Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3.36 — Morrey's inequality for C_c^∞(ℝⁿ), p > n

Proved
HunterPDE.Sobolev.morrey

by mikedeng1 · Sep 28, 2026 · Mathlib 0df444a (Lean v4.33.1)

holder-spacesmorrey-inequalityp2o-batch-books5p2o-gran-per-chapterp2o-plan-bookp2o-v1pdesobolev-spaces

Let n<p<∞n < p < \inftyn<p<∞ and α=1−n/p\alpha = 1 - n/pα=1−n/p. There is a constant C=C(n,p)C = C(n, p)C=C(n,p) such that for all f∈Cc∞(Rn)f \in C_c^\infty(\mathbb{R}^n)f∈Cc∞​(Rn)

[f]α≤C ∥Df∥p(3.12)[f]_\alpha \le C\, \|Df\|_p \qquad (3.12)[f]α​≤C∥Df∥p​(3.12) sup⁡Rn∣f∣≤C ∥f∥W1,p(3.13)\sup_{\mathbb{R}^n} |f| \le C\, \|f\|_{W^{1,p}} \qquad (3.13)Rnsup​∣f∣≤C∥f∥W1,p​(3.13)

where [⋅]α[\cdot]_\alpha[⋅]α​ is the Hölder seminorm [⋅]α,Rn[\cdot]_{\alpha,\mathbb{R}^n}[⋅]α,Rn​ of (1.1), ∣Df∣|Df|∣Df∣ is the Euclidean length of the gradient and ∥⋅∥W1,p\|\cdot\|_{W^{1,p}}∥⋅∥W1,p​ is the norm of Definition 3.23.

For p>np > np>n a gradient in LpL^pLp forces Hölder continuity with exponent 1−n/p1 - n/p1−n/p and boundedness; this is the basis of the embedding W1,p(Rn)↪C0,α(Rn)W^{1,p}(\mathbb{R}^n) \hookrightarrow C^{0,\alpha}(\mathbb{R}^n)W1,p(Rn)↪C0,α(Rn) (Theorem 3.37).

Formalization Note. One constant serves both inequalities, which is equivalent to two constants (take the larger). It is chosen after n,pn, pn,p and before fff. (3.13) is stated pointwise, ∣f(x)∣≤C∥f∥W1,p|f(x)| \le C\|f\|_{W^{1,p}}∣f(x)∣≤C∥f∥W1,p​ for every xxx, which is equivalent to the bound on the supremum. The clause "with α=1\alpha = 1α=1 if p=∞p = \inftyp=∞" of the printed statement is outside the stated range n<p<∞n < p < \inftyn<p<∞ and is not formalized. Values are in [0,∞][0, \infty][0,∞].

Preamble
import Mathlib
import Definitions.Def_HunterPDE_Sobolev_SobolevSpace
import Definitions.Def_HunterPDE_Shared_HolderSeminorm

open MeasureTheory
Formal statement
namespace HunterPDE.Sobolev

/-- Theorem 3.36 of Hunter, *Notes on PDEs* (revised 6/18/2014), p. 68 (Morrey): let `n < p < ∞`
and `α = 1 − n/p`. There is a constant `C = C(n, p)` (chosen before `f`) such that for all
`f ∈ C_c^∞(ℝⁿ)`: `[f]_α ≤ C ‖Df‖_p` (3.12), with `[·]_α` the Hölder seminorm (1.1) on `ℝⁿ` and `|Df|`
the Euclidean gradient norm, and `sup_{ℝⁿ} |f| ≤ C ‖f‖_{W^{1,p}}` (3.13), stated pointwise. -/
theorem morrey {n : ℕ} {p : ℝ} (hnp : (n : ℝ) < p) :
    ∃ C : NNReal, ∀ f : EuclideanSpace ℝ (Fin n) → ℝ,
      ContDiff ℝ ((⊤ : ℕ∞) : WithTop ℕ∞) f → HasCompactSupport f →
        Shared.holderSeminorm (1 - (n : ℝ) / p) Set.univ f ≤
            (C : ENNReal) * eLpNorm (fun x => ‖gradient f x‖) (ENNReal.ofReal p) volume ∧
          ∀ x, ENNReal.ofReal |f x| ≤
            (C : ENNReal) * sobolevNorm 1 (ENNReal.ofReal p) Set.univ f := by sorry

end HunterPDE.Sobolev
Source
Hunter, Notes on Partial Differential Equations (revised 6/18/2014), p. 68, Theorem 3.36
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

The names Shared.holderSeminorm and sobolevNorm come from imported definition files whose text is not part of this declaration.

Let nnn be a natural number and ppp a real number with n<pn < pn<p. Put

α=1−np.\alpha = 1 - \frac{n}{p}.α=1−pn​.

Then there exists a single constant C≥0C \ge 0C≥0, depending only on nnn and ppp and chosen before fff, with the following property. For every C∞C^\inftyC∞ function f:Rn→Rf : \mathbb{R}^n \to \mathbb{R}f:Rn→R with compact support, both of the following hold.

(i) Hölder bound. In [0,∞][0,\infty][0,∞],

holderSeminorm(α, Rn, f)  ≤  C⋅∥ ∣∇f∣ ∥Lp(Rn).\texttt{holderSeminorm}(\alpha,\,\mathbb{R}^n,\,f) \;\le\; C \cdot \big\|\,|\nabla f|\,\big\|_{L^p(\mathbb{R}^n)}.holderSeminorm(α,Rn,f)≤C⋅​∣∇f∣​Lp(Rn)​.

The terms in this bound are:

  • ∇f(x)\nabla f(x)∇f(x) is the gradient of fff at xxx, the vector representing the Fréchet derivative through the Euclidean inner product;
  • ∣∇f(x)∣|\nabla f(x)|∣∇f(x)∣ is its Euclidean length;
  • ∥h∥Lp=(∫∣h∣p dx)1/p\|h\|_{L^p} = \left(\int |h|^p\,dx\right)^{1/p}∥h∥Lp​=(∫∣h∣pdx)1/p with respect to Lebesgue measure, with ppp passed as an extended real.

(ii) Pointwise bound. For every x∈Rnx \in \mathbb{R}^nx∈Rn,

∣f(x)∣  ≤  C⋅sobolevNorm(1, p, Rn, f),|f(x)| \;\le\; C \cdot \texttt{sobolevNorm}(1,\,p,\,\mathbb{R}^n,\,f),∣f(x)∣≤C⋅sobolevNorm(1,p,Rn,f),

compared in [0,∞][0,\infty][0,∞].

Degenerate cases. For n≥1n \ge 1n≥1 the hypothesis forces p>1p > 1p>1.

For n=0n = 0n=0 the hypothesis is only p>0p > 0p>0, so 0<p≤10 < p \le 10<p≤1 is permitted, and α=1\alpha = 1α=1. In that case:

  • R0\mathbb{R}^0R0 is a single point, and every function on it is smooth with compact support;
  • the gradient is 000, so the right side of (i) is 000 and (i) requires the Hölder seminorm to be 000;
  • whether (i) and (ii) then hold depends on the imported definitions, which are not shown here.
Human review
  • Endorsed by Shuze Chen · Oct 2, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 2, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me