🔢 mathematics

Mapping class group action on the cohomology of the $\mathrm{SL}_n$ character variety

This paper describes the mapping class group action on the cohomology of the twisted $\mathrm{SL}_n$ -character variety of a surface by utilizing a relative endoscopic decomposition to reduce the problem to the action on the cohomology of a canonical finite cover, a case previously studied by Looijenga.

Original authors: Anne Larsen

Published 2026-03-16

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Anne Larsen

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: A Dance of Shapes and Symmetries

Imagine you have a piece of dough with a hole in it (a surface with a puncture). You can stretch, twist, and squish this dough in all sorts of ways without tearing it. Mathematicians call the collection of all these possible moves the Mapping Class Group. It's like the "dance moves" available to the surface.

Now, imagine you paint a complex, invisible pattern on this dough. This pattern isn't just a drawing; it's a set of rules describing how things transform as you move around the dough. In math, this is called a Character Variety. It's a giant, multi-dimensional landscape where every point represents a different way to paint these rules on your dough.

The Question: If you perform one of those "dance moves" (twist the dough), how does the invisible pattern change? Specifically, how does the "shape" (cohomology) of the landscape of all possible patterns change?

The Problem: Two Different Landscapes

The paper focuses on a specific type of pattern called an SLn-character variety. Think of this as a very specific, rigid kind of painting where the rules must follow strict symmetry laws (like a kaleidoscope).

For a long time, mathematicians knew how to predict the dance moves for a simpler version of this painting (called the GL version). They knew that the dance moves were controlled by the basic "holes" in the dough.

But for the SL version (the rigid one), things were messy. The landscape was much more complicated. It had "hidden rooms" that the simple version didn't have. The authors wanted to know: What controls the dance moves in these hidden rooms?

The Solution: A Secret Tunnel (The Endoscopic Decomposition)

The paper's main breakthrough is a new tool called a relative endoscopic decomposition. Let's use an analogy to understand this.

Imagine you are trying to understand a massive, confusing library (the SL Character Variety). It's too big to read every book.

The Old Way: You try to read the whole library at once. You get lost.
The New Way (Larsen's Tool): You discover a secret tunnel that connects your library to a much smaller, simpler library (a finite cover of the dough).

This tunnel is based on a previous discovery by Maulik and Shen. Larsen's genius was to upgrade this tunnel so it works while the dough is being twisted. She proved that the complex, messy library (SL) is actually just a collection of copies of the simpler library (GL) and some specific "variant" parts.

Crucially, she showed that the "variant" parts (the hidden rooms) are controlled by a finite cover of the original dough.

The Dough: Your original surface with a hole.
The Cover: A new surface made by tiling the original one $n$ times (like a multi-layered cake).

The Main Discovery: The "Kernel" Connection

The paper proves a specific theorem about the "kernel" of the action. In plain English, the "kernel" is the list of dance moves that don't change the pattern at all.

The Finding: The list of dance moves that leave the complex SL pattern unchanged is almost exactly the same as the list of dance moves that leave the first layer of the "multi-layered cake" (the finite cover) unchanged.

Analogy:
Imagine you have a complex origami crane (the SL variety). You want to know which folds leave the crane looking the same.

Previously, people thought you had to look at the crane's internal structure to know this.
Larsen says: "No! Just look at the flat sheet of paper before you folded it, but imagine that sheet is actually a stack of $n$ sheets glued together."
The moves that don't change the crane are the same moves that don't change the stack of sheets.

Why This Matters: The "Monodromy Group"

The paper concludes by describing the Algebraic Monodromy Group. Think of this as the "fingerprint" of the dance. It tells us the full scope of how the pattern can wiggle and shift.

For the simple case (GL): The fingerprint is a standard, well-known shape (related to the symplectic group).
For the complex case (SL): The fingerprint is a more complex, "commutator subgroup" shape.

The paper shows that for most surfaces (genus $g > 2$ ), the fingerprint of the SL variety is exactly the "commutator subgroup" of the fingerprint of the finite cover.

What does "Commutator Subgroup" mean here?
Imagine a group of dancers.

The Full Group is everyone dancing.
The Commutator Subgroup is the group of dancers who can only do moves that are "purely internal" (moves that cancel out if you do them in a different order).
The paper says: The SL variety's dance is restricted to these "purely internal" moves of the finite cover's dance. It's a more rigid, more symmetrical dance than the cover itself.

The "Why" Behind the Math

Why did the authors have to go through all this trouble with Higgs bundles and non-abelian Hodge correspondence?

The Bridge: The paper uses a bridge called the Non-Abelian Hodge Correspondence. This is like a translator that converts a problem about "twisting dough" (Topology) into a problem about "solving equations on a curve" (Algebraic Geometry).
The Advantage: It is much easier to solve the equation problem than the dough problem. By translating the problem, they could use powerful algebraic tools to prove things about the dough that were impossible to see just by looking at the dough.

Summary in One Sentence

Anne Larsen discovered that the complex, rigid patterns on a twisted surface (SL Character Variety) are controlled by the same rules as a simpler, multi-layered version of that surface, allowing mathematicians to finally predict exactly how these patterns move and change when the surface is twisted.

The "So What?" for a General Audience

This paper is a victory for pattern recognition. It tells us that even in the most complex, high-dimensional mathematical landscapes, there are hidden, simpler structures underneath. By finding the right "tunnel" (the finite cover), we can understand the behavior of the complex system by studying the simple one. It's like realizing that the chaotic weather patterns of a whole continent are actually just a scaled-up version of the wind patterns in a single valley.

1. Problem Statement

The paper investigates the action of the mapping class group $\text{Mod}(\Sigma_g \setminus p)$ on the cohomology of the twisted $SL_n$ -character variety $M^{SL}$ .

Context: Let $\Sigma_g$ be a surface of genus $g \geq 2$ with a puncture $p$ . The character variety $M^{SL}$ consists of $SL_n(\mathbb{C})$ -local systems on $\Sigma_g \setminus p$ with fixed monodromy $e^{2\pi i d/n} \text{Id}_n$ around the puncture, modulo conjugation by $PGL_n(\mathbb{C})$ .
The Challenge: While the mapping class group action on the cohomology of the untwisted or $GL_n$ $G L_{n}$ character varieties is well-understood (generated by "tautological" classes and factoring through the symplectic group $Sp(2g, \mathbb{Z})$ $S p (2 g, Z)$ ), the $SL_n$ $S L_{n}$ case is more complex.
- The cohomology $H^*(M^{SL})$ contains "variant" classes (not invariant under the finite group $\Gamma = H^1(\Sigma_g, \mathbb{Z}/n)$ ) that are not generated by tautological classes.
- The goal is to determine the algebraic monodromy group (the identity component of the Zariski closure of the image of the mapping class group in $GL(H^*(M^{SL}, \mathbb{C}))$ ) and characterize the kernel of the action.

2. Methodology

Larsen employs a multi-step strategy combining Non-Abelian Hodge Theory, Moduli of Higgs Bundles, and Endoscopic Decomposition.

A. Non-Abelian Hodge Correspondence

The author translates the topological problem on the character variety $M^{SL}$ into an algebro-geometric problem on the moduli space of stable $SL_n$ -Higgs bundles of degree $d$ , denoted $\mathcal{M}_{n,d}$ .

This correspondence provides a diffeomorphism between $M^{SL}$ and the Higgs moduli space, allowing the use of powerful tools from algebraic geometry (e.g., Hitchin fibrations, spectral curves).
The problem is lifted to a relative setting over the moduli space of curves $\mathcal{M}_{g,1}$ (and its finite covers), treating the cohomology as a local system.

B. Relative Endoscopic Decomposition (Maulik-Shen)

The core technical tool is a relative version of the endoscopic decomposition established by Maulik and Shen [MS21].

Standard Decomposition: Maulik and Shen related the cohomology of $SL_n$ Higgs bundles to the cohomology of $GL_{n/m}$ Higgs bundles on $m$ -fold covers of the curve.
Larsen's Upgrade: The paper upgrades this isomorphism to a surjection of local systems over a cover of $\mathcal{M}_{g,1}$ . This ensures the isomorphism is compatible with the monodromy (mapping class group) action.
Key Construction:
1. Consider the moduli space of $GL_n$ Higgs bundles on an $m$ -fold étale cover $C_\gamma \to C$ (corresponding to a torsion line bundle).
2. Construct a correspondence between the cohomology of the $SL_n$ moduli space and the $\Gamma$ -invariant part of the cohomology of the $GL_n$ moduli space on the cover.
3. Utilize vanishing cycles and spectral curves to relate the cohomology of the twisted $SL_n$ space to the cohomology of the cover $C_\gamma$ .

C. Reduction to Finite Covers

The main technical achievement is reducing the study of the mapping class group action on $H^*(M^{SL})$ to the action on the cohomology of a canonical finite cover of the surface, denoted $\tilde{\Sigma}_g^n$ .

This cover corresponds to the kernel of the map $\pi_1(\Sigma_g) \to H_1(\Sigma_g, \mathbb{Z}/n)$ .
The action on $H^*(M^{SL})$ is shown to be essentially equivalent to the action on $H^*(\tilde{\Sigma}_g^n)$ .

3. Key Contributions and Results

Main Theorem (Theorem 1.1 / Theorem 3.4)

The kernel of the mapping class group action on the integral cohomology $H^*(M^{SL}, \mathbb{Z})$ is commensurable to the kernel of the action on the first homology of the canonical finite cover $H_1(\tilde{\Sigma}_g^n, \mathbb{Z})$ .

Implication: The "interesting" part of the monodromy is entirely captured by the geometry of the finite abelian cover of the surface, rather than the surface itself.

Algebraic Monodromy Group (Corollary 1.2 / Corollary 3.5)

For $g > 2$ (or $g=2$ with $(n,6)=1$ ), the algebraic monodromy group of $\text{Mod}(\Sigma_g \setminus p)$ acting on $H^*(M^{SL}, \mathbb{C})$ is the commutator subgroup of the group of invertible linear transformations on $H_1(\tilde{\Sigma}_g^n, \mathbb{C})$ that respect:

The intersection form.
The Galois action of the covering group.

Explicitly, this group is a product of symplectic and special linear groups depending on the parity of $n$ :

If $n$ is even: A product involving $SL_{2(g-1)}$ and $Sp_{2g}$ .
If $n$ is odd: A product involving $SL_{2(g-1)}$ and $Sp_{2g}$ .

Comparison with $GL_n$ Case

The paper contrasts this with the $GL_n$ case (Fact 1.3).

$GL_n$ : The monodromy group is $Sp_{2g}(\mathbb{Z})$ , controlled by $H_1(\Sigma_g)$ .
$SL_n$ : The monodromy group is much larger, controlled by $H_1(\tilde{\Sigma}_g^n)$ . This reflects the existence of non-tautological classes in $H^*(M^{SL})$ arising from the $\Gamma$ -variant part of the cohomology.

4. Significance

Resolution of a Topological Gap: The paper provides the first concrete description of the mapping class group action on the cohomology of twisted $SL_n$ character varieties, a space where standard tautological generation fails.
Bridging Fields: It successfully bridges topological questions (mapping class groups, character varieties) with algebro-geometric techniques (Higgs bundles, endoscopy, spectral curves).
Extension of Maulik-Shen: It generalizes the non-relative endoscopic correspondence of Maulik and Shen to a relative setting compatible with monodromy, a necessary step for studying families of curves and mapping class group actions.
Mirror Symmetry Context: The results relate to the "variant" classes discussed by Hausel and Thaddeus in the context of SYZ mirror symmetry, providing a topological interpretation of these classes via finite covers.
Explicit Monodromy: By identifying the algebraic monodromy group as a specific subgroup of a symplectic group acting on a cover, the paper gives a precise structural understanding of how the mapping class group acts on the "hidden" cohomology of these moduli spaces.

In summary, Larsen demonstrates that the complex cohomology of the $SL_n$ character variety is governed by the simpler, albeit larger, geometry of a finite abelian cover of the underlying surface, and the mapping class group action is fully determined by its action on this cover.