Skip to content

Cannon–Thurston map

A Cannon–Thurston map is the continuous boundary map induced, when it exists, by extending a specified inclusion of hyperbolic spaces or groups to their compactifications.

Version
v1 · 2026-10-07 · History
Domain-specific #
13824
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomains
Geometric Group Theory, Hyperbolic Geometry → Mathematics
Aliases
Cannon–Thurston boundary map

Core Idea

A Cannon–Thurston map is the continuous boundary map induced when a specified inclusion of hyperbolic spaces or groups extends to their compactifications. It is tied to the interior inclusion: a continuous function between the same two boundary spaces is not enough unless it agrees with a continuous compactified extension. Existence depends on the setting. Cannon and Thurston prove such an extension for a closed surface fiber in a closed hyperbolic three-manifold; Mitra proves one for a hyperbolic normal subgroup of a hyperbolic group; Baker and Riley give a hyperbolic subgroup inclusion for which no boundary map exists.[ref-57e75b3c784f][ref-fac9b73d98c6][^ref-7365a25e7f55]

The live Continuous Function entry is a strict parent: every existing Cannon–Thurston map is continuous, while most continuous functions do not extend a hyperbolic inclusion. The boundary-at-infinity carriers and induced-extension relation keep this entry domain-specific.[^ref-fac9b73d98c6]

Scope of Application

In the original case, a lifted inclusion H² → H³ of a closed surface fiber extends to closed-ball compactifications, producing a map S¹ → S². In the group case, an inclusion of Cayley graphs can extend to Gromov compactifications, producing a boundary restriction ∂H → ∂G. Always specify the inclusion and boundary models before claiming the map exists. The original paper's punctured figure-eight-knot discussion is evidence for a further case, not a proof of it in that paper.[ref-57e75b3c784f][ref-fac9b73d98c6]

An infinite normal free kernel in a qualified hyperbolic free-group extension has full target boundary image; finite normal subgroups have empty boundary and do not support that surjectivity claim. For a nonnormal subgroup, hyperbolicity of both groups alone is insufficient to infer existence.[ref-fac9b73d98c6][ref-1664414b9520][^ref-7365a25e7f55]

Clarity

Keep three maps separate: the interior inclusion, the extension of compactifications, and its boundary restriction. The last is the Cannon–Thurston map only when the middle map is continuous. Then distinguish the existence result from properties of the boundary restriction, such as image, equivariance, injectivity, or fiber multiplicity. The original map is onto and noninjective; those features do not define every such map.[ref-57e75b3c784f][ref-fac9b73d98c6]

Manages Complexity

The first test reduces many possible boundary-ray identifications to three questions: which inclusion, which compactifications, and what proves continuous extension? Once existence is established, image and fibers become separate case-specific questions. The qualified free-group extension studied by Dowdall, Kapovich, and Taylor has fiber degree at most 2 rank(F); their bound is not universal.[ref-1664414b9520][ref-7365a25e7f55]

Abstract Reasoning

Start with the actual hyperbolic inclusion and choose the relevant visual or Gromov boundaries. Find a theorem or proof showing that the inclusion extends continuously; if that fails, there is no Cannon–Thurston map for that pair. Only then determine what is known about the boundary image and point identifications. Mitra's normal-subgroup theorem supplies a positive route in its stated setting, while Baker–Riley's rank-three free subgroup demonstrates the failure of a blanket existence inference.[ref-fac9b73d98c6][ref-7365a25e7f55]

Knowledge Transfer

A closed surface fiber and a hyperbolic free-group extension have unlike boundary carriers, yet both instantiate an inclusion whose compactification extends continuously and has a boundary restriction. This is literal transfer within hyperbolic geometry. Generic continuity reaches live Continuity through Continuous Function, but the broader interior-to-boundary-extension pattern remains a future-Prime question; it does not carry these hyperbolic existence theorems into unrelated domains.[ref-57e75b3c784f][ref-1664414b9520]

Example

Closed surface fiber. For a closed hyperbolic three-manifold fibering over the circle with closed surface fiber, the lifted H² → H³ inclusion extends to closed balls. Its boundary restriction S¹ → S² is an equivariant Peano curve onto the sphere. Mapped back: the inclusion is the lifted fiber plane; the carriers are the visual circle and sphere; the warrant is the closed-fiber theorem; surjectivity and noninjectivity are consequences of this case. The original paper does not prove the punctured figure-eight-knot case in its separate discussion.[^ref-57e75b3c784f]

Hyperbolic free-group extension. Under the convex-cocompact purely atoroidal hypotheses studied by Dowdall, Kapovich, and Taylor, 1 → F → E_Γ → Γ → 1 has a free kernel of rank at least three and a hyperbolic total group. Mitra's theorem yields the boundary map for the normal kernel inclusion. The source analyzes fibers and proves degree at most 2 rank(F). Mapped back: the inclusion is F → E_Γ; the carriers are ∂F and ∂E_Γ; the warrant is the normal-subgroup theorem under the extension hypotheses; full target image and the degree bound are qualified behavior of this case.[ref-fac9b73d98c6][ref-1664414b9520]

Relationships to Other Abstractions

Local relationship map for Cannon–Thurston mapParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Cannon–Thurston mapDOMAINDomain-specific abstraction: Continuous function — is a kind ofContinuousfunctionDOMAIN

Current abstraction Cannon–Thurston map Domain-specific

Parents (1) — more general patterns this builds on

  • Cannon–Thurston map is a kind of Continuous function Domain-specific

    An existing Cannon–Thurston boundary map is a continuous function on declared boundary spaces.

Hierarchy paths (2) — routes to 2 parentless roots

Neighborhood in Abstraction Space

Cannon–Thurston map sits in a sparse region of the domain-specific corpus (82nd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Manifold Topology & Classification (12 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

A visual or Gromov boundary is a space, not the induced map between spaces. A boundary embedding would require injectivity, which the original S¹ → S² map lacks. Cayley Graph is a carrier in the group example, not a universal direct parent of the lifted-space formulation; the live Topological Boundary entry does not automatically describe boundaries at infinity. Baker–Riley's example has no map at all, rather than a map with pathological fibers.[ref-57e75b3c784f][ref-7365a25e7f55]

References

[^ref-57e75b3c784f]: James W. Cannon and William P. Thurston, “Group invariant Peano curves,” Geometry & Topology 11 (2007), 1315–1355, abstract, Introduction pp. 1315–1317, and §2 for the closed-fibered setting. The Introduction distinguishes the punctured figure-eight-knot discussion from the proved closed case. https://msp.org/gt/2007/11-3/gt-v11-n3-p03-p.pdf

[^ref-fac9b73d98c6]: Mahan Mitra, “Cannon–Thurston Maps for Hyperbolic Group Extensions,” Topology 37 (1998), 527–538, Introduction and main normal-subgroup theorem, PDF pp. 0–1; Theorem 4.3 discussion, PDF p. 3, for finite-subgroup vacuity. https://repository.ias.ac.in/89549/1/29-a.pdf

[^ref-1664414b9520]: Spencer Dowdall, Ilya Kapovich, and Samuel J. Taylor, “Cannon–Thurston Maps for Hyperbolic Free Group Extensions,” author PDF dated 17 December 2015, abstract and Introduction pp. 0–2; Proposition–Definition 2.1, PDF p. 4, for the map; Theorem 4.1, PDF p. 11, for extension hyperbolicity; Theorem 6.3, PDF p. 1, for the 2 rank(F) degree bound. https://math.vanderbilt.edu/dowdalsd/papers/freeCannonThurston.pdf

[^ref-7365a25e7f55]: O. Baker and T. R. Riley, “Cannon–Thurston maps do not always exist,” Forum of Mathematics, Sigma 1 (2013), e3, abstract and Theorem 1, PDF pp. 0–1, DOI 10.1017/fms.2013.4. https://www.cambridge.org/core/services/aop-cambridge-core/content/view/2AD18B19161BECFB2ED7F797A6338227/S2050509413000042a.pdf/cannonthurston_maps_do_not_always_exist.pdf