Rational Hyperbolicity Problem

Ricardo A. E. Mendes, Alessandro Minuzzo, Marco Radeschi · 2025

Read Rational Hyperbolicity Problem by Ricardo A. E. Mendes, Alessandro Minuzzo, Marco Radeschi (2025) 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-2512-23101.txt.

Introduction

Ricardo Mendes, Alessandro Minuzzo, and Marco Radeschi resolve a cluster of questions about the rational homotopy type of manifolds with symmetry. A compact simply connected manifold with a group action is called rationally elliptic when its rational homotopy is finite-dimensional, a strong tameness property. The problem asks when being elliptic can be read off from the shape of its orbit space, the quotient by the action.

Under completeness and mild side conditions, the authors prove the manifold is rationally elliptic exactly when the quotient space is flat. The side conditions include trivial principal isotropy or simply connected principal orbits. That match answers several standing questions. The extra conditions are shown to matter: without them, the authors construct elliptic examples whose quotient admits a hyperbolic metric instead. The paper thus both settles the main statement and maps the boundary where it fails.

Catalog record

CREATORRicardo A. E. Mendes, Alessandro Minuzzo, Marco Radeschi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-23101

Browse the full library · About the Tower of Babel

Full text

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

Rational Hyperbolicity Problem Thanks: R. M. has been supported by DMS-2506409, DMS-2005373, and the Dodge Family College of Arts and Sciences Junior Faculty Summer Fellowship of the University of Oklahoma. A. M. was partally supported by GNSAGA of the INdAM Ricardo A. E. Mendes, Alessandro Minuzzo, Marco Radeschi Abstract . We prove that a compact simply connected manifold M M with a variationally complete G G -action satisfying certain mild conditions (e.g. trivial principal isotropy, or simply connected principal orbits) is rationally elliptic if and only if M / G M/G is flat. This answers several conjectures and problems regarding the rational homotopy of manifolds with symmetries. On the other hand, without the extra conditions we find examples of rationally elliptic G G -manifolds M M where M / G M/G admits a hyperbolic metric. 1. Introduction A compact, simply connected manifold M is called rationally elliptic if \sum_{i>1}\dim(\pi_{i}(M)\otimes\mathbb{Q})<\infty , and rationally hyperbolic otherwise. This notion is known to impose severe topological restrictions to such an M [ 9 ] , and it is important in geometry via the famous Bott–Grove–Halperin conjecture , stating that any compact simply connected manifold of (almost) nonnegative sectional curvature is rationally elliptic. In the past decades, several geometric criteria to detect rational ellipticity have been produced [ 12 , 29 , 14 , 13 , 32 ] . Most of these operate under the assumption that M is acted on by a compact group G of isometries. For example, we have: Theorem 1.1 . Let M be a compact, simply connected manifold with an isometric action by a compact group G . Then: (1) [ 14 ] : If the G action is polar with a flat or spherical section, then M is rationally elliptic. (2) [ 13 ] : if \dim M/G=2 (in which case M/G is a 2-dimensional orbifold by [ 18 ] ) and M/G is an elliptic orbifold then M is rationally elliptic. We recall that an isometric G -action is polar if it admits a section , i.e. a submanifold S with \dim S=\dim M/G which meets all orbits perpendicularly. Recall furthermore (see, for example, [ 5 , Theorem 5.1.5] ) that a compact 2-dimensional orbifold is either bad (i.e. compact simply connected but not a manifold), or covered by a round sphere, by a Euclidean space or by a hyperbolic space. In all but the last case, we say that the orbifold is elliptic. The theorems above were believed to be sharp, in the following sense: Conjecture 1.2 . [ 14 , p. 309] Let M be a compact, simply connected manifold with a polar action by a compact group G . If the section is hyperbolic, then M is rationally hyperbolic. Conjecture 1.3 . [ 13 , Problem 2.6] Let M be a compact, simply connected manifold with an isometric G -action, such that \dim M/G=2 and M/G is covered by the hyperbolic plane. Then M is rationally hyperbolic. The actions in Conjecture 1.2 and 1.3 are special cases of so-called variationally complete actions (see Section 2.2 ), which were proved by Lytchak and Thorbergsson to be equivalent to isometric actions whose quotient space is a good orbifold without conjugate points [ 18 ] . The main goal of this paper is to show that, while in complete generality the conjectures are false, they become true under certain conditions of the action. For this, we say that an isometric action of a Lie group G satisfies condition (P) if any of the following conditions holds: (P)=\left\{\begin{array}[]{l}\textrm{The principal isotropy group $H$ has odd }|\pi_{0}(H)|.\\ \textrm{A principal orbit $L$ has }H^{1}(L;\mathbb{Z}_{2})=0.\end{array}\right. Property (P) is for example satisfied for actions with trivial principal isotropy group, e.g. actions of abelian groups. It is possible to prove using the Leray spectral sequence for the map M\to M/G (although we will not do it here) that property (P) is satisfied for actions where every singular orbit L^{\prime} has \dim L^{\prime}\leq\dim L-2 , where L is a principal orbit. In fact, in [ 14 , page 309] it was essentially observed that Conjecture 1.2 is valid under this condition, by the Lacunary Principle. Theorem A . Let M be a compact, simply connected manifold with a variationally complete G -action satisfying property (P) . Then M is rationally elliptic if and only if M/G is flat. Theorem A uses the following flatness result for manifolds without conjugate points, which might be of independent interest: Lemma 1.4 (Lemma 5.3 ) . Let N be a complete simply connected manifold without conjugate points. Suppose that N has sub-exponential volume growth, and its isometry group contains a discrete, finitely generated co-compact subgroup \Gamma which is either abelian or generated by reflections. Then N is flat. In particular, Theorem A proves Conjecture 1.2 and 1.3 when G satisfies property (P) , since in both situations M/G=N/\Gamma is hyperbolic. On the other hand, if no extra assumption is added to Conjectures 1.2 and 1.3 , we produce counterexamples: Theorem B . (Negative answer to Conjectures 1.2 and 1.3 ) For any \ell\geq 4 (resp. \ell\geq 3 ), there exist rationally elliptic 5 -manifolds (resp. 8-manifolds), with a polar SO(3) -action (resp. SU(2)\times SU(2) -action) whose quotient is isometric to a hyperbolic polygon with \ell edges. Remark 1.5 . We remark that Theorem B is optimal in several ways: • First, the group SO(3) has the lowest possible dimension for a counterexample to Conjectures 1.2 and 1.3 , since Lie groups of dimension <3 are abelian hence they satisfy (P) . • Furthermore, the lowest dimension of a counterexample is indeed 5: in fact, since variationally complete actions do not have exceptional orbits, a simply connected G -manifold ( G connected) of dimension \leq 4 with a hyperbolic quotient M/G must have dimension equal to 4, with principal orbits having dimension 2, singular orbits projecting to boundary strata having dimension 1, and singular orbits projecting to the corners of M/G corresponding to fixed points. At the fixed points, the isotropy representation of G is polar with 2 -dimensional sections, hence G is a torus, thus satisfying the conjectures by Theorem A . • Finally, an 8-dimensional example has quotient a hyperbolic triangle, with the lowest possible number of faces. The counterexamples in Theorem B turn out to provide a negative answer to the following: Problem A (Wilhelm/Radeschi–Samani) . Assume M , M^{\prime} are closed, simply connected manifolds, with isometric actions by Lie groups G on M , and G^{\prime} on M^{\prime} , such that M/G is isometric to M^{\prime}/G^{\prime} . Is it true that M is rationally elliptic if and only if M^{\prime} is rationally elliptic? We actually find infinite families of manifolds with isometric quotients but with different rational homotopy behaviour: Theorem C . (Negative answer to Problem A )There exist infinite pairs of manifolds (M_{\ell}=W^{\#\ell},G=SO(3)) , (N_{\ell}=SU(3)^{\#\ell},H=SU(3)) with M_{\ell} rationally elliptic, N_{\ell} rationally hyperbolic, and M_{\ell}/G isometric to N_{\ell}/H . Rational ellipticity is also related to the concept of topological entropy h_{top}(M,g) of a Riemannian manifold (M,g) , by the work of Paternain and Petean [ 29 ] , who proved in particular that a simply connected manifold that admits a metric with h_{top}(M,g)=0 is rationally elliptic. It

Read the complete text