A finite multisection of a proper surjection onto a smooth curve
ProvedAlgebraicGeometry.SmoothProperCurve.exists_finite_multisection_of_proper_surjectivealgebraic-geometrycurvesnumerical-equivalence
Let be an algebraically closed field, let be a smooth proper integral curve over , and let be a proper surjective morphism. There exist a smooth proper integral curve over and a morphism such that
is finite and surjective. The structure morphism of is the composite of with the structure morphism of .
No smoothness, reducedness, or integrality is assumed for . The map may have positive-dimensional fibres, and need not be separable. This form of the normalized multisection theorem provides finite covers of curve tests through arbitrary proper surjections.
Preamble
import Mathlib.AlgebraicGeometry.Morphisms.Proper import Mathlib.AlgebraicGeometry.Morphisms.Finite import Mathlib.AlgebraicGeometry.Morphisms.Smooth set_option autoImplicit false open AlgebraicGeometry CategoryTheory universe u
Formal statement
theorem AlgebraicGeometry.SmoothProperCurve.exists_finite_multisection_of_proper_surjective
(K : Type u) [Field K] [IsAlgClosed K] {C Z : Scheme.{u}}
(c : C ⟶ Spec (.of K)) [IsIntegral C] [IsProper c]
[SmoothOfRelativeDimension 1 c] (p : Z ⟶ C) [IsProper p] [Surjective p] :
∃ (D : Scheme.{u}) (b : D ⟶ Z),
IsIntegral D ∧ IsProper ((b ≫ p) ≫ c) ∧
SmoothOfRelativeDimension 1 ((b ≫ p) ≫ c) ∧
IsFinite (b ≫ p) ∧ Surjective (b ≫ p) := by
sorrySource
Derived normalized-multisection consequence of the Stacks Project: normalization of varieties, Lemma 33.27.1 (https://stacks.math.columbia.edu/tag/0BXR); normal proper models of curve function fields, Theorem 53.2.6 (https://stacks.math.columbia.edu/tag/0BY1); smoothness over a perfect field, Lemma 53.2.8 (https://stacks.math.columbia.edu/tag/0BY3); and finiteness of nonconstant maps of proper curves, Lemma 53.2.4 (https://stacks.math.columbia.edu/tag/0CCL). Choose a closed point of the nonempty generic fibre of p, take its reduced closure, and normalize. Its residue field is finite over K(C), so the closure has dimension one. Properness and dominance give a surjective curve map. Algebraically closed K is perfect. This states the resulting multisection consequence, not a verbatim transcription of any one lemma.