Intersection pairing on hyperbolic 4-manifolds
Michael Kapovich · 2025
Read Intersection pairing on hyperbolic 4-manifolds by Michael Kapovich (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-2510-16848.txt.
Introduction
Michael Kapovich proves one estimate about shapes in four dimensions in a 2025 mathematics paper. The shapes are hyperbolic 4-manifolds, curved spaces with constant negative bending. He bounds their intersection pairing, a count of how two surfaces inside must cross. The bound uses Thurston norms, which measure the simplest surface carrying a given class. The result ties crossing numbers to surface complexity inside these spaces.
Catalog record
| CREATOR | Michael Kapovich |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-16848
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Intersection pairing on hyperbolic 4-manifolds Michael Kapovich Abstract This paper was written as an MSRI preprint in 1992. For a long time I was planning to revise and clean it up. Instead, I am just posting the paper “as is.” 1. Introduction 1.1 We begin with the following elementary question: How one can estimate the algebraic intersection number <\alpha,\beta> between two closed curves \alpha , \beta on a complete oriented Riemannian surface M ? Suppose that the metric on M is hyperbolic (has curvature (-1) ). Then the answer can be given in the terms of lengths of \alpha,\beta . For example: |<\alpha,\beta>|\leq K(l(\alpha),l(\beta))=2e^{l(\beta)}(\pi/2+l(\alpha))+\frac{\pi/2\ +l(\alpha)}{l(\beta)/2-\log(\exp(l(\beta)/2)-1)} In particular, if |<\alpha,\beta>|\geq 1 then 1\ \leq\ \sinh(l(\beta)/2)\ \sinh(l(\alpha)/2) Note that the right side depends on the metric introduced on M while the left side is purely topological. 1.2. Now let M be an arbitrary complete oriented hyperbolic 4-manifold, \sigma_{j}:\Sigma_{j}\to M\ (j=1,2) be two cycles in Z_{2}(M,\mathbb{Z}) , where \Sigma_{j} are closed oriented connected surfaces. The main aim of this paper is to estimate from above the absolute value of the intersection pairing |<[\sigma_{1}],[\sigma_{2}]>| so that the estimate depends just on the Euler characteristics of \Sigma_{j} . This will be done under certain condition on \sigma_{j} . Definition 1.1 . The maps \sigma_{j} are incompressible if the induced homomorphisms of fundamental groups \sigma_{j*} are injective. Theorem 1.2 . There exists a function C(.,.) such that for any complete hyperbolic 4-manifold M and for any classes [\sigma_{1}],[\sigma_{2}] in H_{2}(M,\mathbb{Z}) which have incompressible representatives, we have: |<[\sigma_{1}],[\sigma_{2}]>|\leq C(|\chi(\Sigma_{1})|,|\chi(\Sigma_{2})|)\ . Theorem 1.3 . The conclusion of Theorem 1.2 holds even if instead of the algebraic intersection number <[\sigma_{1}],[\sigma_{2}]> we take the geometric intersection number i(\sigma_{1},\sigma_{2}):=\inf\{ number of points of intersection between \varphi_{1}(\Sigma_{1}),\varphi_{2}(\Sigma_{2}):\varphi_{j} is homotopic to \sigma_{j},j=1,2\} Thus, Theorem 1.2 shows how the intersection pairing together with simplicial norm on the second homology group provides an obstruction for existence of complete hyperbolic structure on 4-manifolds. The intersection pairing itself is not too interesting invariant for hyperbolic 4-manifolds. If M^{4} is closed then just two symmetric bilinear forms of the given rank can occur. On another hand, any symmetric bilinear form can be realized as the intersection form of compact convex hyperbolic 4-manifold. There are no homological obstructions for hyperbolicity in the dimensions 2 and 3 because the length of curve (unlike the area of surface) is not a topological invariant. 1.3. Consider the special case: suppose M=M(e,g) is homeomorphic to a \mathbb{R}^{2} - bundle over a closed orientable surface F of genus g>1 . Such bundles are characterized by their ”Euler numbers”. In this case it is just the ”selfintersection” number e=<[F],[F]> , where we identified F with the zero section of the bundle. Corollary 1.4 . The condition |e|\leq C(2g-2,2g-2) is necessary for existence of complete hyperbolic structures on M . Denote by {\mathcal{S}}(e,g) an orientable 3-manifold which is a circle bundle over the closed oriented surface of genus g such that the Euler number of the fibration is e . Corollary 1.5 . The condition |e|\leq C(2g-2,2g-2) is necessary for existence on {\mathcal{S}}(e,g) flat conformal structures with nonsurjective development maps. 1.4. The examples of hyperbolic manifolds M=M(e,g) with e=0 are easy to construct and they exist for arbitrary genus g>0 . If the surface F of genus g is uniformized by the group \Gamma\subset PSL(2,\mathbb{C}) then the extension of \Gamma in \mathbb{H}^{4} is the holonomy group for the complete hyperbolic structure on M(0,g) . As the application of the Bieberbach theorem we obtain that for e>0 manifolds M(e,1) can’t have complete hyperbolic structure. The case g>1\ ,\ e>0 is less trivial, first examples of complete hyperbolic structures were obtained in [ 8 , 12 , 15 ] (with the estimate (2g-2)/3 due to Kuiper): Theorem 1.6 . The manifold M=M(e,g) admits a complete hyperbolic structure assuming that 0<e\leq(2g-2)/3 After [ 8 ] M.Anderson, [ 2 ] , proved Theorem 1.7 . Let E\to B be an arbitrary vector bundle with the compact base B of negative sectional curvature. Then E admits a Riemannian metric with strictly negatively pinched sectional curvature: 0>a_{E}>K_{E}>-1 for some constant a_{E} depending on the bundle. Conjecture 1.8 . (See [ 8 ] .) The (Milnor-Wood) inequality 0\leq|e|\leq(2g-2) is the necessary condition for existence of complete hyperbolic structure on the manifold M=M(e,g) . Remark 1.9 . It is important here that M is fibered. N.Kuiper, [ 13 ] , constructed a sequence of complete hyperbolic manifolds M_{g}^{4} which are homotopy equivalent to closed surfaces F_{g} of genus g such that: \lim_{g\to\infty}<[F_{g}],[F_{g}]>/(2g-2)=2/\sqrt{3}>1 Remark 1.10 . (N.Kuiper, [ 14 ] ). Suppose that M is hyperbolic and \Sigma\subset M is an imbedded minimal surface of genus g . Then the Milnor-Wood inequality |<[\Sigma],[\Sigma]>|\leq(2g-2) holds. 1.5. For another example of application of Theorem 1.3 consider a pair of Kleinian groups G_{1},G_{2}\subset SO(4,1) such that G_{j} are isomorphic to \pi_{1}(S_{g_{j}}) (where S_{g} is a closed orientable surface of genus g ); and the limit sets L_{j}=\Lambda(G_{j}) are topological unknots in S^{3} ( j=1,2 ). Suppose that the group G generated by G_{1},G_{2} is discrete and torsion-free. If we do not control the genera g_{j} then L_{1},L_{2} can form an arbitrary link in S^{3} (see for example [GLT]). However, Theorem 1.3 implies that given a pair (g_{1},g_{2}) there exist not more than k(g_{1},g_{2}) links (L_{1},L_{2}) such that the group G is discrete and torsion-free, where k(g_{1},g_{2})\leq\exp(8000(g_{1}+g_{2})/\mu) . 1.6. Probably it’s possible to prove Corollary 1.4 for convex compact 4-manifolds by comparing two \eta -invariants for flat conformal manifolds, [ 16 , 17 ] . A more realistic idea was suggested to author by M. Gromov who proposed to compactify the moduli space of all hyperbolic structures on the given fiber bundle. Formally speaking this idea does not work, since arbitrary large number of self-intersections of zero section can be pinched to point in limit. However, what we are using in this paper are “pre-limit” considerations based on Mumford’s compactness theorem and existence of Margulis constant. The idea of the proof is quite simple. Suppose that we realized the classes [\sigma_{1}],[\sigma_{2}] by piecewise-geodesic surfaces \Sigma_{1},\Sigma_{2} in M such that the number and diameter of simplices in \Sigma_{1}\cup\Sigma_{2} are bounded from above. Then the existence of the universal Margulis constant and the fact that two geodesic planes intersect transversally in not more than one point immediately imply the assertion of Theorem 1.2 . Certainly it’s impossible in general to estimate from above the diameter of simplices since the diameters of the surfaces \Sigma_{1},\Sigma_{2} can be unbounded. However the “long” pieces of \Sigma_{1},\Sigma_{2} are contained in the “thin” part of the manifold M which have very simple topological structure. Then the detailed analysis of the situation in the “thin” part of M (section 2) and correct choice of the surfaces representing [\Sigma_{j}] (sections 3, 4) yield the claimed result. 1.7. Thus, Theorem 1.2 is a “0-th order approximation” to Conjecture 1. The simplest examples of negatively pinched closed manifolds of dimension 4 which do not admit hyperbolic structure are given by complex-hyperbolic manifolds. More sophisticated examples were constructed by Mostow and Siu, [ 22 ] ; there are no (real) hyperbolic structures on any compact Kähler manifold of (real) dimension >2 due to theorem of J.Carlson and D.Toledo, [ 5 ]