# Are there only countably many compact topological manifolds?

Up to homeomorphism, there are 2 one-dimensional topological manifolds and countably many 2- and 3-dimensional compact manifolds, respectively, since each manifold in these dimensions can be triangularized and hence be described by a finite amount of combinatorial data.

In higher dimensions this argument doesn't work anymore. So it still true for $n\ge 4$ that the set of homeomorphism classes of compact, connected topological $n$-manifolds (without boundary) is countable?

(I'd be also interested in the same question for diffeomorphism classes of compact smooth manifolds.)

• For diffeomorphism classes the answer is in the affirmative, as any smooth manifold can be triangulated and within each $PL$-class there are finitely many diffeomorphism classes. Feb 21, 2015 at 11:53
• You also want to assume connectedness, since otherwise any disjoint union of closed manifolds would again be a closed manifold, and these disjoint unions can be arbitrarily big. Feb 21, 2015 at 13:18
• @TobiasFritz They would be finite disjoint unions. Feb 21, 2015 at 13:35
• @Alex Degtyarev: can you give references?
– ACL
Feb 21, 2015 at 14:55
• @Wlodzimierz: what do you mean exactly?
– ACL
Feb 21, 2015 at 14:57

It was shown in

J. Cheeger and J. M. Kister, Counting topological manifolds. Topology 9, 1970 149–151.

that there are only countably many compact manifolds up to homeomorphism (even allowing boundaries).

Here is a link to the article.

• You may google for "homotopy Holsztyński Kister Mather. When I do it then the first reference is to MO (of course). That MO answer provides three references to papers of which the first two are pointed to by: J. Cheeger and J. M. Kister, Counting topological manifolds. Topology 9, 1970 149–151, just mentioned by you. At the time Jim Kister knew very well about my result (and some my related) but didn't tell me about the join paper by him and Cheeger--most like my result preceded theirs but mine took much longer to get it in print. Feb 21, 2015 at 20:32
• The join result by Cheeger and Kister (1970) is much-much sharper than the other three papers by Mather(1965), Kister(1968), Holsztyński (1971). The last one is the most general (and perhaps the simplest too) among the homotopical results. Feb 21, 2015 at 20:42
• If one assumes some quantitative topology, then one indeed also gets quantitative bounds on the number of homotopy type or homeomorphism type. This is the geometric finiteness theorem obtained by Grove-Peterson-Wu 1990. Feb 22, 2015 at 15:21
• A link to the MO answer of Włodzimierz Holsztyński : mathoverflow.net/a/137649
– j.c.
Oct 23, 2015 at 16:58