Tilting modules for reductive algebraic groups: characters and support varieties

Basic Information

• Paper ID: 2511.05063
• Title: Tilting modules for reductive algebraic groups: characters and support varieties
• Authors: Pramod N. Achar (Louisiana State University), Simon Riche (Université Clermont Auvergne)
• Classification: math.RT (Representation Theory)
• Submission Date/Conference: Submitted November 2025, contribution to ICM 2026 proceedings
• Paper Link: https://arxiv.org/abs/2511.05063

Abstract

This paper is a contribution by the authors to the ICM 2026 proceedings, surveying a series of results obtained by the authors and their collaborators on representation theory of reductive algebraic groups over algebraically closed fields of positive characteristic. The research focuses on two core problems concerning tilting modules: character formulas and support varieties.

Research Background and Motivation

Core Problems

1. Character Computation Problem: Over an algebraically closed field of positive characteristic p, how can one compute characters of simple modules and tilting modules for a reductive algebraic group G? This is a fundamental problem in representation theory, analogous to character theory in complex representations of finite groups.
2. Support Variety Description Problem: How can one describe support varieties of tilting modules? This involves the Humphreys conjecture, which predicts a deep connection between support varieties and nilpotent orbits.

Research Significance

• Characters determine the composition factors of representations and their multiplicities, serving as key invariants for understanding representation structure
• Tilting modules play a special role in representation theory: they are closed under tensor products, and their characters can be used to derive characters of simple modules
• These problems are deeply connected to Kazhdan-Lusztig combinatorics of the affine Weyl group, exemplifying the interplay between combinatorics, geometry, and representation theory

Limitations of Existing Methods

1. Lusztig Conjecture: Uses classical Kazhdan-Lusztig polynomials to give character formulas for simple modules, but only holds when p is sufficiently large, and Williamson found counterexamples
2. Andersen Conjecture: Proposes similar formulas for tilting modules, but also requires p ≥ h (where h is the Coxeter number) with restrictions on weights
3. Asymptotic Problem: Classical Kazhdan-Lusztig combinatorics only provides "asymptotic" approximations, failing in general characteristic

Innovations in This Paper

Introduces p-canonical basis to replace classical Kazhdan-Lusztig basis, a new tool developed by Williamson and others that provides exact formulas in more general characteristics (p > h, or even for all p).

Core Contributions

The main contributions of this paper include:

1. Tilting Module Character Formula (Theorem 2.6): Proves the tilting module character formula using p-Kazhdan-Lusztig polynomials for p > h, with versions existing for all p
2. Application of Finkelberg-Mirković Conjecture: When p > h+1, provides a geometric formula for simple module characters through this conjecture (proven by Bezrukavnikov and Riche)
3. Proof of Humphreys Conjecture:
   • Proves traditional and relative Humphreys conjectures when p is sufficiently large
   • Proves relative Humphreys conjecture for all p > h
   • Proves scheme-theoretic version for GLₙ when p > n
4. co-t-structure Theory: Develops co-t-structure theory for the derived category of coherent sheaves on the nilpotent cone, providing a new framework for describing relative cohomology of tilting modules
5. Module Version of Lusztig-Vogan Bijection: Establishes Lusztig-Vogan bijection in positive characteristic, connecting dominant weights with tilting vector bundles on nilpotent orbits

Detailed Methodology

Theoretical Framework

1. Representation Theory Setup

• Reductive Algebraic Groups: G is a connected reductive algebraic group over an algebraically closed field k of characteristic p > 0
• Key Module Classes:
  • Induced modules N(λ): induced from 1-dimensional representations of the Borel subgroup
  • Simple modules L(λ): irreducible representations
  • Tilting modules T(λ): indecomposable modules satisfying special filtration conditions

2. Affine Weyl Group and Dot Action

The affine Weyl group $W_{aff} = W \ltimes \mathbb{Z}\Phi$ acts on the weight lattice via dot action: $(w \ltimes \mu) \bullet \lambda = w(\lambda + p\mu + \rho) - \rho$

Linkage Principle (Andersen): Ext groups of simple modules are nonzero only within the same $W_{aff}$-orbit, leading to block decomposition of the representation category.

Methodology for Character Formulas

Core Tool: p-canonical basis

For the affine Hecke algebra $\mathcal{H}_{aff}$, there exist:

• Classical Kazhdan-Lusztig basis $\{\underline{H}_w : w \in W_{aff}\}$
• p-canonical basis $\{^pH_w : w \in W_{aff}\}$ (depending on prime p)

Define corresponding polynomials: $^pH_w = \sum_{y \in W_{aff}} \,^ph_{y,w} \cdot H_y$

Key Property: When p is sufficiently large, $^pH_w = \underline{H}_w$, but may differ for general p.

Theorem 2.6 (Tilting Module Character Formula)

For finitary subsets $J \subset S_{aff}$ such that $C_J \neq \emptyset$, and $y, w \in W^J_{aff}$: $d^J_{y,w} = \sum_{z \in W} (-1)^{\ell(z)} \cdot \,^ph_{zyw_J, ww_J}(1)$

where $d^J_{y,w}$ are expansion coefficients of tilting module characters in the induced module basis.

Proof Strategy (p > h case)

1. Geometric Realization: Construct degrading functor $\mathcal{F}: D^b\text{Coh}_{G \times \mathbb{G}_m}(\tilde{\mathcal{N}}) \to D^b\text{Rep}_{W_{ext} \bullet 0}(G)$ where $\tilde{\mathcal{N}} = G \times_B \mathfrak{n}$ is the Springer resolution
2. Categorification: Prove existence of Hecke category $\mathcal{H}^k_{aff}$ acting on $\text{Rep}_{W_{aff} \bullet 0}(G)$
3. Koszul Duality: Utilize mixed perverse sheaves on the affine Grassmannian and Koszul duality theory
4. Smith-Treumann Theory: Most recent general proof uses Smith-Treumann theory in geometric Satake equivalence

Methodology for Support Varieties

Frobenius Kernels and Cohomology

• Frobenius Kernel: $G_1 = \ker(\text{Fr}: G \to G^{(1)})$, a finite group scheme
• Cohomology Ring: When p > h, $H^\bullet(G_1; k) \cong k[\mathcal{N}]$ where $\mathcal{N} \subset \mathfrak{g}$ is the nilpotent cone
• Support Variety Definition:
  • $V_{G_1}(M) = \text{supp}(\text{Ext}^\bullet(M,M))$
  • $V^{G_1}(M) = \text{supp}(H^\bullet(G_1; M))$ (relative support variety)

Humphreys Conjecture (Traditional Version, Conjecture 3.3)

For $w \in W^\emptyset_{ext}$, let C be the G-orbit corresponding through the Lusztig bijection (3.2) to the two-sided cell containing w, then: $V_{G_1}(\mathcal{T}(w \bullet 0)) = C$

Proof Outline

Step 1: Geometrization Using the degrading functor, for $\lambda \in X^+$, define: $\mathcal{S}_\lambda := \pi_* \mathcal{T}^{gr}(w_\lambda \bullet 0) \in D^b\text{Coh}_{G \times \mathbb{G}_m}(\mathcal{N})$

Then there is an isomorphism: $H^\bullet(G_1; \mathcal{T}(w_\lambda \bullet 0)) \cong R^\bullet\Gamma(\mathcal{N}, \mathcal{S}_\lambda)$

Step 2: co-t-structure Theory (Achar-Hardesty 5,6)

• Construct co-t-structure on $D^b\text{Coh}_{G \times \mathbb{G}_m}(\mathcal{N})$
• Prove that $\mathcal{S}_\lambda$ is an indecomposable object in the coheart (up to shift)

Step 3: Lusztig-Vogan Bijection Establish bijection: $X^+ \xleftrightarrow{\sim} \Xi_G = \{(C, \sigma) : C \subset \mathcal{N} \text{ is an orbit}, \sigma \in \Sigma_C\}$

where $\Sigma_C$ parametrizes simple equivariant coherent sheaves on orbit C.

Step 4: Support Characterization (Theorem 3.10) If $\lambda$ corresponds to $(C, \sigma)$, then: $\mathcal{S}_\lambda \cong \mathcal{S}(C, \sigma)$ where $\mathcal{S}(C, \sigma)$ is geometrically constructed, satisfying:

• Support on C
• Restriction to C is a tilting vector bundle $\mathcal{T}(\sigma)$

This directly yields the relative Humphreys conjecture.

Experimental Setup

This is a pure mathematics theoretical paper with no traditional "experiments." However, the authors verify the theory through:

Special Case Verification

1. GLₙ Case: First verified the tilting character formula for GLₙ in 45
2. Quantum Group Analogy: Many results have corresponding versions in the quantum group case, which have been independently verified

Computational Verification

• Use computational tools for Kazhdan-Lusztig polynomials and p-canonical basis
• For small rank groups and small characteristic, explicit computations can verify the theory

Theoretical Consistency Checks

• Verify formulas are consistent with known Weyl character formula (2.1)
• Check compatibility with translation functors (2.5)
• Verify consistency with Steinberg tensor product theorem

Experimental Results

Summary of Main Theorems

1. Tilting Module Characters (Theorem 2.6)

Result: For all finitary subsets J and p > h, tilting module characters are completely determined by p-Kazhdan-Lusztig polynomials.

Significance:

• First exact formula for general p
• Automatically reduces to Andersen conjecture as p → ∞
• Versions exist for all p (Riche-Williamson 47)

2. Status of Humphreys Conjecture Proof

Key Progress:

• 8 (2019): Proof for sufficiently large p
• 6 (2024): Proof of relative version via Lusztig-Vogan bijection (p > h)
• 5 (2024): Scheme-theoretic version for GLₙ (p > n)

Theoretical Discoveries

Discovery 1: Central Role of p-canonical basis

Classical Kazhdan-Lusztig combinatorics provides only "asymptotic" approximation; the true controller of representation theory is the p-canonical basis. This explains Williamson's counterexamples.

Discovery 2: Necessity of co-t-structure

In positive characteristic, relative cohomology of tilting modules corresponds not to simple perverse-coherent sheaves, but to indecomposable objects in the coheart of a co-t-structure. This is an essential difference from the quantum group case.

Discovery 3: Geometric Meaning of Finkelberg-Mirković Conjecture

This conjecture (Conjecture 2.10) provides an equivalence between the representation category and the Whittaker category on the affine Grassmannian of the Langlands dual group, meaningful even when p < h (i.e., $C_\emptyset = \emptyset$).

Related Work

Historical Context

1970s: Verma's Prediction

Verma 50 first suggested that positive characteristic representation theory is related to affine Weyl group combinatorics, proposing the linkage principle and character independence conjecture.

1980s: Lusztig Conjecture

Lusztig 37 proposed character formulas for simple modules using Kazhdan-Lusztig polynomials, expecting validity when p ≥ h.

1990s: Andersen's Work

• 15: Proves linkage principle
• 16: Proposes tilting module character conjecture, proves it implies Lusztig conjecture

2000s: Quantum Group Analogy

• Bezrukavnikov 22: Resolves Humphreys conjecture for quantum groups
• Arkhipov-Bezrukavnikov-Ginzburg: Develop geometric methods

2010s: Counterexamples and New Directions

• Williamson 53: Discovers counterexamples to Lusztig conjecture
• Juteau-Mautner-Williamson: Introduce p-canonical basis

2018-2024: Series of Works by Present Authors

• 12 (2018): Construct degrading functor
• 11 (2019): Prove tilting character formula (p > h)
• 8 (2019): Prove Humphreys conjecture (p≫0)
• 47 (2022): New proof using Smith-Treumann theory
• 6 (2024): Complete relative Humphreys conjecture

Comparison with Related Work

Conclusions and Discussion

Main Conclusions

1. p-canonical basis is the correct combinatorial tool: It precisely captures the combinatorial structure of positive characteristic representation theory, rather than classical Kazhdan-Lusztig basis
2. Tilting module characters completely resolved: Theorem 2.6 provides complete answer to tilting character problem (p > h or all p)
3. Humphreys conjecture essentially resolved: Relative version holds for all p > h; traditional version holds for sufficiently large p or A-type
4. Power of geometric methods: Through Springer resolution, affine Grassmannian, and Langlands duality, representation-theoretic problems are transformed into geometric problems

Limitations

1. Characteristic Bound Issues

• Tilting character formula requires p > h (though versions for all p exist)
• Traditional Humphreys conjecture still requires sufficiently large p (except for A-type)
• Validity range of Donkin formula remains unclear (p ≥ 2h-4?)

2. Explicitness of Simple Module Characters

Although derivable from tilting modules in principle:

• Finkelberg-Mirković conjecture gives formulas involving hard-to-compute Euler characteristics
• No simple closed form like Weyl formula exists

3. Cellular Compatibility

• Relationship between p-cells and Kazhdan-Lusztig cells understood only when p is sufficiently large
• This limits proof of traditional Humphreys conjecture

4. Computational Complexity

• Computing p-canonical basis is extremely difficult
• No uniform characteristic bound N(w) exists for all elements w

Future Directions

1. Explicit Formulas for Simple Module Characters

Problem: Can one find more explicit simple module character formulas than Finkelberg-Mirković conjecture?

Possible Approaches:

• Understand Euler characteristics of Iwahori-equivariant intersection cohomology on affine flag varieties
• Develop new combinatorial tools

2. Small Characteristic Case

Problem: What is the representation-theoretic structure when p ≤ h?

Challenges:

• Structure of $H^\bullet(G_1; k)$ unknown
• No "regular block" exists ($C_\emptyset = \emptyset$)
• Finkelberg-Mirković conjecture suggests existence of geometric model

3. Complete Understanding of p-cells

Problem: Relationship between p-cells and Kazhdan-Lusztig cells?

Significance:

• Complete proof of traditional Humphreys conjecture
• Understand classification of tensor ideals (Conjecture 3.5)

4. Higher Frobenius Kernels

Problem: Can one generalize to $G_r = \ker(\text{Fr}^r)$?

Motivation:

• Related to tensor ideal classification
• May reveal new structure of p-cells

5. Computational Methods

Problem: How to effectively compute p-canonical basis?

Requirements:

• Develop new algorithms
• Find explicit bounds for N(w)

In-Depth Evaluation

Strengths

1. Theoretical Depth and Breadth

• Unified Framework: Incorporates character problems and support variety problems into unified geometric-representation-theoretic framework
• Interdisciplinary: Synthesizes algebraic geometry, representation theory, combinatorics, and category theory
• Deep Connections: Reveals role of Langlands duality in positive characteristic representation theory

2. Technical Innovation

• Degrading Functor: Creatively uses Springer resolution to connect geometry and representation theory
• co-t-structure Theory: Discovers new categorical structure in positive characteristic
• Smith-Treumann Theory Application: Introduces algebraic topology tools to representation theory of algebraic groups

3. Thoroughness of Problem Resolution

• Tilting character problem completely resolved (Theorem 2.6)
• Humphreys conjecture relative version completely resolved
• Establishes positive characteristic version of Lusztig-Vogan bijection

4. Writing Quality

• Clear historical review and motivation
• Systematic theoretical framework introduction
• Balanced technical details with intuitive explanations
• Comprehensive literature review and attribution

Weaknesses

1. Extremely High Technical Barrier

• Requires deep background in algebraic geometry, derived categories, and representation theory
• Degrading functor construction is highly technical
• co-t-structure theory difficult for non-specialists

2. Limited Computational Feasibility

• Computing p-canonical basis is extremely difficult
• Character formulas exist but are hard to compute explicitly
• Lacks detailed explicit computations of concrete examples

3. Conditional Results

• Most results require p > h
• Traditional Humphreys conjecture still unresolved
• Simple module character formulas lack explicitness

4. Omitted Technical Details

• As a survey paper, many proof details require consulting original papers
• Complete degrading functor construction not given
• co-t-structure construction only referenced

Impact

Contribution to Field

1. Paradigm Shift: Establishes p-canonical basis as core tool
2. Methodological Innovation: Systematic application of geometric methods to modular representation theory
3. Long-Standing Problems: Resolves Andersen conjecture (tilting characters) and Humphreys conjecture (partially)

Practical Value

• Provides theoretical foundation for concrete computations
• Guides development of computer algebra systems
• Potential applications in mathematical physics (e.g., modular Langlands program)

Reproducibility

• Theoretical Verifiability: High—proofs based on rigorous published mathematics
• Computational Reproducibility: Moderate—computable in principle but practically difficult
• Code Availability: Low—not provided (normal for pure mathematics papers)

Applicable Scenarios

1. Theoretical Research

• Modular representation theory of reductive algebraic groups
• Representation theory of affine Hecke algebras
• Kazhdan-Lusztig theory and generalizations
• Geometric representation theory

2. Related Fields

• Modular Langlands Program: Finkelberg-Mirković conjecture provides geometric realization
• Quantum Group Theory: Many results have quantum group analogues
• Modular Lie Algebra Representations: Connected via Frobenius kernels
• Categorification Theory: Hecke category action

3. Computational Applications

• Character table computation (when p not too large)
• Decomposition number determination
• Understanding tensor product structure

4. Non-Applicable Scenarios

• p = 0 (characteristic zero case completely different)
• p ≤ h requiring explicit formulas (theory not yet complete)
• Non-reductive groups or infinite-dimensional representations

Key References

1. 37 Lusztig (1980): "Some problems in the representation theory of finite Chevalley groups" - Original conjecture
2. 53 Williamson (2017): "Schubert calculus and torsion explosion" - Counterexamples, introduces p-canonical basis
3. 11 Achar-Makisumi-Riche-Williamson (2019): "Koszul duality for Kac-Moody groups and characters of tilting modules" - First proof of tilting character formula
4. 22 Bezrukavnikov (2006): "Cohomology of tilting modules over quantum groups and t-structures" - Humphreys conjecture for quantum groups
5. 8 Achar-Hardesty-Riche (2019): "On the Humphreys conjecture on support varieties" - Humphreys conjecture for sufficiently large p
6. 6 Achar-Hardesty (2024): "Silting complexes of coherent sheaves and the Humphreys conjecture" - Complete relative Humphreys conjecture
7. 47 Riche-Williamson (2022): "Smith-Treumann theory and the linkage principle" - Tilting character formula proof using new methods

Overall Assessment: This is an outstanding survey paper that systematically summarizes over a decade of profound work by the authors and collaborators on positive characteristic representation theory. Through introducing p-canonical basis and developing geometric methods, the authors resolve core problems in the field, advancing representation theory significantly. The paper is technically strong yet clearly organized, valuable for both experts and learners. This work will have lasting impact on modular representation theory and geometric representation theory.