Bounded polynomial separation and linear interpolation
ProvedTranscendenceTheory.bounded_bivariate_relation_separationcommutative-algebraelliptic-functionspolynomials
Let K be a field, and let A and L be commutative K-algebras. Fix x,y in A, u,v in L, and natural numbers b,s. There is a bivariate polynomial P whose outer degree is at most s and whose coefficient degrees are at most b, with P(x,y)=0 but P(u,v) nonzero, if and only if there is no K-linear map T:A→L satisfying T(x^j y^i)=u^j v^i for all 0≤i≤s and 0≤j≤b.
The algebras need not be finite dimensional, reduced, or nontrivial. The bounds b=0 and s=0 are included. The map T is only required to be linear; it is not assumed to be multiplicative.
Preamble
import Mathlib.Algebra.Polynomial.AlgebraMap import Mathlib.LinearAlgebra.Basis.VectorSpace import Mathlib.LinearAlgebra.Isomorphisms noncomputable section open Polynomial open scoped Classical
Formal statement
theorem TranscendenceTheory.bounded_bivariate_relation_separation
(K A L : Type*) [Field K] [CommRing A] [Algebra K A]
[CommRing L] [Algebra K L] (x y : A) (u v : L) (b s : ℕ) :
(∃ P : Polynomial (Polynomial K),
P.natDegree ≤ s ∧ (∀ i, (P.coeff i).natDegree ≤ b) ∧
P.eval₂ (Polynomial.aeval x).toRingHom y = 0 ∧
P.eval₂ (Polynomial.aeval u).toRingHom v ≠ 0) ↔
¬ ∃ T : A →ₗ[K] L,
∀ i ≤ s, ∀ j ≤ b, T (x ^ j * y ^ i) = u ^ j * v ^ i := by sorry
Source
Derived linear separation step for https://prove2.me/theorems/220d1ce8-77e0-4688-b5f4-240f40909f00. Mission context: Senthil Kumar K (2026), Algebraic independence of values of Weierstrass elliptic and zeta functions, Appendix A.2, Theorem A.3 and Proposition A.1; https://doi.org/10.1017/S001309152610145X. This is a derived algebraic tool, not a completion of the paper zero estimate. Primary Lean sources: Mathlib Algebra/Polynomial/AlgebraMap.lean, LinearAlgebra/Basis/VectorSpace.lean and LinearAlgebra/Isomorphisms.lean, revision 0df444a360eaa60ab8c11dca51a86af692955474. Bounded polynomial separation is equivalent to failure to interpolate the corresponding bounded monomials by a linear map. The first isomorphism theorem and extension of linear maps prove the difficult implication, preserving every degree bound.