For a real exponent p>0, the p-variation of a function on an interval is the supremum over all finite partitions of the sums of the pth powers of the absolute increments; it quantifies roughness with a parameter p that weights small and large jumps differently.

Principle

P-variation parametrizes variation by exponent: smaller p amplifies many small increments, larger p emphasizes large jumps; finiteness of p-variation is a scale-dependent regularity property useful for controlling pathwise integrals and roughness.

Demonstration

For f(x)=x on [0,1], the p-variation equals 1^p=1 for any p>0 since all increments sum to 1 and taking pth powers keeps that total at 1; step functions concentrate variation on jumps so their p-variation reflects jump sizes raised to p.

Misapplication

Confusing p-variation with an L^p norm over the interval or treating p-variation as invariant under reparameterization when the path's parameterization changes the partition increments leads to misuse in stochastic or geometric contexts.

Consequence

Finite p-variation provides quantitative control of a path's irregularity: it yields embedding relations with Hölder spaces and determines which integration theories apply (for example Young integration when complementary p-variations match).

Reversal

Taking the formal inverse viewpoint, letting p→1 recovers total variation (summing absolute increments), while letting p→∞ focuses on the largest single jump (essential supremum of increments); these extremes contrast cumulative versus extreme-focused roughness.

Boundary

Defined for functions on ordered domains; sensitive to the choice of exponent p and to parameterization; it should not be conflated with Sobolev seminorms, L^p norms, or with notions that require averaging rather than supremal sums.

Semantic Tension

P-variation sits near concepts such as Hölder regularity and Besov seminorms: all measure regularity but differ in aggregation (supremal partition sums vs. pointwise Hölder bounds vs. frequency-based norms), so the same path can have different characterizations.

Synthesis

P-variation is a one-parameter family of supremal aggregate measures of a function's increments: by raising absolute increments to the pth power before summing, it interpolates between total cumulative change and extreme-jump observables, giving a tunable lens on path roughness.