“The strong do what they can and the weak suffer what they must.” (Thucydides)¹⁾ 

On September 3, 2026, lawyers for Nicolás Maduro filed a motion asking a federal judge in Manhattan to dismiss the narco-terrorism indictment against him, arguing that a sitting head of state enjoys absolute immunity from the criminal jurisdiction of any court but his own. The motion arrives eight months after US forces seized Maduro from the presidential palace in Caracas and delivered him to the same Manhattan courthouse. Commentators reached, understandably, for the vocabulary of regime change and great-power politics when covering the raid itself. Fewer noticed that the doctrinal groundwork for treating a foreign leader’s forcible removal as jurisdictionally unproblematic, immunity motions notwithstanding, had already been laid, more than three decades earlier, in a case about a Mexican physician most people have never heard of.

United States v. Alvarez-Machain, decided by the US Supreme Court in 1992, asked a narrow question: does the forcible abduction of a Mexican national from Mexican soil, carried out at the direction of American law enforcement, deprive an American court of jurisdiction to try him? The Court answered, six to three, that it does not. Read closely, the opinion is less a ruling about one abducted defendant than a template, the doctrinal terminus of a line running back to Ker v. Illinois (1886), and, as this year’s events suggest, a template still very much in use.

The Immunity Objection

Maduro’s lawyers are right that head-of-state immunity is, in the ordinary case, close to absolute: customary international law and the International Court of Justice’s Arrest Warrant decision (2002) recognize broad immunity from foreign criminal jurisdiction for incumbent heads of state, insulating them from the reach of any court but their own. If Maduro were still recognized by Washington as Venezuela’s president, that objection would likely be dispositive, and this piece would have little more to say.

He is not so recognized, at least not consistently. Secretary of State Marco Rubio has said publicly that he does not consider Maduro to be Venezuela’s legitimate president, and US policy toward Venezuela, including a recently announced arrangement over Venezuelan oil reserves, has since been built around Maduro’s former vice president, Delcy Rodríguez, as interim head of state. That matters because head-of-state immunity in US courts is not self-executing: it depends on a “suggestion of immunity” that only the executive branch can make, grounded in the executive’s own recognition of who holds the office. Deny recognition, and the predicate for the immunity claim disappears with it, whatever position the ousted government itself continues to assert.

That move has a precedent, and it is not Alvarez-Machain but United States v. Noriega (S.D. Fla. 1990, aff’d 11th Cir. 1997). Noriega was Panama’s de facto head of state when US forces seized him during the 1989 invasion; his head-of-state immunity claim failed because the United States had never recognized his government as Panama’s legitimate one. The court hearing Maduro’s immunity motion, with oral argument scheduled for November 17, has that template available to it, alongside Ker–Frisbie’s separate holding that the manner of capture is no bar to jurisdiction once a court has it. Immunity and forcible removal are, doctrinally, two independent gates, and Maduro’s lawyers must clear both. This piece is about the second gate, because it is the older, quieter, and in the long run more consequential of the two: the gate through which decades of prior defendants, invisible to headlines, have already been made to pass without ever holding an office capable of generating an immunity argument at all.

The Facts, Briefly

In 1985, DEA agent Enrique Camarena Salazar was abducted, tortured, and murdered in Guadalajara. The resulting investigation implicated Humberto Álvarez-Machain, a physician American investigators believed had helped keep Camarena alive so the torture could continue. When Mexico declined to extradite him, the DEA arranged, through paid Mexican bounty hunters, for Álvarez-Machain to be seized from his medical office and flown to El Paso.

This was not improvisation. The United States had pursued extradition through ordinary channels first. The abduction was a deliberate substitute for a procedure that had already failed. The question the Supreme Court eventually confronted was not whether the US wanted custody, everyone agreed it did, but what legal consequences attach to the method chosen when lawful channels prove too slow.

Treaty Silence as Permission

Writing for the majority, Chief Justice Rehnquist held that because the 1978 US-Mexico extradition treaty did not expressly prohibit forcible abduction as an alternative to its own procedures, the treaty could not be read to bar prosecution of a defendant obtained that way. Silence became permission.

This inverts the ordinary logic of treaty construction. A bilateral extradition treaty is, by definition, a comprehensive substitute for the unilateral use of force to obtain custody of a fugitive, that is its entire function. Article 31 of the Vienna Convention on the Law of Treaties, which the majority never seriously engages, instructs that a treaty be read in light of its object and purpose. It is difficult to square a reading that permits either party to abduct at will, whenever formal procedures move too slowly, with the object and purpose of an instrument meant to be the exclusive lawful mechanism for interstate rendition. Justice Stevens made exactly this argument in dissent, warning that the majority’s approach would let the United States seize a defendant from any of the roughly hundred countries with which it holds extradition treaties, so long as none of those treaties happens to contain the word abduction.

The majority’s method works by separating three planes that a more integrated approach would keep joined: the sovereign violation on the international plane (the seizure itself), the question of treaty breach on the interstate plane, and the question of domestic jurisdiction on the criminal-procedural plane. The Court insulates the third from the first. Even conceding that Mexico’s territorial sovereignty was violated, jurisdiction survives unless the treaty explicitly says otherwise. The violation is acknowledged, even regretted, and then set aside as belonging to a different, non-justiciable conversation.

A Rule of Non-Attribution, Expanded

The majority’s authority for this move is Ker v. Illinois and Frisbie v. Collins, together known as the Ker–Frisbie doctrine: the legality of a prosecution does not depend on the legality of the process that brought the defendant to court. But there is a distinction the majority glosses over. In Ker, a private bounty hunter acted on his own initiative, without government instruction; the state’s involvement was, at most, retrospective ratification. In Alvarez-Machain, the abduction was planned, funded, and directed by DEA officials from its inception. As Michael Glennon has argued, Ker stands, at most, for a rule of non-attribution, a state need not release a defendant delivered by an unauthorized private actor, not for a rule authorizing state-directed abduction. Extending it to cover direct government sponsorship converts a doctrine of judicial restraint into a structural incentive: because the manner of capture carries no jurisdictional cost, the state bears none for choosing the most coercive available method.

Dualism, Operationalized

For a European audience, the most useful way into Alvarez-Machain may not be through American criminal procedure at all, but through the monism/dualism debate that has structured continental engagement with international law since Kelsen. The majority does not deny that Mexico’s sovereignty was violated. It concedes, almost in passing, that the abduction “may be in violation of general international law principles” language that maps directly onto the state-responsibility framework later codified in the Restatement (Third) of Foreign Relations Law § 432. What the Court refuses is any automatic transmission of that violation into the domestic legal order. Diplomatic protest, yes. A jurisdictional bar in an American courtroom, no.

That is dualism in its purest operational form: international law’s prohibition on non-consensual extraterritorial enforcement generates consequences that stay confined to the international plane, unless the domestic legal order chooses, on its own terms, to import them. A monist court, treating the two systems as integrated, could not have reached the majority’s result without confronting the international wrong directly. A dualist court can hold the wrong and the jurisdiction apart. It is precisely that holding-apart that lets coercive jurisdiction function as a system rather than a one-off transgression, the wrong is real, but it is externalized, converted into a grievance between states that leaves the domestic proceeding untouched.

It is worth asking, as one reasonably might, whether any other legal system reaches this same result by dualist reasoning. The honest answer is: not among comparably situated common-law jurisdictions. The House of Lords, confronting a materially similar fact pattern in R v Horseferry Road Magistrates’ Court, ex parte Bennett (1994), stayed the prosecution of a defendant returned to the UK through a process the government had been complicit in circumventing, holding that a court’s inherent power to prevent abuse of its own process can bar jurisdiction regardless of what the underlying treaty says. South Africa’s Appellate Division reached the same result in S v Ebrahim (1991), holding that a state-sponsored abduction from a neighboring country deprived the South African courts of jurisdiction to try the returned defendant. Neither ruling is framed in the vocabulary of monism and dualism; both rest on abuse-of-process and rule-of-law reasoning that a dualist system is entirely free to adopt, and in these two instances did. Ker–Frisbie’s insulation of jurisdiction from the legality of capture is accordingly not what dualism as such requires. It is a specifically American doctrinal choice, one that other non-monist, common-law systems asking the identical question have been unwilling to make.

Eichmann, and the Limits of the Comparison

The obvious comparator is Israel’s 1960 seizure of Adolf Eichmann from Argentina, which produced the same jurisdictional result on structurally identical reasoning: illegality of capture did not deprive the court of jurisdiction to try him. Hannah Arendt, reporting on the trial, treated the manner of Eichmann’s capture as secondary to the gravity of what he had done, while registering unease at the precedent a state-sponsored extraterritorial seizure might set for states with less compelling justification.

The comparison is instructive precisely because of its asymmetry. Eichmann’s case is the most defensible version of extraterritorial abduction imaginable, a state seizing the chief administrator of genocide to answer for crimes against humanity. Álvarez-Machain’s case involves the ordinary machinery of a domestic narcotics prosecution. If an identical jurisdictional doctrine covers both, the doctrine itself is indifferent to the moral weight of the underlying offense, its logic is formal, not substantive, which is exactly what allows it to travel to cases far removed from anything resembling universal jurisdiction over atrocity.

A narrower domestic counter-current exists: in United States v. Toscanino, the Second Circuit held that where a defendant’s presence is procured through conduct that “shocks the conscience,” there, allegations of prolonged torture, due process may require dismissal, notwithstanding Ker and Frisbie. Toscanino has never been overruled. It has simply never again been successfully invoked: a safety valve calibrated, in practice, to remain closed.

The Civil Sequel, and the Closed Loop

Álvarez-Machain was acquitted at trial, the evidence proved too thin. He then sued the United States and the Mexican national who had physically carried out his seizure. The case returned to the Supreme Court as Sosa v. Alvarez-Machain in 2004, which rejected both claims. The Federal Tort Claims Act’s foreign-country exception barred the suit against the government, since the abduction was deemed to have occurred in Mexico for that purpose. Interestingly, the same conduct that generated American criminal jurisdiction in 1992 because it occurred, in relevant part, through American direction, now generated a jurisdictional bar against the plaintiff because it occurred, in relevant part, on Mexican soil. The Alien Tort Statute claim against the individual defendant failed too, on the ground that a single wrongful detention of under a day did not rise to the level of a norm with the definiteness and universal acceptance required to support an ATS claim.

Read together, the two decisions form a closed loop. The victim of an internationally wrongful act was left with his own government’s diplomatic protest and a never-ratified treaty protocol as the international legal order’s entire response, and with nothing at all domestically, acquitted of the crime for which he was seized, uncompensated for the seizure itself. That is not incoherence. It is a system in which jurisdiction is easy to obtain, and consequence is almost impossible to attach to the manner of obtaining it.

Why This Still Matters

Alvarez-Machain is not merely a period piece about one Mexican physician’s ordeal in 1990, nor only a link in the Ker-to-Frisbie chain. It is the moment the American legal system articulated, in the driest possible language of treaty construction, what an expansive theory of extraterritorial enforcement requires: that jurisdictional reach need not be limited by the territorial sovereignty it crosses, so long as the domestic court asked to exercise that jurisdiction is willing to treat the violation of another state’s sovereignty as someone else’s problem. Whatever else changes between Washington and the states whose nationals it wants, the diplomatic relationship, the technology of capture, the political circumstances, that 1992 backstop persists, undisturbed, available for reactivation whenever formal channels prove insufficient. For a European legal culture more instinctively monist, and more accustomed to treaties functioning as genuinely exclusive channels, the case remains a clear specimen of what dualism can be made to do when a court wants it to do a great deal.