A moving lemma for cohomology with support

Basic Information

• Paper ID: 2207.08297
• Title: A moving lemma for cohomology with support
• Author: Stefan Schreieder (Leibniz University Hannover)
• Classification: math.AG (Algebraic Geometry), math.KT (K-Theory)
• Publication Date: 2024, published in Épijournal de Géométrie Algébrique
• Paper Link: https://arxiv.org/abs/2207.08297

Abstract

This paper proves a moving lemma for a class of natural cohomology theories with support (including étale or pro-étale cohomology with appropriate coefficients) on smooth quasi-projective k-varieties admitting smooth projective compactifications (such as in characteristic zero). For such k-varieties and cohomology theories, this yields the following important consequences: local and global generalizations of effacement theorems of Quillen, Bloch-Ogus, and Gabber; a finite level version of the Gersten conjecture in characteristic zero; and generalizations of injectivity properties and codimension j+1 purity theorems for étale cohomology. The results show that the refined unramified cohomology groups in Sch23 are motivic.

Research Background and Motivation

Problem Background

In algebraic geometry, Chow's moving lemma allows moving algebraic cycles to good positions relative to a closed subset S⊂X in the sense of rational equivalence. The essence of the Gersten conjecture is an effacement theorem, which is equivalent to an analogous moving lemma for classes with support in special cases (X is affine, S is a finite set of points).

Research Motivation

1. Theoretical Completeness: Are the effacement theorems of Quillen, Bloch-Ogus, and Gabber special cases of a more general moving lemma?
2. Technical Requirements: Can one move classes with support Z⊂X to support Z' such that Z' is in good position relative to any given closed subset S⊂X?
3. Application Value: Such a moving lemma would yield new applications transcending the original Gersten conjecture.

Limitations of Existing Methods

• Classical effacement theorems hold only in special cases (affine varieties, zero-dimensional support)
• Lack of general theory for handling high-dimensional support S
• Existing methods perform poorly in localization

Core Contributions

1. Main Moving Lemma: Proves a moving lemma for cohomology classes with support on smooth quasi-projective k-varieties admitting smooth projective compactifications
2. Generalization of Effacement Theorems: Provides local and global generalizations of theorems of Quillen, Bloch-Ogus, and Gabber
3. Finite Level Version of Gersten Conjecture: Establishes a finite level version of the Gersten conjecture in characteristic zero
4. Generalization of Purity Theorems: Generalizes injectivity properties and codimension j+1 purity theorems for étale cohomology
5. Motivic Properties: Proves motivic properties of refined unramified cohomology groups

Methodology in Detail

Task Definition

For smooth equidimensional algebraic k-schemes X and closed subsets S,Z⊂X (dimZ < dimX), find closed subsets Z'⊂W⊂X such that:

• Z⊂W, dimZ' = dimZ, dimW = dimZ+1
• Z' and W\Z intersect S appropriately
• For any α∈H_Z(X,n), there exists α'∈H_{Z'}(X,n) such that α and α' have the same image in H*_W(X,n)

Core Technical Framework

1. Twisted Cohomology Theory

Define cohomology theories (X,Z) ↦ H*_Z(X,n) satisfying:

• C1 (Excision): Compatibility with open immersions
• C2 (Pushforwards): Pushforward maps for proper morphisms
• C3 (Long Exact Sequence): Long exact sequence for triples
• C4 (Cycle Action): Action of algebraic cycles
• C5 (Semi-purity): Vanishing under dimension conditions

2. Cycle Action on Open Varieties

For smooth projective equidimensional schemes X,Y and cycle Γ∈Z^c(X×Y), construct the action:

Γ^(W)_*: H^i_Z(U,n) → H^{i+2c-2d_X}_{Z'}(U',n+c-d_X)

where U=X\R, U'=Y\R' are appropriate open subsets.

3. Moving Lemma Proof Strategy

1. Reduction to Projective Case: Utilize smooth projective compactification
2. Apply Chow Moving Lemma: Apply Levine's moving lemma to the diagonal Δ_X⊂X×X
3. Utilize Cycle Action: Move classes through cycle action on cohomology
4. Handle Technical Difficulties: Technical treatment when X is only an open subset of a projective scheme

Technical Innovations

1. Unified Theoretical Framework

Unifies different cohomology theories (étale, pro-étale, etc.) under a single axiomatic framework.

2. Refined Analysis of Cycle Action

Detailed analysis of algebraic cycle action on cohomology of open varieties, particularly:

• Compatibility with rational equivalence
• Localization behavior
• Functoriality properties

3. Globalization of Moving Lemma

Generalizes the classical affine case to general quasi-projective varieties and handles support of arbitrary dimension.

Main Theorems

Theorem 1.1 (Moving Lemma)

Let X be a smooth equidimensional k-scheme admitting a smooth projective compactification, and let S,Z⊂X be closed subsets with dimZ < dimX. Then there exist closed subsets Z'⊂W⊂X satisfying:

• Z⊂W, dimZ' = dimZ, dimW = dimZ+1
• Z' and W\Z intersect S appropriately
• For any α∈H_Z(X,n), there exists α'∈H_{Z'}(X,n) such that α and α' have the same image in H*_W(X,n)

Corollary 1.2 (Global Effacement)

Under the above conditions, if dimS + dimZ < dimX, then there exist a neighborhood U⊂X of S and a closed subset W⊂X such that the composite map:

H*_Z(X,n) → H*_W(X,n) → H*_W(U,n)

is zero.

Corollary 1.5 (Finite Level Version of Gersten Conjecture)

For smooth affine varieties over a characteristic zero field, given a chain of points Z_c = {x} ⊂ Z_ ⊂ ⋯ ⊂ Z_1 ⊂ Z_0 = X_x, after appropriate refinement, the following complex is exact:

0 → H^i(X_x,n) → H^i_{BM}(X_x\Z_1) → ⋯ → H^0_{BM}(Z_i\Z_{i+1}) → 0

Applications and Results

1. Motivic Properties of Refined Unramified Cohomology

Corollary 1.7: For smooth projective equidimensional schemes X,Y, there exists a biadditive pairing:

CH^c(X×Y) × H^i_{j,nr}(X,n) → H^{i+2c-2d_X}_{j+c-d_X,nr}(Y,n+c-d_X)

which is functorial with respect to composition of correspondences.

2. Projective Bundle Formula

Corollary 6.12: There exists a canonical isomorphism:

⊕_{l=0}^{min(j,n)} H^{i-2l}_{j-l,nr}(Y,m-l) ≃ H^i_{j,nr}(Y×ℙ^n_k,m)

3. Birational Invariance

Corollary 6.14: If f: X ⤏ Y is a birational map that is an isomorphism in codimension c, then for j ≤ c:

f*: H^i_{j,nr}(Y,n) ≃ H^i_{j,nr}(X,n)

Technical Details

Pro-étale Cohomology

The paper provides detailed verification that pro-étale cohomology satisfies the required axioms C1-C5. Key techniques include:

• Poincaré Duality: f*_(d)2d ≃ f!
• Purity: i*_(-c)-2c ≃ i!
• Six Functor Formalism: Complete system of pushforward and pullback functors

Cycle Class Map

Constructs the cycle class map:

cl^X_Z(Γ) ∈ H^{2c}_Z(X,ℤ_ℓ(c))

satisfying:

• Linearity
• Compatibility with pushforward and pullback maps
• Cup product formula

Related Work

Classical Results

• Chow Moving Lemma Cho56: Moving of algebraic cycles
• Quillen Qui73: Effacement for K-theory
• Bloch-Ogus BO74: Gersten conjecture for étale cohomology
• Gabber Gab94: Proof in the general case

Modern Developments

• Levine Lev98,Lev05: Moving lemma for higher Chow groups
• Refined Unramified Cohomology Sch23: Author's previous work

Conclusions and Discussion

Main Conclusions

1. Establishes a general moving lemma for cohomology with support
2. Unifies and generalizes multiple classical results
3. Proves motivic properties of refined unramified cohomology
4. Provides a finite level version of the Gersten conjecture

Limitations

1. Requires existence of smooth projective compactification (as in characteristic zero)
2. The moved subset Z' generally does not localize well
3. Verification of certain technical conditions (such as appropriate intersection) may be complex

Future Directions

1. Generalization to more general base schemes
2. Connections with motivic homotopy theory
3. Applications in arithmetic geometry

In-Depth Evaluation

Strengths

1. Theoretical Unification: Unifies multiple seemingly independent results under a single framework
2. Technical Innovation: Refined analysis of cycle action and globalization of moving lemma
3. Broad Applicability: Important applications from abstract theory to concrete computations
4. Complete Proofs: Technical details are handled appropriately, particularly verification of pro-étale cohomology

Weaknesses

1. Technical Complexity: Proofs involve substantial technical details that may affect readability
2. Limited Scope: Certain restrictions on the characteristic of the base field
3. Computational Difficulty: Actual applications may still present computational challenges

Impact

1. Theoretical Contribution: Provides new tools for cohomology theory in algebraic geometry
2. Practical Value: Potential applications in motivic theory and arithmetic geometry
3. Reproducibility: Theoretical results with relatively straightforward verification

Applicable Scenarios

1. Cohomology computations in algebraic geometry
2. Research in motivic theory
3. Applications in arithmetic geometry
4. K-theory and cycle theory

References

Main references include:

• Cho56 W.-L. Chow, "On equivalence classes of cycles in an algebraic variety"
• BO74 S. Bloch and A. Ogus, "Gersten's conjecture and the homology of schemes"
• Qui73 D. Quillen, "Higher algebraic K-theory, I"
• BS15 B. Bhatt and P. Scholze, "The pro-étale topology of schemes"
• Sch23 S. Schreieder, "Refined unramified cohomology of schemes"

This paper makes significant contributions to cohomology theory in algebraic geometry, not only unifying multiple classical results but also providing powerful tools for further research. Its technical innovations and theoretical depth make it an important reference in the field.