Couples stuck in escalation: strategic interventions for the therapy room

One partner raises an issue. The other goes quiet or leaves the room. The first reads the silence as indifference and raises it louder, or more often. The second reads the pressure as attack and withdraws further.

Neither is wrong about what they are seeing. Each is responding reasonably to a sequence they cannot see they are jointly producing.

A safety check that comes first

None of what follows applies where the pattern includes fear of physical harm, coercive control, or a partner who cannot safely disengage from a disagreement.

From a distance, escalation and abuse can look alike. Raised voices, one pursuing, one avoiding. They are not the same clinical picture, and a reframe treating coercive control as a sequence both partners maintain would be a serious error. Ask each partner separately, early and as routine rather than only when something already looks wrong, whether they ever feel unsafe or unable to leave an argument.

Demand-withdraw, and the problem of punctuation

Christensen and Heavey’s (1990) research on demand-withdraw describes this loop directly: one partner pursues discussion, criticises or presses for change while the other avoids, defends or disengages. Each experiences their own behaviour as a reasonable reply to the other’s.

Which is Watzlawick’s point about punctuation. In a circular sequence there is no objective starting point, only the point each participant nominates.

Strategic assessment does not try to settle who started it. It maps what each partner currently does about the relationship’s problem, because by the time a couple reaches your room that attempted solution and the problem are nearly indistinguishable.

Why blame hardens the loop

Naming one partner “the demander” in a way that reads as accusation makes both positions more rigid. A demanding partner told she is too much often demands harder, sensing this is her last chance to be heard. A withdrawing partner told he is avoidant withdraws further, now carrying a diagnosis as well.

The frame that works describes a shared mechanism. Each partner’s solution, applied reasonably, maintains the other’s. Most couples recognise that as accurate about themselves, and the drop in defensiveness is what makes a joint experiment possible.

A small experiment aimed at the sequence

The prescription targets each partner’s own response, never the other’s. The partner who raises concerns immediately agrees to hold them until a scheduled time each day. The partner who goes quiet agrees to say one sentence naming what he feels before leaving the room, instead of leaving in silence.

Neither is asked to become someone else. Both interrupt the point in the sequence that had been escalating fastest.

What the couples research adds

Gottman and Levenson’s (2000) work on divorce timing found that conflict frequency does not distinguish couples who stay from couples who separate. What distinguishes them is how conflict opens and whether small repair attempts get accepted during it. Different research tradition, complementary conclusion: both point at the mechanics of a specific exchange rather than at overall compatibility.

Rohrbaugh and Shoham (2015) add a practical note. Sometimes the most useful session is with one partner alone, when only one is genuinely a customer for change, rather than insisting on conjoint work the couple cannot yet use.

Reading the first signs

The first evidence a prescription is taking is rarely dramatic. Usually one partner mentions in passing that an argument which would normally have escalated simply did not. The underlying issue is unresolved; the sequence that used to carry it there did not fire.

Ask about this specifically. Couples do not think to report a fight that never happened.

The pattern is worked through session by session in the case study of a pursue-withdraw loop.