Quadratic translation families with locally regular parameter coefficients
OpenPhilipponMultiplicity.exists_quadratic_translation_regular_coefficientsLet be an algebraically closed field of characteristic zero and let be a connected commutative algebraic group with the given locally closed projective embedding. There are a projective realization and an algebraic group isomorphism regular in both directions with the following property.
Every pair lies in a Zariski-open neighborhood with fixed nonvanishing chart coordinates and . Write and . There are polynomials
with coefficient functions , each of total degree at most two in , such that the following hold.
For each coefficient in the support of and every , there is an open neighborhood of and polynomials with
Thus each coefficient is locally a regular rational function of the translation parameter. The neighborhoods and fractions may depend on the coefficient and the point. At every , the tuple obtained by evaluating at and its coefficient functions at is nonzero and represents in .
Formalization Note. This is an auxiliary affine-chart version of the quadratic families in the sources, not a verbatim numbered theorem. Coefficient functions are stored in the full function ring; their values outside are unrestricted. The local fraction condition uses only the second input block, and all open sets carry the induced multiprojective Zariski topology. No common neighborhood or rational expression is required for different coefficients. The degree bound permits inhomogeneous quadratic polynomials; zero-dimensional groups are included.
Verified homogeneous-coordinate reduction (8 October 2026). An accepted proof-sketch constructs simultaneous pivot neighborhoods, pulls parameter-only coefficient functions back to pairs, and proves the homogeneous scaling identities that preserve both coefficient fractions and projective output. The sole remaining input is homogeneous quadratic translation families. Their geometric reembedding and existence remain Open; the affine-chart conversion is proved. The original formal statement is unchanged.
import Definitions.Def_PhilipponMultiplicity_AdditionLaws import Definitions.Def_PhilipponMultiplicity_Support set_option autoImplicit false open scoped BigOperators Topology
namespace PhilipponMultiplicity
theorem exists_quadratic_translation_regular_coefficients
(K : Type*) [Field K] [IsAlgClosed K] [CharZero K] :
∀ (E : EmbeddedCommutativeGroup K),
@_root_.IsConnected _ (singleGroupProduct E).zariskiTopology Set.univ →
∃ F : EmbeddedCommutativeGroup K, Nonempty (AlgebraicReembedding E F) ∧
∀ x y : F.Point, ∃ U : Set (F.Point × F.Point),
@IsOpen _ (TopologicalSpace.induced (fun xy => F.additionPair xy.1 xy.2)
(projectiveSquare K F.ambientDimension).zariskiTopology) U ∧
(x,y) ∈ U ∧
∃ c : (projectiveSquare K F.ambientDimension).FactorIndex →
Fin (F.ambientDimension+1),
(∀ xy ∈ U, ∀ i : (projectiveSquare K F.ambientDimension).FactorIndex,
(projectiveSquare K F.ambientDimension).coordinate
(F.additionPair xy.1 xy.2) ⟨i,c i⟩ ≠ 0) ∧
∃ P : Fin (F.ambientDimension+1) →
MvPolynomial (Fin (F.ambientDimension+1)) ((F.Point × F.Point) → K),
(∀ t, ∀ m ∈ (P t).support, (∑ j, m j) ≤ 2) ∧
(∀ t, ∀ m ∈ (P t).support, ∀ xy ∈ U,
∃ V : Set (F.Point × F.Point),
@IsOpen _ (TopologicalSpace.induced (fun zw => F.additionPair zw.1 zw.2)
(projectiveSquare K F.ambientDimension).zariskiTopology) V ∧ xy ∈ V ∧
∃ A B : MvPolynomial (Fin (F.ambientDimension+1)) K,
∀ zw ∈ U ∩ V,
let w : Fin (F.ambientDimension+1) → K := fun j =>
(projectiveSquare K F.ambientDimension).coordinate
(F.additionPair zw.1 zw.2) ⟨(1 : Fin 2),j⟩ /
(projectiveSquare K F.ambientDimension).coordinate
(F.additionPair zw.1 zw.2) ⟨(1 : Fin 2),c (1 : Fin 2)⟩
MvPolynomial.eval w B ≠ 0 ∧
(P t).coeff m zw = MvPolynomial.eval w A / MvPolynomial.eval w B) ∧
∀ xy ∈ U,
let w : Fin (F.ambientDimension+1) → K := fun j =>
(projectiveSquare K F.ambientDimension).coordinate
(F.additionPair xy.1 xy.2) ⟨(0 : Fin 2),j⟩ /
(projectiveSquare K F.ambientDimension).coordinate
(F.additionPair xy.1 xy.2) ⟨(0 : Fin 2),c (0 : Fin 2)⟩
∃ h : (fun t => MvPolynomial.eval₂
(Pi.evalRingHom (fun _ : F.Point × F.Point => K) xy) w (P t)) ≠ 0,
Projectivization.mk K (fun t => MvPolynomial.eval₂
(Pi.evalRingHom (fun _ : F.Point × F.Point => K) xy) w (P t)) h =
(xy.1+xy.2).val := by sorry
end PhilipponMultiplicity