Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Weak solutions of the nonlinear Cauchy–Riemann equation are smooth

Proved
JHolomorphic.contDiffOn_of_isWeakJHolomorphicOn

by Mazecto · Oct 20, 2026 · Mathlib 0df444a (Lean v4.33.1)

elliptic-regularitypseudoholomorphic-curvessobolev-spacessymplectic-geometryunique-continuation

Let p>2p>2p>2, let EEE be a finite-dimensional real vector space and let JJJ be a smooth almost complex structure on an open set O⊆EO\subseteq EO⊆E. Let U⊆CU\subseteq\mathbb{C}U⊆C be open and let u:U→Ou:U\to Ou:U→O be continuous, with u∈Wloc1,p(U)u\in W^{1,p}_{\mathrm{loc}}(U)u∈Wloc1,p​(U) and

∂su+J(u) ∂tu=0almost everywhere on U\partial_su+J(u)\,\partial_tu=0\quad\text{almost everywhere on }U∂s​u+J(u)∂t​u=0almost everywhere on U

for its weak partial derivatives. Then uuu is C∞C^\inftyC∞ on UUU.

Formalization Note The hypothesis is IsWeakJHolomorphicOn p J O U u. Wendl states the result on the unit disc for E=O=CnE=O=\mathbb{C}^nE=O=Cn with the normalization J(0)=iJ(0)=iJ(0)=i, u(0)=0u(0)=0u(0)=0; here the domain is any open UUU, JJJ is only defined on OOO, and no normalization is imposed. Continuity of uuu selects the continuous representative; the conclusion is ContDiffOn ℝ ∞ u U.

Preamble
import Definitions.Def_JHolomorphic_Setting
Formal statement
namespace JHolomorphic

open MeasureTheory Metric Filter Topology DbarRegularity
open scoped ContDiff ENNReal NNReal

/-- Wendl, Theorem 2.22 (1) (p. 35), case `k = 1`: let `p > 2` and let `J` be a smooth almost
complex structure on an open set `O ⊆ E`. A map `u ∈ W^{1,p}_loc(U)` on an open set `U ⊆ ℂ`,
continuous with values in `O`, that satisfies `∂ₛu + J(u) ∂ₜu = 0` almost everywhere, is `C^∞` on
`U`. (Wendl works on the disc with `E = ℂⁿ`, `O = ℂⁿ`, `J(0) = i` and `u(0) = 0`.) -/
theorem contDiffOn_of_isWeakJHolomorphicOn {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] (p : ℝ) (hp : 2 < p)
    (J : E → E →L[ℝ] E) (O : Set E) (hJ : IsAlmostComplexOn J O) (U : Set ℂ) (hU : IsOpen U)
    (u : ℂ → E) (hu : IsWeakJHolomorphicOn p J O U u) :
    ContDiffOn ℝ ∞ u U := by
  sorry

end JHolomorphic
Source
C. Wendl, Lectures on Symplectic Field Theory, arXiv:1612.01009, https://arxiv.org/abs/1612.01009, Lecture 2, Theorem 2.22 (1), p. 35 (case k = 1)
Read-back

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

Read-back: JHolomorphic.contDiffOn_of_isWeakJHolomorphicOn

Data and standing assumptions. Let EEE be a finite-dimensional real normed vector space (any norm; no inner product or complex structure on EEE is assumed). Let ppp be a real number with p>2p > 2p>2. Let J:E→L(E,E)J : E \to \mathcal{L}(E,E)J:E→L(E,E) be any map assigning to each point x∈Ex \in Ex∈E a continuous real-linear operator J(x)J(x)J(x) on EEE. Let O⊆EO \subseteq EO⊆E and U⊆CU \subseteq \mathbb{C}U⊆C be sets, with UUU assumed open. Let u:C→Eu : \mathbb{C} \to Eu:C→E be any function. Throughout, C\mathbb{C}C is regarded as the real plane, a point is written z=s+itz = s + itz=s+it, and "measure" on C\mathbb{C}C means two-dimensional Lebesgue measure.

Hypothesis on JJJ (almost complex on OOO). It is assumed that:

  • OOO is open in EEE;
  • JJJ, viewed as a map E→L(E,E)E \to \mathcal{L}(E,E)E→L(E,E), is C∞C^\inftyC∞ on OOO (smooth, i.e. CkC^kCk for every finite kkk, in the sense of the within-the-set derivatives of JJJ restricted to OOO);
  • for every x∈Ox \in Ox∈O, J(x)∘J(x)=−idEJ(x) \circ J(x) = -\mathrm{id}_EJ(x)∘J(x)=−idE​.

Nothing is assumed about J(x)J(x)J(x) for x∉Ox \notin Ox∈/O.

Hypothesis on uuu (weakly JJJ-holomorphic on UUU with exponent ppp). It is assumed that:

  1. uuu is continuous on UUU (continuity of the restriction u∣Uu|_Uu∣U​);
  2. u(z)∈Ou(z) \in Ou(z)∈O for every z∈Uz \in Uz∈U;
  3. there exist two functions us,ut:C→Eu_s, u_t : \mathbb{C} \to Eus​,ut​:C→E (defined on all of C\mathbb{C}C; their values outside UUU are unconstrained) such that:
    • (weak partial derivatives) for every φ:C→R\varphi : \mathbb{C} \to \mathbb{R}φ:C→R that is C∞C^\inftyC∞ on all of C\mathbb{C}C, has compact support, and whose closed support tsupp⁡φ\operatorname{tsupp}\varphitsuppφ is contained in UUU,
∫C∂sφ(z) u(z) dz=−∫Cφ(z) us(z) dz,∫C∂tφ(z) u(z) dz=−∫Cφ(z) ut(z) dz,\int_{\mathbb{C}} \partial_s\varphi(z)\, u(z)\, dz = -\int_{\mathbb{C}} \varphi(z)\, u_s(z)\, dz, \qquad \int_{\mathbb{C}} \partial_t\varphi(z)\, u(z)\, dz = -\int_{\mathbb{C}} \varphi(z)\, u_t(z)\, dz,∫C​∂s​φ(z)u(z)dz=−∫C​φ(z)us​(z)dz,∫C​∂t​φ(z)u(z)dz=−∫C​φ(z)ut​(z)dz,
 where $\partial_s\varphi(z) = D\varphi(z)[1]$ and $\partial_t\varphi(z) = D\varphi(z)[i]$ are the real directional derivatives of $\varphi$ in the directions $1$ and $i$. These are $E$-valued Bochner integrals over all of $\mathbb{C}$, with the convention that the integral of a non-integrable function is defined to be $0$; the identities are required as stated under that convention, and the definition itself imposes no separate integrability or measurability requirement;
  • (local LpL^pLp) for every compact set K⊆UK \subseteq UK⊆U (including K=∅K = \emptysetK=∅), each of uuu, usu_sus​, utu_tut​ belongs to Lp(K)L^p(K)Lp(K), i.e. is almost-everywhere strongly measurable with respect to Lebesgue measure restricted to KKK and satisfies ∫K∥f∥p<∞\int_K \|f\|^p < \infty∫K​∥f∥p<∞ (here the exponent is the extended non-negative real ppp, which is finite since p>2p > 2p>2);
  • (the equation) for Lebesgue-almost every z∈Uz \in Uz∈U,
us(z)+J(u(z)) ut(z)=0.u_s(z) + J\big(u(z)\big)\, u_t(z) = 0.us​(z)+J(u(z))ut​(z)=0.

Conclusion. Under these assumptions, uuu is C∞C^\inftyC∞ on UUU: it is CkC^kCk on UUU (as a map from the real plane C≅R2\mathbb{C} \cong \mathbb{R}^2C≅R2 to EEE, with within-UUU derivatives) for every finite kkk. Since UUU is open, this is the same as uuu being infinitely real-differentiable near every point of UUU. The smoothness index is the "∞\infty∞" of finite-order smoothness, not real-analyticity. Nothing is asserted about uuu outside UUU, and nothing is asserted about usu_sus​ or utu_tut​ (for instance, the conclusion does not say that they agree with the classical partial derivatives).

Edge cases and remarks.

  • If U=∅U = \emptysetU=∅, every hypothesis on uuu holds trivially (take us=ut=0u_s = u_t = 0us​=ut​=0) and the conclusion is trivially true.
  • OOO may be empty. Then condition 2 forces U=∅U = \emptysetU=∅, which is the trivial case above.
  • If E={0}E = \{0\}E={0}, the condition J(x)2=−idJ(x)^2 = -\mathrm{id}J(x)2=−id holds automatically, since every operator is 0=−00 = -00=−0.
  • If EEE is nonzero of odd real dimension, no operator satisfies J(x)2=−idEJ(x)^2 = -\mathrm{id}_EJ(x)2=−idE​ (taking determinants gives det⁡(J)2=(−1)dim⁡E=−1\det(J)^2 = (-1)^{\dim E} = -1det(J)2=(−1)dimE=−1). So the hypothesis on JJJ can hold only when O=∅O = \emptysetO=∅, which again forces U=∅U = \emptysetU=∅.
  • The exponent ppp is a real number strictly greater than 222. The case p=∞p = \inftyp=∞ is not covered, and neither is p≤2p \le 2p≤2.
  • The equation is required only almost everywhere on UUU, and it uses JJJ only at the points u(z)∈Ou(z) \in Ou(z)∈O.
  • The statement opens the DbarRegularity namespace, but none of its definitions (the maximal function, closed cubes, the operator ∂ˉ\bar\partial∂ˉ, the Sobolev norm) appear in the hypotheses or the conclusion. Of the JHolomorphic definitions, only "almost complex on OOO" and "weakly JJJ-holomorphic on UUU" (together with its weak-partials clause) are used. The pointwise operator ∂ˉJ\bar\partial_J∂ˉJ​, strong JJJ-holomorphicity and Cloc∞C^\infty_{\mathrm{loc}}Cloc∞​ convergence do not appear.
Human review
  • Endorsed by Shuze Chen · Oct 21, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Mazecto · Oct 21, 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