Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Exponential decay of derivatives from quadratically many zeros

Proved
TranscendenceTheory.finite_zeros_exponential_derivative_bound

by tomasz · Sep 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

asymptoticscomplex-analysiselliptic-functionstranscendence

Fix real constants α,β,K>0\alpha,\beta,K>0α,β,K>0 and B≥0B\ge0B≥0. For all sufficiently large positive integers NNN, the following holds uniformly over all the remaining data.

Let GGG be an entire complex function, let S⊂CS\subset\mathbb CS⊂C be finite, and prescribe nonnegative integer multiplicities mam_ama​ at its points. Suppose

G(j)(a)=0(a∈S, 0≤j<ma),E:=∑a∈Sma≥αN2.G^{(j)}(a)=0\quad(a\in S,\ 0\le j<m_a),\qquad E:=\sum_{a\in S}m_a\ge\alpha N^2.G(j)(a)=0(a∈S, 0≤j<ma​),E:=a∈S∑​ma​≥αN2.

Let 0<r<R0<r<R0<r<R, and assume

∣a∣≤r(a∈S),2rR≤N−β,max⁡∣z∣=R∣G(z)∣≤eBN2.|a|\le r\quad(a\in S),\qquad \frac{2r}{R}\le N^{-\beta},\qquad \max_{|z|=R}|G(z)|\le e^{BN^2}.∣a∣≤r(a∈S),R2r​≤N−β,∣z∣=Rmax​∣G(z)∣≤eBN2.

For every complex www and nonnegative integer nnn satisfying

∣w∣+1≤r,n≤KNlog⁡N,|w|+1\le r,\qquad n\le K\frac{N}{\log N},∣w∣+1≤r,n≤KlogNN​,

one has

∣G(n)(w)∣≤exp⁡ ⁣(−αβ2N2log⁡N).|G^{(n)}(w)|\le \exp\!\left(-\frac{\alpha\beta}{2}N^2\log N\right).∣G(n)(w)∣≤exp(−2αβ​N2logN).

Moreover, suppose fff and ψ\psiψ are analytic near www, with G=ψfG=\psi fG=ψf there, and

f(j)(w)=0(0≤j<n),ψ(w)≠0,∣ψ(w)∣−1≤eBN2.f^{(j)}(w)=0\quad(0\le j<n),\qquad \psi(w)\ne0,\qquad |\psi(w)|^{-1}\le e^{BN^2}.f(j)(w)=0(0≤j<n),ψ(w)=0,∣ψ(w)∣−1≤eBN2.

Then the same bound holds before regularization:

∣f(n)(w)∣≤exp⁡ ⁣(−αβ2N2log⁡N).|f^{(n)}(w)|\le \exp\!\left(-\frac{\alpha\beta}{2}N^2\log N\right).∣f(n)(w)∣≤exp(−2αβ​N2logN).

The threshold depends only on the four fixed constants. In particular, the functions, nodes, multiplicities, radii, evaluation point, and derivative order may all vary with NNN. Order zero, zero multiplicities, and the identically zero function are allowed. The statement does not assert that the bounded derivative is nonzero. This is an abstract supporting estimate for the source's passage from Schwarz--Cauchy bounds to exponentially small auxiliary values.

Preamble
import Mathlib.Analysis.Analytic.Order
import Mathlib.Analysis.SpecialFunctions.Pow.Real

open Filter Metric Set
open scoped Topology
Formal statement
theorem TranscendenceTheory.finite_zeros_exponential_derivative_bound (α β B K : ℝ)
    (hα : 0 < α) (hβ : 0 < β) (hB : 0 ≤ B) (hK : 0 < K) :
    ∀ᶠ N : ℕ in atTop,
      ∀ (G : ℂ → ℂ), AnalyticOnNhd ℂ G univ →
      ∀ (s : Finset ℂ) (m : ℂ → ℕ),
        (∀ a ∈ s, ∀ j < m a, iteratedDeriv j G a = 0) →
      ∀ r R : ℝ, 0 < r → r < R →
        (∀ a ∈ s, ‖a‖ ≤ r) →
        α * (N : ℝ) ^ 2 ≤ (∑ a ∈ s, m a : ℕ) →
        2 * r / R ≤ (N : ℝ) ^ (-β) →
        (∀ z ∈ sphere (0 : ℂ) R, ‖G z‖ ≤ Real.exp (B * (N : ℝ) ^ 2)) →
      ∀ (w : ℂ), ‖w‖ + 1 ≤ r →
      ∀ n : ℕ, (n : ℝ) ≤ K * ((N : ℝ) / Real.log N) →
        ‖iteratedDeriv n G w‖ ≤
          Real.exp (-(α * β / 2) * (N : ℝ) ^ 2 * Real.log N) ∧
        ∀ f ψ : ℂ → ℂ, AnalyticAt ℂ f w → AnalyticAt ℂ ψ w →
          G =ᶠ[𝓝 w] (fun z => ψ z * f z) →
          (∀ j < n, iteratedDeriv j f w = 0) → ψ w ≠ 0 →
          ‖ψ w‖⁻¹ ≤ Real.exp (B * (N : ℝ) ^ 2) →
          ‖iteratedDeriv n f w‖ ≤
            Real.exp (-(α * β / 2) * (N : ℝ) ^ 2 * Real.log N) := by sorry
Source
Abstract supporting estimate for Senthil Kumar K (2026), Lemma 6 and Section 5, equations (30)-(35), especially the factorial estimate in (30), the radius choice (32), and the resulting small derivative estimates (33),(35). General constants alpha,beta,B,K and the factor 1/2 are explicit formalization choices; this is not a verbatim source theorem. https://doi.org/10.1017/S001309152610145X

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