Finiteness of maximal geodesic submanifolds in hyperbolic hybrids (Q824408)

The main motivation for this paper is the following question which was posed by Reid and McMullen: if a hyperbolic manifold of dimension at least 3 and of finite volume contains infinitely many geodesic hypersurfaces, does it have to be arithmetic? This can be seen as a partial geometric strengthening of Margulis' commensurator criterion for arithmeticity. Recall that the commensurator of a subgroup \(\Gamma\) in \(G\) is the subgroup of those \(g \in G\) such that the intersection of \(g\Gamma g^{-1}\) with \(\Gamma\) has finite index in both groups; for example, the commensurator of \(\mathrm{PSL}_2(\mathbb Z)\) in \(\mathrm{PGL}_2(\mathbb R)\) is \(\mathrm{PGL}_2(\mathbb Q)\). If a lattice has infinite index in its commensurator then Margulis' criterion states that it is arithmetic; on the other hand, if this is the case and it contains a single geodesic hypersurface, its images under elements of the commensurator give infinitely many. In a subsequent work by a subset of the authors and other collaborators they give a complete solution to a generalised version of this, namely they prove that a real hyperbolic manifold \(M\) is arithmetic if and only if it contains infinitely many maximal geodesic submanifolds of dimension \(k\) for some \(2 \le k < \dim(M)\). See [\textit{U. Bader} et al., Ann. Math. (2) 193, No. 3, 837--861 (2021; Zbl 07353243)]. In this paper the authors study this question for a particular family of hyperbolic manifolds, namely the ``hybrid manifolds'' constructed by cutting and gluing arithmetic manifolds along totally geodesic hypersurfaces. These were introduced by \textit{M. Gromov} and \textit{I. I. Piatetski-Shapiro} [Publ. Math., Inst. Hautes Étud. Sci. 66, 93--103 (1988; Zbl 0649.22007)] as the first examples of non-arithmetic hyperbolic manifolds in sufficiently high dimensions. While some variations on their construction have been studied since, there are still no other examples known in large dimensions. The authors prove that hybrid manifolds contain only finitely many maximal geodesic submanifolds of dimension \(>1\). A hybrid manifold is constructed as follows: starting from two arithmetic manifolds which contain isometric non-separating embedded geodesic hypersurfaces, one obtains manifolds with isometric totally geodesic boundary by removing these hypersurfaces from both; one then glues these manifolds according to an isometry between the boundaries. The noncommensurability of the original arithmetic manifolds ensures that the resulting hybrid manifold is non-arithmetic. The main ingredient in the proof of the finiteness result in this paper is what the authors call ``angle rigidity'', which states that in a hybrid manifold any geodesic submanifold of dimension \(>1\) intersects the boundary between two non-commensurable blocks perpendicularly (if it is not contained in it). They deduce from this that the geodesic submanifolds cannot be dense in the tangent Grasmannian bundle; the finiteness result then follows from Ratner-type theorems (for hypersurfaces they can use a theorem of \textit{N. A. Shah} [in: Group theory from a geometrical viewpoint. Proceedings of a workshop, held at the International Centre for Theoretical Physics in Trieste, Italy, 26 March to 6 April 1990. Singapore: World Scientific. 718--732 (1991; Zbl 0846.53041)]; in general they have to prove new results). Another interesting consequence of angle rigidity is that it gives a geometric criterion to decide that a hyperbolic manifold cannot be commensurable to a hybrid manifold. The authors apply this in particular to a Coxeter group in dimension 5 which they show is not commensurable to any hybrid manifold; it is also not commensurable to the ``inbred'' manifolds of Agol (see [\textit{M. V. Belolipetsky} and \textit{S. A. Thomson}, Algebr. Geom. Topol. 11, No. 3, 1455--1469 (2011; Zbl 1248.22004)] for a description of these) since it is not quasi-arithmetic.