Table of Contents
Fetching ...

Relative free splitting and free factor complexes I: Hyperbolicity

Michael Handel, Lee Mosher

TL;DR

The paper proves that the relative free splitting complex $\mathcal{FS}(\Gamma;\mathcal{A})$ and the relative free factor complex $\mathcal{FF}(\Gamma;\mathcal{A})$ are hyperbolic for any finitely generated group $\Gamma$ relative to a free factor system $\mathcal{A}$, with parallel results for $F_n$ relative to $\mathcal{A}$ and general $\Gamma$. It adapts the Masur–Minsky framework to the relative setting by constructing coarse-path families via Stallings fold paths, defining free splitting units, and establishing projection diagrams to verify the Coarse Retract, Coarse Lipschitz, and Strong Contraction Axioms; the core technical tool is the Big Diagram argument, which translates combinatorial fold behavior into global hyperbolicity constants that depend only on the corank and the size of $\mathcal{A}$. The work also proves uniform quasigeodesic parameterizations of fold paths using free splitting units and shows that the relative free factor complex inherits hyperbolicity via a Kapovich–Rafi transfer from the relative splitting complex, with the projection map controlling geodesic behavior. Together, these results provide a robust, deformation-space–inspired framework for understanding the large-scale geometry of outer automorphism groups of freely decomposable groups, and establish a foundation for analyzing dynamics of $\mathrm{Out}(\Gamma;\mathcal{A})$ in both absolute and relative contexts.

Abstract

We study the large scale geometry of the relative free splitting complex and the relative free factor complex of the rank $n$ free group $F_n$, relative to the choice of a free factor system of $F_n$, proving that these complexes are hyperbolic. Furthermore we present the proof in a general context, obtaining hyperbolicity of the relative free splitting complex and of the relative free factor complex of a general group $Γ$, relative to the choice of a free factor system of $Γ$. The proof yields information about coarsely transitive families of quasigeodesics in each of these complexes, expressed in terms of fold paths of free splittings.

Relative free splitting and free factor complexes I: Hyperbolicity

TL;DR

The paper proves that the relative free splitting complex and the relative free factor complex are hyperbolic for any finitely generated group relative to a free factor system , with parallel results for relative to and general . It adapts the Masur–Minsky framework to the relative setting by constructing coarse-path families via Stallings fold paths, defining free splitting units, and establishing projection diagrams to verify the Coarse Retract, Coarse Lipschitz, and Strong Contraction Axioms; the core technical tool is the Big Diagram argument, which translates combinatorial fold behavior into global hyperbolicity constants that depend only on the corank and the size of . The work also proves uniform quasigeodesic parameterizations of fold paths using free splitting units and shows that the relative free factor complex inherits hyperbolicity via a Kapovich–Rafi transfer from the relative splitting complex, with the projection map controlling geodesic behavior. Together, these results provide a robust, deformation-space–inspired framework for understanding the large-scale geometry of outer automorphism groups of freely decomposable groups, and establish a foundation for analyzing dynamics of in both absolute and relative contexts.

Abstract

We study the large scale geometry of the relative free splitting complex and the relative free factor complex of the rank free group , relative to the choice of a free factor system of , proving that these complexes are hyperbolic. Furthermore we present the proof in a general context, obtaining hyperbolicity of the relative free splitting complex and of the relative free factor complex of a general group , relative to the choice of a free factor system of . The proof yields information about coarsely transitive families of quasigeodesics in each of these complexes, expressed in terms of fold paths of free splittings.

Paper Structure

This paper contains 52 sections, 50 theorems, 84 equations, 10 figures.

Key Result

Theorem 1.1

For any nonmaximal free factor system $\mathcal{A}$ of $F_n$, the complex $\mathcal{FS}(F_n;\mathcal{A})$ is nonempty, connected, and hyperbolic.

Figures (10)

  • Figure 1: The Extension Lemma \ref{['LemmaExtension']} shows that for each extension $\textcolor{red}{\mathcal{A}} \sqsubset \mathcal{A}'$ of free factor systems of $\Gamma$, and for any realization of $\mathcal{A}'$, there exists a free factorization with terms as depicted which simultaneously incorporates the following: the given realization of $\mathcal{A}' = \{[A'_1],\ldots,[A'_K]\}$ with cofactor $B'$; for each $j \le J$ a realization of a free factor system of $A'_j$, namely $\mathcal{A}'_j = \{[\textcolor{red}{A_{j1}}],\ldots,[\textcolor{red}{A_{jk_j}}]\}$, with cofactor $B_j$; and a realization of $\textcolor{red}{\mathcal{A}} = \{[\textcolor{red}{A_{11}}],\ldots,[\textcolor{red}{A_{1k_1}}],\ldots\ldots,[\textcolor{red}{A_{1J}}],\ldots,[\textcolor{red}{A_{1k_J}}]\}$ with cofactor $B = B_1 * \cdots * B_J * A'_{J+1} * \cdots * A'_K * B'$. In fact these conclusions can be reformulated even when $\mathcal{A}'$ is only a weak free factor system, but one then deduces that $\mathcal{A}'$ is actually a (strong) free factor system.
  • Figure 2: An augmented projection diagram of depth $J$ from $T$ to $S_I\mapsto\cdots\mapsto S_K$.
  • Figure 3: The Big Diagram, Step 0. Certain columns $L=L_0$, $L_1$, … are emphasized, using free splitting units along the fold path $T^0_J \to \cdots \to T^0_L$. As the Big Diagram evolves, and up until nearly the end of the evolution, the original projection diagram atop which the diagram is built, which involves the $S'$ and $S$ rows, will not change. Those rows will be suppressed in the meantime, returning only in the Penultimate Diagram of Figure \ref{['FigurePentDiagram']}.
  • Figure 4: Factoring the combing rectangle between rows $2,3$ and columns $0,\ldots,L_1$.
  • Figure 5: The Big Diagram, step 0.1
  • ...and 5 more figures

Theorems & Definitions (94)

  • Theorem 1.1
  • Theorem 1.2
  • Theorem 1.3
  • Theorem 1.4
  • Lemma 2.1
  • Lemma 2.2
  • proof
  • Lemma 2.3
  • proof
  • Definition 2.4: Free factor systems.
  • ...and 84 more