Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Analytic extraction with sigma regularizer inverse bounds supplied

Proved
WeierstrassEllipticZeta.exists_complex_auxiliary_systems_from_sigma_inverse_bounds

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

complex-analysiselliptic-functionstranscendence

Let LLL be a complex period pair, and let ω,u1,u2∈C\omega,u_1,u_2\in\mathbb Cω,u1​,u2​∈C have regular auxiliary-grid data. Thus integer grid points are distinct, their lattice congruences are determined by their first two coordinates, and every point shifted by u1/2u_1/2u1​/2 is outside the lattice. The finite grids have cardinality A1A2A3A_1A_2A_3A1​A2​A3​ and shifted radius at most

A1∣u1∣+A2∣u2∣+A3∣ω∣+∣u1∣/2.A_1|u_1|+A_2|u_2|+A_3|\omega|+|u_1|/2.A1​∣u1​∣+A2​∣u2​∣+A3​∣ω∣+∣u1​∣/2.

The data also include the elliptic periodicity formulas and

ζ(z+nω)=ζ(z)+nη(ω)(n∈Z, z∉Ω).\zeta(z+n\omega)=\zeta(z)+n\eta(\omega) \qquad(n\in\mathbb Z,\ z\notin\Omega).ζ(z+nω)=ζ(z)+nη(ω)(n∈Z, z∈/Ω).

Assume the canonical zeta derivative identity and the multiplied elliptic and zeta addition identities at regular arguments. Let θ\thetaθ be transcendental over Q\mathbb QQ, let ν∈C\nu\in\mathbb Cν∈C, and let g∈Z[X,Y]g\in\mathbb Z[X,Y]g∈Z[X,Y] be monic of positive YYY-degree rrr with

p(θ,ν)=0  ⟺  g∣p.p(\theta,\nu)=0\iff g\mid p.p(θ,ν)=0⟺g∣p.

For a fixed d∈Z[X]d\in\mathbb Z[X]d∈Z[X] with d(θ)≠0d(\theta)\ne0d(θ)=0, assume the eighteen auxiliary values admit presentations of YYY-degree less than rrr after multiplication by d(θ)d(\theta)d(θ). Those values are

g2/4,g3/4,ω,η(ω),u1/2,u2,ζ(u1/2),℘(u1/2),℘′(u1/2),℘′′(u1/2),g_2/4,g_3/4,\omega,\eta(\omega),u_1/2,u_2, \zeta(u_1/2),\wp(u_1/2),\wp'(u_1/2),\wp''(u_1/2),g2​/4,g3​/4,ω,η(ω),u1​/2,u2​,ζ(u1​/2),℘(u1​/2),℘′(u1​/2),℘′′(u1​/2), ℘(uj),℘′(uj),℘′′(uj),ζ(uj)(j=1,2).\wp(u_j),\wp'(u_j),\wp''(u_j),\zeta(u_j)\quad(j=1,2).℘(uj​),℘′(uj​),℘′′(uj​),ζ(uj​)(j=1,2).

Assume reduced arithmetic jet-system data as follows.

Fix a period pair LLL, complex numbers θ,ν\theta,\nuθ,ν, and g∈Z[X][Y]g\in\mathbb Z[X][Y]g∈Z[X][Y]. For a polynomial, ℓ\ellℓ denotes the sum of the absolute values of all integer coefficients.

Reduced arithmetic jet-system data mean that there exist positive integers B,HB,HB,H with the following property. For any nonnegative M,L0,TM,L_0,TM,L0​,T, let

I={0,…,L0}×{0,…,M}2,k(n)=(L0,5M,5M,5M,Kn,Kn,Kn,Kn),Kn=L0+5M+n.I=\{0,\ldots,L_0\}\times\{0,\ldots,M\}^2,\qquad k(n)=(L_0,5M,5M,5M,K_n,K_n,K_n,K_n),\quad K_n=L_0+5M+n.I={0,…,L0​}×{0,…,M}2,k(n)=(L0​,5M,5M,5M,Kn​,Kn​,Kn​,Kn​),Kn​=L0​+5M+n.

Choose eight numerator and denominator polynomials Sa,Qa∈Z[X,Y]S_a,Q_a\in\mathbb Z[X,Y]Sa​,Qa​∈Z[X,Y], whose total degrees are at most dad_ada​ and lengths at most hah_aha​, respectively. There exists a matrix

R=(Rn,i)0≤n<Ti∈IR=(R_{n,i})_{\substack{0\le n<T\\i\in I}}R=(Rn,i​)0≤n<Ti∈I​​

of integer bivariate polynomials. Writing Dn=∑a=07ka(n)daD_n=\sum_{a=0}^7 k_a(n)d_aDn​=∑a=07​ka​(n)da​, every entry satisfies

deg⁡YRn,i<deg⁡Yg,deg⁡X[Yj]Rn,i≤BDn(j≥0),\deg_Y R_{n,i}<\deg_Yg,\qquad \deg_X[Y^j]R_{n,i}\le BD_n\quad(j\ge0),degY​Rn,i​<degY​g,degX​[Yj]Rn,i​≤BDn​(j≥0), ℓ(Rn,i)≤n! 241(L0+M+n)(∏a=07haka(n))HDn+1.\ell(R_{n,i})\le n!\,2^{41(L_0+M+n)}\left(\prod_{a=0}^7h_a^{k_a(n)}\right)H^{D_n+1}.ℓ(Rn,i​)≤n!241(L0​+M+n)(a=0∏7​haka​(n)​)HDn​+1.

This matrix precedes the following points and coefficient vectors. For any v,z,z+vv,z,z+vv,z,z+v outside the lattice, suppose the coordinate presentations satisfy

Sa(θ,ν)=Qa(θ,ν)Jv(z)a,S_a(\theta,\nu)=Q_a(\theta,\nu)J_v(z)_a,Sa​(θ,ν)=Qa​(θ,ν)Jv​(z)a​,

where

Jv(z)=(z+v,ζ(v),℘(v),℘′(v),ζ(z),℘(z),℘′(z),℘′′(z)).J_v(z)=(z+v,\zeta(v),\wp(v),\wp'(v),\zeta(z),\wp(z),\wp'(z),\wp''(z)).Jv​(z)=(z+v,ζ(v),℘(v),℘′(v),ζ(z),℘(z),℘′(z),℘′′(z)).

For i=(i0,i2,i3)i=(i_0,i_2,i_3)i=(i0​,i2​,i3​), set

Gi(w)=(w+v)i0[2(℘(v)−℘(w))]3M℘(w+v)i2ζ(w+v)i3,Δn=∏a=07Qa(θ,ν)ka(n).G_i(w)=(w+v)^{i_0}[2(\wp(v)-\wp(w))]^{3M} \wp(w+v)^{i_2}\zeta(w+v)^{i_3},\qquad \Delta_n=\prod_{a=0}^7Q_a(\theta,\nu)^{k_a(n)}.Gi​(w)=(w+v)i0​[2(℘(v)−℘(w))]3M℘(w+v)i2​ζ(w+v)i3​,Δn​=a=0∏7​Qa​(θ,ν)ka​(n).

The exact entry evaluations are

Rn,i(θ,ν)=ΔnGi(n)(z).R_{n,i}(\theta,\nu)=\Delta_nG_i^{(n)}(z).Rn,i​(θ,ν)=Δn​Gi(n)​(z).

If additionally z−vz-vz−v is regular and all eight evaluated denominators are nonzero, then for every complex vector c=(ci)i∈Ic=(c_i)_{i\in I}c=(ci​)i∈I​,

(∀ 0≤n<T, ∑iRn,i(θ,ν)ci=0) ⟺ (∀ 0≤n<T, Fc(n)(z+v)=0),\left(\forall\,0\le n<T,\ \sum_iR_{n,i}(\theta,\nu)c_i=0\right) \ \Longleftrightarrow\ \left(\forall\,0\le n<T,\ F_c^{(n)}(z+v)=0\right),(∀0≤n<T, i∑​Rn,i​(θ,ν)ci​=0) ⟺ (∀0≤n<T, Fc(n)​(z+v)=0),

where

Fc(w)=∑i∈Iciwi0℘(w)i2ζ(w)i3.F_c(w)=\sum_{i\in I}c_iw^{i_0}\wp(w)^{i_2}\zeta(w)^{i_3}.Fc​(w)=i∈I∑​ci​wi0​℘(w)i2​ζ(w)i3​.

The exact evaluation identities also hold with zero coordinate denominators; nonzero denominators are required only for kernel equivalence. Constants are uniform in all cutoffs, orders, and presentations. The property assembles bounded reduced derivative presentations into finite linear equations. It does not produce grid-coordinate presentations, prove that their denominators are nonzero, choose asymptotic parameters, or establish a small nonzero test value.

Also assume auxiliary-grid interpolation data as follows.

Fix complex numbers ω,u1,u2\omega,u_1,u_2ω,u1​,u2​. For a triple of nonnegative integers A=(A1,A2,A3)A=(A_1,A_2,A_3)A=(A1​,A2​,A3​), let

SA={a1u1+a2u2+a3ω+u1/2:0≤ai<Ai, ai∈Z},S_A=\{a_1u_1+a_2u_2+a_3\omega+u_1/2: 0\le a_i<A_i,\ a_i\in\mathbb Z\},SA​={a1​u1​+a2​u2​+a3​ω+u1​/2:0≤ai​<Ai​, ai​∈Z}, qA=A1∣u1∣+A2∣u2∣+A3∣ω∣+∣u1∣/2.q_A=A_1|u_1|+A_2|u_2|+A_3|\omega|+|u_1|/2.qA​=A1​∣u1​∣+A2​∣u2​∣+A3​∣ω∣+∣u1​∣/2.

Auxiliary-grid interpolation data mean the following property. Choose a nonnegative integer TTT, functions f,G,ψ:C→Cf,G,\psi:\mathbb C\to\mathbb Cf,G,ψ:C→C, with GGG entire and f,ψf,\psif,ψ analytic at every point of SAS_ASA​. Suppose G=ψfG=\psi fG=ψf in a neighborhood of each grid point and

f(j)(x)=0(x∈SA, 0≤j<T).f^{(j)}(x)=0\qquad(x\in S_A,\ 0\le j<T).f(j)(x)=0(x∈SA​, 0≤j<T).

For all 0<r<R0<r<R0<r<R and C∈RC\in\mathbb RC∈R with qA≤rq_A\le rqA​≤r and ∣G(z)∣≤C|G(z)|\le C∣G(z)∣≤C on ∣z∣=R|z|=R∣z∣=R, put NA=TA1A2A3N_A=TA_1A_2A_3NA​=TA1​A2​A3​. Then

∣G(z)∣≤C(2rR)NA(∣z∣≤r),|G(z)|\le C\left(\frac{2r}{R}\right)^{N_A}\qquad(|z|\le r),∣G(z)∣≤C(R2r​)NA​(∣z∣≤r),

and, whenever ρ>0\rho>0ρ>0 and ∣w∣+ρ≤r|w|+\rho\le r∣w∣+ρ≤r,

∣G(n)(w)∣≤n!ρn C(2rR)NA(n≥0).|G^{(n)}(w)|\le\frac{n!}{\rho^n}\,C \left(\frac{2r}{R}\right)^{N_A}\qquad(n\ge0).∣G(n)(w)∣≤ρnn!​C(R2r​)NA​(n≥0).

The multiplier need not be nonzero. Empty grids and T=0T=0T=0 are included. This predicate records the effect of analytic regularization and interpolation once the regularized entire function and its outer-circle bound have been provided. It does not assert existence or growth estimates for a Weierstrass sigma regularizer.

Also assume the following elliptic regularization data.

Fix a period pair LLL and complex numbers ω,u1,u2\omega,u_1,u_2ω,u1​,u2​. Write

Φ0=ζL,Φ1=℘L,Φ2=℘L′.\Phi_0=\zeta_L,\qquad\Phi_1=\wp_L,\qquad\Phi_2=\wp'_L.Φ0​=ζL​,Φ1​=℘L​,Φ2​=℘L′​.

Elliptic regularization data mean the following property. Supply entire functions σ,S0,S1,S2\sigma,S_0,S_1,S_2σ,S0​,S1​,S2​ satisfying

Sj(z)=σ(z)j+1Φj(z)(z∉L, 0≤j≤2).S_j(z)=\sigma(z)^{j+1}\Phi_j(z)\qquad(z\notin L,\ 0\le j\le2).Sj​(z)=σ(z)j+1Φj​(z)(z∈/L, 0≤j≤2).

For any finite index set III, complex shift vvv, complex coefficients cic_ici​, and nonnegative integers ℓi,eij,D,K\ell_i,e_{ij},D,Kℓi​,eij​,D,K with

ℓi≤D,ei0+2ei1+3ei2≤K,\ell_i\le D,\qquad e_{i0}+2e_{i1}+3e_{i2}\le K,ℓi​≤D,ei0​+2ei1​+3ei2​≤K,

put

f(z)=∑ici(z+v)ℓi∏j=02Φj(z)eij.f(z)=\sum_i c_i(z+v)^{\ell_i}\prod_{j=0}^2\Phi_j(z)^{e_{ij}}.f(z)=i∑​ci​(z+v)ℓi​j=0∏2​Φj​(z)eij​.

There is an entire function GGG, chosen independently of the radius, grid, and bounds, with G=σKfG=\sigma^KfG=σKf outside the lattice. If B≥1B\ge1B≥1 bounds the absolute values of σ,S0,S1,S2\sigma,S_0,S_1,S_2σ,S0​,S1​,S2​ throughout ∣z∣≤R|z|\le R∣z∣≤R, define

CR=(∑i∣ci∣)max⁡(1,R+∣v∣)DBK.C_R=\left(\sum_i|c_i|\right)\max(1,R+|v|)^D B^K.CR​=(i∑​∣ci​∣)max(1,R+∣v∣)DBK.

Then ∣G(z)∣≤CR|G(z)|\le C_R∣G(z)∣≤CR​ on ∣z∣≤R|z|\le R∣z∣≤R. Moreover, choose nonnegative integers A1,A2,A3,TA_1,A_2,A_3,TA1​,A2​,A3​,T and radii 0<r<R0<r<R0<r<R such that

A1∣u1∣+A2∣u2∣+A3∣ω∣+∣u1∣/2≤r.A_1|u_1|+A_2|u_2|+A_3|\omega|+|u_1|/2\le r.A1​∣u1​∣+A2​∣u2​∣+A3​∣ω∣+∣u1​∣/2≤r.

Suppose all derivatives of fff below order TTT vanish on

{a1u1+a2u2+a3ω+u1/2:0≤ai<Ai}.\{a_1u_1+a_2u_2+a_3\omega+u_1/2:0\le a_i<A_i\}.{a1​u1​+a2​u2​+a3​ω+u1​/2:0≤ai​<Ai​}.

Then, with Q=TA1A2A3Q=TA_1A_2A_3Q=TA1​A2​A3​,

∣G(z)∣≤CR(2r/R)Q(∣z∣≤r),|G(z)|\le C_R(2r/R)^Q\qquad(|z|\le r),∣G(z)∣≤CR​(2r/R)Q(∣z∣≤r), ∣G(n)(z)∣≤n!ρnCR(2r/R)Q(ρ>0, ∣z∣+ρ≤r, n≥0).|G^{(n)}(z)|\le \frac{n!}{\rho^n}C_R(2r/R)^Q \qquad(\rho>0,\ |z|+\rho\le r,\ n\ge0).∣G(n)(z)∣≤ρnn!​CR​(2r/R)Q(ρ>0, ∣z∣+ρ≤r, n≥0).

The definition supplies the analytic extension and its interpolation consequences conditional on the four entire factors and their bounds. It does not assert existence of those factors or an order-two growth bound for them. At lattice points the entire extension need not equal the product formed from totalized meromorphic values.

Also assume cleared-addition entire data, as follows.

Let LLL be a complex period pair, with lattice Ω\OmegaΩ and canonical functions ζ,℘,℘′\zeta,\wp,\wp'ζ,℘,℘′. Assume their multiplied addition identities at regular arguments. Supply entire functions σ,S0,S1,S2\sigma,S_0,S_1,S_2σ,S0​,S1​,S2​ satisfying

S0=σζ,S1=σ2℘,S2=σ3℘′outside Ω.S_0=\sigma\zeta,\qquad S_1=\sigma^2\wp,\qquad S_2=\sigma^3\wp' \qquad\text{outside }\Omega.S0​=σζ,S1​=σ2℘,S2​=σ3℘′outside Ω.

Let III be any finite index set, let v∉Ωv\notin\Omegav∈/Ω, and choose complex coefficients cic_ici​ and nonnegative integers ℓ0i,ℓ2i,ℓ3i,D,M\ell_{0i},\ell_{2i},\ell_{3i},D,Mℓ0i​,ℓ2i​,ℓ3i​,D,M with

ℓ0i≤D,ℓ2i,ℓ3i≤M.\ell_{0i}\le D,\qquad\ell_{2i},\ell_{3i}\le M.ℓ0i​≤D,ℓ2i​,ℓ3i​≤M.

Define the translated, cleared auxiliary sum

f(z)=∑i∈Ici(z+v)ℓ0i[2(℘(v)−℘(z))]3M℘(z+v)ℓ2iζ(z+v)ℓ3i.f(z)=\sum_{i\in I}c_i(z+v)^{\ell_{0i}} [2(\wp(v)-\wp(z))]^{3M}\wp(z+v)^{\ell_{2i}}\zeta(z+v)^{\ell_{3i}}.f(z)=i∈I∑​ci​(z+v)ℓ0i​[2(℘(v)−℘(z))]3M℘(z+v)ℓ2i​ζ(z+v)ℓ3i​.

There exists a single entire function GGG, chosen before any radius or bounds, with

G(z)=σ(z)15Mf(z)(z,z+v∉Ω).G(z)=\sigma(z)^{15M}f(z)\qquad(z,z+v\notin\Omega).G(z)=σ(z)15Mf(z)(z,z+v∈/Ω).

Put

Vv=1+∣ζ(v)∣+∣℘(v)∣+∣℘′(v)∣,Kv=36Vv3.V_v=1+|\zeta(v)|+|\wp(v)|+|\wp'(v)|,\qquad K_v=36V_v^3.Vv​=1+∣ζ(v)∣+∣℘(v)∣+∣℘′(v)∣,Kv​=36Vv3​.

For every R∈RR\in\mathbb RR∈R and B≥1B\ge1B≥1, if all four basic entire functions are bounded in absolute value by BBB on ∣z∣≤R|z|\le R∣z∣≤R, then

∣G(z)∣≤CR:=(∑i∣ci∣)max⁡(1,R+∣v∣)DKv15MB90M(∣z∣≤R).|G(z)|\le C_R:=\left(\sum_i|c_i|\right) \max(1,R+|v|)^D K_v^{15M}B^{90M}\qquad(|z|\le R).∣G(z)∣≤CR​:=(i∑​∣ci​∣)max(1,R+∣v∣)DKv15M​B90M(∣z∣≤R).

The finite set may be empty, coefficients may vanish, and D=M=0D=M=0D=M=0 is allowed. Neither σ\sigmaσ nor the addition factor is assumed nonzero. The identity is restricted to regular arguments; the entire extension supplies its own values at poles. This is a conditional version of the construction in Lemma 6(ii), with a coarse explicit growth constant. It does not construct the basic sigma factors or assert the source's sharper displayed numerical bound.

For fixed L,ω,u1,u2L,\omega,u_1,u_2L,ω,u1​,u2​, cleared-addition entire data assert the preceding entire-extension and disk-bound property for every choice of the four basic factors, regular shift, finite coefficients, and exponents. The data also assert the following interpolation consequences for the same entire function GGG.

Let A=(A1,A2,A3)A=(A_1,A_2,A_3)A=(A1​,A2​,A3​) and TTT be nonnegative integers. Suppose 0<r<R0<r<R0<r<R, B≥1B\ge1B≥1, the four factors are bounded by BBB on the closed radius-RRR disk, and

A1∣u1∣+A2∣u2∣+A3∣ω∣+∣u1∣/2≤r.A_1|u_1|+A_2|u_2|+A_3|\omega|+|u_1|/2\le r.A1​∣u1​∣+A2​∣u2​∣+A3​∣ω∣+∣u1​∣/2≤r.

Write

ΓA={a1u1+a2u2+a3ω+u1/2:0≤ai<Ai}.\Gamma_A=\{a_1u_1+a_2u_2+a_3\omega+u_1/2:0\le a_i<A_i\}.ΓA​={a1​u1​+a2​u2​+a3​ω+u1​/2:0≤ai​<Ai​}.

Assume x+v∉Ωx+v\notin\Omegax+v∈/Ω for every x∈ΓAx\in\Gamma_Ax∈ΓA​, and

f(t)(x)=0(x∈ΓA, 0≤t<T).f^{(t)}(x)=0\qquad(x\in\Gamma_A,\ 0\le t<T).f(t)(x)=0(x∈ΓA​, 0≤t<T).

Then

∣G(z)∣≤CR(2r/R)TA1A2A3(∣z∣≤r),|G(z)|\le C_R(2r/R)^{TA_1A_2A_3}\qquad(|z|\le r),∣G(z)∣≤CR​(2r/R)TA1​A2​A3​(∣z∣≤r), ∣G(n)(z)∣≤n!ρnCR(2r/R)TA1A2A3(ρ>0, ∣z∣+ρ≤r, n≥0).|G^{(n)}(z)|\le \frac{n!}{\rho^n}C_R(2r/R)^{TA_1A_2A_3} \qquad(\rho>0,\ |z|+\rho\le r,\ n\ge0).∣G(n)(z)∣≤ρnn!​CR​(2r/R)TA1​A2​A3​(ρ>0, ∣z∣+ρ≤r, n≥0).

The entire function precedes all choices of grids, radii and bounds. The data do not provide the four basic factors, establish their nonvanishing or growth, or prove that a test value is nonzero.

Also assume the following period arithmetic jet-system data at z=u1/2z=u_1/2z=u1​/2.

For fixed L,ω,z,θ,νL,\omega,z,\theta,\nuL,ω,z,θ,ν, a monic relation g∈Z[X,Y]g\in\mathbb Z[X,Y]g∈Z[X,Y], and d∈Z[X]d\in\mathbb Z[X]d∈Z[X], the period arithmetic jet-system data assert the following uniform property. Write δ=d(θ)\delta=d(\theta)δ=d(θ) and e=deg⁡Yge=\deg_Yge=degY​g. There exist positive integers B,HB,HB,H such that for every integer aaa and nonnegative integers M,L0,TM,L_0,TM,L0​,T, with

I={0,…,L0}×{0,…,M}2,Kn=L0+2M+n,I=\{0,\ldots,L_0\}\times\{0,\ldots,M\}^2,\qquad K_n=L_0+2M+n,I={0,…,L0​}×{0,…,M}2,Kn​=L0​+2M+n,

there is a matrix R=(Rn,i)0≤n<T, i∈IR=(R_{n,i})_{0\le n<T,\,i\in I}R=(Rn,i​)0≤n<T,i∈I​ over Z[X,Y]\mathbb Z[X,Y]Z[X,Y] satisfying

deg⁡YRn,i<e,deg⁡X([Yj]Rn,i)≤BKn(j≥0),\deg_YR_{n,i}<e,\qquad \deg_X([Y^j]R_{n,i})\le BK_n\quad(j\ge0),degY​Rn,i​<e,degX​([Yj]Rn,i​)≤BKn​(j≥0), L(Rn,i)≤n! 24Kn(1+∣a∣)L0+MHKn+1,\mathscr L(R_{n,i})\le n!\,24^{K_n}(1+|a|)^{L_0+M}H^{K_n+1},L(Rn,i​)≤n!24Kn​(1+∣a∣)L0​+MHKn​+1, Rn,i(θ,ν)=δ7Kndndwn(wi0℘(w)i2ζ(w)i3)∣w=z+aω.R_{n,i}(\theta,\nu)=\delta^{7K_n} \left.\frac{d^n}{dw^n}\big(w^{i_0}\wp(w)^{i_2}\zeta(w)^{i_3}\big) \right|_{w=z+a\omega}.Rn,i​(θ,ν)=δ7Kn​dwndn​(wi0​℘(w)i2​ζ(w)i3​)​w=z+aω​.

The same matrix, chosen before all coefficient vectors, satisfies for every c∈CIc\in\mathbb C^Ic∈CI

[∑iRn,i(θ,ν)ci=0  (0≤n<T)]  ⟺  [dndwn∑iciwi0℘(w)i2ζ(w)i3∣w=z+aω=0  (0≤n<T)].\left[\sum_iR_{n,i}(\theta,\nu)c_i=0\ \ (0\le n<T)\right] \iff \left[\left.\frac{d^n}{dw^n}\sum_i c_iw^{i_0}\wp(w)^{i_2}\zeta(w)^{i_3} \right|_{w=z+a\omega}=0\ \ (0\le n<T)\right].[i∑​Rn,i​(θ,ν)ci​=0  (0≤n<T)]⟺[dwndn​i∑​ci​wi0​℘(w)i2​ζ(w)i3​​w=z+aω​=0  (0≤n<T)].

The denominator exponent 7Kn7K_n7Kn​ is a uniform padding from coordinatewise clearing. It is explicitly coarser than the exponent in source equation (28); the degree and logarithmic height still have the required linear dependence on the degree and derivative parameters. The data concern only shifts by integer multiples of ω\omegaω. They neither supply coordinates at other grid points nor prove the zero estimate.

Also assume the following rounded-parameter and actual-grid estimates.

For a nonnegative integer NNN, define

m=⌊N/log⁡N⌋,ℓ=⌊Nlog⁡N⌋,s=⌊N3/16⌋,q=⌊N5/8log⁡N/64⌋,R=N49/72.m=\lfloor N/\log N\rfloor,\qquad \ell=\lfloor\sqrt{N\log N}\rfloor,\qquad s=\lfloor N^{3/16}\rfloor,\qquad q=\lfloor N^{5/8}\log N/64\rfloor,\qquad R=N^{49/72}.m=⌊N/logN⌋,ℓ=⌊NlogN​⌋,s=⌊N3/16⌋,q=⌊N5/8logN/64⌋,R=N49/72.

Only sufficiently large NNN enter the estimates. The formal definitions use nonnegative integer floors, which agree with the displayed floors there. The constant 1/641/641/64 is one explicit small choice for the source's freely chosen grid constant.

Write Ω\OmegaΩ for the period lattice and put

U=∣u1∣+∣u2∣+∣ω∣+1,r=4qU,U=|u_1|+|u_2|+|\omega|+1,\qquad r=4qU,U=∣u1​∣+∣u2​∣+∣ω∣+1,r=4qU, ΓN={a1u1+a2u2+a3ω+u1/2:0≤a1,a2<s, 0≤a3<q, ai∈Z},\Gamma_N=\{a_1u_1+a_2u_2+a_3\omega+u_1/2: 0\le a_1,a_2<s,\ 0\le a_3<q,\ a_i\in\mathbb Z\},ΓN​={a1​u1​+a2​u2​+a3​ω+u1​/2:0≤a1​,a2​<s, 0≤a3​<q, ai​∈Z}, ΓN(3)={a1u1+a2u2+a3ω+u1/2:0≤a1,a2<3s, 0≤a3<3q, ai∈Z}.\Gamma_N^{(3)}=\{a_1u_1+a_2u_2+a_3\omega+u_1/2: 0\le a_1,a_2<3s,\ 0\le a_3<3q,\ a_i\in\mathbb Z\}.ΓN(3)​={a1​u1​+a2​u2​+a3​ω+u1​/2:0≤a1​,a2​<3s, 0≤a3​<3q, ai​∈Z}.

The grid parameter data assert that, for every fixed B>0B>0B>0, all sufficiently large NNN satisfy

m,ℓ,q≥1,2≤s≤q,m,\ell,q\ge1,\qquad 2\le s\le q,m,ℓ,q≥1,2≤s≤q, 8(m+1)∣ΓN∣≤(m+1)(ℓ+1)2,N2512≤(m+1)∣ΓN∣≤N232,8(m+1)|\Gamma_N|\le(m+1)(\ell+1)^2,\qquad \frac{N^2}{512}\le(m+1)|\Gamma_N|\le\frac{N^2}{32},8(m+1)∣ΓN​∣≤(m+1)(ℓ+1)2,512N2​≤(m+1)∣ΓN​∣≤32N2​, mlog⁡N≤N,ℓs2≤m,R≥1,r>0,m\log N\le N,\qquad \ell s^2\le m,\qquad R\ge1,\qquad r>0,mlogN≤N,ℓs2≤m,R≥1,r>0, z∈ΓN(3) ⟹ z∉Ω and ∣z∣+1≤r,z\in\Gamma_N^{(3)}\ \Longrightarrow\ z\notin\Omega\ \text{and}\ |z|+1\le r,z∈ΓN(3)​ ⟹ z∈/Ω and ∣z∣+1≤r, 2rR≤N−1/36,BℓR2≤N2.\frac{2r}{R}\le N^{-1/36},\qquad B\ell R^2\le N^2.R2r​≤N−1/36,BℓR2≤N2.

Here the grids are the actual finite sets, so their cardinalities count distinct points. All shifted points are regular, including shifts whose unshifted point lies in the lattice. The counts allow vanishing conditions on the entire shifted grid; the data do not assert that an auxiliary function satisfying those conditions has already been constructed. Basic sigma factors, arithmetic coordinate presentations, coefficient envelopes, and the zero estimate are separate obligations.

Also assume the following uniform decay data on translated grids.

For fixed complex ω,u1,u2\omega,u_1,u_2ω,u1​,u2​, put

m=⌊N/log⁡N⌋,s=⌊N3/16⌋,q=⌊N5/8log⁡N/64⌋,R=N49/72,m=\lfloor N/\log N\rfloor,\quad s=\lfloor N^{3/16}\rfloor,\quad q=\lfloor N^{5/8}\log N/64\rfloor,\quad R=N^{49/72},m=⌊N/logN⌋,s=⌊N3/16⌋,q=⌊N5/8logN/64⌋,R=N49/72, r=4q(∣u1∣+∣u2∣+∣ω∣+1),ΓN={a1u1+a2u2+a3ω+u1/2:0≤a1,a2<s, 0≤a3<q, ai∈Z}.r=4q(|u_1|+|u_2|+|\omega|+1),\qquad \Gamma_N=\{a_1u_1+a_2u_2+a_3\omega+u_1/2: 0\le a_1,a_2<s,\ 0\le a_3<q,\ a_i\in\mathbb Z\}.r=4q(∣u1​∣+∣u2​∣+∣ω∣+1),ΓN​={a1​u1​+a2​u2​+a3​ω+u1​/2:0≤a1​,a2​<s, 0≤a3​<q, ai​∈Z}.

The decay data assert the following proposition for every fixed B≥0B\ge0B≥0 and K>0K>0K>0, for all sufficiently large integers NNN, uniformly in all subsequent choices.

Let ∣v∣≤r|v|\le r∣v∣≤r and write Δ=ΓN−v\Delta=\Gamma_N-vΔ=ΓN​−v. Supply functions f,G,ψf,G,\psif,G,ψ such that GGG is entire, while f,ψf,\psif,ψ are analytic near every point of Δ\DeltaΔ and satisfy G=ψfG=\psi fG=ψf locally there. Assume

f(j)(x)=0(x∈Δ, 0≤j<m+1),max⁡∣z∣=R∣G(z)∣≤eBN2.f^{(j)}(x)=0\quad(x\in\Delta,\ 0\le j<m+1),\qquad \max_{|z|=R}|G(z)|\le e^{BN^2}.f(j)(x)=0(x∈Δ, 0≤j<m+1),∣z∣=Rmax​∣G(z)∣≤eBN2.

For every complex www and nonnegative integer nnn with

∣w∣+1≤2r,n≤Km,|w|+1\le 2r,\qquad n\le Km,∣w∣+1≤2r,n≤Km,

the regularized derivative satisfies

∣G(n)(w)∣≤exp⁡ ⁣(−N2log⁡N73728).|G^{(n)}(w)|\le \exp\!\left(-\frac{N^2\log N}{73728}\right).∣G(n)(w)∣≤exp(−73728N2logN​).

If f,ψf,\psif,ψ are also 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

∣f(n)(w)∣≤exp⁡ ⁣(−N2log⁡N73728).|f^{(n)}(w)|\le \exp\!\left(-\frac{N^2\log N}{73728}\right).∣f(n)(w)∣≤exp(−73728N2logN​).

The double inner radius accommodates arbitrary translates with ∣v∣≤r|v|\le r∣v∣≤r. The constant 1/737281/737281/73728 is a coarse explicit choice, not a quoted constant from the source. This predicate supplies the interpolation implication; construction and growth of the basic sigma factors, the inverse multiplier bound, and a nonzero test derivative remain separate obligations. No regularity or nonvanishing is inferred from totalized meromorphic values.

Also assume the following bounded period-jet data at z=u1/2z=u_1/2z=u1​/2.

Fix a complex period pair LLL, complex numbers ω,z,θ,ν\omega,z,\theta,\nuω,z,θ,ν, and polynomials g∈Z[X,Y]g\in\mathbb Z[X,Y]g∈Z[X,Y], d∈Z[X]d\in\mathbb Z[X]d∈Z[X]. Write e=deg⁡Yge=\deg_Yge=degY​g, δ=d(θ)\delta=d(\theta)δ=d(θ), and let L(P)\mathscr L(P)L(P) be the sum of the absolute values of all integer coefficients of PPP.

For sufficiently large positive integers NNN, use the auxiliary parameters

m=⌊N/log⁡N⌋,ℓ=⌊Nlog⁡N⌋,q=⌊N5/8log⁡N/64⌋,I={0,…,m}×{0,…,ℓ}2.m=\lfloor N/\log N\rfloor,\qquad \ell=\lfloor\sqrt{N\log N}\rfloor,\qquad q=\lfloor N^{5/8}\log N/64\rfloor,\qquad I=\{0,\ldots,m\}\times\{0,\ldots,\ell\}^2.m=⌊N/logN⌋,ℓ=⌊NlogN​⌋,q=⌊N5/8logN/64⌋,I={0,…,m}×{0,…,ℓ}2.

Bounded auxiliary period-jet data mean the following property. For every fixed nonnegative integer KKK, there is a real constant A>0A>0A>0 such that, for all sufficiently large NNN, one can choose a nonnegative integer DDD with

D≤Am,(m+1)(ℓ+1)2(e+1)(D+1)≤eAN.D\le Am,\qquad (m+1)(\ell+1)^2(e+1)(D+1)\le e^{AN}.D≤Am,(m+1)(ℓ+1)2(e+1)(D+1)≤eAN.

This same degree bound works for every integer aaa satisfying ∣a∣≤3q|a|\le3q∣a∣≤3q. For each such aaa, there is a matrix

R=(Rn,i)0≤n≤Km, i∈Iover Z[X,Y]R=(R_{n,i})_{0\le n\le Km,\ i\in I}\quad\text{over }\mathbb Z[X,Y]R=(Rn,i​)0≤n≤Km, i∈I​over Z[X,Y]

such that, for every entry and every nonnegative jjj,

deg⁡YRn,i<e,deg⁡X([Yj]Rn,i)≤D,L(Rn,i)≤eAN.\deg_YR_{n,i}<e,\qquad \deg_X([Y^j]R_{n,i})\le D,\qquad \mathscr L(R_{n,i})\le e^{AN}.degY​Rn,i​<e,degX​([Yj]Rn,i​)≤D,L(Rn,i​)≤eAN.

Writing Jn=m+2ℓ+nJ_n=m+2\ell+nJn​=m+2ℓ+n, the exact evaluations are

Rn,i(θ,ν)=δ7Jndndwn(wi0℘L(w)i2ζL(w)i3)∣w=z+aω.R_{n,i}(\theta,\nu)=\delta^{7J_n} \left.\frac{d^n}{dw^n}\big(w^{i_0}\wp_L(w)^{i_2}\zeta_L(w)^{i_3}\big) \right|_{w=z+a\omega}.Rn,i​(θ,ν)=δ7Jn​dwndn​(wi0​℘L​(w)i2​ζL​(w)i3​)​w=z+aω​.

Moreover, for every complex coefficient vector c=(ci)i∈Ic=(c_i)_{i\in I}c=(ci​)i∈I​, set

Fc(w)=∑i∈Iciwi0℘L(w)i2ζL(w)i3.F_c(w)=\sum_{i\in I}c_iw^{i_0}\wp_L(w)^{i_2}\zeta_L(w)^{i_3}.Fc​(w)=i∈I∑​ci​wi0​℘L​(w)i2​ζL​(w)i3​.

The same matrix satisfies

[∑iRn,i(θ,ν)ci=0(0≤n≤Km)]⟺[Fc(n)(z+aω)=0(0≤n≤Km)].\left[\sum_iR_{n,i}(\theta,\nu)c_i=0\quad(0\le n\le Km)\right] \quad\Longleftrightarrow\quad \left[F_c^{(n)}(z+a\omega)=0\quad(0\le n\le Km)\right].[i∑​Rn,i​(θ,ν)ci​=0(0≤n≤Km)]⟺[Fc(n)​(z+aω)=0(0≤n≤Km)].

The constant and threshold precede the period index, the matrix precedes the vector, and DDD is common to all period indices. The length bound also bounds each individual integer coefficient. Order zero, K=0K=0K=0, and both signs of aaa are included. The padded denominator exponent is inherited from the earlier period-jet construction; it is coarser than the source's exponent in (28). These data do not assert a nonzero derivative or any estimates at nonlattice translates.

Also assume the following conditional nonlattice matrix bounds.

Fix a complex period pair LLL, complex numbers θ,ν\theta,\nuθ,ν, and g∈Z[X,Y]g\in\mathbb Z[X,Y]g∈Z[X,Y]. Write e=deg⁡Yge=\deg_Yge=degY​g and let L(P)\mathscr L(P)L(P) denote the sum of the absolute values of all integer coefficients of PPP.

For all sufficiently large positive integers NNN, put

m=⌊N/log⁡N⌋,ℓ=⌊Nlog⁡N⌋,s=⌊N3/16⌋,I={0,…,m}×{0,…,ℓ}2.m=\lfloor N/\log N\rfloor,\qquad \ell=\lfloor\sqrt{N\log N}\rfloor,\qquad s=\lfloor N^{3/16}\rfloor,\qquad I=\{0,\ldots,m\}\times\{0,\ldots,\ell\}^2.m=⌊N/logN⌋,ℓ=⌊NlogN​⌋,s=⌊N3/16⌋,I={0,…,m}×{0,…,ℓ}2.

For nonnegative integers K,CK,CK,C, define the eight coordinate degree and logarithmic length bounds

d=(C,Cs2,Cs2,Cs2,C,C,C,C),d=(C,Cs^2,Cs^2,Cs^2,C,C,C,C),d=(C,Cs2,Cs2,Cs2,C,C,C,C), b=(Clog⁡N,C(s2+log⁡N),C(s2+log⁡N),C(s2+log⁡N),C,C,C,C).b=(C\log N,C(s^2+\log N),C(s^2+\log N),C(s^2+\log N),C,C,C,C).b=(ClogN,C(s2+logN),C(s2+logN),C(s2+logN),C,C,C,C).

The derivative weights are

k(n)=(m,5ℓ,5ℓ,5ℓ,Jn,Jn,Jn,Jn),Jn=m+5ℓ+n.k(n)=(m,5\ell,5\ell,5\ell,J_n,J_n,J_n,J_n),\qquad J_n=m+5\ell+n.k(n)=(m,5ℓ,5ℓ,5ℓ,Jn​,Jn​,Jn​,Jn​),Jn​=m+5ℓ+n.

Bounded auxiliary nonlattice jet data assert that, for every K,CK,CK,C, there is A>0A>0A>0 such that, for all sufficiently large NNN, one can choose a nonnegative integer DDD satisfying

D≤Am,(m+1)(ℓ+1)2(e+1)(D+1)≤eAN.D\le Am,\qquad (m+1)(\ell+1)^2(e+1)(D+1)\le e^{AN}.D≤Am,(m+1)(ℓ+1)2(e+1)(D+1)≤eAN.

This degree bound is chosen before the following coordinate presentations. Choose eight numerator and denominator polynomials Pa,Qa∈Z[X,Y]P_a,Q_a\in\mathbb Z[X,Y]Pa​,Qa​∈Z[X,Y] and nonnegative integer length bounds hah_aha​ such that

deg⁡Pa,deg⁡Qa≤da,L(Pa),L(Qa)≤ha≤eba(0≤a<8).\deg P_a,\deg Q_a\le d_a,\qquad \mathscr L(P_a),\mathscr L(Q_a)\le h_a\le e^{b_a}\qquad(0\le a<8).degPa​,degQa​≤da​,L(Pa​),L(Qa​)≤ha​≤eba​(0≤a<8).

There is a polynomial matrix

R=(Rn,i)0≤n≤Km, i∈IR=(R_{n,i})_{0\le n\le Km,\ i\in I}R=(Rn,i​)0≤n≤Km, i∈I​

with

deg⁡YRn,i<e,deg⁡X([Yj]Rn,i)≤D(j≥0),L(Rn,i)≤eAN.\deg_Y R_{n,i}<e,\qquad \deg_X([Y^j]R_{n,i})\le D\quad(j\ge0),\qquad \mathscr L(R_{n,i})\le e^{AN}.degY​Rn,i​<e,degX​([Yj]Rn,i​)≤D(j≥0),L(Rn,i​)≤eAN.

The matrix is chosen before any complex evaluation points v,zv,zv,z. Suppose v,z,z+vv,z,z+vv,z,z+v lie outside the period lattice and

Pa(θ,ν)=Qa(θ,ν)Jv(z)a,P_a(\theta,\nu)=Q_a(\theta,\nu)\mathcal J_v(z)_a,Pa​(θ,ν)=Qa​(θ,ν)Jv​(z)a​, Jv(z)=(z+v,ζL(v),℘L(v),℘L′(v),ζL(z),℘L(z),℘L′(z),℘L′′(z)).\mathcal J_v(z)=(z+v,\zeta_L(v),\wp_L(v),\wp'_L(v), \zeta_L(z),\wp_L(z),\wp'_L(z),\wp''_L(z)).Jv​(z)=(z+v,ζL​(v),℘L​(v),℘L′​(v),ζL​(z),℘L​(z),℘L′​(z),℘L′′​(z)).

Put

Gi(w)=(w+v)i0[2(℘L(v)−℘L(w))]3ℓ℘L(w+v)i2ζL(w+v)i3,Δn=∏a=07Qa(θ,ν)ka(n).G_i(w)=(w+v)^{i_0}[2(\wp_L(v)-\wp_L(w))]^{3\ell} \wp_L(w+v)^{i_2}\zeta_L(w+v)^{i_3},\qquad \Delta_n=\prod_{a=0}^7Q_a(\theta,\nu)^{k_a(n)}.Gi​(w)=(w+v)i0​[2(℘L​(v)−℘L​(w))]3ℓ℘L​(w+v)i2​ζL​(w+v)i3​,Δn​=a=0∏7​Qa​(θ,ν)ka​(n).

Then the exact entry evaluations are

Rn,i(θ,ν)=ΔnGi(n)(z).R_{n,i}(\theta,\nu)=\Delta_nG_i^{(n)}(z).Rn,i​(θ,ν)=Δn​Gi(n)​(z).

If in addition z−vz-vz−v lies outside the lattice and every Qa(θ,ν)Q_a(\theta,\nu)Qa​(θ,ν) is nonzero, then, for every complex vector c=(ci)i∈Ic=(c_i)_{i\in I}c=(ci​)i∈I​, writing

Fc(w)=∑i∈Iciwi0℘L(w)i2ζL(w)i3,F_c(w)=\sum_{i\in I}c_iw^{i_0}\wp_L(w)^{i_2}\zeta_L(w)^{i_3},Fc​(w)=i∈I∑​ci​wi0​℘L​(w)i2​ζL​(w)i3​,

the same matrix satisfies

[∑iRn,i(θ,ν)ci=0(0≤n≤Km)]⟺[Fc(n)(z+v)=0(0≤n≤Km)].\left[\sum_i R_{n,i}(\theta,\nu)c_i=0\quad(0\le n\le Km)\right] \quad\Longleftrightarrow\quad \left[F_c^{(n)}(z+v)=0\quad(0\le n\le Km)\right].[i∑​Rn,i​(θ,ν)ci​=0(0≤n≤Km)]⟺[Fc(n)​(z+v)=0(0≤n≤Km)].

The constant and threshold precede all coordinate polynomials; DDD is common to every presentation and derivative order. The matrix precedes the points and vectors. Both K=0K=0K=0 and C=0C=0C=0 are allowed. Zero denominators are permitted for the exact evaluation identity, and nonzero evaluated denominators are explicit hypotheses of the kernel equivalence. This is a conditional bound for nonlattice derivative matrices: existence of coordinate presentations satisfying the profiles, their denominator nonvanishing, and a nonzero test derivative are separate obligations.

Also assume assembled grid jet-matrix data, including the retained coordinate presentations, as follows.

Fix a complex period pair with lattice Ω\OmegaΩ, and complex numbers ω,u1,u2,θ,ν\omega,u_1,u_2,\theta,\nuω,u1​,u2​,θ,ν. Write z0=u1/2z_0=u_1/2z0​=u1​/2 and let L(P)\mathscr L(P)L(P) be the sum of the absolute values of the integer coefficients of a polynomial PPP. For a nonnegative integer NNN, put

s=⌊N3/16⌋,q=⌊N5/8log⁡N/64⌋,s=\lfloor N^{3/16}\rfloor,\qquad q=\lfloor N^{5/8}\log N/64\rfloor,s=⌊N3/16⌋,q=⌊N5/8logN/64⌋, Γ3={a1u1+a2u2+a3ω:0≤a1,a2<3s, 0≤a3<3q, ai∈Z}.\Gamma_3=\{a_1u_1+a_2u_2+a_3\omega: 0\le a_1,a_2<3s,\ 0\le a_3<3q,\ a_i\in\mathbb Z\}.Γ3​={a1​u1​+a2​u2​+a3​ω:0≤a1​,a2​<3s, 0≤a3​<3q, ai​∈Z}.

For a nonnegative integer CCC, use the coordinate degree and logarithmic length profiles

dC=(C,Cs2,Cs2,Cs2,C,C,C,C),d_C=(C,Cs^2,Cs^2,Cs^2,C,C,C,C),dC​=(C,Cs2,Cs2,Cs2,C,C,C,C), bC=(Clog⁡N,C(s2+log⁡N),C(s2+log⁡N),C(s2+log⁡N),C,C,C,C).b_C=(C\log N,C(s^2+\log N),C(s^2+\log N),C(s^2+\log N),C,C,C,C).bC​=(ClogN,C(s2+logN),C(s2+logN),C(s2+logN),C,C,C,C).

A coordinate presentation at v,zv,zv,z consists of eight numerator polynomials PaP_aPa​, denominator polynomials QaQ_aQa​ in Z[X,Y]\mathbb Z[X,Y]Z[X,Y], and nonnegative integers hah_aha​, such that

deg⁡Pa,deg⁡Qa≤dC,a,L(Pa),L(Qa)≤ha≤ebC,a,\deg P_a,\deg Q_a\le d_{C,a},\qquad \mathscr L(P_a),\mathscr L(Q_a)\le h_a\le e^{b_{C,a}},degPa​,degQa​≤dC,a​,L(Pa​),L(Qa​)≤ha​≤ebC,a​, Qa(θ,ν)≠0,Pa(θ,ν)=Qa(θ,ν)Jv(z)a,Q_a(\theta,\nu)\ne0,\qquad P_a(\theta,\nu)=Q_a(\theta,\nu)J_v(z)_a,Qa​(θ,ν)=0,Pa​(θ,ν)=Qa​(θ,ν)Jv​(z)a​, Jv(z)=(z+v,ζ(v),℘(v),℘′(v),ζ(z),℘(z),℘′(z),℘′′(z)).J_v(z)=(z+v,\zeta(v),\wp(v),\wp'(v),\zeta(z),\wp(z),\wp'(z),\wp''(z)).Jv​(z)=(z+v,ζ(v),℘(v),℘′(v),ζ(z),℘(z),℘′(z),℘′′(z)).

Auxiliary nonlattice coordinate data assert

∃C∈N ∀N sufficiently large ∀v∈Γ3∖Ω,there exists a coordinate presentation at v,z0.\exists C\in\mathbb N\ \forall N\text{ sufficiently large}\ \forall v\in\Gamma_3\setminus\Omega, \quad\text{there exists a coordinate presentation at }v,z_0.∃C∈N ∀N sufficiently large ∀v∈Γ3​∖Ω,there exists a coordinate presentation at v,z0​.

The constant and threshold precede all points. Every evaluated denominator is nonzero. The four fixed-coordinate bounds are independent of NNN, the ordinary coordinate has fixed degree and polynomial length in NNN, and the three moving elliptic coordinates have degree O(s2)O(s^2)O(s2) and logarithmic length O(s2+log⁡N)O(s^2+\log N)O(s2+logN). No derivative matrices or analytic estimates are part of this coordinate property.

With the coordinate presentations and notation just specified, let g∈Z[X,Y]g\in\mathbb Z[X,Y]g∈Z[X,Y], d∈Z[X]d\in\mathbb Z[X]d∈Z[X], e=deg⁡Yge=\deg_Yge=degY​g, and δ=d(θ)\delta=d(\theta)δ=d(θ). Put

m=⌊N/log⁡N⌋,ℓ=⌊Nlog⁡N⌋,I={0,…,m}×{0,…,ℓ}2,m=\lfloor N/\log N\rfloor,\qquad \ell=\lfloor\sqrt{N\log N}\rfloor,\qquad I=\{0,\ldots,m\}\times\{0,\ldots,\ell\}^2,m=⌊N/logN⌋,ℓ=⌊NlogN​⌋,I={0,…,m}×{0,…,ℓ}2, Γ={a1u1+a2u2+a3ω:0≤a1,a2<s, 0≤a3<q, ai∈Z}.\Gamma=\{a_1u_1+a_2u_2+a_3\omega:0\le a_1,a_2<s,\ 0\le a_3<q,\ a_i\in\mathbb Z\}.Γ={a1​u1​+a2​u2​+a3​ω:0≤a1​,a2​<s, 0≤a3​<q, ai​∈Z}.

Auxiliary grid jet-matrix data assert that there is a single nonnegative integer CCC such that, for each nonnegative integer KKK, there is A>0A>0A>0 with the following property for every sufficiently large NNN. There are a nonnegative integer DDD, a polynomial matrix family

R=(Rv,n,i)v∈Γ3, 0≤n≤Km, i∈I,R=(R_{v,n,i})_{v\in\Gamma_3,\ 0\le n\le Km,\ i\in I},R=(Rv,n,i​)v∈Γ3​, 0≤n≤Km, i∈I​,

and coordinate presentations with the profiles dC,bCd_C,b_CdC​,bC​ at every nonlattice point of Γ3\Gamma_3Γ3​, satisfying

D≤Am,∣I∣(e+1)(D+1)≤eAN,8(m+1)∣Γ∣≤∣I∣,D\le Am,\qquad |I|(e+1)(D+1)\le e^{AN},\qquad 8(m+1)|\Gamma|\le |I|,D≤Am,∣I∣(e+1)(D+1)≤eAN,8(m+1)∣Γ∣≤∣I∣, deg⁡YRv,n,i<e,deg⁡X([Yj]Rv,n,i)≤D (j≥0),L(Rv,n,i)≤eAN.\deg_Y R_{v,n,i}<e,\qquad \deg_X([Y^j]R_{v,n,i})\le D\ (j\ge0),\qquad \mathscr L(R_{v,n,i})\le e^{AN}.degY​Rv,n,i​<e,degX​([Yj]Rv,n,i​)≤D (j≥0),L(Rv,n,i​)≤eAN.

Define the ordinary monomials and cleared translates by

fi(w)=wi0℘(w)i2ζ(w)i3,Gv,i(w)=(w+v)i0[2(℘(v)−℘(w))]3ℓ℘(w+v)i2ζ(w+v)i3.f_i(w)=w^{i_0}\wp(w)^{i_2}\zeta(w)^{i_3},\qquad G_{v,i}(w)=(w+v)^{i_0}[2(\wp(v)-\wp(w))]^{3\ell} \wp(w+v)^{i_2}\zeta(w+v)^{i_3}.fi​(w)=wi0​℘(w)i2​ζ(w)i3​,Gv,i​(w)=(w+v)i0​[2(℘(v)−℘(w))]3ℓ℘(w+v)i2​ζ(w+v)i3​.

Write Qv,aQ_{v,a}Qv,a​ for the denominators in the retained presentation at a nonlattice point, and put

k(n)=(m,5ℓ,5ℓ,5ℓ,Jn,Jn,Jn,Jn),Jn=m+5ℓ+n,Δv,n=∏a=07Qv,a(θ,ν)ka(n).k(n)=(m,5\ell,5\ell,5\ell,J_n,J_n,J_n,J_n),\quad J_n=m+5\ell+n,\qquad \Delta_{v,n}=\prod_{a=0}^7 Q_{v,a}(\theta,\nu)^{k_a(n)}.k(n)=(m,5ℓ,5ℓ,5ℓ,Jn​,Jn​,Jn​,Jn​),Jn​=m+5ℓ+n,Δv,n​=a=0∏7​Qv,a​(θ,ν)ka​(n).

The entries retain their exact evaluations in both cases:

Rv,n,i(θ,ν)={δ7(m+2ℓ+n)fi(n)(z0+v),v∈Ω,Δv,nGv,i(n)(z0),v∉Ω.R_{v,n,i}(\theta,\nu)= \begin{cases} \delta^{7(m+2\ell+n)}f_i^{(n)}(z_0+v),&v\in\Omega,\\ \Delta_{v,n}G_{v,i}^{(n)}(z_0),&v\notin\Omega. \end{cases}Rv,n,i​(θ,ν)={δ7(m+2ℓ+n)fi(n)​(z0​+v),Δv,n​Gv,i(n)​(z0​),​v∈Ω,v∈/Ω.​

For every point v∈Γ3v\in\Gamma_3v∈Γ3​ and every complex coefficient vector c=(ci)i∈Ic=(c_i)_{i\in I}c=(ci​)i∈I​, writing Fc=∑icifiF_c=\sum_i c_i f_iFc​=∑i​ci​fi​, the same matrix family satisfies

[∑iRv,n,i(θ,ν)ci=0(0≤n≤Km)]⟺[Fc(n)(z0+v)=0(0≤n≤Km)].\left[\sum_iR_{v,n,i}(\theta,\nu)c_i=0\quad(0\le n\le Km)\right] \Longleftrightarrow \left[F_c^{(n)}(z_0+v)=0\quad(0\le n\le Km)\right].[i∑​Rv,n,i​(θ,ν)ci​=0(0≤n≤Km)]⟺[Fc(n)​(z0​+v)=0(0≤n≤Km)].

The matrices and retained presentations precede every coefficient vector. Their degree and length bounds are uniform in the grid point, derivative order, and monomial. Order zero and K=0K=0K=0 are included. The dimension gap refers to the smaller grid with m+1m+1m+1 equations per point; it does not assert a dimension gap for the whole enlarged derivative family. This interface supplies matrices and their derivative interpretation, while existence and estimates for a nonzero analytic test value remain separate.

In addition, suppose a normalized entire sigma differential solution σ\sigmaσ and three entire functions S0,S1,S2S_0,S_1,S_2S0​,S1​,S2​ are supplied, with

σ(0)=0,σ′(0)=1,σ′(z)=ζ(z)σ(z)(z∉Ω),\sigma(0)=0,\quad\sigma'(0)=1,\quad \sigma'(z)=\zeta(z)\sigma(z)\quad(z\notin\Omega),σ(0)=0,σ′(0)=1,σ′(z)=ζ(z)σ(z)(z∈/Ω), S0=σζ,S1=σ2℘,S2=σ3℘′outside Ω.S_0=\sigma\zeta,\quad S_1=\sigma^2\wp,\quad S_2=\sigma^3\wp' \quad\text{outside }\Omega.S0​=σζ,S1​=σ2℘,S2​=σ3℘′outside Ω.

Suppose a single real constant A>0A>0A>0 bounds these four entire functions globally:

∣σ(z)∣≤eA(1+∣z∣2),∣Sj(z)∣≤eA(1+∣z∣2)(z∈C, j=0,1,2).|\sigma(z)|\le e^{A(1+|z|^2)},\qquad |S_j(z)|\le e^{A(1+|z|^2)} \quad(z\in\mathbb C,\ j=0,1,2).∣σ(z)∣≤eA(1+∣z∣2),∣Sj​(z)∣≤eA(1+∣z∣2)(z∈C, j=0,1,2).

Assume also that σ(z)≠0\sigma(z)\ne0σ(z)=0 outside Ω\OmegaΩ. For all sufficiently large integers NNN, suppose that the two sigma regularizers obey

∣σ(u1/2)15ℓ∣−1≤eN2,|\sigma(u_1/2)^{15\ell}|^{-1}\le e^{N^2},∣σ(u1​/2)15ℓ∣−1≤eN2, ∣σ(u1/2+v)3ℓ∣−1≤eN2(v∈Γ(3S,3S,3S3)∩Ω).|\sigma(u_1/2+v)^{3\ell}|^{-1}\le e^{N^2} \quad\bigl(v\in\Gamma(3S,3S,3S_3)\cap\Omega\bigr).∣σ(u1​/2+v)3ℓ∣−1≤eN2(v∈Γ(3S,3S,3S3​)∩Ω).

Here ℓ=⌊Nlog⁡N⌋\ell=\lfloor\sqrt{N\log N}\rfloorℓ=⌊NlogN​⌋ is the auxiliary degree, S=⌊N3/16⌋S=\lfloor N^{3/16}\rfloorS=⌊N3/16⌋, and S3=⌊N5/8log⁡N/64⌋S_3=\lfloor N^{5/8}\log N/64\rfloorS3​=⌊N5/8logN/64⌋, as in the supplied grid data.

Then there exist a,c>0a,c>0a,c>0 such that every sufficiently large NNN admits a complex auxiliary system with

b=(3+∣θ∣+∣ν∣)a,E<deg⁡Yg.b=(3+|\theta|+|\nu|)a,\qquad E<\deg_Y g.b=(3+∣θ∣+∣ν∣)a,E<degY​g.

The system has the degree, height, size, and dimension-gap bounds in ComplexAuxiliarySystem. Every nonzero vector in its evaluated equation kernel with coordinate magnitudes at most ebNe^{bN}ebN has a nonzero test value of magnitude at most e−cN2log⁡Ne^{-cN^2\log N}e−cN2logN. The matrices and finite test family precede the choice of vector.

The entire sigma factors, their finite quadratic exponential growth, and the two reciprocal-power estimates are inputs to this continuation. The remaining work includes the zero estimate producing a bounded-order nonzero derivative, the cleared nonlattice arithmetic estimates, and assembling a finite test family with the required decay. Every original geometric, arithmetic, matrix, regularization, and growth hypothesis and the exact complex-auxiliary-system conclusion are retained.

Preamble
import Definitions.Def_WeierstrassEllipticZeta_SigmaDifferential
import Definitions.Def_WeierstrassEllipticZeta_GridJetMatrices
import Definitions.Def_WeierstrassEllipticZeta_NonlatticeJetBounds
import Definitions.Def_WeierstrassEllipticZeta_PeriodJetBounds
import Definitions.Def_WeierstrassEllipticZeta_InterpolationDecay
import Definitions.Def_WeierstrassEllipticZeta_AuxiliaryParameters
import Definitions.Def_WeierstrassEllipticZeta_ReducedJetSystems
import Definitions.Def_WeierstrassEllipticZeta_GridInterpolation
import Definitions.Def_WeierstrassEllipticZeta_EntireRegularization
import Definitions.Def_WeierstrassEllipticZeta_ClearedEntire
import Definitions.Def_WeierstrassEllipticZeta_PeriodJets
import Definitions.Def_WeierstrassEllipticZeta_AuxiliaryGrids
import Definitions.Def_TranscendenceTheory_ComplexAuxiliarySystem
import Mathlib.RingTheory.Algebraic.Defs
import Mathlib.Algebra.Polynomial.AlgebraMap

open scoped Polynomial
open Filter

open WeierstrassEllipticZeta
Formal statement
theorem WeierstrassEllipticZeta.exists_complex_auxiliary_systems_from_sigma_inverse_bounds
    (L : PeriodPair) (ω u₁ u₂ : ℂ)
    (h_grid : RegularAuxiliaryGridData L ω u₁ u₂)
    (h_parameters : AuxiliaryGridParameterData L ω u₁ u₂)
    (h_decay : AuxiliaryGridDecayData ω u₁ u₂)
    (h_zeta_deriv : ∀ z : ℂ, z ∉ L.lattice →
      HasDerivAt (weierstrassZeta L) (-L.weierstrassP z) z)
    (h_zeta_addition : ∀ z v : ℂ,
      z ∉ L.lattice → v ∉ L.lattice → z + v ∉ L.lattice →
      2 * (L.weierstrassP v - L.weierstrassP z) * weierstrassZeta L (z + v) =
        2 * (weierstrassZeta L z + weierstrassZeta L v) *
          (L.weierstrassP v - L.weierstrassP z) +
        L.derivWeierstrassP v - L.derivWeierstrassP z)
    (h_wp_addition : ∀ z v : ℂ,
      z ∉ L.lattice → v ∉ L.lattice → z + v ∉ L.lattice →
      4 * (L.weierstrassP v - L.weierstrassP z) ^ 2 * L.weierstrassP (z + v) =
        -4 * (L.weierstrassP z + L.weierstrassP v) *
          (L.weierstrassP v - L.weierstrassP z) ^ 2 +
        (L.derivWeierstrassP v - L.derivWeierstrassP z) ^ 2)
    (θ : ℂ) (hθ : Transcendental ℚ θ)
    (ν : ℂ)
    (g : ℤ[X][X]) (hg_monic : g.Monic) (hg_degree : 0 < g.natDegree)
    (hg_kernel : ∀ p : ℤ[X][X],
      p.eval₂ (Polynomial.aeval θ).toRingHom ν = 0 ↔ g ∣ p)
    (h_jet_systems : ReducedArithmeticJetSystemData L θ ν g)
    (h_nonlattice_bounds : BoundedAuxiliaryNonlatticeJetData L θ ν g)
    (h_interpolation : AuxiliaryGridInterpolationData ω u₁ u₂)
    (h_regularization : EllipticRegularizationData L ω u₁ u₂)
    (h_cleared_entire : ClearedAdditionEntireData L ω u₁ u₂)
    (d : ℤ[X]) (hd : Polynomial.aeval θ d ≠ 0)
    (h_period_jets : PeriodArithmeticJetSystemData L ω (u₁ / 2) θ ν g d)
    (h_period_bounds : BoundedAuxiliaryPeriodJetData L ω (u₁ / 2) θ ν g d)
    (h_grid_matrices : AuxiliaryGridJetMatrixData L ω u₁ u₂ θ ν g d)
    (h_data : ∀ i : Fin 18, ∃ p : ℤ[X][X],
      p.natDegree < g.natDegree ∧
      p.eval₂ (Polynomial.aeval θ).toRingHom ν = Polynomial.aeval θ d *
        (![L.g₂/4, L.g₃/4, ω, zetaQuasiPeriod L ω, u₁/2, u₂,
        weierstrassZeta L (u₁/2), L.weierstrassP (u₁/2),
        L.derivWeierstrassP (u₁/2), deriv L.derivWeierstrassP (u₁/2),
        L.weierstrassP u₁, L.derivWeierstrassP u₁, deriv L.derivWeierstrassP u₁,
        weierstrassZeta L u₁, L.weierstrassP u₂, L.derivWeierstrassP u₂,
        deriv L.derivWeierstrassP u₂, weierstrassZeta L u₂] i))
    (D : EllipticSigmaDifferentialData L) (S : Fin 3 → ℂ → ℂ)
    (h_factors_entire : ∀ j, AnalyticOnNhd ℂ (S j) Set.univ)
    (h_factors_eq : ∀ z : ℂ, z ∉ L.lattice → ∀ j : Fin 3,
      S j z = D.sigma z ^ (j.val + 1) * ellipticPoleCoordinates L z j)
    (A : ℝ) (hA : 0 < A)
    (h_sigma_growth : ∀ z : ℂ, ‖D.sigma z‖ ≤ Real.exp (A * (1 + ‖z‖ ^ 2)))
    (h_factor_growth : ∀ (z : ℂ) (j : Fin 3),
      ‖S j z‖ ≤ Real.exp (A * (1 + ‖z‖ ^ 2)))
    (h_sigma_nonzero : ∀ z : ℂ, z ∉ L.lattice → D.sigma z ≠ 0)
    (h_sigma_inverse : ∀ᶠ N : ℕ in Filter.atTop,
      ‖D.sigma (u₁ / 2) ^ (15 * auxiliaryL N)‖⁻¹ ≤ Real.exp ((N : ℝ) ^ 2) ∧
      ∀ v ∈ auxiliaryGrid u₁ u₂ ω
          ![3 * auxiliaryS N, 3 * auxiliaryS N, 3 * auxiliaryS3 N],
        v ∈ L.lattice →
          ‖D.sigma (u₁ / 2 + v) ^ (3 * auxiliaryL N)‖⁻¹ ≤
            Real.exp ((N : ℝ) ^ 2)) :
    ∃ a c : ℝ, 0 < a ∧ 0 < c ∧
      ∀ᶠ N : ℕ in Filter.atTop,
        ∃ S : TranscendenceTheory.ComplexAuxiliarySystem
          θ ν a ((3 + ‖θ‖ + ‖ν‖) * a) c N,
          S.yDegree < g.natDegree := by sorry
Source
Senthil Kumar K (2026), Section 5, Lemmas 8–10, equation (35), and the period case concluding Lemma 10, https://doi.org/10.1017/S001309152610145X. This continuation receives the actual two sigma inverse estimates in addition to the existing entire-factor and growth hypotheses; zero estimates and finite-test extraction remain obligations.

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