Non-amenability of mapping class groups of infinite-type surfaces and graphs

This paper establishes the non-amenability of mapping class groups for infinite-type surfaces and locally finite infinite graphs of higher ranks, provides an example of a non-amenable point stabilizer in a specific hyperbolic Polish group, and identifies a class of amenable mapping class groups associated with trees and rank-one graphs.

The Gist

Imagine the mathematical world as a vast landscape of shapes and connections. In this landscape, there are two main characters: Surfaces (like sheets of rubber that can be stretched infinitely) and Graphs (like networks of roads or trees that branch out forever).

Mathematicians study the "Mapping Class Groups" of these shapes. Think of a Mapping Class Group as the club of all possible ways you can twist, turn, and shuffle a shape without tearing it, and then group together all the moves that look the same after you're done.

The big question this paper answers is: Is this club "tame" (amenable) or "wild" (non-amenable)?

In math-speak, "amenable" means the group is orderly enough that you can assign a fair, balanced "probability" to every possible move, like a perfectly weighted coin. "Non-amenable" means the group is chaotic, wild, and contains hidden structures (like free groups) that make it impossible to balance the scales.

Here is the breakdown of the paper's findings, using everyday analogies:

1. The Infinite Surfaces: The "Wild" Club

The paper looks at surfaces that are infinite in size (like a plane with infinitely many holes or a sphere with a Cantor set of holes).

• The Finding: The author proves that every Mapping Class Group of an infinite-type surface is non-amenable (wild).
• The Analogy: Imagine a giant, infinite dance floor where everyone is dancing. The author shows that no matter how you try to organize a small, quiet corner of this dance floor (an "open subgroup"), you can never make it truly peaceful. There is always a hidden, chaotic sub-group of dancers doing a complex, unbalanced routine that prevents the whole floor from being "fair" or "tame."
• Why it matters: Previously, mathematicians knew these groups were weird, but this paper proves they are inherently chaotic. You can't find a "safe zone" inside them.

2. The Infinite Graphs: The "Mixed Bag"

Next, the author looks at infinite graphs (networks of lines and dots).

• The Finding: Here, the story is different. It depends on the shape of the graph.
  • If the graph is complex (has "loops" or "rings" like a bicycle wheel, with a rank of 2 or more): The group is non-amenable (wild). It's like a tangled ball of yarn that you can't untangle without chaos.
  • If the graph is a simple tree (no loops, just branches): The group can be amenable (tame).
• The Analogy:
  • Complex Graphs: Imagine a city with many roundabouts and loops. The traffic flow (the group) is so interconnected that you can't find a quiet, orderly neighborhood. It's always chaotic.
  • Trees: Imagine a family tree or a river delta with no loops. If the "ends" (the tips of the branches) are few and countable (like a finite number of leaves), the group is tame. You can organize the movement perfectly.
  • The Catch: If the tree has a "Cantor set" of ends (an infinitely complex, fractal-like collection of tips), it can suddenly become wild again. It's like a tree that looks simple at first, but if you zoom in on the leaves, they form a chaotic, infinite pattern.

3. The "Point at Infinity" Surprise

The paper also tackles a specific problem about "Hyperbolic Polish Groups" (a fancy term for groups with a specific kind of geometry that stretches out to infinity).

• The Finding: In finite geometry, the "guards" at the edge of the world (stabilizers of points at infinity) are usually tame. But in this infinite world, the author found a wild guard.
• The Analogy: Imagine a lighthouse at the edge of an infinite ocean. Usually, the lighthouse keeper is calm and follows a strict routine. This paper found a lighthouse where the keeper is actually running a chaotic, unbalanced operation. This breaks a long-held rule that "guards at the edge are always calm."

4. The "Cantor Set" Connection

A major tool used in the paper is the Cantor Set (a mathematical object that looks like a dust of infinite points).

• The Insight: The author realized that studying the "ends" of these infinite trees is exactly the same as studying how you can shuffle a closed set of points from the Cantor Set.
• The Rule of Thumb:
  • If the set of points is countable (you can list them like 1, 2, 3...), the shuffling group is tame.
  • If the set of points is uncountable and has a "perfect" core (like a solid chunk of the Cantor Set), the shuffling group is wild.

Summary for the General Audience

Think of this paper as a guidebook to the chaos of infinite shapes.

1. Infinite Surfaces: Always chaotic. No matter how you slice them, they are too wild to be "fair."
2. Infinite Graphs: It depends. If they have loops, they are chaotic. If they are simple trees with simple ends, they are orderly. But if the ends are fractal-like, they become chaotic again.
3. The Big Lesson: Just because a group is "infinite" doesn't mean it's automatically chaotic, but if it has certain complex features (like loops or fractal ends), it will inevitably be wild.

The author, Yusen Long, has essentially drawn a map showing exactly where the "order" ends and the "chaos" begins in these infinite mathematical worlds.

Technical Summary

Here is a detailed technical summary of the paper "Non-amenability of mapping class groups of infinite-type surfaces and graphs" by Yusen Long.

1. Problem Statement and Context

The paper addresses the amenability of mapping class groups (MCGs) associated with infinite-type surfaces and locally finite infinite graphs.

• Background: In the context of finite-type surfaces and graphs, the Tits Alternative holds: any subgroup either contains a non-abelian free group ($F_n, n \ge 2$) or is virtually solvable (and thus amenable). Consequently, non-amenability in these discrete groups is equivalent to containing a non-abelian free subgroup.
• The Shift to Infinite-Type: For "big" mapping class groups (infinite-type surfaces) and infinite graphs, the Tits Alternative fails; these groups can contain non-amenable subgroups that do not contain free groups (e.g., Thompson's group or certain Burnside groups).
• The Core Question: While the Tits Alternative fails, does a similar dichotomy hold for amenability in the context of Polish topology (where these groups are equipped with the compact-open topology)? Specifically:
  1. Are big mapping class groups themselves amenable?
  2. What characterizes the amenable closed subgroups of these groups?
  3. How does amenability behave for mapping class groups of infinite graphs, particularly trees and rank-one graphs?

2. Methodology

The author employs a combination of geometric group theory, topological group theory, and descriptive set theory (specifically Cantor-Bendixson analysis).

• Topological Group Properties: The paper relies on the definition of amenability for topological groups: a group $G$ is amenable if every continuous action on a compact space admits an invariant probability measure. Key hereditary properties used include:
  • Open subgroups of amenable groups are amenable.
  • If a group maps continuously onto a non-amenable group, the domain is non-amenable.
  • Directed unions of amenable groups are amenable.
• Geometric Construction (Surfaces): The author constructs continuous epimorphisms from open subgroups of big mapping class groups to known non-amenable discrete groups (specifically, mapping class groups of finite-type surfaces). This is achieved by identifying finite-type subsurfaces fixed by pointwise stabilizers of finite curve collections.
• Graph Theory and Ends: For graphs, the analysis focuses on the end space ($\partial \Gamma$). The mapping class group of a tree is isomorphic to the homeomorphism group of its end space.
• Cantor-Bendixson Analysis: For trees, the end space is a closed subset of the Cantor set. The author decomposes this space into a perfect kernel (Cantor set) and a countable scattered part. Amenability is analyzed via transfinite induction on the Cantor-Bendixson rank of the end space.

3. Key Contributions and Results

A. Infinite-Type Surfaces

Theorem 1.1: Let $S$ be an infinite-type surface. Every open subgroup of the mapping class group $MCG(S)$ is non-amenable. Consequently, $MCG(S)$ itself is non-amenable.

• Significance: This is a strong result. It implies that any amenable closed subgroup of a big mapping class group must be nowhere dense. This holds for both orientable and non-orientable surfaces.
• Proof Strategy: For any open subgroup $G$, the author finds a finite collection of curves $F$ such that the pointwise stabilizer $MCG(S)(F)$ is contained in $G$. This stabilizer maps continuously onto the mapping class group of a finite-type subsurface $\Sigma$ (which is discrete and non-amenable), proving $G$ is non-amenable.

B. Hyperbolic Polish Groups and Stabilizers

Theorem 1.4: There exists a coarsely bounded (CB) generated hyperbolic Polish group (specifically, a big mapping class group of a specific surface) such that the stabilizer of a point on its Gromov boundary is non-amenable.

• Context: In finitely generated hyperbolic groups, stabilizers of boundary points are amenable. This result shows that this property fails for the broader class of CB-generated hyperbolic Polish groups.
• Example: The group $MCG(S^2 \setminus (C \cup \{p\}))$, where $C$ is a Cantor set and $p$ is an isolated puncture.

C. Infinite Graphs (General Case)

Theorem 1.5: Let $\Gamma$ be a locally finite infinite graph of genus (rank) at least 2.

• $MCG(\Gamma)$ is non-amenable.
• If $\Gamma$ is of infinite type, every open subgroup of $MCG(\Gamma)$ is non-amenable.
• Proof Strategy: Similar to surfaces, the author constructs an open subgroup that maps onto $Out(F_n)$ (for $n \ge 2$), which is non-amenable.

D. Infinite Graphs (Trees and Rank-One)

The situation changes significantly for graphs with rank $\le 1$ (trees and graphs with one loop).
Theorem 1.6 & 1.8: Let $T$ be a locally finite infinite tree with end space $E$ (a closed subset of the Cantor set).

• Amenable Case: If the end space $E$ is countable, then $MCG(T)$ (isomorphic to $Homeo(E)$) is amenable.
• Non-Amenable Case: If there exists a clopen subset $K \subset E$ such that the setwise stabilizer $Homeo(E)_K$ admits no invariant probability measure on $K$, then $MCG(T)$ is non-amenable.
• Characterization: The paper provides a partial characterization. If the end space is uncountable and contains a "large" perfect kernel (e.g., the full Cantor set), the group is generally non-amenable. However, amenability is not solely determined by countability; specific topological configurations of the end space matter.

Corollary 1.7: For a locally finite infinite graph $\Gamma$ with rank $\le 1$, if the end space $\partial \Gamma$ is countable, then $MCG(\Gamma)$ is amenable.

4. Significance and Implications

1. Resolution of the Amenability Dichotomy: The paper establishes a sharp distinction between finite-type and infinite-type mapping class groups regarding amenability. While finite-type groups are discrete and non-amenable (containing $F_2$), infinite-type groups are non-discrete Polish groups that are also non-amenable, but for different structural reasons (related to their action on the end space and the existence of non-amenable open subgroups).
2. Failure of Classical Hyperbolic Properties: The result regarding CB-generated hyperbolic Polish groups (Theorem 1.4) challenges the extension of classical geometric group theory results (like the amenability of boundary stabilizers) to the infinite-type/Polish setting.
3. Topological vs. Discrete Amenability: The work highlights that amenability in topological groups is a subtle property dependent on the topology. A group can contain non-amenable discrete subgroups (like $F_2$) yet be amenable as a topological group (e.g., if the topology is trivial or if the group is a directed union of amenable groups). Conversely, big mapping class groups are non-amenable in their natural Polish topology.
4. Cantor-Bendixson Analysis: The application of Cantor-Bendixson ranks to determine the amenability of homeomorphism groups of closed subsets of the Cantor set provides a new tool for analyzing the dynamics of infinite-type surfaces and graphs.

In summary, Long's paper provides a comprehensive classification of the amenability of mapping class groups for infinite-type objects, proving that "big" mapping class groups are inherently non-amenable (except for specific low-rank graph cases with countable ends), and demonstrating that the geometric intuition from finite-type groups does not directly translate to the infinite-type Polish setting.