Phase transitions
Why change is sometimes abrupt — and what that means for recognising it.
A phase transition is a sudden qualitative change in the behaviour of a system. Water that freezes. A population that tips into collective action. A patient who suddenly collapses — or suddenly recovers.
In physics this is also called a tipping point or a bifurcation. In the catastrophe theory of René Thom and Christopher Zeeman it is described mathematically. In psychology, Günter Schiepek has worked for decades on rendering it measurable in psychotherapy.
First-order and second-order phase transitions
Not all phase transitions are equally abrupt. A first-order phase transition is the sharp variant: at a specific temperature, pressure or load, the system tips at once. Water at 0°C or 100°C. A spoked wheel that suddenly buckles. A depression that in days can shift from deep despair to recovered energy — or the reverse.
A second-order phase transition is smoother. The system gradually changes in character without any single moment being "the" turning point. Many developmental processes are of this kind. But if one looks at the system before and after the transition, one nevertheless sees that it has become qualitatively something else.
Hysteresis: there and back are not the same
A fascinating feature of many phase transitions is hysteresis: the turning point on the way out does not lie at the same place as the turning point on the way back. Marten Scheffer's classical example are the shallow lakes. Under an increasing phosphorus load, a clear lake tips into turbid. But to make the turbid lake clear again, the phosphorus load must drop far below the original tipping point — sometimes so far that it is no longer feasible.
"The dogma that the cause of the problem is the key to its solution does not necessarily hold for complex systems." — Van der Maas (2024)
For depression and other chronic conditions this is an important insight. The system that once tipped into depression under stress does not return to the healthy state on its own when the stress disappears. The hysteresis works against recovery. We must push further than seems logical — or, as proved necessary with the lakes, seek a quite different intervention (remove the fish, do not reduce the phosphorus).
Early warning signals
The most exciting part of phase-transition research is that they often announce themselves. Just before a system tips, it shows characteristic signs — what we call early warning signals or critical slowing down: the system recovers more slowly from small perturbations, variance increases, autocorrelation rises, oscillations become slower.
For the consulting room: when one has time-series data from a patient, one can in principle see when she is approaching a phase transition. Sometimes for the good — a breakthrough on its way. Sometimes for the worse — a relapse announcing itself. The catastrophe flags (sudden jump, multimodality, divergence, hysteresis, critical slowing down, and a few others) are methodological tools for diagnosing this in real data.
The most important of all this is perhaps this: change need not be gradual. We are used in the consulting room to thinking in dose-response — more therapy yields more recovery. But with phase transitions it works differently. Long periods of little effect, and then suddenly a leap. One who does not expect this thinks too early that the treatment does not work. One who does expect it stays put until the moment arrives.