Grief-Cycle Geopolitics

The question a desk asks first determines most of what follows. Asked what a state wants, an analyst produces a schedule of interests, ranks them by value, and models the counterparty as a machine for maximising the schedule. The method is tractable, teachable and largely correct about ordinary weeks. Its weakness is specific and it recurs: it under-predicts behaviour that costs the actor more than it returns.

That residue is not noise. A state that strikes a position it cannot hold, walks out of a forum it built, or declines a settlement materially better than one it accepted a decade earlier is not failing at arithmetic. It is doing something the arithmetic was never built to see. The working question here is therefore not what does this state want but what does this state feel, and the strategies that follow from the second question are not the strategies that follow from the first.

The arc, and what each stage looks like from outside

States acting from perceived betrayal, exclusion or humiliation do not behave randomly. They move through a sequence, and each position in it has an observable signature and a different available response.

Denial shows as public minimisation while the injury is privately registered. Communiqués describe the relationship as unchanged. Delegations are sent at accustomed level. Nothing is demanded, because demanding would concede that something happened. The response is quiet: let the narrative settle, and do not force an acknowledgement the other side is not yet able to receive.

Anger is where the cost is made legible. Rhetoric escalates, measures become punitive, and action is taken whose object is visibility rather than victory. The signature is a demand with a deadline attached and no described end state. The correct response is to shield without escalating, because retaliation re-energises the phase, and to market the restraint to every audience watching, because restraint that is not seen is read as weakness.

Bargaining announces itself in ambiguity: back-channel feelers, conditional ceasefires, signals that can be disowned. The state is testing whether a resolution exists that it can survive domestically. This window is narrow and it does not reopen on request. The off-ramp has to have been designed during the anger phase, when nobody wanted to hear about it.

Depression appears as withdrawal from international engagement while domestic pressure rises and internal fracture becomes visible in personnel and in budget. Measured economic leverage does more here than at any other point, and discreet support for the voices arguing internally for re-engagement does more than leverage.

Acceptance is pragmatic recalibration — the exploration of a new relational framework on terms the state can justify to itself. The architecture offered at this stage must preserve the dignity of both parties, and a settlement that does not is not an ending but a deferral.

The expensive misreading

The single most costly error in this field is to treat the anger phase as a negotiating failure. It is not a failure. It is a phase, and phases have durations that are not set by the quality of the offer in front of them.

An adversary in that position cannot be reasoned with, and the attempt itself does damage: it reads as an instruction to be calm, which is heard as a second dismissal by a party whose entire grievance is that it was dismissed. Firms and ministries lose years here, cycling through improved offers into a wall, then concluding the counterparty is irrational and reaching for pressure. Pressure applied in the anger phase does not shorten it. It extends it, because it supplies exactly the evidence the phase runs on.

Timing is not incidental to diplomacy. It is the discipline.

A second-order consequence follows for whoever does the work. Because the bargaining window is narrow and cannot be requested, the off-ramp has to be designed while the anger phase is still running — which is to say at the precise moment when proposing it looks like appeasement to every domestic audience on both sides. That is an uncomfortable place for a foreign ministry to stand, and it is the structural reason quiet intermediaries exist at all: a party with no public position of its own can hold a draft that neither principal can be seen to hold. Designing the off-ramp is not the difficult part. Carrying it without either side having to admit that the carrying is happening is the difficult part, and it is not a capability that can be assembled after the window opens.

A closed case

The record offers a settlement that ran the arc in the open. The Treaty of Versailles fixed responsibility in a text: Part VIII, Article 231, states that Germany accepts responsibility for causing all the loss and damage of a war imposed by its aggression. Whatever the drafting intent — the article was a legal predicate for the reparation claim in the articles that follow it — its operative function was different. It required one party to sign its own humiliation as a condition of peace, and to do so in a document it had no part in negotiating.

What followed is legible as denial and then anger: the settlement described publicly as no settlement at all, then a decade in which the demand was not for better terms but for the sentence to be unsaid. The instrument that eventually worked did not revisit the reparations arithmetic. The Locarno treaties of October 1925 produced a mutual guarantee of the western frontier negotiated between the parties rather than imposed on one of them, and Germany's admission to the League followed the next year. The material terms improved modestly. The relational terms — negotiated rather than dictated, guaranteed mutually rather than unilaterally — changed completely, which is the point.

Locarno did not hold. That is also the point, and any use of this case that stops before saying so is selling the framework rather than describing it.

What the framework is not

The five-stage structure is borrowed from a clinical model, and honesty requires saying that the model is contested in the discipline that produced it. The Yale Bereavement Study followed 233 bereaved individuals in Connecticut between 2000 and 2003 and tested the five indicators against the sequence stage theory predicts. It did not find it. Disbelief was not the dominant opening indicator; acceptance was the most frequently endorsed item throughout the twenty-four months; yearning, not anger, was the dominant negative indicator, and the negative indicators peaked at around six months rather than proceeding in orderly succession. The authors' own later framing was that these are states rather than stages.

So the framework borrows a structure, not a validated mechanism. It is not a schedule, it does not predict durations, and a state can occupy two positions at once or return to one it appeared to have left. What it supplies is a set of questions that a rational-actor model does not generate: what does this posture cost the actor, who is it addressed to, and what would have to be true for it to be worth paying.

The framework also carries a standing restriction in how it is applied. It is a reading of institutional behaviour from published conduct — communiqué language, delegation level, the presence or absence of a deadline, attendance and non-attendance. It is not a reading of any individual, and it produces no view on whether a government is entitled to hold the position it holds. A state that understands its counterpart's position holds an advantage; a state that mistakes that understanding for endorsement, or for a timetable, holds nothing.