Arithmetic zero-singularity theorem for algebraic combinations of rational E-series
OpenArithmeticE.algebraic_e_combination_zero_singularityLet have rational factorial-normalized coefficients with exponential size and common-denominator bounds. Let , and put . Suppose has a minimal complex polynomial differential equation of positive order , with leading coefficient . At a nonzero algebraic point , assume
Then .
The equation supplies holonomicity of ; the coefficient hypotheses supply its E-function arithmetic and convergence. The positive-order and minimality assumptions exclude the identically zero function. This is the zero-singularity consequence of the established André–Beukers theory, stated for the combinations needed in linear lifting. It remains a substantial open formalization task. It does not follow from ordinary analytic ODE existence alone.
Established supporting results and remaining bridge
The following auxiliary theorems now have accepted complete Lean proofs:
- Division preserves minimal differential order: a least-order equation for transfers to a least-order equation of the same order for , with transformed leading polynomial .
- Minimal-operator coefficient-field descent: a complex minimal equation for a formal series over a subfield descends to , at the same order. This is coefficientwise formal algebra; it does not assert that a discontinuous linear projection commutes with analytic evaluation.
- Irreducible spanning-orbit factor lemma: a pointwise vanishing finite product of linear evaluations on spanning orbits, with closed zero loci over an irreducible space, forces one evaluation to vanish identically.
These are supporting results, not yet a checked reduction of the present theorem. The missing classical bridge must construct the relevant E-function solution spaces and conjugate orbits, prove the irreducibility/closedness/spanning/product hypotheses in that setting, and supply André's full holomorphic-basis theorem at finite nonzero points. Merely knowing that a solution vanishes does not force its differential equation to be singular. No proof dependency is being claimed until those mathematical interfaces are instantiated.
The links record completed work available to a future proof of this leaf. They introduce no new conjectural arithmetic assumption and do not close this known-but-unformalized classical theorem.
import Definitions.Def_beukersLiftingData open ArithmeticE
theorem ArithmeticE.algebraic_e_combination_zero_singularity
(m : ℕ) (f : Fin m → PowerSeries ℂ) (harith : ∀ i, RationalSeriesArithmetic (f i))
(P : Fin m → Polynomial ℂ) (hP : ∀ i k, IsAlgebraic ℚ ((P i).coeff k))
(ξ : ℂ) (hξ : IsAlgebraic ℚ ξ) (hξ0 : ξ ≠ 0)
(hz : ∑ i, (P i).eval ξ * seriesValue (f i) ξ = 0)
(p : ℕ → Polynomial ℂ) (n : ℕ) (hn : 0 < n)
(hmin : MinimalEquation p n (∑ i, (P i : PowerSeries ℂ) * f i)) :
(p n).eval ξ = 0 := by sorry