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.

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

The problem is:

Let $E$ be a normed space, let $x\in E$, let $\left(x_{n}\right)_{n\in\mathbb{N}}\subset E$ be a sequence. I need to show that $\left(x_{n}\right)$ converges weakly to $x$ if and only if $\left(x_{n}\right)$ converges to $x$ in $\left(E,\,\sigma\left(E,\, E^{*}\right)\right)$, $\sigma\left(E,\, E^{*}\right)$ denoting the weak topology on $E$.

My efforts:

I have that $x_{n}\rightarrow x$ in $\left(E,\,\sigma\left(E,\, E^{*}\right)\right)$ is equivalent to: $\forall U\in\sigma\left(E,\, E^{*}\right),\, x\in U\,:\quad\left\{ n\in\mathbb{N}:\, x_{n}\notin U\right\}$ is finite.

My question:

It is unclear to me how to pass from the weak convergence in $E$ to the convergence in the weak topology on $E$.

How can I prove both directions?

Thanks, Franck.

share|cite|improve this question
What is $X$ (capital $X$) The same as $E$? And what is $\sigma$? Not everyone has access to the index of notation to your book or lecture. – user20266 Jun 10 '12 at 15:54
Assume $x_n \to 0$ weakly. This means that for fixed $\varepsilon$ you have for every $f \in E^\star$ an index $N_f$ such that $\lvert f(x_n)\rvert <\varepsilon$ for every $n > N_f$. This is what you have. Now let $U(f_1 \ldots f_k; \varepsilon)$ denote a weak neighborhood of the origin like this: $$\{x\in E \ :\ \lvert f_j(x)\rvert < \varepsilon,\ j=1, 2\ldots k\}.$$ Playing a bit with what you have you can prove that there exists an index $N$ such that $x_n \in U(f_1 \ldots f_k; \varepsilon)$ for every $n\ge N$. (...) – Giuseppe Negro Jun 10 '12 at 17:47
So $x_n\in U(f_1 \ldots f_k; \varepsilon)$ except, at most, for a finite number of indices. This is enough to infer that $x_n \to 0$ in $\sigma (E; E^\star)$, since every weakly open set $U$ containing the origin of $E$ contains a neighborhood of the form $U(f_1 \ldots f_k; \varepsilon)$ for some $f_1\ldots f_k \in E^\star$ and $\varepsilon >0$. (This is what "base of neighborhoods" means). – Giuseppe Negro Jun 10 '12 at 17:51
Try reasoning in a similar way to prove the other implication, namely: $x_n \to 0$ in $\sigma(E;E^\star) \Rightarrow x_n \to 0$ weakly. – Giuseppe Negro Jun 10 '12 at 17:54
What you have is not "almost equivalent" to $x_n \to 0$ weakly: indeed, it is stronger. $x_n \to 0$ weakly means exactly that for every $f \in E^\star$ and $\varepsilon > 0$ we have $x_n \in U(f; \varepsilon)$ except at most for a finite number of indices. And you already have that: just set $k=1$. – Giuseppe Negro Jun 10 '12 at 21:15
up vote 1 down vote accepted

WLOG, set $x=0$ (weak translation is continuous):

We have to show that $x_{n}\rightharpoonup0\Leftrightarrow x_{n}\rightarrow0$ in $\left(E,\sigma\left(E,E^{\star}\right)\right).$

$$ \Rightarrow\forall\epsilon\quad\exists N_{\epsilon}\in\mathbb{N}\quad\forall f\in E^{\star}:\quad\left|f(x_{n})\right|<\epsilon\quad\forall n\geq N_{\epsilon}. $$

Let $U\left(f_{1},...,f_{k};\epsilon\right)$ denote a weak neighborhood of the origin:

$$ U\left(f_{1},...,f_{k};\epsilon\right):=\left\{ x\in E:\,\left|f_{i}\left(x\right)\right|<\epsilon,\, i=1,...,k\right\} . $$

So $x_{n}\in U\left(f_{1},...,f_{k};\epsilon\right)$, except at most for a finite number of indices. This is enough to infer that $x_{n}\rightarrow0$ in $\left(E,\sigma\left(E,E^{\star}\right)\right)$, since every weakly open set $U$ containing the origin of $E$ contains a neighborhood of the form $U\left(f_{1},...,f_{k};\epsilon\right)$ for some $f_{1},...,f_{k}\in E^{\star}$ and $\epsilon>0$.

For the reverse:

$x_{n}\rightarrow0$ in $\left(E,\sigma\left(E,E^{\star}\right)\right)\Rightarrow x_{n}\in U\left(f_{1},...,f_{k};\epsilon\right)$, except for at most a finite number of indices, $f_{i}\in E^{\star},\,\epsilon>0$.

$\Rightarrow f_{i}(x_{n})\rightarrow0$ for $f_{i}\in E^{\star},\, i=1,...,k$.

Set $k=1\Rightarrow\forall f\in E^{\star}:\, f\left(x_{n}\right)\rightarrow f\left(0\right)=0\Rightarrow x_{n}\rightharpoonup0$


share|cite|improve this answer
Good, thank you! I should have made our comments into an answer but I'm busy right now, you saved me some precious time. – Giuseppe Negro Jun 13 '12 at 17:13

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.