Take the 2-minute tour ×
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.

Define $$ X := \left \{ (a_n)_{n=0}^\infty :\sum_{n =0}^\infty |a_n| < \infty \right \} $$ with metrics $$ d_{l^1}((a_n)_{n=0}^\infty,(b_n)_{n=0}^\infty) := \sum_{n=0}^\infty |a_n - b_n| $$ $$ d_\infty((a_n)_{n=0}^\infty,(b_n)_{n=0}^\infty) := \sup_{n\geq 0} |a_n-b_n| $$

I want to show that these metrics are not equivalent. And here is my "proof" of that. I am not sure if this is correct.

Let $(x^{(n)})_{n=0}^\infty$ be a sequence in $X$ such that $$ x^{(n)}_k := \begin{cases} \frac 1 n & \text{ if } k \leq n \\ 0 & \text{ if k > n } \end{cases} $$ Define further $\tilde{x} \in X$ with $\tilde{x}_k = 0 $ for all $k \in \mathbb N$. Then $$ \lim_{n \rightarrow \infty} d_\infty (x^{(n)},\tilde{x}) = \lim_{n \rightarrow \infty} |1/n| = 0 $$ such that $(x^{(n)})_{n=0}^\infty$ converges to $\tilde{x}$ with respect to $d_\infty$. Assume now that there exists some (another) sequence $\tilde{x} \in X$ such that $\lim_{n \rightarrow \infty} d_{l^1}(x^{(n)},\tilde{x}) = 0$. Then we have $$ 0=\lim_{n \rightarrow \infty} \sum_{k=0}^n |1/n - \tilde{x}_k| + \sum_{k=n+1}^\infty |\tilde{x}_k| $$

Now we can split the limit and the right term will be $0$ because $\tilde{x} \in X$ and thus by the zero-test the claim follows. But I am not sure how to show that the left term is not $0$.

share|improve this question
Note: Maybe is this approach completly incorrect :) –  André Jan 4 '13 at 22:55

2 Answers 2

up vote 4 down vote accepted

Let me slightly redefine your $x^{(n)}$ by letting $x^{(n)}_k=\frac1n$ if $k<n$ and $0$ otherwise. Then

$$d_{\ell^1}\left(x^{(n)},\tilde x\right)=\sum_{k\ge 0}\left|x^{(n)}_k\right|=n\cdot\frac1n=1$$

for all $n\in\Bbb N$, so $\left\langle x^{(n)}:n\in\Bbb N\right\rangle\not\to\tilde x$ in the topology generated by $d_{\ell^1}$. Thus, $d_{\ell^1}$ and $d_\infty$ generate different topologies on $X$, since (as you showed) $\left\langle x^{(n)}:n\in\Bbb N\right\rangle\to\tilde x$ in the topology generated by $d_\infty$. It’s not necessary to show that $\left\langle x^{(n)}:n\in\Bbb N\right\rangle$ isn’t convergent at all with respect to $d_{\ell^1}$, which is what you seem to be trying to do.

In fact it’s true that $\left\langle x^{(n)}:n\in\Bbb N\right\rangle$ isn’t convergent with respect to $d_\infty$. Suppose that $\tilde x\ne\hat x\in X$, so that $\hat x_m\ne 0$ for some $m\in\Bbb N$. Then for all $n\ge m$ we have

$$d_{\ell^1}\left(x^{(n)},\hat x\right)\ge\left|\frac1n-\hat x_m\right|\;,$$

and $$\left|\frac1n-\hat x_m\right|~\underset{n\to\infty}\longrightarrow~\left|\hat x_m\right|>0\;,$$

so the sequence does not converge to any non-zero sequence, either.

share|improve this answer
I dont get why $d_{l^1}(x^{(n)},\tilde{x})$ is what you say. Where are the terms $\tilde{x}_k$ left ? –  André Jan 4 '13 at 23:20
@André: I’m using your definition of $\tilde x$: $\tilde x_k=0$ for all $k\in\Bbb N$. –  Brian M. Scott Jan 4 '13 at 23:25
Yeah. Thanks. Now I get it. Thank you for your help :) –  André Jan 4 '13 at 23:26
@André: You’re welcome. –  Brian M. Scott Jan 4 '13 at 23:27

You are on the right track until near the end. Note that $|x^{(n)}_k-\tilde{x}_k|\leq d_{\ell^1}(x^{(n)},\tilde{x})$ for each $k$, so $x^{(n)}_k\to \tilde x_k$ for each $k$, which implies $\tilde x_k=0$ for each $k$. So what is $d_{\ell^1}(x^{(n)},\tilde x)$?

share|improve this answer
(And as Brian points out, it is enough to show that $x^{(n)}$ does not converge to the same $\tilde x$. Your approach, which I picked up here, shows that in fact the sequence doesn't converge at all in the $d_{\ell^1}$ distance.) –  Jonas Meyer Jan 4 '13 at 23:04

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.