A cybernetics principle stating that effective regulation of a system requires a controller whose repertoire of possible responses (variety) is at least as large as the variety of disturbances or states the system may present; without sufficient controller variety, some disturbances cannot be absorbed or countered.

Principle

Control succeeds only if the controller can generate responses that map onto the range of perturbations; therefore one must either increase controller variety, reduce environmental variety, or constrain interactions (e.g., via buffering, hierarchy, or rules) to achieve regulation.

Demonstration

Illustrative scenario — Situation: A simple on/off regulator designed for two discrete environmental states is applied to an environment with continuous fluctuations. Recognition: The regulator lacks sufficient response variety to stabilize all states. Action: Designer replaces the regulator with a proportional controller or adds buffering to reduce effective disturbance range. Consequence: Regulation succeeds because controller variety or environmental variety is adjusted to restore the required relationship.

Misapplication

Assuming mere addition of components or options always solves variety mismatch (i.e., equating more complexity with more effective control) mistakes raw variety for coordinated, usable variety; the error is neglecting integration, observability, and mapping between disturbances and responses.

Consequence

Designs guided by the law focus effort on matching controller capabilities to environmental complexity: increase decision alternatives, improve sensing, reduce disturbance scope, or impose structure. Failure to respect the law yields persistent unregulated variability or brittle controllers.

Reversal

When disturbances are stochastic and objectives are probabilistic, exact one‑to‑one matching of variety is neither necessary nor practical; stochastic control, robust design, or acceptance of bounded error can provide effective regulation without matching raw variety.

Boundary

Clearly within: feedback control systems and managerial control problems where disturbances are describable as varieties of states and controller responses are enumerable. Boundary case: adaptive systems where the controller learns over time — requisite variety must be considered in a dynamic rather than static sense. Clearly outside: optimization problems that do not involve regulation against disturbance varieties (pure cost minimization absent external perturbations).

Semantic Tension

Robustness (ability to absorb disturbances) ↔ Simplicity (minimal, comprehensible controllers): increasing variety can improve robustness but at the cost of complexity, observability and maintainability.

Synthesis

Effective regulation requires matching the actionable diversity of responses to the diversity of disturbances; practitioners can either expand coordinated response options, reduce environmental variety by design, or rearrange system architecture (hierarchies, buffers) so that the requisite relationship holds in practice.