Stability-Plasticity Tradeoff

principle

Adaptive systems need fast-moving parts that explore and slow-moving parts that preserve coherence. Remove either side and the system either fossilizes or churns without accomplishing anything.

The force obstructing change may also be what lets change amount to something. The slow layer can rust into irrelevance—but without it, the fast layer simply races from novelty to novelty and never gets anything done.

E1

Change needs something to push against

Plasticity generates variation: new practices, ideas, and responses can be tried quickly. Stability retains what works, coordinates activity over time, and gives experiments a durable platform. Each restrains the other's characteristic failure. Too much retention turns yesterday's solution into a fossil; too much variation prevents learning from accumulating into capability. The resulting Productive Friction is therefore not merely resistance—it is the negotiation through which novelty becomes usable without dissolving continuity.

E1

Not every resistance is useful

The tradeoff does not imply that stability and change should always receive equal weight. A slow layer that can no longer preserve anything valuable is merely fossilized; a fast layer that cannot carry an experiment through is merely restless. The relevant balance depends on whether continuity is still enabling coordinated achievement and whether novelty is producing learning rather than motion.

E1

Separate the platform from the experiment

For your next proposed change, name one element that must remain fixed long enough to coordinate around and one element that can vary rapidly. Then give the experiment a defined interval before judging it, so adaptation has room to occur without turning the whole system into churn.

E1

Episodes that teach this