Sign up ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $E$ be a not normable locally convex space, define $$F: E'\times E\to \mathbb R$$ $$(f,e)\to f(e)$$ I have to show that $F$ is not continuous when $E'\times E$ is given product topology.

I was reading an article and i came across with this fact.. Please give me atleast a hint to start..

My try: I know that $E$ is normable if and only if origin has a convex bounded neighborhood. So i was trying to produce any such neighborhood to contradict to assumption. Assume $F$ is continuous, then we have $\{(f,e): a<f(e)<b\}$ is open in product topology of $E'\times E$, for any $a,b\in \mathbb R$. This means there is some open set $U'$ in $E'$ and $U$ in $E$ such that $$U'\times U\subset \{(f,e): a<f(e)<b\}$$
Now let $V:=\{e\in E: a<f(e)<b;\forall f\in U'\}$, this is open convex neighborhood of origin, but how to prove this is bounded. Or we have any other way to produce such a neighborhood.

Thanks for your time.

share|cite|improve this question
I find the presence of $a,b$ distracting. Since you are constructing a neighborhood of the origin, it's convenient to consider $\{(f,e) : |f(e)|<1\}$ and accordingly $V=\{e: |f(e)|<1 \ \forall f\in U'\}$. // More importantly, you have not really used the assumption that $E$ is locally convex: it carries a family of seminorms, which we can assume directed. // Take any open nbhd of zero $W\subset E$: it contains a nbhd of the form $\{p(x)<r\}$ where $p$ is a seminorm. Now you need a functional $f\in U'$ such that $|f(x)|<\epsilon \implies p(x)<r$. Think of Hahn-Banach. –  user53153 Dec 31 '12 at 7:11
What is $E'$? What topology do you take on $E'$? –  Neutral Element Oct 26 '14 at 10:28

1 Answer 1

In fact it turns out that the evaluation map is discontinuous for any locally convex topology on $E'$ (if $E$ is not normable).

You can find a concise sketch of the argument on the top of p.2 of Kriegl and Michors book: The convenient setting of global analysis (available online 1): To prove your statement one argues by contradiction and assumes that $ev$ was continuous, i.e. $ev (V \times U) \subseteq [-1,1]$. Then $U$ is a bounded zero-neighbourhood $U$ in $E$ since it is contained in the polar of $V$. Thus $E$ must be normable.

share|cite|improve this answer

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.