS was understood to allow civil suits for violations of international law even without any express causes of action. They also support the conclusion that the ATS was not understood to be limited to the Blackstone three: None of these early incidents involved safe conducts, ambassadors, or piracy at sea. Instead, they involved violations of other international-law principles, including neutrality and prohibitions on “plunder” and “robbery.” 2 E. de Vattel, The Law of Nations, ch. VI §78 (J. Chitty ed., 6th ed. 1844); see Franchise Tax Bd. of Cal. v. Hyatt, 587 U. S. 230, 239 (2019) (describing de Vattel as “the founding era’s foremost expert on the law of nations”). As Sosa did, the founding generation thus understood the ATS to leave the door open to other causes of action rooted in the law of nations beyond the Blackstone three.

The majority does not explain why implied causes of action for the Blackstone three are permissible but implied causes of action for any other norm of international law, which either existed in 1791 or exist today, are not. The majority instead turns down a different road, one paved not with the opinions of the Court, which are the law, but with the opinions of individual Justices, which are not. Despite claiming that its “starting point is Sosa’s key insight,” the majority veers off with its very next paragraph. Ante, at 7. It observes that Justice Scalia, who concurred in part and concurred in the judgment in Sosa, “would have stopped” with just the Blackstone three, ante, at 7, and then in the next paragraph, cites four more individual opinions expressing sympathy for that view, but none of which was the controlling majority opinion. See ante, at 8 (citing Jesner, 584 U. S., at 274 (Thomas, J., concurring); id., at 280–293 (Gorsuch, J., concurring in part and concurring in judgment); Nestlé, 593 U. S., at 634–640 (opinion of Thomas, J.); id., at 658 (Alito, J., dissenting)).

These five opinions, however, are not, and never have been, the law. The Sosa majority expressly rejected the call to stop at the Blackstone three, dismissing Justice Scalia’s position as “particularly unconvincing in light of what we know about congressional understanding bearing on this issue.” 542 U. S., at 729–730. Nor did any of the subsequent opinions garner a majority, as the Court in each case rejected liability under the ATS on narrower grounds than the separate writings advocated. See Nestlé, 593 U. S., at 634 (rejecting extraterritorial application of the ATS); Jesner, 584 U. S., at 270–272 (holding that Sosa’s second step counseled against extending ATS liability to foreign corporations). To say something three times (or five) does not make it true. See L. Carroll, The Hunting of the Snark (1876).

Setting that aside, the majority distills two points from its tour through separate writings: (1) judicial authority to recognize causes of action is “ ‘narrow’ ” because ATS suits “by their nature implicate foreign policy”; and (2) “the power to create causes of action belongs to Congress.” Ante, at 8–9.

Neither point explains why it respects Congress to recognize causes of action for the Blackstone three but not for other international-law principles that are sufficiently similar to those causes of action. To the extent the majority thinks that the Blackstone three are uniquely supported by the historical record, that is doubtful for the reasons given above. See supra, at 13–14. There also is no indication that Congress intended to “tra[p]” the ATS “in amber.” United States v. Rahimi, 602 U. S. 680, 691 (2024).

In any event, this Court already resolved the majority’s objections in Sosa itself. First, Sosa explained that foreign-policy concerns can arise from the failure to provide remedies for international-law violations just as they can arise from the imposition of liability for such violations. The founding generation was deeply concerned with foreign tensions resulting from “the Continental Congress’s incapacity to deal with,” and States’ failures to provide remedies for, international-law violations, especially when an American defendant was involved. 542 U. S., at 716–718. The ATS was a direct response to those “anxieties of the preconstitutional period.” Id., at 719. Indeed, the incidents discussed above all came to the Federal Government’s attention because foreign nations complained to the United States about them. See, e.g., Jefferson Opinion 693 (noting complaints from Spanish and French diplomats); 1 Op. Atty. Gen., at 58 (referring to complaint submitted by British minister).

Today, too, providing an American forum for international-law violations can improve foreign relations, as evidenced by the fact that foreign sovereigns have sometimes filed briefs expressing support for ongoing ATS litigation. See, e.g., In re Estate of Ferdinand Marcos Hum. Rights Litigation, 94 F. 3d 539, 547 (CA9 1996) (noting that the Philippines “urged” an ATS “suit to proceed in American courts” because “ ‘relations may well be improved if Filipino citizens see that justice is available in U. S. courts’ ”); Sarei v. Rio Tinto, PLC, 671 F. 3d 736, 756 (CA9 2011) (noting that Papua New Guinea “expressly urged” that an ATS aiding-and-abetting case “ ‘be heard by courts in the United States’ ” because it would not “ ‘adversely affec[t] any relations between [Papua New Guinea] and the United States’ ”), judgt. vacated, 569 U. S. 945 (2013). Former United States Ambassadors-at-Large for War Crimes also explain that authorizing aiding-and-abetting liability under the ATS “is consonant with U. S. foreign policy interests” in light of the United States’ “consisten[t] promot[ion of] aiding and abetting as an essential means for holding perpetrators of international crimes to account.” Brief for Former U. S. Ambassadors-at-Large for War Crimes Issues/Global Criminal Justice et al. as Amici Curiae 10, 12; see id., at 9, 11–29 (describing the United States’ role in supporting aiding-and-abetting liability in proceedings before the Nuremberg Tribunal and other international criminal courts).

Given that there are strong foreign-policy interests on both sides of the scale, Sosa correctly rejected the argument that fear of “adverse foreign policy consequences” should categorically preclude judicial involvement in this area. 542 U. S., at 728. Instead, although it stressed that courts must act “with great caution” when finding implied causes of action, Sosa left the door open to doing so. Id., at 728, 731.

Second, Sosa also fully addressed the majority’s charge that, according to modern legal sensibilities, Congress, not courts, should create causes of action. Like today’s Court, the Sosa Court was well aware that “this Court has recently and repeatedly said that a decision to create a private right of action is one better left to legislative judgment in the great majority of cases.” Id., at 727; see ante, at 9. Even so, Sosa explained that the First Congress, which passed the ATS, did not see things the same way and that it would be “particularly unconvincing” to impose present-day legal theories on a Congress from two centuries ago. 542 U. S., at 730. It would be “unreasonable,” Sosa continued, to fault Congress for failing in 1791 to anticipate the prevailing legal fashion two centuries down the line and to calcify a statute that Congress intended to be flexible. Ibid.

For all these reasons, Sosa correctly held that courts may find implied private rights of action cognizable under the ATS, so long as they proceed with “great caution” and do not recognize causes of action dissimilar to the Blackstone three: that is, those “for violations of any international law norm with less definite content and acceptance . . . than the historical paradigms familiar when [the ATS] was enacted.” Id., at 728, 732.

B

Even if Sosa was wrongly decided, that still would not justify the result the majority reaches today. The rule of stare decisis, or “[a]dherence to precedent,” this Court has explained, is “ ‘a “foundation stone of the rule of law,” ’ ” Kisor v. Wilkie, 588 U. S. 558, 586 (2019). This “important doctrine . . . ensure[s] that the law will not merely change erratically, but will develop in a principled and intelligible fashion.” Vasquez v. Hillery, 474 U. S. 254, 265 (1986). It also “permits society to presume that bedrock principles are founded in the law rather than in the proclivities of individuals, and thereby contributes to the integrity of our constitutional system of government, both in appearance and in fact.” Id., at 265–266. Overturning a prior decision therefore requires “special justification”; “it is not alone sufficient that we would decide a case differently now than we did then.” Kimble v. Marvel Entertainment, LLC, 576 U. S. 446, 455–456 (2015).

The majority, however, nowhere mentions stare decisis. Indeed, the Court’s opinion does not even state the necessary consequence of what it decides today: that Sosa is overruled. Disturbingly, this is not the first time in recent years that the Court has been “willing to overrule precedent without even acknowledging it is doing so, much less providing any special justification.” Jones v. Mississippi, 593 U. S. 98, 144 (2021) (Sotomayor, J., dissenting) (describing the Court’s implicit overruling of Miller v. Alabama, 567 U. S. 460 (2012), and Montgomery v. Louisiana, 577 U. S. 190 (2016)). “How low this Court’s respect for stare decisis has sunk.” 593 U. S., at 144.

The majority’s apparent disregard for stare decisis is particularly lamentable here because this Court has long held that stare decisis has “special force” as to decisions interpreting congressional statutes. Patterson v. McLean Credit Union, 491 U. S. 164, 172 (1989). That is because Congress, not this Court, has the primary responsibility to “correct any mistake[s]” in this Court’s statutory decisions. Kimble, 576 U. S., at 456. Accordingly, this Court has consistently required a “superspecial justification to warrant reversing” a statutory precedent. Id., at 458. This Court has applied this heightened standard “even when a decision has,” like Sosa, “announced a ‘judicially created doctrine’ designed to implement a federal statute.” 576 U. S., at 456.

When deciding whether to overturn a decision like Sosa, this Court will “most often” consider “subsequent legal developments—‘either the growth of judicial doctrine or further action taken by Congress’—that have removed the basis for a decision.” 576 U. S., at 458. Here, however, this “primary reason” for overruling statutory precedents, Patterson, 491 U. S., at 173, is completely absent. As discussed above, this Court’s modern push to limit implied causes of action was well underway by the time Sosa was decided. See supra, at 17.

The majority says that “[s]ince Sosa was decided, we have firmly committed to the view that judicially created causes of action offend the separation of powers in almost every circumstance.’ ” Ante, at 10. It relies primarily on cases concerning not implied causes of action under the ATS, but cases from this Court’s Bivens jurisprudence. See ante, at 10 (citing, among others, Egbert v. Boule, 596 U. S. 482 (2022); Ziglar v. Abbasi, 582 U. S. 120 (2017)). Those cases are far afield: They involved implying causes of action under the Constitution, not the ATS. There are, of course, many differences between the Constitution and the ATS, chief among which is that Congress passed the ATS specifically to furnish a path for plaintiffs to recover monetary damages. See Sosa, 542 U. S., at 724. Setting those differences aside, by the time this Court decided Sosa in 2004, it was already “firmly committed,” ante, at 10, to limiting implied rights of action. Indeed, years before Sosa, this Court had already “sworn off ” the habit of implying private causes of action outside the ATS. Alexander v. Sandoval, 532 U. S. 275, 287 (2001); see Sosa, 542 U. S., at 727 (citing Alexander). Even so, Sosa confronted that issue and held that courts may find implied rights of action under the ATS. The cases the majority cites from the Bivens context therefore did not unearth some new objection not considered by the Sosa Court, “remov[e] the basis for [Sosa],” Kimble, 576 U. S., at 458, or render Sosa “irreconcilable with competing legal doctrines or policies,” Patterson, 491 U. S., at 173.

Nor has Congress amended the ATS or passed other legislation that undercuts Sosa’s holding. Sosa itself “welcome[d] any congressional guidance” over the interpretation of the ATS and how courts proceed when defining causes of action. 542 U. S., at 731. It also reiterated that “Congress may” close the door to new causes of action “at any time (explicitly, or implicitly by treaties or statutes that occupy the field), just as it may modify or cancel any judicial decision so far as it rests on recognizing an international norm as such.” Ibid. Yet Congress has done none of this in response to Sosa.

The other circumstance in which this Court will sometimes overturn a statutory precedent is if its standard has “prove[n] unworkable.” Kimble, 576 U. S., at 459. Here, there is no reason to think that is true of Sosa’s two-step framework. No one argues before this Court that Sosa’s steps are unclear or hard to apply. Instead, Cisco argues that Sosa and the ATS have allowed causes of action to proliferate, ensnaring defendants and courts in litigation. See Brief for Petitioners 26.

No evidence supports Cisco’s worries. Federal courts have been restrained in recognizing new causes of action under Sosa. One study found that, as of 2022, just 300 ATS suits had ever been filed in federal courts. C. Ewell, O. Hathaway, & E. Nohle, Has the Alien Tort Statute Made a Difference?: A Historical, Empirical, and Normative Assessment, 107 Cornell L. Rev. 1205, 1240 (2022) (Ewell). Fifty-two cases succeeded, and of the 248 cases that were dismissed, over one-third of them were dismissed at Sosa’s first step because the plaintiffs had not identified an actionable violation of international law to begin with. Ewell 1241.

Two observations follow from this study. First, there is no real reason to think that ATS suits are clogging federal courts’ dockets, as 300 cases across multiple centuries reflect a tiny fraction of the hundreds of thousands of civil cases that are filed in federal courts each year. Second, this Court’s cases narrowing ATS liability have had a significant, and effective, role in reining in federal courts and controlling the proliferation of implied causes of action for violations of international law. For example, the number of filed ATS cases fell precipitously after this Court held that the ATS does not apply extraterritorially. Ewell 1237–1238 (discussing Kiobel v. Royal Dutch Petroleum Co., 569 U. S. 108 (2013)). Putting these points together, there is no reason to think that Sosa’s framework has become, or is at risk of becoming, unmanageable.

Again, the majority largely skips over the stare decisis analysis and prefers instead to decree that Justice Scalia was right in Sosa after all. Ante, at 7–12. In so doing, the majority forgets that while creating causes of action may be a task primarily for Congress, so too is correcting erroneous interpretations of statutes. See Kimble, 576 U. S., at 460. In nonetheless reaching out to correct a perceived error in Sosa, the Court does not honor the constitutional division of labor between Congress and this Court. Instead, it disrespects it. Proper respect for Congress should have led to a different outcome today.

Moreover, even if Justice Scalia did get it right in 2004, “[l]ost arguments are not grounds to overrule a case.” Students for Fair Admissions, Inc. v. President and Fellows of Harvard College, 600 U. S. 181, 352 (2023) (Sotomayor, J., dissenting). “When proponents of [previously rejected] arguments, greater now in number on the Court, return to fight old battles anew, it betrays an unrestrained disregard for precedent.” Id., at 352–353. That is exactly what the majority again exhibits today.

C

The majority wraps up its discussion of Sosa by asserting that “it is hard to see how correcting Sosa’s unfulfilled prediction would upset reliance interests” because “it is a stretch to believe that we ever would have” found an implied right of action under Sosa. Ante, at 11. The majority, however, overlooks the 107 other federal courts. Many of them fulfilled this Court’s prediction by finding implied causes of action under the ATS in individual cases, just as the First Congress intended. Those courts were not “ ‘indulg[ing a] fiction.’ ” Ibid. Rather, they were dutifully applying this Court’s decision in Sosa, and their decisions are proof that Sosa “ ‘operated in practice.’ ” Ante, at 11. It is therefore unclear what the majority means when it says that Sosa’s promise went unfulfilled, or that no one, not even the plaintiffs who sued under the ATS, was entitled to rely on it.

What is clear is that the majority today forecloses future reliance on Sosa and shuts the courthouse doors to almost any claimed violation of international law under the ATS. That includes torture. See, e.g., Filartiga v. Pena-Irala, 630 F. 2d 876, 878 (CA2 1980) (ATS suit brought by plaintiff alleging that his son had been kidnapped and tortured to death in retaliation for the plaintiff ’s political beliefs). It includes forced labor. See, e.g., Licea v. Curacao Drydock Co., 584 F. Supp. 2d 1355, 1359 (SD Fla. 2008) (ATS suit alleging that the defendant trafficked Cubans to Curacao, held them in captivity, and forced them to work repairing ships and oil platforms). It also includes perhaps the most universally condemned crime in the modern era: genocide. See, e.g., Kadic v. Karadzic, 70 F. 3d 232, 236–237 (CA2 1995) (ATS suit alleging “brutal acts of rape, forced prostitution, forced impregnation, torture, and summary execution, carried out by Bosnian-Serb military forces as part of a genocidal campaign”). “Like the pirates of the 18th century,” whose conduct so concerned Blackstone and the First Congress, “today’s torturers, slave traders, and perpetrators of genocide are ‘hostis humani generis, an enemy of all mankind.’ ” Nestlé, 593 U. S., at 647 (opinion of Sotomayor, J.).

As to each of these offenses, as to each of these enemies of mankind, the majority decides that there is simply no way that a suit could possibly proceed without offending Congress. Noticeably absent from the majority’s analysis is any evidence that Congress would be offended by these suits. Of course, there may be reasons why allowing an individual ATS suit to proceed would be unwise. That possibility, however, should be addressed on a case-by-case basis. Indeed, that option was available in this very case. Even though the analysis of the dissenting judge below was mistaken for the reasons given above, see supra, at 5–10, it at least applied Sosa and concluded that allowing this specific suit to proceed would not be an appropriate exercise of judicial discretion, see 73 F. 4th, at 748–751. The possibility that one case should not be allowed through the door is no reason to weld the door shut and lock out all others.

IV

The majority also errs by holding that the TVPA does not allow plaintiffs to sue for the aiding and abetting of torture. The TVPA states that “[a]n individual who . . . subjects an individual to torture shall, in a civil action, be liable for damages to that individual.” Note following 28 U. S. C. §1350 (Establishment of Civic Action) (§1350 note). This Court has already acknowledged that this language extends beyond those who “personally execute the torture.” Mohamad v. Palestinian Authority, 566 U. S. 449, 458 (2012). This Court has also explained that aiding and abetting requires that “the defendant consciously and culpably ‘participate[d]’ in a wrongful act so as to help ‘make it succeed.’ ” Twitter, Inc. v. Taamneh, 598 U. S. 471, 493 (2023).

Putting the terms “subjects” and “aiding and abetting” together yields a simple answer. In ordinary meaning, a person “subjects” another to torture if he “consciously and culpably ‘participate[s]’ in” the torture “so as to help ‘make it succeed.’ ” Ibid. Imagine an individual who rounds people up and delivers them to a black site to be tortured by another person. Or a company that sets up the torture chamber with all the necessary tools. Or an informant who provides sensitive information about the victim so that another person can tailor the torture to the victim’s weaknesses. None of these actors conducted the torture themselves, but they have all “subjected” those people to torture by “culpably ‘participat[ing]’ in” it and helping to “ ‘make it succeed.’ ” Ibid.

Dictionary definitions reinforce this understanding. Dictionaries define “subjects” to mean “to cause to undergo or submit to” or to “expose.” Webster’s Third New International Dictionary 2275 (1993) (capitalization modified); see American Heritage Dictionary 807 (3d ed. 1992) (“[t]o expose to something,” “[t]o cause to experience”). A person plainly causes, exposes, or otherwise makes someone submit to torture when she “consciously and culpably ‘participat[es]’ in” it. Twitter, 598 U. S., at 493; see also Restatement (Second) of Torts §876, Comment d (1977) (explaining that, under the common law, aiding and abetting requires proving the “assistance is a substantial factor in causing the resulting tort”).

The majority’s narrow interpretation of the word “subjects” is also incompatible with another form of secondary liability that all agree is covered by the TVPA: command responsibility. See Brief for Petitioners 41; Brief for United States as Amicus Curiae 31–32. Command responsibility holds superior officers vicariously liable for the actions of their subordinates. It generally requires showing that the superior “knew, or should have known,” that subordinates had or were about to commit torture and “failed to take all necessary and reasonable measures to prevent” it. Chavez v. Carranza, 559 F. 3d 486, 499 (CA6 2009); see Mohamad, 566 U. S., at 458 (approvingly citing Chavez). It requires showing neither “proximate cause” nor that the officer participated in the torture. See Chavez, 559 F. 3d, at 499. That the TVPA extends such attenuated liability to someone who did not participate in torture suggests that it also holds liable someone who culpably participates in torture, such as someone who provides tools to the torturer. Contra, ante, at 14, n. 4.

Finally, the majority’s reliance on Central Bank is misplaced. As discussed above, that case interpreted a statute imposing liability on someone who “ ‘directly or indirectly’ ” engaged in proscribed acts. 511 U. S., at 176. That language is narrower than the TVPA’s operative language that extends liability to anyone who “subjects” another to torture. Yet by extending Central Bank in this way, the majority effectively creates a magic-words test and “impose[s] a ‘clarity tax’ on Congress by demanding that it speak unequivocally if it wants to” impose aiding-and-abetting liability. Biden v. Nebraska, 600 U. S. 477, 508 (2023) (Barrett, J., concurring) (quoting J. Manning, Clear Statement Rules and the Constitution, 110 Colum. L. Rev. 399, 403 (2010)). Even in contexts where this Court requires a clear statement, like waivers of sovereign immunity, this Court has emphasized that “ ‘no magic words are required’ ” because the “fact that Congress chose to use certain language” in one statute “hardly means it was ‘foreclose[d] . . . from using different language to accomplish th[e] same goal.’ ” Department of Agriculture Rural Development Rural Housing Service v. Kirtz, 601 U. S. 42, 52 (2024); see Sebelius v. Auburn Regional Medical Center, 568 U. S. 145, 153 (2013) (emphasizing Congress need not “incant magic words” to make clear that a statutory limitation is jurisdictional). The majority offers no reason why magic words would be required in this context.

At bottom, the TVPA expressly authorizes victims of torture to sue any “individual who . . . subjects [them] to torture.” §1350 note. Because the plain text of this statute includes individuals who aid and abet the victim’s torture, I would affirm the Ninth Circuit’s judgment on this score as well.

V

The Court’s decision today is yet another notch in its belt, unabashedly remaking the law in its preferred image. The majority jettisons two decades of settled precedent and breathes new life into two decades of rejected legal theories. It does all this in the name of respecting Congress and the separation of powers. Yet in fixating on the separation of powers, the majority ignores the foundations of its own and shakes the public’s confidence in the stable and predictable development of the law. Because resolving this case did not require inflicting these wounds, I respectfully dissent.