Sensitive dependence on initial conditions

Sensitive dependence on initial conditions — popularly known as the butterfly effect — is the defining property of chaotic dynamical systems. Two trajectories that start arbitrarily close together diverge exponentially fast, making long-term prediction impossible even when the governing equations are perfectly known and deterministic. Mathematically, this is quantified by a positive Lyapunov exponent: the average rate at which nearby orbits separate.

This property challenges the classical Laplacian view of determinism, in which perfect knowledge of initial conditions implies perfect predictability. In chaotic systems, perfect predictability would require infinite precision in measuring initial conditions — a physical impossibility. The implication is profound: deterministic systems can be inherently unpredictable, not because of randomness or missing information, but because of the geometry of the dynamics itself.

Sensitive dependence is not sufficient for chaos. A system must also exhibit topological mixing and dense periodic orbits to be classified as chaotic in the sense of Devaney. But sensitive dependence is the property that most directly connects chaos to the limits of prediction and knowledge.