Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 1.2.2 — Cauchy formula for derivatives, coefficients, Cauchy estimates

Proved
LeblSCV.Holomorphic.cauchy_derivative_formula_and_estimates

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

cauchy-estimatescauchy-formulap2o-batch-books5p2o-gran-per-chapterp2o-plan-bookp2o-v1power-seriesseveral-complex-variables

Let Δ=Δρ(a)⊂Cn\Delta = \Delta_\rho(a) \subset \mathbb{C}^nΔ=Δρ​(a)⊂Cn be a polydisc (all ρk>0\rho_k > 0ρk​>0) with distinguished boundary Γ\GammaΓ, and let f:Δ‾→Cf : \overline{\Delta} \to \mathbb{C}f:Δ→C be continuous and holomorphic in Δ\DeltaΔ. Then:

  1. for z∈Δz \in \Deltaz∈Δ and every multi-index α∈N0n\alpha \in \mathbb{N}_0^nα∈N0n​,
∂∣α∣f∂zα(z)=1(2πi)n∫Γα! f(ζ)(ζ−z)α+1 dζ;\frac{\partial^{|\alpha|} f}{\partial z^\alpha}(z) = \frac{1}{(2\pi i)^n} \int_\Gamma \frac{\alpha!\, f(\zeta)}{(\zeta - z)^{\alpha + 1}}\, d\zeta;∂zα∂∣α∣f​(z)=(2πi)n1​∫Γ​(ζ−z)α+1α!f(ζ)​dζ;
  1. if fff is given on Δ\DeltaΔ by a power series f(z)=∑αcα(z−a)αf(z) = \sum_\alpha c_\alpha (z-a)^\alphaf(z)=∑α​cα​(z−a)α converging uniformly absolutely on compact subsets of Δ\DeltaΔ (expansion (1.3)), then
cα=1α!∂∣α∣f∂zα(a),c_\alpha = \frac{1}{\alpha!} \frac{\partial^{|\alpha|} f}{\partial z^\alpha}(a),cα​=α!1​∂zα∂∣α∣f​(a),

and the Cauchy estimates ∣cα∣≤∥f∥Γ/ρα|c_\alpha| \le \|f\|_\Gamma / \rho^\alpha∣cα​∣≤∥f∥Γ​/ρα hold.

Here α!=α1!⋯αn!\alpha! = \alpha_1! \cdots \alpha_n!α!=α1​!⋯αn​!, ρα=ρ1α1⋯ρnαn\rho^\alpha = \rho_1^{\alpha_1} \cdots \rho_n^{\alpha_n}ρα=ρ1α1​​⋯ρnαn​​, (ζ−z)α+1=∏k(ζk−zk)αk+1(\zeta - z)^{\alpha+1} = \prod_k (\zeta_k - z_k)^{\alpha_k + 1}(ζ−z)α+1=∏k​(ζk​−zk​)αk​+1, dζ=dζ1∧⋯∧dζnd\zeta = d\zeta_1 \wedge \cdots \wedge d\zeta_ndζ=dζ1​∧⋯∧dζn​, and ∥f∥Γ=sup⁡ζ∈Γ∣f(ζ)∣\|f\|_\Gamma = \sup_{\zeta \in \Gamma} |f(\zeta)|∥f∥Γ​=supζ∈Γ​∣f(ζ)∣. In particular the coefficients depend only on the derivatives of fff at aaa.

Formalization Note. ∂∣α∣/∂zα\partial^{|\alpha|}/\partial z^\alpha∂∣α∣/∂zα is the iterated Wirtinger derivative wirtingerIter; the integral over Γ\GammaΓ is torusIntegral (see Theorem 1.1.4). The Cauchy estimate is stated for every real MMM with ∣f∣≤M|f| \le M∣f∣≤M on Γ\GammaΓ: ∣cα∣≤M/ρα|c_\alpha| \le M/\rho^\alpha∣cα​∣≤M/ρα; taking M=∥f∥ΓM = \|f\|_\GammaM=∥f∥Γ​ (finite, as fff is continuous on the compact set Γ\GammaΓ) gives the book's form, and the book's form implies this one. Expansion (1.3) is restated as hypotheses: uniform absolute convergence on every compact K⊂ΔK \subset \DeltaK⊂Δ, and HasSum of the terms to f(z)f(z)f(z) at every z∈Δz \in \Deltaz∈Δ.

Preamble
import Mathlib
import Definitions.Def_LeblSCV_Shared_polydisc
import Definitions.Def_LeblSCV_Holomorphic_distinguishedBoundary
import Definitions.Def_LeblSCV_Holomorphic_IsHolomorphicOn
import Definitions.Def_LeblSCV_Holomorphic_wirtingerIter
import Definitions.Def_LeblSCV_Holomorphic_powerSeriesTerm

open Complex
Formal statement
namespace LeblSCV.Holomorphic

/-- Proposition 1.2.2 (Lebl, p. 22). Let `Δ = Δ_ρ(a)` be a polydisc with distinguished boundary `Γ`,
and `f` continuous on `closure Δ` and holomorphic in `Δ`. Then:
(1) for `z ∈ Δ` and every `α ∈ ℕ₀ⁿ`,
`∂^{|α|} f / ∂z^α (z) = (2πi)^{-n} ∫_Γ α! f(ζ) / (ζ - z)^{α + 1} dζ`;
(2) if `f(z) = ∑_α c_α (z - a)^α` on `Δ` with uniform absolute convergence on compact subsets
of `Δ` (expansion (1.3)), then `c_α = (1/α!) ∂^{|α|} f / ∂z^α (a)`, and the Cauchy estimates
`|c_α| ≤ ‖f‖_Γ / ρ^α` hold, stated here for every bound `M` of `|f|` on `Γ` (the least such
`M` being `‖f‖_Γ = sup_Γ |f|`). Here `α! = α_1! ⋯ α_n!`, `ρ^α = ρ_1^{α_1} ⋯ ρ_n^{α_n}`,
`(ζ - z)^{α+1} = ∏_k (ζ_k - z_k)^{α_k + 1}`. -/
theorem cauchy_derivative_formula_and_estimates {n : ℕ} (a : Fin n → ℂ) (ρ : Fin n → ℝ)
    (hρ : ∀ k, 0 < ρ k) {f : (Fin n → ℂ) → ℂ}
    (hcont : ContinuousOn f (closure (LeblSCV.Shared.polydisc a ρ)))
    (hf : IsHolomorphicOn f (LeblSCV.Shared.polydisc a ρ)) :
    (∀ z ∈ LeblSCV.Shared.polydisc a ρ, ∀ α : Fin n → ℕ,
      wirtingerIter α f z = ((2 * Real.pi * I) ^ n)⁻¹ *
        torusIntegral (fun ζ => ((∏ k : Fin n, (α k).factorial : ℕ) : ℂ) * f ζ /
          ∏ k : Fin n, (ζ k - z k) ^ (α k + 1)) a ρ) ∧
    (∀ c : (Fin n → ℕ) → ℂ,
      (∀ K ⊆ LeblSCV.Shared.polydisc a ρ, IsCompact K → ConvergesUniformlyAbsolutelyOn c a K) →
      (∀ z ∈ LeblSCV.Shared.polydisc a ρ, HasSum (fun α => powerSeriesTerm c a α z) (f z)) →
      (∀ α : Fin n → ℕ,
        c α = (((∏ k : Fin n, (α k).factorial : ℕ) : ℂ))⁻¹ * wirtingerIter α f a) ∧
      (∀ M : ℝ, (∀ ζ ∈ distinguishedBoundary a ρ, ‖f ζ‖ ≤ M) →
        ∀ α : Fin n → ℕ, ‖c α‖ ≤ M / ∏ k : Fin n, ρ k ^ α k)) := by sorry

end LeblSCV.Holomorphic
Source
Lebl, Tasty Bits of Several Complex Variables, version 4.4 (2026), p. 22, Proposition 1.2.2
Read-back

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

Setup. Let n∈Nn \in \mathbb{N}n∈N, a∈Cna \in \mathbb{C}^na∈Cn, and ρ∈Rn\rho \in \mathbb{R}^nρ∈Rn with ρk>0\rho_k > 0ρk​>0 for all kkk. Let Δ\DeltaΔ be the set given by the imported definition LeblSCV.Shared.polydisc a ρ. Its code is not included in the supplied text, so its exact membership condition cannot be stated here. Let Δ‾\overline{\Delta}Δ be its closure.

Hypotheses. Let f:Cn→Cf : \mathbb{C}^n \to \mathbb{C}f:Cn→C be:

  • continuous on Δ‾\overline{\Delta}Δ; and
  • holomorphic on Δ\DeltaΔ: locally bounded relative to Δ\DeltaΔ near each point of Δ\DeltaΔ, and at each z∈Δz \in \Deltaz∈Δ complex-differentiable in each coordinate separately (the map ξ↦f(z+ξek)\xi \mapsto f(z + \xi e_k)ξ↦f(z+ξek​) is complex-differentiable at 000).

Notation. For a multi-index α∈Nn\alpha \in \mathbb{N}^nα∈Nn:

  • α!=α1!⋯αn!\alpha! = \alpha_1!\cdots\alpha_n!α!=α1​!⋯αn​!, computed in N\mathbb{N}N and viewed in C\mathbb{C}C.
  • Dαf=W1α1(W2α2(⋯Wnαnf))D^\alpha f = W_1^{\alpha_1}(W_2^{\alpha_2}(\cdots W_n^{\alpha_n} f))Dαf=W1α1​​(W2α2​​(⋯Wnαn​​f)), with WnW_nWn​ applied innermost. Here (Wkg)(z)=12(ddt∣0 g(z+tek)−i ddt∣0 g(z+itek))(W_k g)(z) = \tfrac12\big(\tfrac{d}{dt}|_{0}\, g(z + te_k) - i\,\tfrac{d}{dt}|_{0}\, g(z + i t e_k)\big)(Wk​g)(z)=21​(dtd​∣0​g(z+tek​)−idtd​∣0​g(z+itek​)) for real ttt, and each derivative is replaced by 000 if it does not exist.
  • For h:Cn→Ch : \mathbb{C}^n \to \mathbb{C}h:Cn→C,
I(h)=∫[0,2π]n(∏kiρkeiθk) h(a1+ρ1eiθ1,…,an+ρneiθn) dθ.I(h) = \int_{[0,2\pi]^n} \Big(\prod_k i\rho_k e^{i\theta_k}\Big)\, h(a_1+\rho_1e^{i\theta_1},\dots,a_n+\rho_ne^{i\theta_n})\,d\theta.I(h)=∫[0,2π]n​(k∏​iρk​eiθk​)h(a1​+ρ1​eiθ1​,…,an​+ρn​eiθn​)dθ.

This is a Lebesgue/Bochner integral, and it equals 000 if the integrand is not integrable. Division by 000 in C\mathbb{C}C yields 000.

  • Γ={ζ:∣ζk−ak∣=ρk ∀k}\Gamma = \{\zeta : |\zeta_k - a_k| = \rho_k \ \forall k\}Γ={ζ:∣ζk​−ak​∣=ρk​ ∀k}.
  • Tα(z)=cα∏k(zk−ak)αkT_\alpha(z) = c_\alpha\prod_k (z_k - a_k)^{\alpha_k}Tα​(z)=cα​∏k​(zk​−ak​)αk​, with 00=10^0 = 100=1.

The statement asserts both parts below.

Part (1). For every z∈Δz \in \Deltaz∈Δ and every α∈Nn\alpha \in \mathbb{N}^nα∈Nn,

(Dαf)(z)=((2πi)n)−1 I ⁣(ζ↦α! f(ζ)∏k(ζk−zk)αk+1).(D^\alpha f)(z) = \big((2\pi i)^n\big)^{-1}\, I\!\left(\zeta \mapsto \frac{\alpha!\, f(\zeta)}{\prod_{k}(\zeta_k - z_k)^{\alpha_k+1}}\right).(Dαf)(z)=((2πi)n)−1I(ζ↦∏k​(ζk​−zk​)αk​+1α!f(ζ)​).

Part (2). Take any coefficient family c:Nn→Cc : \mathbb{N}^n \to \mathbb{C}c:Nn→C satisfying two hypotheses:

  • (i) For every compact K⊆ΔK \subseteq \DeltaK⊆Δ, the series converges uniformly absolutely on KKK. That is, there is S:Cn→RS : \mathbb{C}^n \to \mathbb{R}S:Cn→R such that for every ε>0\varepsilon>0ε>0 there is a finite s0⊆Nns_0 \subseteq \mathbb{N}^ns0​⊆Nn with ∣S(z)−∑α∈s∣Tα(z)∣∣<ε\big|S(z) - \sum_{\alpha\in s}|T_\alpha(z)|\big| < \varepsilon​S(z)−∑α∈s​∣Tα​(z)∣​<ε for all finite s⊇s0s \supseteq s_0s⊇s0​ and all z∈Kz \in Kz∈K.
  • (ii) For every z∈Δz \in \Deltaz∈Δ, the family (Tα(z))α(T_\alpha(z))_{\alpha}(Tα​(z))α​ is unconditionally summable with sum f(z)f(z)f(z). That is, its finite partial sums, directed by inclusion, converge to f(z)f(z)f(z).

Then:

  • (a) for every α\alphaα,   cα=(α!)−1 (Dαf)(a)\;c_\alpha = (\alpha!)^{-1}\,(D^\alpha f)(a)cα​=(α!)−1(Dαf)(a);
  • (b) for every real MMM such that ∣f(ζ)∣≤M|f(\zeta)| \le M∣f(ζ)∣≤M for all ζ∈Γ\zeta \in \Gammaζ∈Γ, and every α\alphaα,
∣cα∣≤M∏kρkαk.|c_\alpha| \le \frac{M}{\prod_k \rho_k^{\alpha_k}}.∣cα​∣≤∏k​ρkαk​​M​.

Degenerate cases.

  • Since ρk>0\rho_k > 0ρk​>0, Γ\GammaΓ is nonempty, so any admissible MMM is ≥0\ge 0≥0. The denominator ∏kρkαk\prod_k \rho_k^{\alpha_k}∏k​ρkαk​​ is positive, so no division by zero occurs in (b).
  • Whether a∈Δa \in \Deltaa∈Δ is not visible from the supplied code. Part (a) evaluates DαfD^\alpha fDαf at aaa regardless.
  • If n=0n = 0n=0: there is only the empty multi-index, with α!=1\alpha! = 1α!=1 and Dαf=fD^\alpha f = fDαf=f. The integral reduces to evaluation at the unique point, so Part (1) reads f(z)=f(z)f(z) = f(z)f(z)=f(z). Part (2) asserts c∅=f(a)c_\emptyset = f(a)c∅​=f(a) and ∣c∅∣≤M|c_\emptyset| \le M∣c∅​∣≤M.
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