For an Abelian topological group $G$ by $G^{\wedge}$ we denote the Pontryagin dual of $G$, i.e. the group of continuous homomorphisms $G\to\mathbb T:=\{z\in\mathbb C:|z|=1\}$. The group $G^{\wedge}$ is endowed with the topology of uniform convergence on compact subsets of $G$. A topological group $G$ is called Pontryagin reflexive if the canonical homomorphism $G\to (G^\wedge)^\wedge$ is a topological isomorphism.

Problem. Let $G$ be a metrizable Abelian topological group. Is the Pontriagin dual $G^{\wedge}$ of $G$ Pontryagin reflexive?

(This problem was posed 21.09.2017 by Lydia Aussenhofer on page 71 of Volume 1 of the Lviv Scottish Book).

  • $\begingroup$ Is it clear that $G^{\wedge}$ is metrizable? $\endgroup$ – YCor Sep 25 '17 at 22:40
  • $\begingroup$ @YCor, maybe I miss the point of your question, but $\hat G$ doesn't have to be metrisable to ask whether it is reflexive, right? $\endgroup$ – LSpice Sep 25 '17 at 22:53
  • 2
    $\begingroup$ If G is metrizable then $G^\wedge$ is a hemicompact $k$--space. This implies that the canonical mapping $G^\wedge\to G^{\wedge\wedge\wedge}$ is continuous (even an embedding). Only the surjectivity of this mapping is a problem. $\endgroup$ – LDA Sep 26 '17 at 9:06
  • 4
    $\begingroup$ I hope you know that in the case when $G$ is a (metrizable) locally convex space the answer is "yes", see e.g. link.springer.com/article/10.1023%2FA%3A1020929201133 or ams.org/mathscinet-getitem?mr=2037513 $\endgroup$ – Sergei Akbarov Sep 26 '17 at 10:08
  • 1
    $\begingroup$ In my previous comment I forgot to indicate that $G$ must be complete, I am sorry. $\endgroup$ – Sergei Akbarov Sep 26 '17 at 18:51

Not an answer to this problem.
For a non-reflexive $G$ see example: Exercise (23.32) in Hewitt & Ross, Abstract Harmonic Analysis I (Springer 1963).

Consider a topological vector space $L^p(\mathbb R)$, $0 < p < 1$. They show that the (separable metrizable) topological group $G = L^p(\mathbb R)$ with operation of addition has no nonzero continuous characters. So $\widehat{G} = \{0\}$ and $\widehat{\widehat{G}} = \{0\}$, not isomorphic to $G$.

Facts used (Hewitt & Ross include outlines of proofs):

(Day, 1940) this topological vector space has no nonzero continuous linear functionals

(Hewitt & Zuckerman, 1950) every continuous character $\chi$ on $G$ has the form $\chi(x) = \exp(2\pi i f(x))$ for some continuous linear functional $f$.

| cite | improve this answer | |
  • 2
    $\begingroup$ This says that this $G$ isn't reflexive, but it doesn't seem to provide an example of a non-reflexive Pontryagin dual. I apologize if I'm missing something... $\endgroup$ – Alon Amit Sep 26 '17 at 0:17
  • $\begingroup$ You are right. Not an answer to the original problem. $\endgroup$ – Gerald Edgar Sep 26 '17 at 0:19

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.