8
$\begingroup$

Is it true that if $ E,F$ are two topological vector spaces (or say Banach spaces) over $\mathbb{R}$ such that they have nonempty open subsets $U\subset E, V\subset F$ which are homeomorphic, then the two vector spaces are isomorphic? If false, then what can we say if the two open subsets are $\mathcal{C}^1$-diffeomorphic?

$\endgroup$
4
  • $\begingroup$ I would guess that the first question is false, simply on the grounds that it is pretty trivially true for finite dimensional ones, so it's probably false for arbitrary vector spaces. $\endgroup$
    – Simon Rose
    Sep 30, 2015 at 9:49
  • $\begingroup$ My guess is that it is true in all spaces, since it is true for finite-dimentional ones. But I cannot give a proof or a disproof. $\endgroup$
    – usr203050
    Sep 30, 2015 at 9:51
  • 1
    $\begingroup$ My guess for the second part is that the differential at some point (or may be at any point) can give an isomorphism between the two spaces. $\endgroup$
    – usr203050
    Sep 30, 2015 at 10:06
  • 2
    $\begingroup$ Yes. The interesting question between those two extremes is the case of Lipschitz equivalence (for Banach spaces). Again much work has been done on this case, e.g. by Lindenstrauss et al. $\endgroup$
    – shasta
    Sep 30, 2015 at 10:19

1 Answer 1

11
$\begingroup$

This is false. All separable Banach spaces, for example, are homeomorphic.Indeed, there is a considerable body of work on when topological vector spaces are homeomorphic (see Bessaga and Pelczynski "Selected Topics in infinite-dimensional Topology" for starters).

$\endgroup$
5
  • $\begingroup$ Ok thank you. So it is far from evident in any case. $\endgroup$
    – usr203050
    Sep 30, 2015 at 9:59
  • $\begingroup$ And what if the subsets are smoothly diffeomorphic? (The second question). $\endgroup$
    – usr203050
    Sep 30, 2015 at 10:03
  • 1
    $\begingroup$ Isn't the diffeomorphic part answered by the comment of usr203050? $\endgroup$ Sep 30, 2015 at 10:17
  • 2
    $\begingroup$ If diffeomorphism is something that has differential as a bounded invertible operator between tangent spaces, then this very operator is linear isomorphism between spaces. If not, what concretely do you mean? $\endgroup$ Sep 30, 2015 at 10:17
  • $\begingroup$ That is what I mean! :-) Also, the comment of shasta gives something more. $\endgroup$
    – usr203050
    Sep 30, 2015 at 10:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.