Etale and crystalline companions, I

Basic Information

• Paper ID: 1811.00204
• Title: Etale and crystalline companions, I
• Author: Kiran S. Kedlaya (University of California San Diego)
• Classification: math.NT math.AG (Number Theory, Algebraic Geometry)
• Published Journal: Épijournal de Géométrie Algébrique, Volume 6 (2022), Article No. 20
• Paper Link: https://arxiv.org/abs/1811.00204

Abstract

This paper investigates coefficient objects on smooth schemes over finite fields and their companion theories. For a smooth scheme $X$ over a finite field $k$ of characteristic $p$, we consider coefficient objects of locally constant rank in $\ell$-adic Weil cohomology: lisse Weil $\mathbb{Q}_\ell$-sheaves in étale cohomology when $\ell \neq p$, and overconvergent $F$-isocrystals in rigid cohomology when $\ell = p$. Using the Langlands correspondence for global function fields in both the étale and crystalline settings (due to Lafforgue and Abe respectively), we know that in the curve case, any coefficient object in one category has a "companion" in other categories with matching Frobenius characteristic polynomials at closed points. This paper generalizes Drinfeld's method to prove that any crystalline coefficient object has an étale companion.

Research Background and Motivation

Core Problem

The core problem addressed in this paper is Deligne's companion conjecture (Conjecture 1.1) posed during his study of the Weil conjectures, namely to prove the existence of correspondence relations between coefficient objects in different $\ell$-adic cohomology theories.

Research Significance

1. Theoretical Unification: Establishing bridges between different cohomology theories, unifying the study of étale cohomology and rigid cohomology
2. Langlands Program: Closely related to the Langlands correspondence over function fields, providing important tools for arithmetic geometry
3. Weight Theory: Extending Deligne's weight theory to the $p$-adic case, perfecting cohomology theories in algebraic geometry

Limitations of Existing Methods

1. Dimensional Restrictions: Previous results were mainly limited to the curve case
2. Coefficient Restrictions: Lack of effective treatment for the $\ell = p$ case
3. Technical Obstacles: Inability to directly apply representation-theoretic methods to crystalline coefficient objects

Core Contributions

1. Main Theorem: Proves that any crystalline coefficient object has an étale companion (part of Theorem 1.3)
2. Methodological Innovation: Generalizes Drinfeld's method to the crystalline case, avoiding direct use of representation theory
3. Technical Breakthrough: Establishes a systematic method for handling the $\ell = p$ case
4. Theoretical Completion: Proves most of Deligne's conjecture (except part (vi))

Detailed Methodology

Task Definition

Given a smooth scheme $X$ over a finite field $k$, we study companion relations between coefficient objects in different categories:

• Input: A coefficient object $E$ in some category (lisse Weil $\mathbb{Q}_\ell$-sheaf or overconvergent $F$-isocrystal)
• Output: Companion objects in other categories with matching Frobenius characteristic polynomials
• Constraints: Preserving algebraicity, weights, and other important properties

Core Methodological Architecture

1. Foundation in the Curve Case (Section 3)

Establishing complete theory on curves using the Langlands correspondence: $\text{Coefficient objects} \leftrightarrow \text{Automorphic representations}$

Theorem 3.4: There exists a bijection between irreducible coefficient objects and cuspidal automorphic representations, preserving local properties.

2. Weight Theory (Section 4.1)

Generalizing Deligne's weight theory:

Theorem 4.11:

• (a) Every irreducible coefficient object is pure of some weight
• (b) There exists a unique weight filtration: $0 = E_0 \subset \cdots \subset E_l = E$

3. Lefschetz Slice Principle (Section 4.2)

Lemma 4.14: For a geometrically irreducible coefficient object $E$ and a closed point $x \in X^{\circ}$, there exist a positive integer $n$ and a curve $C$ in $X_n$ containing $x$ such that $E|_C$ is irreducible.

4. Uniqueness of Companions (Section 4.3)

Theorem 4.17 (Tsuzuki): Let $E_1, E_2$ be algebraic coefficient objects that are companions of each other. Then:

• (a) If $E_1$ is irreducible, then $E_2$ is also irreducible
• (b) If $E_1, E_2$ are in the same category, they have the same semisimplification

Technical Innovations

1. Avoiding Representation Theory: Replacing the Chebotarëv density theorem for mod-$\ell^n$ representations with cohomological methods
2. Semistable Reduction: Utilizing the semistable reduction theorem for overconvergent $F$-isocrystals to handle wild ramification at the boundary
3. Weight Arguments: Using weight theory to uniformly handle different characteristic cases

Experimental Setup

This is a pure theoretical mathematics paper with no numerical experiments; results are verified through rigorous mathematical proofs.

Verification Methods

1. Curve Verification: Verifying general theory through known results in the curve case
2. Consistency Checks: Verifying matching of $L$-functions between different categories
3. Local-Global Compatibility: Checking consistency between local and global properties

Main Results

Core Theorem

Theorem 1.3: Parts (i)-(v) of Deligne's conjecture hold:

• (i) Purity: $E$ is a pure object of weight 0
• (ii) Algebraicity: $E$ is absolutely algebraic
• (iii) $p$-Triviality: Trivial valuations at primes not dividing $p$
• (iv) Valuation Bounds: Bounded valuations at primes above $p$
• (v) Existence of Étale Companion: An étale companion exists

Key Technical Results

Theorem 4.28: Every algebraic coefficient object has an étale companion, preserving irreducibility and finite determinant properties.

Corollary 4.29: When $\ell' \neq p$, the companion conjecture holds.

Related Work

Historical Development

1. Deligne (1981): Proposes the companion conjecture, establishes foundations of weight theory
2. Lafforgue (2002): Establishes the étale Langlands correspondence for $\text{GL}_n$
3. Drinfeld (2012): Proves that étale coefficient objects have étale companions
4. Abe (2018): Extends the Langlands correspondence to the crystalline case

Contributions of This Paper

Compared to prior work, this paper:

• Generalizes Drinfeld's method to the crystalline case
• Establishes a unified weight theory framework
• Avoids direct dependence on representation theory

Conclusions and Discussion

Main Conclusions

1. Successfully proves the existence of étale companions for crystalline coefficient objects
2. Establishes a systematic method for handling higher-dimensional cases
3. Lays the foundation for subsequent construction of crystalline companions

Limitations

1. Part (vi) Unresolved: The existence of crystalline companions is left for future work
2. Smoothness Assumption: Requires $X$ to be smooth; additional techniques are needed for normal but non-smooth schemes
3. Finite Field Restriction: Methods depend on special properties of finite fields

Future Directions

1. Crystalline Companion Construction: Addressed in the subsequent paper Ked22b
2. Non-smooth Cases: Extending to more general schemes
3. Application Development: Further applications in arithmetic geometry and representation theory

In-Depth Evaluation

Strengths

1. Methodological Innovation: Successfully circumvents representation-theoretic obstacles and establishes a unified framework
2. Theoretical Depth: Profound application of algebraic geometry, number theory, and cohomology theory
3. Technical Rigor: Complete proofs with clear logical structure
4. Far-reaching Impact: Provides important tools for related fields

Weaknesses

1. Technical Complexity: Requires extensive background knowledge with high entry barriers
2. Incompleteness: The final part of the core conjecture remains unresolved
3. Limited Applications: Primarily theoretical contributions with relatively limited direct applications

Influence

1. Academic Value: Advances cross-disciplinary research between algebraic geometry and number theory
2. Methodological Contribution: Provides new methods for handling $p$-adic cases
3. Foundation for Future Work: Establishes important foundations for related research

Applicable Scenarios

1. Arithmetic Geometry Research: Particularly $p$-adic cohomology theory
2. Langlands Program: In-depth study of the function field case
3. Algebraic Geometry: Cohomology theory on smooth schemes

References

The paper cites numerous important references, including:

• Del81 Deligne's Weil II
• Laf02 Lafforgue's Langlands correspondence
• Dri12 Drinfeld's work on Deligne's conjecture
• Abe18a Abe's crystalline Langlands correspondence
• AE19 Abe-Esnault's related work

These references form important support for the theoretical foundations of this paper.