Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Homogeneous quadratic translation families with regular parameter coefficients

Open
PhilipponMultiplicity.exists_quadratic_homogeneous_translation_families

by tomasz · Oct 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

algebraic-geometryalgebraic-groupsphilippon-multiplicityproof-frontier

Let KKK be algebraically closed of characteristic zero and let EEE be a connected commutative algebraic group with a locally closed projective embedding. There is a projective realization F⊆PNF\subseteq\mathbf P^NF⊆PN and an algebraic group isomorphism E→FE\to FE→F regular in both directions such that every pair (x,y)∈F×F(x,y)\in F\times F(x,y)∈F×F has an open neighborhood UUU with the following data.

There are forms

Pt(X;y)=∑matm(y)Xm(0≤t≤N),P_t(X;y)=\sum_m a_{tm}(y)X^m\qquad(0\le t\le N),Pt​(X;y)=m∑​atm​(y)Xm(0≤t≤N),

homogeneous of degree two in XXX, whose coefficients are functions of the translation parameter y∈Fy\in Fy∈F alone. For every coefficient in a support and every point of UUU, there is a smaller open neighborhood VVV and homogeneous polynomials A,B∈K[Y0,…,YN]A,B\in K[Y_0,\ldots,Y_N]A,B∈K[Y0​,…,YN​] of the same degree, such that on U∩VU\cap VU∩V,

B(Y(y′))≠0,atm(y′)=A(Y(y′))/B(Y(y′)).B(Y(y'))\ne0,\qquad a_{tm}(y')=A(Y(y'))/B(Y(y')).B(Y(y′))=0,atm​(y′)=A(Y(y′))/B(Y(y′)).

For every (x′,y′)∈U(x',y')\in U(x′,y′)∈U, the tuple of evaluated forms is nonzero and satisfies

[P0(X(x′);y′):⋯:PN(X(x′);y′)]=x′+y′.[P_0(X(x');y'):\cdots:P_N(X(x');y')]=x'+y'.[P0​(X(x′);y′):⋯:PN​(X(x′);y′)]=x′+y′.

This gives local translation families directly in homogeneous coordinates, without selecting input-chart pivots. Equal numerator and denominator degrees make the coefficient ratios independent of scaling the parameter coordinates.

Formalization Note. This is the group-restricted coordinate formulation of the source's homogeneous translation families, not a verbatim source theorem. The coefficient functions are stored on all group points and are unrestricted away from the local parameter domains; the fraction identities are imposed on the indicated pair neighborhoods. The displayed coordinates are the fixed projective representatives in the existing definitions. All opens use the induced multiprojective Zariski topology. The geometric reembedding and homogeneous families remain to be constructed. The parent proves the passage to normalized affine coordinates, including preservation of nonvanishing, coefficient regularity and the represented sum.

Verified localized-coefficient reduction (8 October 2026). An accepted proof-sketch derives the required homogeneous coefficient fractions and projective specialization from quadratic forms over localized parameter coordinate rings. It proves the fraction presentation, common-degree homogenization, coefficient-function extension and preservation of the represented sum. The geometric existence of those forms and the reembedding remain Open. The original formal statement is unchanged.

Preamble
import Definitions.Def_PhilipponMultiplicity_AdditionLaws
import Definitions.Def_PhilipponMultiplicity_Support
set_option autoImplicit false
open scoped BigOperators Topology
Formal statement
namespace PhilipponMultiplicity

theorem exists_quadratic_homogeneous_translation_families
    (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 ∧
          ∃ P : Fin (F.ambientDimension+1) →
              MvPolynomial (Fin (F.ambientDimension+1)) (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 ∧
              ∃ (d : ℕ) (A B : MvPolynomial (Fin (F.ambientDimension+1)) K),
                (∀ a ∈ A.support, (∑ j, a j) = d) ∧
                (∀ a ∈ B.support, (∑ j, a j) = d) ∧
                ∀ zw ∈ U ∩ V,
                  MvPolynomial.eval zw.2.val.rep B ≠ 0 ∧
                    (P t).coeff m zw.2 = MvPolynomial.eval zw.2.val.rep A /
                      MvPolynomial.eval zw.2.val.rep B) ∧
          ∀ xy ∈ U,
            ∃ h : (fun t => MvPolynomial.eval₂
                (Pi.evalRingHom (fun _ : F.Point => K) xy.2) xy.1.val.rep (P t)) ≠ 0,
              Projectivization.mk K (fun t => MvPolynomial.eval₂
                (Pi.evalRingHom (fun _ : F.Point => K) xy.2) xy.1.val.rep (P t)) h =
                  (xy.1+xy.2).val := by sorry

end PhilipponMultiplicity
Source
L. Rovelli, Explicit equivariant compactification and Riemann-Roch for algebraic groups, ETH dissertation 14704 (2002), Section 3.3: definition of families described by forms, pp.63–64; Theorem 3.3.2, p.64; rational coefficients in projective parameter coordinates, p.66 immediately before Corollary 3.3.4; suitable projective embedding in Theorem 3.4.5 and Theorem 3.4.6(2), p.72. https://doi.org/10.3929/ethz-a-004445245 ; https://www.research-collection.ethz.ch/bitstreams/63ac9c62-da67-437e-9dcb-413054fb5550/download . See H. Lange, Families of translations of commutative algebraic groups, J. Algebra 109(1) (1987), 260–265, DOI 10.1016/0021-8693(87)90174-8. Auxiliary restriction to the group with the translated point first and parameter second. Regular functions on projective locally closed charts are represented locally by homogeneous fractions of equal degree. Extending their values outside the parameter domain is immaterial. The reembedding and comparison of the source geometry with the concrete point model remain Open; the parent proves chart refinement and homogeneous normalization, not these geometric inputs.

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