On 5 February 2026, the last treaty limiting the strategic nuclear arsenals of the United States and Russia expired. Since 1972 such a limit had always existed. It exists no longer.

Both sides have stated that they intend to observe the old numbers for the time being. Nobody signed anything. There is no inspection, no deadline, no document anyone could invoke. The intention remained, the binding disappeared.

I consider this the clearest event our time has to offer for anyone trying to understand why there is no world community. No enemy tore down an order. Two governments let an obligation lapse, because they could and because nobody was in a position to stop them.

That possibility is the real subject of this brief. It is not the failing of one particular government. It is built into every form of rule we know, including the ones we consider better.

What the Four Orders Can Do

Before I name the weakness, the strength has to stand, otherwise analysis turns into indictment.

Liberal democracy is the only order that can correct its own mistakes without violence. Those who govern can be voted out and nobody dies. In the history of political orders this is a rare invention, and it is routinely underrated because its slowness is mistaken for weakness. That slowness is the price of being correctable.

Democracy with a developed welfare state can absorb distribution shocks. It holds a society together when markets turn. This is more than social generosity. A society that comes apart internally cannot promise anything outward that outlasts the next crisis.

Autocratic order decides quickly. It can carry a project across twenty years without justifying it anew every four. Anyone who has watched an infrastructure project ground down by shifting majorities knows this is not an empty argument.

Even closed rule can do something democracies find hard. It can commit for decades without an election intervening.

The numbers show how unevenly these forms are distributed. At the end of 2025, research counted 92 autocracies and 87 democracies. Among them 31 liberal democracies and 56 electoral ones, alongside 57 electoral and 35 closed autocracies. Roughly three quarters of humanity, about six billion people, live in autocracies. Seven percent live in liberal democracies.

Whoever Can Commit Cannot Be Held

Each of these strengths has a reverse side, and the reverse sides fit together in a troubling way.

Democracy is accountable. That is exactly why its clock ends with the election. A government cannot promise anything that binds its successor, and it should not be able to, because otherwise the election would mean nothing. Correctability and the capacity to commit sit in the same mechanism and exclude one another.

Autocracy can commit. But nobody can hold it to the commitment. There is no internal body that compels the ruler to honour his own word. And what a closed order signs does not outlive the ruler, because the only body that gave the assurance disappears with him.

That produces the finding I am after. The capacity to commit and the capacity to be held accountable are inversely distributed. Those who are accountable cannot commit. Those who can commit are not accountable.

No combination of these forms yields an assurance one could build on across generations. This is why the world order consists of declarations of intent rather than obligations. Not because the participants are dishonest. Because none of the available designs yields anything else.

Self-Release

I call what happens here self-release. A state binds itself and keeps the release of that binding in its own hand.

You can see it in the shape international commitments now take. The ratified treaty has become rare. In its place stand declarations, frameworks, accords and joint statements. This is not a fashion in diplomatic language. Ratification was the binding part, because it involves a parliament and makes an exit visible. Avoiding it avoids the binding and keeps the effect.

You can also see it in the machinery meant to administer those bindings. The regular United Nations budget for 2026 stands at 3.24 billion dollars, 15.1 percent below the previous year. Posts fall by 18.8 percent, with 2,681 positions eliminated. Part of that is overdue streamlining. Another part is the result of unpaid contributions, and not paying is the quietest way to release oneself.

Self-release is not malice. It is the rational act of a party that knows nobody will hold it. As long as this is true for everyone, it would be unwise to be the only one who stays.

The Layer That Runs Through All Four

There is a fifth way to rule, and it appears in no classification of state forms. It consists of running an operation others depend on.

Whoever operates a constellation a country communicates through decides on that country's connections. Whoever holds the compute on which models are trained decides what can be built. Whoever owns the cables the traffic runs through decides on the traffic. None of this is a form of state, and all of it is rule.

This layer does not replace the four orders. It runs across them, in four variants.

In the United States and Europe it is hosted. The state could intervene and does not, because it depends on the operator itself. In Russia it is commissioned. The state orders the function instead of permitting it, and keeps its grip. In China it is fused. Party and operator coincide, with reach and capability at once.

The fourth variant covers most countries on Earth. They are exposed. They can neither build the function nor commission it nor regulate it. They can only use it. Some 150 governments hold no position at all in this layer, regardless of whether they are democratic or autocratic.

The Russian case reveals the underlying rule. Since 2020 a Moscow company has been building its own constellation. In March 2026 the first sixteen production satellites launched, with a second batch following in the summer. The plan calls for 250 satellites by 2027 and over 900 by 2035, funded with roughly 5.7 billion dollars. A state with the strongest available means of coercion against its own entrepreneurs needs a decade and billions to replace a function a private provider supplies over its own war zone.

From this follows a sentence that applies to every country. Whoever wants to pull the plug must be able to replace the function. Anything else is switching off your own line.

The Dependency That Does Not Bind

At this point a mistake suggests itself, and I made it myself for a long time. One might think mutual dependency substitutes for missing treaties. Those who depend on each other, the hope runs, do not attack each other.

This holds for one kind of dependency and not at all for another.

A shared reference does not shrink when one more party uses it. A common reckoning of time, a common orbital catalogue, a common protocol, a common language. Leaving it damages you as much as everyone else. Such dependencies genuinely bind, and they bind both sides equally.

A chokepoint in one hand works the other way round. A single water connection for many houses is not something held in common, it is a lever. Whoever holds it becomes freer. Whoever hangs on it becomes less free.

European gas before 2022 was such a connection. It did not bind, it made the buyer vulnerable to pressure. And the lever is precisely the tool by which a self-release is carried out at no cost.

China is demonstrating the same pattern more precisely. About ninety percent of global rare earth processing sits there, against roughly a third of the deposits. The chokepoint was built, it is not geological. And it is operated not with bans but with permits that can be delayed. Approval rates for European firms have fallen below a quarter in some areas, and prices rose at times to six times their level. There is no breach of contract anyone could challenge.

It does not run in one direction, however. The West holds chokepoints of its own, in chip manufacturing equipment, in advanced fabrication stages, in engines, in payment clearing. What actually exists is mutual hostage-holding. It produces restraint, but the restraint of a standoff rather than of a rule. It holds as long as both sides believe they would lose the exchange.

And it destroys itself. The sensible response to such a standoff is to escape it. That is exactly what is under way. Europe intends to hold forty percent of processing within its own space by 2030 and to draw no more than sixty-five percent from any single third country. Others pursue the same goal by other means.

So both stabilisers are falling away in the same decade. The treaties are lapsing, and the entanglement that has absorbed their absence is being dismantled by plan.

What Can Be Built

Four things can be done, and I order them by what is possible now and what takes decades.

First, a public record of self-releases. Every treaty contains its withdrawal clause, yet nobody keeps account of who released themselves from what and when, and what took its place. Today each such event vanishes within a week of news. A dated, comparable record turns it into a balance sheet across twenty years. It requires nobody's consent.

Then the construction of shared references instead of new treaties. They bind only under two conditions. The rules must be open, and the operator must be replaceable. A dependency with both properties binds without ruling. A constellation in private ownership meets neither.

This also resolves the duplication problem. Nobody needs to be forbidden from building their own infrastructure. But the orbital data, the time reference, the conjunction warning and the ground measurements must be the same, otherwise the duplicate does not work. The reference is the one thing that cannot sensibly be duplicated.

Third, substitution capacity, held jointly. No medium-sized country reaches it alone. In August 2026 the European Union agreed the implementation of its own constellation, 348 satellites, first launches in 2029, with four billion euros of private participation. What matters is what happens with that capacity. A credible second source that stands ready and remains unused turns coercion into negotiation. A second source that goes fully into operation ends the relationship. Europe's error with gas was not the dependency, it was dependency without an option.

Finally, a short list of crossings nobody can take back alone. Each of them needs a measurable definition, a body that measures, and a named consequence for whoever crosses. The work lies in the definition, not in the penalty. For most of these thresholds nobody today can say where exactly they lie, and therefore nobody can establish that they have been crossed.

Tasks Instead of Conviction

None of this yet amounts to a world community. It prevents damage, it does not build belonging.

Large communities have so far emerged along four paths. By conquest, which produced an empire every time and never a community. Through a common enemy, which is unavailable for the world as a whole, because there is no outside. After a catastrophe that struck everyone at once, and this path worked in 1945, at the price of what preceded it.

The fourth path is the shared task, and it is the only one that works without war. The Universal Postal Union of 1874 carried letters across borders long before states were politically close. The Telegraph Union of 1865 did the same for messages. The joint administration of coal and steel became Europe. Nobody resolved on a European conviction first.

Two examples are running now. On 17 January 2026 the High Seas Treaty entered into force, covering ocean areas beyond national jurisdiction, a substantial share of the planet's surface, with 89 parties to date. Nearly two decades of work for a single shared object.

The second example teaches more. The treaty banning nuclear tests never entered into force. Its monitoring network runs regardless, with over three hundred stations worldwide, and its data are used by states that never ratified. The measurement outlived the treaty. That is the most practical piece of information this subject has to offer: a rule requires consent, a measurement does not.

From this follows the order of operations. You begin with objects rather than values, because values divide and objects do not. You measure before you regulate, because whoever measures first determines what is later argued about. And you create belonging through participation rather than declaration, since people belong to what they work on.

That leaves the timing. Every large world institution was founded after a catastrophe, because the willingness came only then. That willingness will return, and it will return suddenly. The point is that something finished should be standing at that moment, something that can be reached for.

A world community does not mean agreement. In every country millions of people argue about nearly everything and remain a community. What holds them together is not consensus but the common frame within which the argument takes place. So the goal is not that the peoples of the world come to share one view. The goal is that their argument acquires a common frame.

Which task could you share with someone whose order you reject, without either of you having to give ground?

Website: https://planet-futures.org