Mapping Class Groups in Topology

Updated 23 January 2026

Mapping class groups are groups defined by isotopy classes of orientation-preserving homeomorphisms on surfaces, essential for understanding moduli and deformation theory.
They exhibit rich algebraic and topological structures with generators from Dehn twists, Nielsen–Thurston classifications, and stable cohomology behaviors.
The study extends to infinite-type surfaces, where big mapping class groups show uncountable, non-finitely generated structures with complex dynamics.

A mapping class group is a group that encodes the isotopy classes of orientation-preserving homeomorphisms of a surface or, more generally, a manifold, often fixing some additional structure (such as a set of marked points or a boundary). Mapping class groups are central objects in low-dimensional topology, geometric group theory, Teichmüller theory, topological quantum field theory, and algebraic geometry, where they govern moduli of surfaces, dynamics on moduli spaces, and automorphism groups of various geometric and combinatorial structures.

1. Formal Definition and Variants

Given an orientable surface $S$ , possibly with boundary $\partial S$ and finitely many marked points $P$ , the mapping class group is defined as

$\Map(S,P) = \Homeo^+(S, \partial S, P) / \Homeo_0(S, \partial S, P)$

where $\Homeo^+$ is the group of orientation-preserving self-homeomorphisms fixing $\partial S$ pointwise and permuting (or fixing) the marked set $P$ , and $\Homeo_0$ is the identity component, i.e., homeomorphisms isotopic to the identity rel boundary and marked points. For closed surfaces and no marked points, this reduces to the classical case $\Map(S)$.

For infinite-type surfaces—those not of finite type, i.e., with non-finitely generated fundamental group—the same definition applies, but the resulting group exhibits dramatically different structure and topology, justifying the term "big mapping class group" (Bellek, 19 Dec 2025, Aramayona et al., 2020).

Variants include:

Pure mapping class groups $\PMap(S)$, consisting of classes fixing every end and/or marked point.
Mapping class groups of manifolds in higher dimensions, such as simply connected 4-manifolds with or without boundary (Orson et al., 2022).
Orbifold mapping class groups, where $S$ is replaced by an orbifold $\mathcal{O}$ arising as a global quotient $M/G$ with suitable $G$ -action; these are studied via $G$ -equivariant homeomorphisms up to $G$ -equivariant isotopy (Flechsig, 2023).

2. Structure for Finite-Type Surfaces

For closed, orientable, finite-type surfaces $S_{g,n}^b$ of genus $g$ with $n$ punctures and $b$ boundary components, the mapping class group $\Map(S_{g,n}^b)$ is discrete and finitely generated. Fundamental results include:

Generation by Dehn twists about simple closed curves: the Lickorish–Humphries generators suffice for $g\ge2$ , with $2g+1$ nonseparating curves generating the pure group (Bellek, 19 Dec 2025).
Presentation in terms of Dehn twists, their commutation when supports are disjoint, the braid relations when curves intersect once, and higher relations such as lantern and chain relations (Petyt, 2021).
Nielsen–Thurston classification: every element is periodic, reducible (preserves a multicurve), or pseudo-Anosov.
The groups have important normal subgroups: the Torelli group (kernel of action on $H_1(S_{g,n};\mathbb{Z})$ ), Johnson kernel, and so on.
Cohomology and stability: classical results (Harer, Madsen–Weiss) that rational cohomology stabilizes as $g\to\infty$ , and detailed understanding in low genus (Cohen et al., 2014).

3. Big Mapping Class Groups: Infinite-Type Surfaces

For infinite-type (big) surfaces, mapping class groups are uncountable, Polish (separable, completely metrizable) topological groups with non-discrete and highly non-locally compact topologies (Aramayona et al., 2020, Bellek, 19 Dec 2025, Mann et al., 2019). Key features include:

Topological generation: They are topologically generated by the closure of Dehn twists (from compactly supported mapping classes) plus a set of handle-shifts—mapping classes shifting genus towards nonplanar ends (Bellek, 19 Dec 2025). Dehn twists alone suffice if and only if at most one end is accumulated by genus.
Algebraic structure: They split as a semidirect product $\PMap(S) = \overline{\PMap_c(S)} \rtimes \mathbb{Z}^r$, where $r$ is the rank of homology generated by separating curves; this is determined by the structure of the ends of $S$ (Bellek, 19 Dec 2025).
Non-finite generation and simplicity: In general, $\Map(S)$ is not compactly generated and is highly non-simple, with large abelianizations except in special cases (e.g., some perfect groups for Cantor-tree surfaces) (Aramayona et al., 2020, Vlamis, 2020).
Polish group properties: The natural topology is the quotient of the compact-open topology and is homeomorphic to the Baire space $\mathbb{N}^\mathbb{N}$ ; every compact subset has empty interior.
Normal subgroups and simplicity breakdown: The compactly supported subgroup $\PMap_c(S)$ is always closed and normal in $\PMap(S)$; the quotient is a free abelian group of (possibly infinite) rank.

Infinite-type mapping class groups reveal structural phenomena absent in the finite-type case, including the necessity of handle-shifts for topological generation and the appearance of uncountable combinatorial complexity due to the topology of ends (Bellek, 19 Dec 2025, Aramayona et al., 2020).

4. Combinatorial and Geometric Actions

Mapping class groups of finite type act naturally by automorphisms on various complexes:

Curve graph: Vertices are isotopy classes of essential simple closed curves; the mapping class group acts by simplicial automorphisms. For finite-type surfaces, the curve graph is Gromov-hyperbolic and strongly rigid (Petyt, 2021), providing a means to transfer geometric group theory tools.
CAT(0) cube complexes: Closed mapping class groups with finitely many marked points are quasiisometric to finite-dimensional CAT(0) cube complexes; these cube complexes admit combinatorial (cubical) metrics and the mapping class group acts on them via quasimedian quasi-isometries (Petyt, 2021).
Infinite-type variants: For infinite-type surfaces, the classical curve graph trivializes (diameter 2), but various subgraphs (ray graph, arc graph, nonseparating curve graphs) are hyperbolic and support complicated dynamics. Actions on these enable the construction of nontrivial quasimorphisms and detection of mapping class types (e.g., finite-type pseudo-Anosovs act loxodromically) (Aramayona et al., 2020, Rasmussen, 2019).
WWPD and quasimorphisms: On surfaces with isolated punctures, elements which are pseudo-Anosov on finite-type witnesses are precisely the WWPD (weakly weakly properly discontinuous) elements for the action on loop graphs (Rasmussen, 2019).

5. Cohomology, Representations, and Quantum Structures

The (co)homological properties of mapping class groups are intensely studied:

Finite-type/low genus: Integral and mod- $p$ cohomology of mapping class groups and their marked point variants are well understood for low genus and few marked points; the stable cohomology ring is polynomially generated by the Mumford–Morita–Miller classes (Cohen et al., 2014).
Central extensions: The quantum Teichmüller theory produces a canonical central extension of the mapping class group, whose class is $12$ times the Meyer signature class plus the Euler classes of the punctures (Funar et al., 2010). This extension encodes the projectivity obstructions in the quantum representations.
Quantum representations: Representations arising from Drinfeld doubles of finite groups yield mapping class group representations with finite image, acting by permutations on 2g-conjugacy classes of group elements with prescribed relations (Fjelstad et al., 2015). For non-abelian groups and $g\ge3$ , such representations restrict nontrivially to the Torelli subgroup.
Representations in Chern-Simons and TQFT: In $U(1)$ Chern-Simons theory on a closed orientable surface $\Sigma$ , the quantum Hilbert space carries a unitary representation of the mapping class group. The representation splits into clock algebras for holonomies and large gauge transformations, with a $k\leftrightarrow 1/k$ duality, and the mapping class group acts via precisely determined unitary matrices up to framing phases (Chen, 2011).

6. Connections with Braids, Covers, and Orbifolds

Mapping class groups are closely related to braid groups and configuration spaces:

Braid-mapping class group homomorphisms: Classification results dictate that for homomorphisms $B_n \to \Mod(S_g)$ with $g<n-2$ , only cyclic and standard (chain twist) representations arise; for $g=n-2$ , generalized super-hyperelliptic covers are possible (Chen et al., 2020). This is exact for $n\ge 23$ and classifies surface bundles over configuration spaces.
Covers with boundary and braid embeddings: The liftable and symmetric mapping class groups of finite covers with boundary coincide with the full mapping class group only in certain cases (Burau covers of the disk), yielding infinite families of non-geometric braid group embeddings into mapping class groups, where standard braid generators correspond to chain twists (Ghaswala et al., 2018).
Orbifold mapping class groups: For 2-orbifolds, orbifold mapping class groups are defined in terms of $G$ -equivariant homeomorphisms. Presentations generalize the classical Birman exact sequence, with additional generators and relations for arcs and curves affected by the orbifold structure (Flechsig, 2023).

7. Higher-Dimensional and Cobordism Aspects

Mapping class groups generalize to higher dimensions, yielding further phenomena:

Simply connected 4-manifolds: The mapping class group of a simply connected, compact 4-manifold with boundary is classified in terms of variations of the intersection form and spin obstructions. The stable smooth mapping class group coincides with automorphism groups of the stabilized intersection lattice (Orson et al., 2022).
Cobordism invariance: The mapping class group is not an h-cobordism invariant of high-dimensional manifolds; there exist $h$ -cobordant closed manifolds with mapping class groups of different cardinality (Muñoz-Echániz, 2022). This is demonstrated using the moduli of block bundles and the Whitehead group.

Mapping class groups thus serve as fundamental invariants encoding the global deformation theory of surfaces and manifolds, providing a unifying framework for problems in low-dimensional topology, geometric group theory, moduli of geometric structures, and quantum topology. Their algebraic, geometric, and cohomological properties—finite and infinite type alike—continue to drive research at the intersection of many mathematical disciplines.