A Hadamard theorem in transversely affine geometry with applications to affine orbifolds
Francisco C. Caramello Jr, H. A. Puel Martins, Ivan P. Costa e Silva · 2025
Read A Hadamard theorem in transversely affine geometry with applications to affine orbifolds by Francisco C. Caramello Jr, H. A. Puel Martins, Ivan P. Costa e Silva (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-2503-06344.txt.
Introduction
Hadamard's classic theorem says that a complete surface curving the right way, with no conjugate points, must look like a simple curved sheet at large scale. Francisco Caramello and co-authors carry this idea into a newer setting: manifolds sliced into leaves, where the geometry across the leaves is affine, meaning built from straight line structure rather than distance. They first pin down what a transversely affine foliation should mean, comparing their notion with earlier ones in the literature.
Under transverse versions of the classic conditions, plus compact leaves with finite holonomy, they prove that the universal cover of such a manifold splits as a product. One factor is a contractible space, and the other is the universal cover of a leaf. The same machinery yields a Hadamard style result for affine orbifolds, extending a theorem once proved for affine manifolds. The paper sits in pure differential geometry.
Catalog record
| CREATOR | Francisco C. Caramello Jr, H. A. Puel Martins, Ivan P. Costa e Silva |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 138 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-06344
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Hadamard theorem in transversely affine geometry with applications to affine orbifolds Francisco C. Caramello Jr, H. A. Puel Martins, Ivan P. Costa e Silva Abstract . We introduce and investigate a novel notion of transversely affine foliation, comparing and contrasting it to the previous ones in the literature. We then use it to give an extension of the classic Hadamard’s theorem from Riemannian geometry to this setting. Our main result is a transversely affine version of a well-known “Hadamard-like” theorem by J. Hebda for Riemannian foliations. Alternatively, our result can be viewed as a foliation-theoretic analogue of the Hadamard’s theorem for affine manifolds proven by Beem and Parker. Namely, we show that under the transverse analogs of pseudoconvexity and disprisonment for the family of geodesics in the transverse affine geometry, together with an absence of transverse conjugate points, the universal cover of a manifold endowed with a transversely affine foliation whose leaves are compact and with finite holonomy is diffeomorphic to the product of a contractible manifold with the universal cover of a leaf. This also leads to a Beem–Parker-type Hadamard-like theorem for affine orbifolds. 2020 Mathematics Subject Classification 53C12, 53C50 1. Introduction Let (M,h) be a connected Riemannian manifold. The classic Hadamard’s theorem in Riemannian geometry [ 28 , 29 ] states that the exponential map \exp^{h}_{p}\mathrel{\mathop{\mathchar 58\relax}}T_{p}M\rightarrow M at any point p\in M is a (universal) covering map, provided (M,h) is geodesically complete and its sectional curvature is nonpositive. Since this result implies such a nice control over the underlying manifold’s topology, it naturally motivates attempts of generalizations to a broader class of geometries — say semi-Riemannian or affine manifolds — where the notion of an exponential map still makes sense. At first, such attempts would seem misguided, because geodesic completeness and nonpositivity of sectional curvatures, which are natural in Riemannian geometry (see. e.g., [ 24 , 25 ] ) are much less so in these broader contexts. For example, it is well-known that semi-Riemannian metrics do not, in general, define any distance function, and no analogue of the Hopf-Rinow theorem exists [ 4 , 28 ] . In particular, compactness of the underlying manifold does not imply either geodesic completeness or geodesic connectedness in this context [ 28 , Example 7.16, Exercise 9.12] . In addition, in the semi-Riemannian context it is well-known that both geodesic completeness and how the existence of conjugate points is affected by sectional curvatures depend on the so-called causal character of the geodesics [ 3 ] , and furthermore many of the most interesting geometric examples in Lorentzian geometry — arguably the most important semi-Riemannian type of geometry apart from the Riemannian one — are geodesically incomplete [ 16 , Ch. 5] . These issues notwithstanding, there have been a number of concrete alternative results in these broader contexts that still can be regard as “Hadamard-like” theorems. For example, Flaherty [ 12 ] , studying the particular case of Lorentzian manifolds, applied the hypotheses of geodesic completeness and sign for the sectional curvatures only in timelike directions, though at the cost of introducing the fairly strong hypothesis of future 1-connectedness . i.e., that any two timelike curves starting and ending at the same points can be homotopically deformed through timelike intermediate curves. Theorem 1.1 (Flaherty) . Let M be a time-oriented, causally future-complete Lorentzian manifold which is also future 1-connected. Further suppose that the sectional curvature is nonpositive for all timelike planes. Then the exponential map regularly embeds the future timecone at each point into M . Observe that having hypotheses only in the timelike directions entails that the result affects only timecones, a kind of “partial covering map” result. A different approach was taken by Beem and Parker [ 5 ] , who obtained a result applicable to the broader class of affine manifolds (M,\nabla) . As hypotheses, they directly impose the absence of pairs of conjugate points along all geodesics — a property that follows from nonpositivity of sectional curvatures in the Riemannian case — and ensure a suitable path-lifting property in lieu of completeness, via the concepts of disprisonment and pseudoconvexity of the family of all geodesics of (M,\nabla) . These are standard geometric constraints meant to control the behavior of the geodesic flow and other dynamical systems. (See the references in [ 5 ] for an extensive list of classic applications). However, since (a transversal version of) disprisonment and pseudoconvexity will turn out to be important for us in this paper, we briefly recall the latter notions here. Definition 1.2 . Let (M,\nabla) be an affine manifold and consider a family \mathcal{C} of non-constant \nabla -geodesics. We say that \mathcal{C} is • pseudoconvex if for every compact set K\subset M there is another compact set K^{*}\subset M such that if any segment of \gamma\in\mathcal{C} has its endpoints in K , then that segment is entirely contained in K^{*} ; • disprisoning if given the maximal extension \gamma\mathrel{\mathop{\mathchar 58\relax}}(a,b)\rightarrow M of a geodesic in \mathcal{C} and any t_{0}\in(a,b) , both \gamma|_{(a,t_{0}]} and \gamma|_{[t_{0},b)} fail to have compact closure. If the class \mathcal{C} of all geodesics is pseudoconvex and disprisoning, then we say that (M,\nabla) itself is pseudoconvex and disprisoning. Theorem 1.3 (Beem–Parker) . Let (M^{n},\nabla) be both pseudoconvex and disprisoning, and suppose that no inextendible geodesic therein has a pair of conjugate points. Then (M,\nabla) is geodesically connected and, for each p\in M , the exponential map \exp_{p}\mathrel{\mathop{\mathchar 58\relax}}\mathcal{D}_{p}\subset T_{p}M\rightarrow M is a smooth covering map. In particular, the universal covering manifold of M is diffeomorphic to \mathbb{R}^{n} . Our goal in this paper is to give a foliation-theoretic analogue of the Beem–Parker result described above. A key antecedent to our work is a well-known result by Hebda [ 17 ] , who obtained a transverse version of Hadamard’s theorem for Riemannian foliations, i.e. a foliated manifold endowed with a Riemannian bundle-like metric , which is a Riemannian metric h such that h(X,Y) is a basic function whenever X,Y are projectable fields g -orthogonal to \mathcal{F} . Theorem 1.4 (Hebda) . Let (M,\mathcal{F}) be a Riemannian, codimension q foliation with a complete bundle-like metric h . If the sectional curvature function of h is nonpositive, then the universal cover \tilde{M} of M fibers over a complete simply-connected Riemannian q -manifold N of nonpositive curvature. Moreover, each fiber in this fibration is diffeomorphic to the universal covering of a leaf of \mathcal{F} . In particular, because the fibration is trivial since N\simeq\mathbb{R}^{q} is contractible, \tilde{M} is diffeomorphic to \tilde{L}\times N . Observe that Hebda’s result reduces to the standard Hadamard’s theorem for a Riemannian manifold (M,h) when the foliation is zero-dimensional, that is, when it is the trivial foliation of M by its points. Our work can be alternatively construed as an affine-geometric version of Hebda’s result. However, as we shall see in more detail below, there are in the literature many alternative, not necessarily equivalent notions of what a transversely affine foliation should be. Therefore, we divide the paper in two parts. In the first part, we give an extended discussion of, and develop a new machinery to deal with, affine aspects of the transverse geometry of foliations. In particular, since we are interested in (a suitable notion)