On totally hyperbolic non-Fuchsian type-preserving representations
Inyoung Ryu
TL;DR
The paper constructs explicit counterexamples to Bowditch's question by exhibiting uncountably many totally hyperbolic but non-Fuchsian type-preserving representations of punctured surfaces. These representations have relative Euler class $e(φ)=n$ with $n$ in the Milnor–Wood-satisfying range and, for certain sign data $s$, their PSL$(2,ℝ)$-conjugacy classes occupy a full-measure subset of $2p$ components of the relative character variety, demonstrating that total hyperbolicity does not force Fuchsian behavior on punctured surfaces. A key feature is that, although globally non-Fuchsian, these representations restrict to Fuchsian (holonomy) representations on suitable subsurfaces, revealing an almost-Fuchsian structure on pieces of $Σ$. The results illuminate the dynamics of the mapping class group action on relative character varieties and provide new insights into Goldman-type conjectures in the punctured setting.
Abstract
We identify type-preserving representations $φ: π_1(Σ)\to \mathrm{PSL}(2,\mathbb{R})$ of the fundamental group of every punctured surface $Σ= Σ_{g,p}$ that are not Fuchsian yet send all non-peripheral simple closed curves to hyperbolic elements, which give a negative answer to a question of Bowditch. These representations have relative Euler class $e(φ) = \pm (χ(Σ) + 1)$, and their $\mathrm{PSL}(2,\mathbb{R})$-conjugacy classes form a full-measure subset of $2p$ connected components of the relative character variety. We further show that, while these representations are not Fuchsian, their restrictions to certain subsurfaces of $Σ$ are Fuchsian.
