MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Pontryagin Duality for locally compact abliean groups gives plenty of continuous (unitary) characters $\chi : A \to \mathbb{R} / \mathbb{Z}$, but if we do not assume local compactness, can anything be said? In particular, is the following true?

Every abelian Hausdorff topological group has a nontrivial continuous (unitary) character

I believe this to be false, but cannot find a concrete example.

share|cite|improve this question
The trivial group is a concrete example. – YCor Oct 28 '13 at 9:48
up vote 14 down vote accepted

The infinite-dimensional sphere $S^\infty$ (the evident colimit of finite-dimensional spheres) admits a topological group structure whose underlying abelian group is a torsion group (in fact a structure of $\mathbb{F}_2$-vector space). See this answer here:

If $\phi: S^\infty \to S^1$ is a continuous character, then $\phi(S^\infty)$ is evidently a connected torsion subgroup of $S^1$, which can only be trivial since the torsion subgroup $\mathbb{Q}/\mathbb{Z}$ of $S^1$ is totally disconnected.

share|cite|improve this answer

If a topological vector space $X$ is not locally convex, then it usually has not non-zero linear continuous functionals, and this means that there are no non-trivial continuous characters on $X$. For example, you can take the space of all measurable functions on $[0,1]$ with the metrics $$ d(x,y)=\int_0^1\frac{|x(t)-y(t)|}{1+|x(t)-y(t)|}d t. $$

share|cite|improve this answer
This is a very elementary example. Nice :-) – Mariano Suárez-Alvarez Oct 28 '13 at 4:38
I guess we could also use a Banach space such that the existence of nonzero linear functionals on it is independent of ZF? – Qiaochu Yuan Oct 28 '13 at 4:41
I thought that each Banach space has enough linear continuous functionals (by the Hahn-Banach theorem). :) – Sergei Akbarov Oct 28 '13 at 4:44
@SergeiAkbarov, that depends on the Axiom of Choice, which ZF does no include. (The two things are not equivalent, though) – Mariano Suárez-Alvarez Oct 28 '13 at 4:47

I'd think that a linearly ordered group extension of   $\mathbb R$,   especially any non-standard real field wouldn't have any non-trivial (Pontryagin) character.

share|cite|improve this answer
Which topology you have in mind? For the topologies I can think of, there are proper open subgroups, and hence the discrete quotient admits Pontryagin characters. – YCor Oct 28 '13 at 12:05
I thought of the linearly ordered topology (induced by the standard inequality relation). – Włodzimierz Holsztyński Oct 29 '13 at 1:33
Then the convex hull of the standard reals is an open subgroup and is not the whole group. So the quotient is a nontrivial (indeed infinite) discrete abelian group. Thus this topological group admits plenty of characters. – YCor Oct 29 '13 at 22:47
Thank you, @Yves. – Włodzimierz Holsztyński Oct 31 '13 at 7:54

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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