Geometric-series equivalence of truncated interpolation rank tests
ProvedTranscendenceTheory.truncated_geometric_rank_equivalenceelliptic-functionslinear-algebrapolynomials
Let be a field, let be finite index sets, let , and let . Take arbitrary polynomial families for , , and . Define
All matrices below have rows indexed by . Define matrices with columns by
and matrices with columns by
Then the two augmented-rank tests are equivalent:
The result has no degree restrictions or nonemptiness assumptions and includes and . It removes the candidate parameter from the left-hand matrix when the input families are fixed. The parameter then enters only through the explicitly specified finite geometric transformation of the right-hand sides.
Preamble
import Mathlib.LinearAlgebra.Matrix.Rank import Mathlib.Data.Matrix.ColumnRowPartitioned import Mathlib.LinearAlgebra.FiniteDimensional.Lemmas import Mathlib.LinearAlgebra.Finsupp.LinearCombination import Mathlib.Algebra.Ring.GeomSum import Mathlib.Algebra.Polynomial.BigOperators import Mathlib.Algebra.Polynomial.Div import Mathlib.Tactic.Ring open scoped BigOperators
Formal statement
theorem TranscendenceTheory.truncated_geometric_rank_equivalence
(K α ι κ : Type*) [Field K] [Fintype α] [Fintype ι] [Fintype κ]
(s : ℕ) (c : K) (V : α → ι → Polynomial K) (B : α → κ → Polynomial K) :
let G : Polynomial K := ∑ k ∈ Finset.range (s + 1), (Polynomial.C c * Polynomial.X) ^ k
let Ac : Matrix (α × Fin (s + 1)) ι K := fun e i =>
((1 - Polynomial.C c * Polynomial.X) * V e.1 i).coeff e.2.val
let Bc : Matrix (α × Fin (s + 1)) κ K := fun e k => (B e.1 k).coeff e.2.val
let A₀ : Matrix (α × Fin (s + 1)) ι K := fun e i => (V e.1 i).coeff e.2.val
let B₀ : Matrix (α × Fin (s + 1)) κ K := fun e k => (G * B e.1 k).coeff e.2.val
(Ac.rank < (Matrix.fromCols Ac Bc).rank) ↔
A₀.rank < (Matrix.fromCols A₀ B₀).rank := by sorry
Source
Derived explicit geometric-series reduction for the interpolation frontier https://prove2.me/theorems/37582c08-3ae4-434c-b78d-a8083d516763. Primary sources at Mathlib revision 0df444a360eaa60ab8c11dca51a86af692955474: Algebra/Ring/GeomSum.lean, geom_sum_mul_neg, line 240: https://github.com/leanprover-community/mathlib4/blob/0df444a360eaa60ab8c11dca51a86af692955474/Mathlib/Algebra/Ring/GeomSum.lean#L240; Algebra/Polynomial/Div.lean, Polynomial.X_pow_dvd_iff, line 44: https://github.com/leanprover-community/mathlib4/blob/0df444a360eaa60ab8c11dca51a86af692955474/Mathlib/Algebra/Polynomial/Div.lean#L44; LinearAlgebra/Matrix/Rank.lean, Matrix.rank_eq_finrank_span_cols, line 398: https://github.com/leanprover-community/mathlib4/blob/0df444a360eaa60ab8c11dca51a86af692955474/Mathlib/LinearAlgebra/Matrix/Rank.lean#L398. The complete theorem is derived here from a finite geometric inverse modulo X^(s+1) and the finite column-span criterion; it is not a verbatim theorem from the mission paper. Mission context: Senthil Kumar K, Algebraic independence of values of Weierstrass elliptic and zeta functions (2026), Appendix A.2, Theorem A.3 and Proposition A.1, https://doi.org/10.1017/S001309152610145X. The connecting reduction removes the candidate parameter from both left-hand matrices, retaining the same row counts, fields, geometric hypotheses, witnesses, root conditions and numerical bound. Both directions are checked in Lean. The uniform geometric estimate remains open.