A norm on the space of bounded linear operators between normed vector spaces defined as the supremum of the operator's output norms over all inputs of unit norm; equivalently the smallest constant C such that ||T x|| ≤ C ||x|| for all x.

Principle

Organize operator size by the maximal stretching effect on unit vectors; this induced norm is submultiplicative and compatible with the underlying vector norms.

Demonstration

For a matrix seen as a linear map on R^n with the Euclidean norm, the operator norm equals the largest singular value; numerically, ||A||2 = sup_{||x||2=1} ||Ax||2 = σ_max(A).

Misapplication

Treating an entrywise matrix norm (e.g., maximum absolute row sum) as identical without specifying the underlying vector norms, or attempting to compute an operator norm for an unbounded operator without domain considerations.

Consequence

A finite operator norm implies continuity and boundedness of the operator; norms give operator topology, yield submultiplicativity ||ST|| ≤ ||S|| ||T||, and control error propagation in numerical schemes.

Reversal

Replacing the supremum by an infimum (minimal amplification over unit vectors) produces a quantity that is typically zero and fails to measure global operator size; conversely, considering the norm of an inverse highlights invertibility rather than magnitude of forward action.

Boundary

Defined for bounded linear operators between normed spaces; does not apply to nonlinear or unbounded operators without domain restrictions, and its value depends on the choice of underlying vector norms.

Semantic Tension

Tension with entrywise or componentwise matrix norms: operator norm measures action on vectors relative to chosen vector norms, while entrywise norms measure coefficients and can give different orderings of operator size.

Synthesis

The operator norm condenses the maximal effect of a bounded linear map with respect to given vector norms into a single submultiplicative scalar that characterizes continuity, stability, and composition behavior.