We picture a truce as total. But you can also choose to silence the guns in a single domain (the sky, the sea, energy) while the rest of the conflict continues. This selective de-escalation looks like a neutral gesture. It never is : the domain you freeze favours one side, its definition becomes the stake of the negotiation, and its enforcement rests on a verification that is almost always missing. This notion takes the mechanism apart.
1 The definition
Reducing violence in a single space.
The definition
Putting out one fire without putting out the others
De-escalation by domain means lowering violence in a single space of the conflict (the air, the sea, energy, one type of weapon) rather than demanding a global ceasefire. You freeze one domain and let the others run their course. The appeal is speed : civilians or infrastructure can be spared without waiting for a comprehensive deal, and without forcing either side to recognise the front line as a border. But reducing violence in one place does not stop it everywhere : the conflict does not vanish, it moves to the domain left open.
2 The perimeter
Defining the frozen domain is already negotiating.
The hidden stake
The boundary of the truce is the real text of the deal
Announcing that you freeze “the air” is not enough : you still have to say what that covers. Drones only, or missiles too ? Any strike, or only those aimed at cities ? Out to what distance from the front line ? Each dividing line strips a capability from one side and not the other. Negotiating the perimeter of a partial truce is therefore not settling a technicality : it is deciding, weapon by weapon, who keeps what. The perimeter, more than the duration, is the heart of the deal.
3 The asymmetry
Each side dominates a different space.
The imbalance
Freezing a domain favours whoever dominates elsewhere
A partial truce is never symmetrical, because the belligerents are not strong in the same domains. Freezing the space where one side dominates strips it of an advantage ; freezing the one where it is weak relieves it. So the same truce does not cost both the same. Suspending one domain amounts to freezing the balance of power in the others, in favour of whoever is advancing there. That is why the choice of the domain you silence is never harmless : it points, by omission, to the war you agree to let continue.
4 Verification
Without monitoring, the truce dissolves.
The condition for survival
A truce is only worth what can be observed
De-escalation is only worth it if you can verify it is being kept. A ground ceasefire can be read on a map ; a ceasefire in a domain like the air requires tracking strikes in real time, attributing them, telling an accident from a deliberate violation. Without a shared verification mechanism (sensors, imagery, a third party accepted by both sides), the truce comes down to one side's word against the other's, and the slightest strike becomes proof of bad faith. Ill-defined, unchecked partial truces do not die at a stroke : they dissolve into crossed accusations.
The idea to keep
A total truce negotiates when the guns fall silent ; a partial truce first negotiates which ones. That choice decides who keeps the initiative : the perimeter is the heart of the deal, and verification its condition for survival.
5 Takeaways
Worth remembering.
✓
De-escalation by domain : reducing violence in a single space (air, sea, energy) rather than everywhere.
✓
The perimeter : defining what you freeze amounts to negotiating, weapon by weapon, who keeps which capability.
✓
The asymmetry : freezing a domain favours the side strong in the others ; no partial truce is neutral.
✓
Verification : without shared monitoring, a partial truce dissolves into crossed accusations.
This notion informs an analysis
First published : 5 August 2026