Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Let $\ell^\infty$ be the Banach space of bounded sequences with the usual norm and let $c,c_0$ be the subspaces of sequences that are convergent, resp. convergent to zero. Show that:

  • The linear functional $\ell_0\colon c\rightarrow \mathbb{C}$ defined for $x = (x_n) \in c$ by $$ \ell_0(x) = \lim_{n\rightarrow \infty} x_n$$ extends to a continuous functional on $\ell^\infty$
  • if $L$ denotes the set of all continuous extensions of the functional $\ell_0$ from (1), then a sequence $x = (x_n) \in \ell^\infty$ belongs to $c_0$ iff $$\ell(x) = 0 \;\; \forall \ell \in L$$
  • Describe $c$ in a similar way My try:

(1): This follows by Banach limits.

(2): $(\Rightarrow)$ follows by extension

$(\Leftarrow)$ Here Im a bit unsure, assume $x \not \in c_0$ if $x \in c$ we get an contradiction. But if $x\not \in c$ what happens then, can we use a subsequence? since we have bounded functionals? can we use $\ell x = \lim_{k \rightarrow \infty} x_{n_k}$ or something like that, would $\ell \in L$?

(3): same as two I suppose, can we use subseqeunces?

Please correct what I'm missed

share|cite|improve this question
(2) is a direct application of the Hahn-Banach theorem: As $c_0$ is a closed subspace of $l^\infty$, for $x\in l^\infty \setminus c_0$ exists an $l\in L$ with $l(x)\neq 0$. – Vobo Dec 29 '12 at 12:08
thanks, but I don't see the importance of $c_0$ being closed. can you please expand? – Johan Dec 29 '12 at 13:53
Hahn-Banach can only seperate a point and a closed set. – Vobo Dec 29 '12 at 18:01
Yes you are correct! Would it work for $c$ as well? is it also closed? – Johan Dec 29 '12 at 18:11
It is indeed. But you should think why, it is the same reason as for $c_0$. – Vobo Dec 29 '12 at 19:08

In the general case you have the equivalence: $x\in c$ if and only if there exist a constant $a\in\mathbb{R}$ such that $l(x)=a$ for all $l\in L$.

One is easy, so suppose that $x\in \ell^\infty\setminus c$. Consider the space $$c+x\mathbb{R}=\{z\in\ell^\infty:\ z=y+\lambda x,\ y\in c,\ \lambda\in\mathbb{R} \}$$

Let $b\neq a$ and define a bounded linear functional $\tilde{l}:c+x\mathbb{R}\rightarrow\mathbb{R}$ by $$\tilde{l}(z)=l_0(y)+\lambda b$$

Note that $\tilde{l}$ is a bounded linear functional defined in a subspace of $\ell^\infty$ and such that $\tilde{l}(x)=b\neq a$. From here you can conclude.

share|cite|improve this answer
thanks a lot! But this $\tilde{l}$ is it defined for $\ell^\infty$? Or is $c + x\mathbb{R} = \ell^\infty$? – Johan Dec 29 '12 at 14:53
No, but now you can extend it by using Hahn-Banach and then you conclude. I think that there is a more straightforward approach. – Tomás Dec 29 '12 at 15:02
So we can extend two times? – Johan Dec 29 '12 at 15:04
There is no problem in extend it again. But as I say maybe there is a more straightforward argument. – Tomás Dec 29 '12 at 15:14

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.