Multi-stability

Multi-stability is the property of a dynamical system that possesses two or more distinct stable states — attractors — that can persist indefinitely without external forcing. A multi-stable system does not converge to a unique equilibrium; instead, which long-run state it reaches depends on its history, initial conditions, or the nature of past perturbations. The coexistence of multiple attractors, each with its own basin of attraction, is the formal definition.

Multi-stability appears wherever positive feedback operates in conjunction with saturation: each attractor is stabilized by feedback that amplifies displacement toward it, limited by some ceiling or floor that prevents indefinite runaway. Bistable systems — the simplest case, with exactly two attractors — include: the flip-flop in digital circuits, the action potential in neurons (firing vs. resting), ice-albedo feedback in climate (glaciated vs. ice-free states), and polarized political equilibria in social systems.

The practical significance of multi-stability is that intervention must account for basin geometry. A system near the boundary between two basins can be shifted by a small push; a system deep within one basin requires a large intervention — and even then, removing the intervention may not return the system to its previous attractor. The history of failed ecosystem restoration efforts is largely a history of underestimating how deep the degraded state's basin had become before intervention was attempted.