The One Move That Requires No Concession

In an escalation the instruments that get discussed are the ones that can be seen. Force, sanction, the formal demarche, the resolution tabled and vetoed. Each is legible, each can be reported as an action taken, and each is priced by every audience watching. The instrument with the highest leverage in a particular and common configuration is none of these, and its defining property is that it cannot be seen at all.

The configuration is betrayal. Not defeat, not encirclement, not material loss — betrayal, which is a different injury with a different remedy, and which conventional analysis persistently misreads because the demands it generates are stated in the vocabulary of the injuries it is not.

The model

Where a conflict is rooted in perceived alliance betrayal, the useful diagnostic is relational rather than strategic, and the parallel to intimate betrayal is not decoration. It predicts behaviour with more precision than the interest calculation does.

It begins with an exclusivity that was never written down. A state believed that non-alignment implied loyalty to a shared regional or ideological position — an assumption deeply held and nowhere codified, which is precisely why its violation cannot be litigated and must instead be felt. Then comes the discovery: a deepening arrangement elsewhere, normalisation, military cooperation, overflight rights, something that makes the prior assumption impossible to keep holding.

What follows is the behaviour that rational-actor models cannot place. Action designed not to win but to register the cost, with visibility as the objective rather than victory. Then the conditional gesture: a ceasefire whose shelf life runs in days rather than months, because the condition underneath it is unresolved. Then conditional forgiveness, explicitly or implicitly contingent on the cessation of the activating conduct — the logistical facilitation, the overflight, the joint exercise — that constitutes the betrayal in the aggrieved party's framework rather than in anybody else's.

Read that sequence as a series of demands for compensation and none of it computes; the compensation sought is never proportionate to the loss claimed, and improving the offer does not close it. Read it as a series of demands for acknowledgement and it resolves completely. The injury in a betrayed alliance is rarely the material loss. It is the discovery that the relationship was not what it was understood to be, and the party that made that discovery wants it witnessed.

The move, and why it has to be invisible

What structurally reduces an aggrieved state as a continuing threat is not a concession. It is an acknowledgement: a statement, delivered credibly, that the exclusion is seen, that the other party is not an enemy, that it cannot control every act of its partners, and that a framework in which the aggrieved state's voice in regional security is formally recognised is available for discussion.

Notice what that costs. It concedes no position, abandons no alliance, transfers no asset and admits no fault. It offers the one thing no weapons system can secure, which is being heard.

Notice also why it cannot be said out loud. A public acknowledgement is not received as acknowledgement; it is received as a concession, and priced as one immediately — by the domestic audience of the party giving it, by the partners whose arrangements prompted the grievance, and by the aggrieved party itself, which now holds a public admission and is obliged to seek more. The identical words, delivered privately, cost nothing anyone can tally. This is not a presentational nicety. It is the whole mechanism: the acknowledgement works because it is deniable, and it stops working the moment it is not.

That produces the structural argument for private diplomacy as a category. An acknowledgement is credible only if carried by a party whose own position is not at stake in the carrying. A state intermediary has interests that make it a participant. An institution with a public mandate has to report. What the transaction requires is a party with no public position of its own, able to hold a form of words that neither principal can be seen to hold, and to establish that both would accept it before either has committed to anything.

The form is the argument

The clearest instance in the record is visible in the shape of the instrument rather than in its contents. The Algiers Accords of 19 January 1981 open by reciting that the Government of Algeria, having been requested by the Governments of the Islamic Republic of Iran and the United States of America to serve as an intermediary in seeking a mutually acceptable resolution of the crisis, had consulted extensively with both.

The document is not a treaty between the two states. It is a declaration by the Government of Algeria, declaring that interdependent commitments have been made by the two governments on the basis of formal adherences received from each. Neither principal signed anything with the other. Each adhered to a text held by a third party.

That architecture was not ceremonial. It was adopted because the parties could not settle mutually agreeable terms directly, and it allowed each to undertake obligations independently rather than bilaterally. It is also what made Point I possible — a pledge that it is and from now on will be the policy of the United States not to intervene, directly or indirectly, politically or militarily, in Iran's internal affairs. As a bilateral undertaking given to an adversary, that sentence was unsayable. As an adherence to a declaration held in Algiers, it was sayable, and it was said.

The failure mode

An acknowledgement that leaks is worse than no acknowledgement at all, and the reason is not embarrassment. A private acknowledgement is an instrument of relationship. A public one is a document, and documents get scored. Point I has been argued over ever since as a question of compliance, which is what a written acknowledgement becomes once it exists in a form that can be produced and checked. The gain it purchased was real. The permanent audit it created was the price, and it was paid by the party that gave it.

There is a second failure mode, quieter and more common: an acknowledgement offered to the wrong phase. Delivered while the aggrieved party is still making the cost legible, it is not received as acknowledgement but as confirmation that the pressure is working, and what it purchases is escalation.

None of this is a capability that can be assembled at the point of need. The credibility that allows a form of words to be carried between two parties who will not speak to each other is built across years in which nothing is being carried at all, and it can be destroyed by a single disclosure. That is an uncomfortable economic proposition, and it is the correct one.