There is a joke that left-wing intellectuals are fond of. What is international law? Something that powerful states obey when they wish to, and disregard when they do not. Every time the United States bypasses the United Nations to launch a war, the joke is trotted out once more, its bearer wearing the satisfied air of one who has seen through to the truth of the world.
In 2023, Perry Anderson published a long essay in NLR that promoted the joke into an academic theory. His verdict was crisp: international law is "neither particularly international, nor truly law." The UN's principle of sovereign equality is a veil, and beneath the veil lies a thoroughgoing hierarchy. The International Court of Justice has no enforcement power; the International Criminal Court tries only weak states. From Suez to Vietnam to Iraq — when has the United States ever taken international law seriously?
Anderson is right — about the part he describes. Koskenniemi's response in NLR 154 opens by graciously conceding this much. Then he turns: everything you critique does indeed exist, but you have seen only the tip of the iceberg above the waterline, and your gaze has never dropped below.
What lies beneath the water? An entire legal infrastructure that you live within every day and almost never notice. It is not weak; on the contrary, it is so powerful as to be omnipresent. It is not occasionally violated — it is domination itself.
Law Is Not a Set of Rules; It Is a Language
To extract the real insight from Koskenniemi, one must first discard a deeply rooted misconception: that law is a set of rules, either obeyed or violated.
It is not. Law, Koskenniemi says, is a language — a language that translates raw reality into the binary code of "legal" and "illegal." The vocabulary of this translation consists of "rights," "obligations," "competences," "privileges."
The formulation sounds abstract, but the moment you press it against reality, it raises goosebumps.
A group of people gathers outside a heavily fortified border. In the physical sense, they are simply a group of people standing there. Then the language of law goes to work — they become "asylum seekers," attempting to cross an "international boundary" and enter another "sovereign state," to apply for something called "the right to asylum." Once this translation is complete, when you look at the group again you no longer see people; you see an assemblage of legal categories. Whether they are "legal" or "illegal" depends on whether the forms they filled out, the procedures they followed, and the stories they told conform to a set of standards they had no part in drafting.
The border itself is a creature of law. Without law, there is no "border" — only a line. It is law that turns that line into a cutting surface of power, severing those who "may remain here" from those who "may not."
Consider another case. A group of well-dressed people meets in a building to discuss economic sanctions against a certain country. In the physical sense, they are people speaking in a room. After legal translation, they become "representatives" of an entity called "the European Union," exercising a power called "decision-making," directed at a legal person called "Russia." The violence of the entire process — sanctions will make people go hungry, go without medicine, die — is filtered out by this layer of legal language, leaving behind a clean, neutral, procedurally proper technical decision.
The power of law resides in the objective and neutral meaning it carries. When something is treated in legal terms, the history of struggle and violence that Foucault spoke of — the origin of law in real battles, victories, massacres, and conquests — is suspended, forgotten. The reality of law is what it is, and everyone must obey.
This is Koskenniemi's deepest insight. The power of law lies not in coercion but in making what it prescribes appear as the natural order. It wipes the traces of power clean, leaving you with the feeling that the world has always worked this way.
"Deregulation" Is a Grammatical Error
If you have grasped the point above, then the next argument Koskenniemi tosses off feels like a gift — but it is brilliant enough to be savored on its own.
He says: in the strict sense, there is no such thing as "deregulation."
Every time you remove a regulatory rule, you are not creating a "free blank space." You are establishing a new rule, authorizing someone to do what was previously forbidden. Abolishing financial regulation is not "letting the law go away"; it is vesting finance capital with a new sphere of action through a fresh legal framework. "Deregulation" and "re-regulation" are in fact the same thing, distinguished only by their propaganda.
Every removal of a regulatory obligation is accompanied by the establishment of a rule that authorizes someone to do what was previously forbidden.
The beauty of this argument lies in its extreme simplicity and its irrefutability. You cannot produce a single example of "deregulation" that cannot be redescribed as "laying down a new rule for the other side." Relaxing the regulation of banks = granting banks new freedoms. Weakening the bargaining position of unions = granting employers new privileges. Law never disappears; it merely changes whom it serves.
Anderson Stares at the Front Desk; the Real Domination Is in the Back Office
Back to Anderson. The things he critiques — the UN, the International Court of Justice, multilateral treaties — are all "public, diplomatic law." They are indeed often feeble, and indeed frequently manipulated by the powerful. The left's criticism of them is like complaining that a restaurant's front desk service is poor.
Koskenniemi invites you to walk past the front desk and into the kitchen.
What runs in the kitchen is a different legal machinery: trade law, investment law, debt-management law, intellectual-property law, financial regulation. These laws are not weak. An arbitral tribunal of the International Centre for Settlement of Investment Disputes (ICSID) can order a sovereign state to pay a transnational corporation billions of dollars, and enforce that order. The WTO's dispute settlement mechanism may be paralyzed, but private contract arbitration operates efficiently — and in total secrecy. Sovereign debt restructuring has no proper legal framework? No matter: the "collective action clauses" embedded in bond contracts will do the job for you — drafted, needless to say, by the creditors' lawyers.
Koskenniemi cites the research of Katharina Pistor to drive the point home: capital is not a natural phenomenon; it is a creature of law. A building, a plot of land, a patent — in the physical sense, they are simply themselves. It is law that endows them with "priority" (who gets paid first in bankruptcy), "enforceability" (capable of being protected by courts), "transferability" (capable of being traded across jurisdictions), "universality" (binding on all), and "convertibility" (capable of being transformed into other forms) — it is these five legal attributes that turn a physical existence into "capital."
Public international law is supplemented, and at times overridden, by a wide-ranging network of unequally distributed private rights, powers, and privileges.
So when you stare at the UN Security Council and sigh that "international law is useless," a transnational corporation is, through an ISDS tribunal, siphoning hundreds of millions of dollars of public funds out of some developing country — pursuant to an investment treaty signed twenty years ago, drafted by a Washington law firm, adjudicated in Washington, under Washington law. This process requires no Security Council approval and no "public" procedure of any kind. It is quiet, efficient, legal. It is the real domination.
The Brussels Effect: Even Law Is a Kind of Export
Koskenniemi is not yet satisfied with exposing the public-law/private-law swindle; he goes on to demonstrate how law reproduces itself.
Take the European Union. Brussels sets a high standard of regulatory rules — data protection, product safety, environmental standards. Transnational corporations that wish to enter the EU market must comply. And having already invested in compliance, they simply apply the EU standard as a global standard — maintaining two separate systems would cost more. Thus companies in non-EU countries, having never voted for a single Brussels official, are forced to accept the EU's legal order.
Koskenniemi's tone here is almost a bitter humor:
When such "regulation" is produced by powerful actors such as the United States or the EU — the latter notorious for its "Brussels effect" — it is hard not to see neo-colonial implications.
The term "neo-colonialism" once referred to gunboats and viceroys. Now it refers to a compliance document you must sign in order to do business. Colonization no longer needs to show up in person — the market does it for you.
What role does the state play in this system? Koskenniemi's diagnosis is cold and precise: the state has been reduced to "the local manager of global processes." Sovereignty remains, the flag still flies, parliament still convenes — but the real rules are made elsewhere: in the technical committees of the ISO, in the expert meetings of the OECD, in the "Ease of Doing Business" rankings of the World Bank. These standards "have no formal binding force," yet they are tied to loan conditions, market access, and international reputation — and who would dare not comply?
The Discourse of Rights: It Hands You the Vocabulary, and Takes Away the Weapon
One of the most brilliant passages in the entire essay is the dissection of "rights culture." This section will make many liberal readers uncomfortable.
Koskenniemi does not deny that the discourse of rights has a progressive history — in the 1970s, it was a weapon against authoritarian governments. But what follows is a less pleasing story. As rights discourse spread across the globe, it translated every social conflict into a "conflict of rights." Freedom of speech against hate speech, the right to security against the right to privacy, property rights against the right to subsistence. Once translated, the conflict was dispatched to the courts.
This is precisely the problem. What a court can give you is only "a legally available remedy." It will not give you structural change. When a social movement concerning distributive justice is translated into a set of rights claims, it is funneled into a bureaucratic channel, and the only possible output is a judgment. The judgment may rule in your favor, but what it wins is a specific, limited remedy — not any fundamental change to "the social conditions that produced these rights violations."
More cunningly, the meaning of a right is itself indeterminate. It depends on how legal institutions "balance" — and the standard of balance is often nothing other than the priorities and prejudices of those institutions. Rights first took the stage as a "trumping" force (Dworkin said rights can "override" the preferences of the majority), but gradually they became just another administrative vocabulary for managing social conflict, subordinated to the very institutions they were meant to constrain.
Rights advocacy channels social conflict into bureaucratic avenues such as the courts, with outcomes confined to legally available remedies, while excluding large-scale transformation of the social conditions that give rise to the gravest rights violations.
This is not to say rights are useless. It is to say that when the only tool in your hand is the language of rights, the scope of what you can demand is already circumscribed. You are permitted to be angry — but only in the ways the law permits.
[Editor's Note] At this point, Koskenniemi's analysis has touched a political judgment deeper than jurisprudence: establishment reformism is destined to fall into a bureaucratic trap. The thing this precision-engineered legal system does best is not resolve conflicts but dismantle them — breaking an irreconcilable confrontation over who should own what into innumerable tiny cases that can be litigated, mediated, "balanced." Each case offers you a verdict; added together, they change nothing. The revolutionary fervor of a roaring fire is thus dissolved into the helpless resignation of a frog in slowly warming water. You think you are fighting for your rights through the law; in truth, you are merely helping the machine prove that it still runs.
The Legal Empire: No Emperor, but Omnipresent
Finally, Koskenniemi borrows the concept of "Empire" from Hardt and Negri and gives this system a name: the Legal Empire.
It is not a power radiating outward from a single center; there is no emperor, no command tower. It is a hierarchical network in which power is embedded in expertise — trade-law experts push for more trade, environmental experts push for more protection, security experts push for more surveillance, human-rights experts hope for less surveillance. Each expert sincerely believes that their own project ought to become a global project. They are not in collusion; they often quarrel with one another. What they share is this: they all speak the same language (law), and they all point in the same direction (more legalization, more technocratic governance, less political contestation).
The most formidable feature of this empire is that compliance with it is non-optional for any single actor. You cannot opt out of the global trade-law system any more than you can opt out of gravity. You may oppose any particular rule, but the only way to oppose it is to invoke another rule. The boundary of the Legal Empire is the boundary of imaginable political action.
Trump's Paradox
The essay closes on Trump — an unexpected yet apt landing point.
Looking back from 2025, the golden age of international law in the 1990s seems to be drawing to a close. The WTO is paralyzed, the assembly line of multilateral treaties has stopped, the United States is attacking and ignoring international institutions. It looks as though the "Legal Empire" is coming apart.
Koskenniemi says: not so fast.
Yes, it is still there. Everything that has carried Trump into the presidency was initiated, coordinated, and implemented by law.
Trump opposes "the international rule of law," but he does not oppose the rule of law. Every one of his actions — executive orders, travel bans, tariffs — sits within the American legal framework, will be reviewed by courts, and is implemented through legal procedure. What he attacks is a particular kind of law (public international law, human-rights treaties); what he relies upon is another particular kind of law (property law, contract law, immigration law). The latter is precisely the core of the "legal infrastructure" that Koskenniemi's entire essay has been arguing for.
And then comes the most subversive line of the whole piece:
Authoritarians have their rule of law too. You may have every reason to abhor it, but to deny it the status of "law" would be an analytical and political error.
This sentence will sting many people. For liberals have long treated "the rule of law" as their exclusive badge, as though the rule of law were naturally equivalent to the protection of human rights and the checking of power. Koskenniemi says: no. The rule of law is simply a particular kind of legal order, serving particular distributional outcomes. The authoritarian's rule of law protects property rights, enforces contracts, maintains order — it simply does not protect the rights of the disobedient. Both are "law"; they simply serve different masters.
So the real question was never "is there the rule of law or not." The question is the one in Koskenniemi's final line:
Who must submit to whose legal reality?
This is not a rhetorical question. It has a concrete answer — written into the clauses of your employment contract, into the user agreement you click "I agree" to every day, into the jurisdiction clauses of the venue where your country's debt contracts are signed. The answer is not in the halls of the United Nations, nor in the trial records of the International Court of Justice. It lies between the lines of texts you have never read, yet are bound by every day.