The operator norm of a bounded linear operator T between normed vector spaces is the supremum of the output norm over all unit input vectors: ||T|| = sup{||Tx|| : ||x|| = 1}, equivalently the smallest constant C such that ||Tx|| ≤ C||x|| for all x.

Principle

It is the norm induced by the underlying vector norm and organizes linear maps by their maximal stretching factor; it is submultiplicative (||ST|| ≤ ||S||·||T||) and provides Lipschitz-type bounds.

Demonstration

For a 2×2 matrix A with respect to the Euclidean norm, the operator norm equals the largest singular value. Example: A = [[3,0],[0,1]] has ||A||2 = 3 because the unit vector e1 is stretched by factor 3 and no unit vector is stretched more.

Misapplication

Treating the operator norm as if it were the Frobenius norm or spectral radius in contexts where those differ; for instance, using the Frobenius norm to bound iterative growth when operator-submultiplicativity or tight spectral information is required.

Consequence

Correct use yields concrete stability and error bounds, a Lipschitz constant for linear maps, and control of convergence rates in numerical methods and perturbation estimates.

Reversal

The conceptual inverse is to consider the minimal nonzero output over unit inputs (the minimal norm or the smallest singular value), which measures injectivity rather than maximal stretching.

Boundary

Defined for bounded linear operators between normed spaces; does not directly apply to unbounded operators without domain and graph-norm considerations, nor to arbitrary nonlinear maps except via linearization.

Semantic Tension

Tension exists between operator norm, spectral radius, and entrywise norms (e.g., Frobenius): they all measure size but differ in invariance and tightness, so choosing one affects conclusions about growth and stability.

Synthesis

The operator norm is the induced measure of a linear operator's largest amplification on vectors in a given norm, yielding sharp Lipschitz constants and interacting with spectral and other matrix norms according to context.