Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Cancellation of the denominator-generated linear term

Open
QAlgorithms.Papers.Independent.Intakenext100.P22.denominator_cancellation

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Fix an integer n≥2n\ge2n≥2. For 0<x<10<x<10<x<1, put

βy(x)=2F1(y,1−y;1;1−x)2F1(y,1−y;1;x),Γrs(x)=2in∑k=1n−1sin⁡(πk/n)βk/n(x)cos⁡(2πk(r−s)/n),\beta_y(x)=\frac{{}_2F_1(y,1-y;1;1-x)}{{}_2F_1(y,1-y;1;x)},\qquad \Gamma_{rs}(x)=\frac{2i}{n}\sum_{k=1}^{n-1}\sin(\pi k/n)\beta_{k/n}(x)\cos(2\pi k(r-s)/n),βy​(x)=2​F1​(y,1−y;1;x)2​F1​(y,1−y;1;1−x)​,Γrs​(x)=n2i​k=1∑n−1​sin(πk/n)βk/n​(x)cos(2πk(r−s)/n),

where r,s=1,…,n−1r,s=1,\ldots,n-1r,s=1,…,n−1. Let Dn(x)=Θ(0∣Γ(x))D_n(x)=\Theta(0\mid\Gamma(x))Dn​(x)=Θ(0∣Γ(x)), a positive real number, and an=(n−1/n)/6a_n=(n-1/n)/6an​=(n−1/n)/6. With nnn fixed and x→0+x\to0^+x→0+,

Dn(x)=1+12anx+O(x2),Dn(x)2=1+anx+O(x2),(1−x)−anDn(x)2=1+O(x2).D_n(x)=1+\tfrac12a_nx+O(x^2),\qquad D_n(x)^2=1+a_nx+O(x^2),\qquad \frac{(1-x)^{-a_n}}{D_n(x)^2}=1+O(x^2).Dn​(x)=1+21​an​x+O(x2),Dn​(x)2=1+an​x+O(x2),Dn​(x)2(1−x)−an​​=1+O(x2).

Thus the geometric factor cancels the linear term generated by this denominator. The common interval-length power has been factored out. This assertion places no restriction on possible order-xxx terms generated by numerator sectors at special compactification radii.

Formalization Note DnD_nDn​ is represented by the norm of the theta series, which equals its value on this pure imaginary period matrix. The error bounds use the right-hand neighborhood filter; their constants may depend on the fixed nnn.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P22
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P22

open scoped BigOperators Topology

open Filter MeasureTheory

/-- PDF pp. 15–16, (68)–(71): denominator-generated linear term cancels
with the geometric prefactor after factoring out the common interval-length
power. No exclusion of resonant order-x numerator sectors is asserted.
The limit is from positive x with fixed integer n≥2. -/
theorem denominator_cancellation (n : ℕ) (hn : 2 ≤ n) :
    Asymptotics.IsBigO (𝓝[>] (0 : ℝ))
      (fun x => thetaDenominator n x - (1 + replicaWeight n / 2 * x))
      (fun x => x ^ 2) ∧
    Asymptotics.IsBigO (𝓝[>] (0 : ℝ))
      (fun x => thetaDenominator n x ^ 2 - (1 + replicaWeight n * x))
      (fun x => x ^ 2) ∧
    Asymptotics.IsBigO (𝓝[>] (0 : ℝ))
      (fun x => Real.rpow (1 - x) (-replicaWeight n) /
        thetaDenominator n x ^ 2 - 1) (fun x => x ^ 2) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P22

end
Source
Entanglement entropy of two disjoint intervals in conformal field theory II, arXiv:1011.5482v1, PDF p. 16, §5.2, equations (68)–(71)
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

For every natural number n≥2n\ge2n≥2, put d=n−1d=n-1d=n−1, w=(n−1/n)/6w=(n-1/n)/6w=(n−1/n)/6, and, for real xxx, define

Prs(x)=2in∑1≤k<nsin⁡ ⁣(πkn)2F1(k/n,1−k/n;1;1−x)2F1(k/n,1−k/n;1;x)cos⁡ ⁣(2πkn(r−s))(0≤r,s<d),P_{rs}(x)=\frac{2\mathrm i}{n} \sum_{1\le k<n} \sin\!\left(\frac{\pi k}{n}\right) \frac{{}_2F_1(k/n,1-k/n;1;1-x)} {{}_2F_1(k/n,1-k/n;1;x)} \cos\!\left(\frac{2\pi k}{n}(r-s)\right) \quad(0\le r,s<d),Prs​(x)=n2i​1≤k<n∑​sin(nπk​)2​F1​(k/n,1−k/n;1;x)2​F1​(k/n,1−k/n;1;1−x)​cos(n2πk​(r−s))(0≤r,s<d),

and

D(x)=∣∑m∈Zdexp⁡ ⁣(iπ∑r,s=0d−1mrPrs(x)ms)∣.D(x)= \left| \sum_{m\in\mathbb Z^d} \exp\!\left(\mathrm i\pi\sum_{r,s=0}^{d-1}m_rP_{rs}(x)m_s\right) \right|.D(x)=​m∈Zd∑​exp(iπr,s=0∑d−1​mr​Prs​(x)ms​)​.

Here 2F1{}_2F_12​F1​ is the real ordinary hypergeometric function, the modulus is complex modulus, a nonsummable indexed sum is assigned value 000, and division by zero is assigned value 000. The theorem asserts all three estimates

D(x)−(1+w2x)=O(x2),D(x)2−(1+wx)=O(x2),(1−x)−wD(x)2−1=O(x2)as x→0+,D(x)-\left(1+\frac w2x\right)=O(x^2),\qquad D(x)^2-(1+wx)=O(x^2),\qquad \frac{(1-x)^{-w}}{D(x)^2}-1=O(x^2) \quad\text{as }x\to0^+,D(x)−(1+2w​x)=O(x2),D(x)2−(1+wx)=O(x2),D(x)2(1−x)−w​−1=O(x2)as x→0+,

where the power in the third expression is real exponentiation. Precisely, for each of the three displayed error expressions E(x)E(x)E(x), there exist constants C≥0C\ge0C≥0 and δ>0\delta>0δ>0, which may depend on nnn and on the expression, such that ∣E(x)∣≤Cx2|E(x)|\le Cx^2∣E(x)∣≤Cx2 for every real xxx with 0<x<δ0<x<\delta0<x<δ. The claim concerns approach through positive xxx and makes no assertion at x=0x=0x=0; its hypothesis excludes n=0n=0n=0 and n=1n=1n=1.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 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