№ 10

Early warning signals

How a system may announce that it is about to tip, and why there are two ways of looking at this.

One of the intriguing findings of complexity research in recent decades is that systems seem to announce themselves before they tip. Shortly before a shift, something changes in the way a system moves. The change is subtle, but it may be measurable in a time series.

The literature describes two families of early warning signals. They are quite often confused, although they rest on different assumptions about the kind of system one is dealing with. For those who practise psychotherapy, the distinction is more than an academic matter, because it determines what one pays attention to.

Critical slowing down

The concept of critical slowing down comes from ecology and climate science (Scheffer et al., 2009). It assumes a system in stable equilibrium that slowly loses resilience because an external condition changes gradually. An example is a clear lake in which the amount of nutrients increases year after year, until one day the water turns turbid.

Shortly before such a tipping point, the system recovers more and more slowly from small disturbances. In a time series, this can be seen as:

  • Increasing autocorrelation: the value at one moment comes to resemble that of the previous moment more and more, because the system recovers less quickly
  • Increasing variance: the fluctuations around the mean become larger
  • Stronger coupling between variables that previously moved relatively independently of one another

In mental health care this has been studied mainly in depression, in particular by the group around Marieke Wichers in Groningen (van de Leemput et al., 2014; Wichers et al., 2020). The results are mixed. In recent research these signals preceded only about a third of relapses (Smit et al., 2025), and the researchers themselves advocate caution before the method is used in clinical practice (Helmich et al., 2024).

Critical fluctuations

Synergetics, the theory of Hermann Haken that Günter Schiepek translated to psychotherapy, looks at change in a different way. Here a person in therapy is understood not as a system at rest, but as an open system far from equilibrium, kept in continual motion by the therapy, by relationships and by what happens in life. From this perspective, change is not a slow slide out of equilibrium but a self-organised transition from one pattern to another.

Such a transition is then preceded not by slowing down but by increasing instability: stronger and more irregular fluctuations that cover a larger part of the possible values. The old pattern loses its hold while the new one has not yet formed. In the Synergetic Navigation System this is measured as dynamic complexity. In patients with mood disorders, such critical fluctuations were found to precede sudden improvements and deteriorations during treatment (Olthof et al., 2020; Schiepek & Strunk, 2010).

Why the distinction matters

Both approaches predict unrest before a shift, but for different reasons and with different measures. Critical slowing down presupposes a system close to equilibrium, a slowly changing environment and long, stable time series. Psychotherapy rarely meets these conditions: circumstances keep changing along with it, and the series are short (Schiepek et al., 2020). Applying the assumptions of one model to a system that behaves according to the other carries the risk of looking for the wrong signals.

What it means in the consulting room

A patient who becomes more changeable, whose mood and complaints swing back and forth more strongly, is not necessarily deteriorating. It may be a sign that the system is reorganising itself (see also critical instability). It is precisely in such a period that an intervention may have more effect than usual. This calls for attention, and for looking together with the patient at what is happening, rather than for damping the unrest as quickly as possible.

The aim is not to predict the future, for which a person is too complex, but to become more sensitive to the moments at which change becomes possible.