It is known, and not too difficult to prove that if a dual of a normed space is separable, it is separable itself. Does the same hold for Frechet spaces?

(By the dual I mean the dual endowed with the strong topology.)

  • $\begingroup$ Did you try to transfer the usual proof in case of a normed vector space? $\endgroup$ – gerw Apr 9 '18 at 6:49
  • $\begingroup$ @gerw The proof that I know relies on the norm heavily: you should be able to choose "small" functionals and "large" vectors. If you know a transferable proof, could you please give a reference? $\endgroup$ – erz Apr 9 '18 at 7:02
  • $\begingroup$ Do you consider the weak*-topology or the strong topology on the dual space? Note that $F^*$ with the strong dual is in general not metrizable. In fact, this happens only if $F$ is already normable, i.e. a banach space. In this situation we can apply the result for banach spaces. $\endgroup$ – p4sch Apr 9 '18 at 8:39
  • $\begingroup$ @p4sch I mean the strong topology. I know that $F^*$ is not always metrizable, but in my specific case it is separable, and I want to conclude that $F$ is then separable too, without assuming that $F$ is normed. $\endgroup$ – erz Apr 9 '18 at 8:48

I think so (no books at hand for a reference). All we need for the normed proof to go through, is some sort of Hahn-Banach theorem that gives us enough functionals to works with. And Fréchet spaces have enough functionals.

  • $\begingroup$ we also need to choose "small" functionals and "large" vectors (at least in the proof that I know). Perhaps you could elaborate? $\endgroup$ – erz Apr 9 '18 at 7:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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