Generic curve fibrations admit dominating smooth covering families over a normal projective base model
OpenAlgebraicGeometry.exists_normal_projective_base_and_dominating_smooth_curve_cover_familyLet be a field, let be an integral proper projective -scheme with a specified projective embedding, and let be a proper dominant morphism with integral. Fix the actual generic point of , and assume
There is a finite intermediate extension inside the prescribed algebraic closure of , a normal integral projective -scheme , and a proper surjective map with
A dense normal affine chart is retained, where is an actual finitely generated -subalgebra of with fraction field . Its generic spectrum map is exactly the one induced by , and the projective embedding of is compatible with its map to .
On a dense open there is a smooth proper family with an exhibited projective ambient model and an actual proper surjective covering map
The target family is flat and proper, and the composite is dominant. The projective ambient model is a closed subscheme of projective space over a principal localization of , and its base change to is exactly .
The actual generic source is retained: there is a smooth projective -scheme of dimension at most one with a proper surjection over the finite field extension. The map factors through , and is cartesian over it. Its composite with is exactly followed by the original generic-fibre inclusion.
The theorem works in arbitrary universes and characteristic and does not assume normality of or algebraic closedness of . It provides the finite-base and smooth-family preparation for the curve-fibration alteration argument. It asserts relative smoothness over the dense base open; semistable compactification and a global smooth projective alteration remain separate.
Retirement note (25 October 2026). This expanded presentation is replaced by the equivalent named-predicate theorem AlgebraicGeometry.hasNormalProjectiveBaseSmoothCurveCoverFamily_of_dimension_le_one_genericFiber, https://prove2.me/theorems/ae594a3b-8d61-41b8-8e72-b6a40e8b9e2f, whose proof is ACCEPTED. The named predicate unfolds to exactly this proposition; the local equivalence check is Iff.rfl. The expanded presentation repeatedly reached the server whole-type comparison timeout despite structurally identical compiled types in local checks. Retirement removes a duplicate presentation, not a mathematical obligation.
import Mathlib import Definitions.Def_AlgebraicGeometry_ProjSpace set_option autoImplicit false open AlgebraicGeometry CategoryTheory CategoryTheory.Limits TopologicalSpace universe u attribute [local instance] MvPolynomial.gradedAlgebra
theorem AlgebraicGeometry.exists_normal_projective_base_and_dominating_smooth_curve_cover_family
(k : Type u) [Field k] {S X : Scheme.{u}} [IsIntegral S] [IsIntegral X]
(π : S ⟶ Spec (.of k)) [IsProper π]
(m : ℕ) (i : S ⟶ Proj (MvPolynomial.homogeneousSubmodule (Fin (m + 1)) k))
[IsClosedImmersion i] (hi : i ≫ ProjSpace.π k m = π)
(q : X ⟶ S) [IsProper q] [IsDominant q]
(η : S) (hη : IsGenericPoint η Set.univ) (hdim : topologicalKrullDim (q.fiber η) ≤ 1) :
letI : Algebra k (S.residueField η) :=
((Scheme.ΓSpecIso (.of k)).inv ≫ π.appTop ≫ S.presheaf.germ ⊤ η trivial ≫ S.residue η).hom.toAlgebra
∃ E : IntermediateField (S.residueField η) (AlgebraicClosure (S.residueField η)),
FiniteDimensional (S.residueField η) E ∧
letI : Algebra k E := ((algebraMap (S.residueField η) E).comp (algebraMap k (S.residueField η))).toAlgebra
∃ B : Subalgebra k E, Algebra.FiniteType k B ∧ IsIntegrallyClosed B ∧ IsFractionRing B E ∧
∃ (N : Scheme.{u}) (qN : N ⟶ S) (t : Spec (.of B) ⟶ N),
IsIntegral N ∧ (∀ x : N, IsIntegrallyClosed (N.presheaf.stalk x)) ∧
IsProper qN ∧ Surjective qN ∧ IsOpenImmersion t ∧ IsDominant t ∧
topologicalKrullDim N = topologicalKrullDim S ∧
t ≫ (qN ≫ π) = Spec.map (CommRingCat.ofHom (algebraMap k B)) ∧
Spec.map (CommRingCat.ofHom (algebraMap B E)) ≫ t ≫ qN =
Spec.map (CommRingCat.ofHom (algebraMap (S.residueField η) E)) ≫ S.fromSpecResidueField η ∧
(∃ (nN : ℕ) (iN : N ⟶ Proj (MvPolynomial.homogeneousSubmodule (Fin (nN + 1)) k)),
IsClosedImmersion iN ∧ iN ≫ ProjSpace.π k nN = qN ≫ π) ∧
∃ a : B, a ≠ 0 ∧ ∃ (nM : ℕ) (M : Scheme.{u})
(iM : M ⟶ Proj (MvPolynomial.homogeneousSubmodule (Fin (nM + 1)) (Localization.Away a))),
IsClosedImmersion iM ∧
∃ (U : Scheme.{u}) (j : U ⟶ Spec (.of (Localization.Away a))),
IsOpenImmersion j ∧ DenseRange j ∧
let u : U ⟶ N := (j ≫ Spec.map (CommRingCat.ofHom (algebraMap B (Localization.Away a)))) ≫ t
IsOpenImmersion u ∧ IsDominant u ∧
∃ (Y Z : Scheme.{u}) (pY : Y ⟶ U) (pZ : Z ⟶ U)
(tY : Y ⟶ M) (tZ : Z ⟶ X) (h : Y ⟶ Z),
IsPullback tY pY (iM ≫ ProjSpace.π (Localization.Away a) nM) j ∧
IsPullback tZ pZ q (u ≫ qN) ∧ Flat pZ ∧ IsProper pZ ∧
h ≫ pZ = pY ∧ IsProper h ∧ Surjective h ∧
Smooth pY ∧ IsProper pY ∧ IsDominant (h ≫ tZ) ∧
∃ (C₀ : Scheme.{u}) (p₀ : C₀ ⟶ Spec (.of E)) (a₀ : C₀ ⟶ q.fiber η),
Smooth p₀ ∧ IsProper p₀ ∧ topologicalKrullDim C₀ ≤ 1 ∧
IsProper a₀ ∧ Surjective a₀ ∧
a₀ ≫ q.fiberToSpecResidueField η = p₀ ≫
Spec.map (CommRingCat.ofHom (algebraMap (S.residueField η) E)) ∧
(∃ i₀ : C₀ ⟶ Proj (MvPolynomial.homogeneousSubmodule (Fin (nM + 1)) E),
IsClosedImmersion i₀ ∧ i₀ ≫ ProjSpace.π E nM = p₀) ∧
∃ (b : Spec (.of E) ⟶ U) (eC : C₀ ⟶ Y),
b ≫ u = Spec.map (CommRingCat.ofHom (algebraMap B E)) ≫ t ∧
IsPullback eC p₀ pY b ∧ eC ≫ h ≫ tZ = a₀ ≫ q.fiberι η := by sorry