A nonnegative extended real number that quantifies the exponential growth rate of the number of distinguishable orbit segments of a continuous self-map on a compact topological (or metric) space; it can be defined via open covers, separated/spanning sets, or (for shifts) growth of admissible words.

Principle

Measure complexity by counting distinguishable dynamical behaviors at finer and finer scales over time: exponential growth of distinguishable orbits per unit time yields positive entropy, while subexponential growth gives zero entropy.

Demonstration

The full shift on k symbols has topological entropy ln(k) (base-e) because the number of length-n distinct words grows like k^n; many chaotic interval maps have positive topological entropy, whereas an isometry or a rotation on a compact group has entropy zero.

Misapplication

Confusing topological entropy with measure-theoretic entropy for a particular invariant measure, using the definitions on noncompact spaces without proper modification, or treating small Lyapunov exponents as ensuring small topological entropy.

Consequence

Topological entropy is invariant under topological conjugacy, lower bounds measure-theoretic entropy via the variational principle, and positive topological entropy implies existence of complicated orbit structures (e.g., horseshoes, exponential orbit separation) in many settings.

Reversal

The reversal is systems of zero topological entropy, which exhibit limited orbit complexity (equicontinuous systems, rotations, and many minimal isometries) and typically lack exponential orbit separation.

Boundary

Standard definitions assume compactness (or use compact invariant subsets) and continuity of the map; for noncompact spaces, one must use growth rates relative to compacta or other adapted notions; different equivalent definitions require a metric or generating cover.

Semantic Tension

Tension appears with measure-theoretic entropy and metric invariants (Lyapunov exponents): topological entropy is a topological invariant of the map, while measure entropy depends on a chosen invariant measure; they relate by the variational principle but capture different aspects of complexity.

Synthesis

Topological entropy compresses a continuous map's orbit complexity into a single growth rate: count distinguishable orbit pieces at scale ε for time n and take exponential growth in n as ε→0; a positive value signals exponential proliferation of distinct dynamical behaviors.