Typical hyperbolic surfaces have an optimal spectral gap

Laura Monk · 2026

Read Typical hyperbolic surfaces have an optimal spectral gap by Laura Monk (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-15157.txt.

Introduction

Laura Monk proves that a typical curved surface is hard to split in two. Her 2026 arXiv paper deals with hyperbolic surfaces, shapes with constant negative curvature. The key gauge is the spectral gap, the first nonzero eigenvalue of the Laplace operator.

A large gap means the surface is well connected. Such a surface has a small diameter and mixes fast. Monk shows that typical hyperbolic surfaces have an optimal spectral gap. In other words, the generic surface is about as hard to cut apart as the theory of such surfaces allows.

Catalog record

CREATORLaura Monk
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE113 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-15157

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Typical hyperbolic surfaces have an optimal spectral gap Laura Monk Abstract The first non-zero Laplace eigenvalue of a hyperbolic surface, or its spectral gap, measures how well-connected the surface is: surfaces with a large spectral gap are hard to cut in pieces, have a small diameter and fast mixing times. For large hyperbolic surfaces (of large area or large genus g g , equivalently), we know that the spectral gap is asymptotically bounded above by 1 4 \frac{1}{4} . The aim of these talks is to present joint work with Nalini Anantharaman, where we prove that most hyperbolic surfaces have a near-optimal spectral gap. That is to say, we prove that, for any ϵ > 0 \epsilon>0 , the Weil–Petersson probability for a hyperbolic surface of genus g g to have a spectral gap greater than 1 4 − ϵ \frac{1}{4}-\epsilon goes to one as g g goes to infinity. This statement is analogous to Alon’s 1986 conjecture for regular graphs, proven by Friedman in 2003. I will present our approach, which shares many similarities with Friedman’s work, and introduce new tools and ideas that we have developed in order to tackle this problem. 1 Introduction: the spectral gap of a compact hyperbolic surface 1.1 Hyperbolic surfaces and the high-genus limit The objects we will study throughout these lectures are compact hyperbolic surfaces of large genus . For a integer g\geq 2 , by the classification of surfaces, we know that there exists one and only one oriented, connected, closed (i.e. compact without boundary) surface of genus g ; we denote as S_{g} such a surface. We are interested in hyperbolic surfaces , i.e. Riemannian metrics on the topological surface S_{g} of constant curvature -1 . Such a hyperbolic surface is always isometric to a quotient X=\Gamma\diagdown\mathbb{H} of the hyperbolic plane \mathbb{H}=\{x+iy,y>0\}\quad\text{equipped with}\quad\,\mathrm{d}s^{2}=\frac{\,\mathrm{d}x^{2}+\,\mathrm{d}y^{2}}{y^{2}} (1) by a Fuchsian group \Gamma , a discrete, co-compact subgroup of \mathrm{PSL}_{2}(\mathbb{R}) . The set of all hyperbolic metrics on S_{g} , up to isometry, is the moduli space \mathcal{M}_{g} . By the Gauss–Bonnet theorem, the area of a hyperbolic surface of genus g is always 2\pi(2g-2) . We will be interested in the large-genus regime, which can also be interpreted as a large-scale regime. One of the main characters that will appear in our story are closed geodesics , and their lengths. A loop \mathbf{c} on the surface X is a piece-wise smooth map \mathbf{c}:\mathbb{R}\diagup\mathbb{Z}\rightarrow X with nowhere vanishing derivatives. Note that, in particular, all of our loops are oriented . Two loops \mathbf{c}_{0} , \mathbf{c}_{1} are said to be homotopic if there exists a continuous map [0,1]\times\mathbb{R}\diagup\mathbb{Z}\rightarrow X which matches with \mathbf{c}_{0} and \mathbf{c}_{1} when the first variable is fixed to be 0 and 1 respectively. A loop is called a curve if it is simple (no self-intersections) and non-contractible. We extend these definitions to multi-loops \mathbf{c}=(\mathbf{c}_{1},\ldots,\mathbf{c}_{\mathfrak{n}}) and multi-curves (in the latter case we asssume that the whole multi-loop has no self-intersection, and that for i\neq j , \mathbf{c}_{i} is not homotopic to \mathbf{c}_{j} nor \mathbf{c}_{j}^{-1} ). Due to the negative curvature, in each non-trivial homotopy class on X admits a unique length-minimizer, which is a closed geodesic. We denote as \mathcal{G}(X) the set of primitive closed geodesics on X , i.e. closed geodesics which are not iterates of a closed geodesic. For \gamma\in\mathcal{G}(X) , we denote as \ell_{X}(\gamma) the length of the closed geodesic \gamma on X . The set \{\ell_{X}(\gamma),\gamma\in\mathcal{G}(X)\} is called the length spectrum ; it contains a lot of information on the hyperbolic surface X . 1.2 The spectral gap of the Laplacian The core focus of these notes is the spectrum of the Laplace–Beltrami operator, or Laplacian. The Laplacian is a positive unbounded operator on the space of square-integrable functions L^{2} . One of its defining properties is that it is invariant by isometries. The spectrum of the Laplacian \Delta is defined as the set of complex numbers \lambda such that the operator \Delta-\lambda\mathrm{Id} does not have a continuous inverse defined on L^{2} . On the hyperbolic plane \mathbb{H} , the Laplacian can be written simply in coordinates as \Delta=-y^{2}\big(\frac{\partial^{2}}{\partial x^{2}}+\frac{\partial^{2}}{\partial y^{2}}\big) . Its spectrum is purely continuous and exactly equal to the interval [\frac{1}{4},+\infty) . This special value \frac{1}{4} , central to this discussion, arises from the observation that, for any smooth L^{2} function f on \mathbb{H} , we have \int_{\mathbb{H}}f\Delta f\geq\frac{1}{4}\int_{\mathbb{H}}f^{2} by a simple application of Cauchy–Schwarz and integration by parts. In other words, the operator \Delta-\frac{1}{4}\mathrm{Id} is positive on L^{2}(\mathbb{H}) [ McK70 ] . Considering the Laplacian on a compact hyperbolic surface X=\Gamma\diagdown\mathbb{H} corresponds to restricting the domain of the Laplacian on \mathbb{H} to \Gamma -periodic functions. Because the surfaces we study are compact, the spectrum in this case is a discrete sequence of eigenvalues 0=\lambda_{0}<\lambda_{1}\leq\ldots\leq\lambda_{n}\underset{n\rightarrow+\infty}{\longrightarrow}+\infty. (2) The first eigenvalue, \lambda_{0}=0 , corresponds to constant eigenfunctions (which belong to L^{2}(X) because the surface has finite area). Its multiplicity is one because the dimension of the space of constant functions is the number of connected components, here, one. The spectral gap is the first non-trivial eigenvalue, \lambda_{1}>0 . It is very easy to construct examples of hyperbolic surfaces of any genus with a small \lambda_{1} by pinching a separating closed geodesic, as represented in Fig. 1 . This is what we often refer to as a dumbbell in spectral geometry. The idea is that, at the limit, the surface has two connected components, and hence \lambda_{1} goes to 0 as the length \epsilon of the pinched closed geodesic goes to 0 . Figure 1 : The dumbbell example. The dumbbell surface is poorly connected (it is easy to disconnect it by cutting it with scissors). Its diameter goes to infinity as \epsilon goes to 0 , and, hence, some points in the surface become very far apart. Hyperbolic surfaces are interesting objects to study dynamically as they are simple examples of chaotic systems, due to their negative curvature. For a small \epsilon , the surface has a bottle-neck, which means that particles or random walks starting on one side of the dumbbell will take a very long time to travel to the other side; this means that the mixing time is unusually long. All of these observations are actually characterizations of surfaces with a small spectral gap, and we have the following relationships between all these notions (see the references for precise statements). \lambda_{1}(X) is big X is well-connected \diam(X) is small X has good dynamical properties [ Mag20 ] [ Bus82 , Che70 ] [ Rat87 , GK19 ] Figure 2 : Relationship between the spectral gap and other properties of the surface. 1.3 Surfaces with a large spectral gap Now that we know that having a large spectral gap is a desirable property, our next question is: how big can a spectral gap be? We believe that the hyperbolic surface with the largest spectral gap is the Bolza surface, obtained by gluing the opposite sides of a regular hyperbolic octagon with all angles equal to \pi/4 , as represented in Figure 3 . For this surface, we have a very precise numerical approximation by Strohmaier and Uski [ SU13 ] , \lambda_{1}\approx 3.8388872588421995185866224504354645970819150157. Whilst the fact that this maximises \lambda_{1} over all hyperbolic surfaces is

Read the complete text