Skip to content

Rademacher's theorem

A locally Lipschitz map between finite-dimensional Euclidean spaces has a total first-order linear derivative at Lebesgue-almost every point.

Version
v1 · 2026-10-17 · History
Domain-specific #
14272
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomains
Real Analysis, Geometric Measure Theory → Mathematics
Aliases
Rademacher differentiability theorem

Core Idea

Rademacher's theorem says that a locally Lipschitz map between finite-dimensional Euclidean spaces has a total first-order linear derivative at almost every point of its open domain. The exceptions have Lebesgue measure zero, so a Lipschitz map may still have a corner where no derivative exists. This is a proved implication from the Lipschitz condition, rather than a definition of it.[ref-aec7f6225028][ref-e2992282a492]

Scope of Application

The domain must be open and Euclidean, the map Euclidean-valued and locally Lipschitz, and “almost every” refers to Lebesgue measure; other metric settings need separate results.[^ref-aec7f6225028]

  • Hinge and null exception: The Lipschitz map u(x) = max(0, x) has no classical derivative at its single corner, 0.[^ref-e2992282a492]
  • Weak derivative bridge: For that hinge, compare its weak derivative with the classical derivative away from the corner.[^ref-e2992282a492]
  • Vector-valued maps: The same theorem covers locally Lipschitz maps into a higher-dimensional Euclidean space.[ref-aec7f6225028][ref-1fc777f4d445]
  • Transport potential: Under Gangbo and McCann's convex-cost hypotheses, a locally Lipschitz potential has a gradient almost everywhere in the interior of its effective domain; further transport conclusions require additional assumptions.[^ref-28e7ac11b7b4]

Clarity

“Differentiable” here means more than having separate directional or coordinate derivatives. The conclusion requires one total linear approximation that works as the displacement approaches zero from every direction. The multidimensional proof needs an additional step after the one-dimensional and Fubini arguments to establish this total derivative.[ref-aec7f6225028][ref-1fc777f4d445]

Likewise, “almost every” is a measure statement rather than “every point we care about.” If an application needs a derivative at one specified point, that point must be checked separately.

Manages Complexity

Once the local Lipschitz estimate is known, an analyst need not prove almost-everywhere differentiability again for each new map. The remaining work is to verify the estimate, keep track of which measure makes the exceptional set negligible, and check any extra conditions needed by an application.[ref-aec7f6225028][ref-28e7ac11b7b4]

Abstract Reasoning

To use the theorem, first name the open Euclidean domain, codomain, and Lebesgue measure. Then establish local Lipschitz control. The theorem gives a total derivative outside some null set. Use that derivative only where the application's other assumptions permit it; it does not automatically give a derivative at a corner or a complete transport map.[ref-aec7f6225028][ref-e2992282a492][^ref-28e7ac11b7b4]

Knowledge Transfer

The same implication can be used for unlike maps within analysis: a piecewise hinge and a transport potential both meet local regularity conditions and yield an almost-everywhere derivative. Their downstream uses differ. The name should not be transferred by analogy to unrelated systems without a corresponding Euclidean map, Lipschitz estimate, measure, and total derivative.[ref-e2992282a492][ref-28e7ac11b7b4]

Neighborhood in Abstraction Space

Rademacher's theorem sits in a sparse region of the domain-specific corpus (65th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Complex & Holomorphic Function Theory (9 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-17

References

[^ref-aec7f6225028]: Harvard University (2005), Rademacher's theorem. Math 212b handout dated 28 January 2005, especially Theorems 6–7 and proof, PDF pp. 4–7. https://abel.math.harvard.edu/archive/212b_spring_05/handouts/Rademacher.pdf [^ref-1fc777f4d445]: Tao, Terence. 2006. Lecture Notes 3 for Math 247A. Theorems 2.13 and 2.15 and proof, PDF pp. 17–18. https://www.math.ucla.edu/~tao/247a.1.06f/notes3.pdf [^ref-e2992282a492]: Guermond, Jean-Luc. 2023. Weak derivatives and Sobolev spaces. Part I, Chapter 2, Example 2.5(i) and Theorem 2.7, PDF pp. 2–3. https://people.tamu.edu/~guermond/M661_FALL_2023/chap02.pdf [^ref-28e7ac11b7b4]: Gangbo, Wilfrid, and Robert J. McCann. 1996. The Geometry of Optimal Transportation. Theorem 3.3, Proposition 3.4 and proof, PDF pp. 18–20; further transport conditions in Proposition 3.6 and Theorem 3.7. https://www.math.ucla.edu/~wgangbo/publications/geo.pdf