/- Copyright 2026 The Formal Conjectures Authors. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. -/ module public import FormalConjecturesUtil

Resolution of singularities

A variety $X$ over a field $k$ admits a resolution of singularities if there is a smooth $k$-variety $Y$ and a proper birational morphism $Y \to X$. Hironaka proved that every variety over a field of characteristic zero admits one. In positive characteristic this is known over a perfect field in dimension at most three, and is open from dimension four on. Perfectness of $k$ cannot be dropped from the statement in this form: over an imperfect field it fails already in dimension zero, see exists_not_hasResolution_of_not_perfectField below. Over an arbitrary field one asks instead that $Y$ be regular.

References:

@[expose] public sectionopen CategoryTheoryuniverse unamespace AlgebraicGeometry

A scheme X over a base S with structure morphism sX : X ⟶ S admits a resolution of singularities if there is an integral scheme Y, smooth over S, together with a proper morphism f : Y ⟶ X that is birational over X, i.e. that restricts to an isomorphism between dense open subschemes of Y and of X.

def Scheme.HasResolution {S X : Scheme.{u}} (sX : X ⟶ S) : Prop := ∃ (Y : Scheme.{u}) (f : Y ⟶ X), IsIntegral Y ∧ IsProper f ∧ Smooth (f ≫ sX) ∧ Scheme.BirationalOver f (𝟙 X)

A scheme that is already smooth over the base is its own resolution.

@[category test, AMS 14] theorem Scheme.hasResolution_of_smooth {S X : Scheme.{u}} (sX : X ⟶ S) [IsIntegral X] [Smooth sX] : Scheme.HasResolution sX := ⟨X, 𝟙 X, ‹_›, inferInstance, S:SchemeX:SchemesX:X ⟶ Sinst✝¹:IsIntegral Xinst✝:Smooth sX⊢ Smooth (𝟙 X ≫ sX) rwa [S:SchemeX:SchemesX:X ⟶ Sinst✝¹:IsIntegral Xinst✝:Smooth sX⊢ Smooth sXS:SchemeX:SchemesX:X ⟶ Sinst✝¹:IsIntegral Xinst✝:Smooth sX⊢ Smooth sX, .refl _⟩

A proper birational morphism onto a scheme is surjective. In particular a resolution of singularities is onto, which is what rules out the inclusion of the smooth locus of X as a trivial solution.

All goals completed! 🐙

Resolution of singularities in positive characteristic. Let $k$ be a perfect field of characteristic $p > 0$ and let $X$ be an integral scheme that is separated and of finite type over $k$. Then there is an integral scheme $Y$ that is smooth over $k$ together with a proper birational morphism $Y \to X$.

This is open from dimension four on; see [Hau2010]. Perfectness of $k$ is needed for the conclusion as stated; see exists_not_hasResolution_of_not_perfectField.

@[category research open, AMS 14] theorem resolution_of_singularities (k : Type u) [Field k] [PerfectField k] (p : ℕ) [Fact p.Prime] [CharP k p] {X : Scheme.{u}} (sX : X ⟶ Spec (.of k)) [IsIntegral X] [LocallyOfFiniteType sX] [QuasiCompact sX] [IsSeparated sX] : Scheme.HasResolution sX := k:Type uinst✝⁷:Field kinst✝⁶:PerfectField kp:ℕinst✝⁵:Fact (Nat.Prime p)inst✝⁴:CharP k pX:SchemesX:X ⟶ Spec (CommRingCat.of k)inst✝³:IsIntegral Xinst✝²:LocallyOfFiniteType sXinst✝¹:QuasiCompact sXinst✝:IsSeparated sX⊢ Scheme.HasResolution sX All goals completed! 🐙

Hironaka's theorem. Every integral scheme that is separated and of finite type over a field of characteristic zero admits a resolution of singularities. This is the main theorem of [Hir1964]; see [Kol2007], Theorem 36, for a statement in the generality used here.

@[category research solved, AMS 14] theorem resolution_of_singularities_of_charZero (k : Type u) [Field k] [CharZero k] {X : Scheme.{u}} (sX : X ⟶ Spec (.of k)) [IsIntegral X] [LocallyOfFiniteType sX] [QuasiCompact sX] [IsSeparated sX] : Scheme.HasResolution sX := k:Type uinst✝⁵:Field kinst✝⁴:CharZero kX:SchemesX:X ⟶ Spec (CommRingCat.of k)inst✝³:IsIntegral Xinst✝²:LocallyOfFiniteType sXinst✝¹:QuasiCompact sXinst✝:IsSeparated sX⊢ Scheme.HasResolution sX All goals completed! 🐙

Resolution of singularities holds in dimension at most three over any perfect field. In characteristic zero this is [Hir1964]. In positive characteristic, dimension at most two is [Lip1978] and dimension three is due to Cossart and Piltant: [CP2008] and [CP2009] prove it for quasi-projective varieties over a field that is differentially finite over a perfect subfield, and [CP2019], Theorem 1.1, removes both restrictions, resolving every reduced separated Noetherian quasi-excellent scheme of dimension at most three in any characteristic. That theorem gives an everywhere regular source, which over a perfect field is the same as a smooth one. The problem is open in positive characteristic from dimension four on.

@[category research solved, AMS 14] theorem resolution_of_singularities_of_topologicalKrullDim_le_three (k : Type u) [Field k] [PerfectField k] {X : Scheme.{u}} (sX : X ⟶ Spec (.of k)) [IsIntegral X] [LocallyOfFiniteType sX] [QuasiCompact sX] [IsSeparated sX] (hX : topologicalKrullDim X ≤ 3) : Scheme.HasResolution sX := k:Type uinst✝⁵:Field kinst✝⁴:PerfectField kX:SchemesX:X ⟶ Spec (CommRingCat.of k)inst✝³:IsIntegral Xinst✝²:LocallyOfFiniteType sXinst✝¹:QuasiCompact sXinst✝:IsSeparated sXhX:topologicalKrullDim ↥X ≤ 3⊢ Scheme.HasResolution sX All goals completed! 🐙

Resolution of singularities fails over imperfect fields. If the field $k$ is not perfect, then there is an integral scheme $X$ of dimension zero that is separated and of finite type over $k$ and has no resolution of singularities: no integral scheme smooth over $k$ admits a proper birational morphism to $X$.

A witness: $k$ has characteristic $p > 0$ and there is some $a \in k \setminus k^p$. Take $X = \operatorname{Spec} k(a^{1/p})$. Every integral scheme birational to $X$ has function field $k(a^{1/p})$, which is not separable over $k$, whereas the function field of an integral scheme smooth over $k$ is separable over $k$.

So the hypothesis PerfectField k cannot be dropped from resolution_of_singularities or from resolution_of_singularities_of_topologicalKrullDim_le_three.

@[category textbook, AMS 14] theorem exists_not_hasResolution_of_not_perfectField (k : Type u) [Field k] (hk : ¬ PerfectField k) : ∃ (X : Scheme.{u}) (sX : X ⟶ Spec (.of k)), IsIntegral X ∧ LocallyOfFiniteType sX ∧ QuasiCompact sX ∧ IsSeparated sX ∧ topologicalKrullDim X = 0 ∧ ¬ Scheme.HasResolution sX := k:Type uinst✝:Field khk:¬PerfectField k⊢ ∃ X sX, IsIntegral X ∧ LocallyOfFiniteType sX ∧ QuasiCompact sX ∧ IsSeparated sX ∧ topologicalKrullDim ↥X = 0 ∧ ¬Scheme.HasResolution sX All goals completed! 🐙end AlgebraicGeometry