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

The following question came up during a reading of Rudin's functional analysis. I have not been able to find any information through searching online, but I apologise if the answer is obvious, or the question is trivial.

Consider $\Phi$ to be the space of sequences that have finitely many non-zero terms. The space is not closed in $\ell_1$, therefore $\ell_1/\Phi$ with the quotient topology is not Hausdorff, and so it cannot be metrizable. However, does there exist a metric on $\ell_1/\Phi$ that gives rise to a non-trivial topology? Furthermore, is $\ell_1/\Phi$ normable?

share|cite|improve this question
You should explain what you mean by non-trivial. If you do not specify a way the topology is to be related to the topology of $\ell^1$ and to the quotient map, then of course there are a lot of non-trivial toplogies on this quotient set -- just find a interesting topological space with the same cardinal. – Benoît Kloeckner Sep 9 '11 at 19:40
What Benoit said. Do you want this non-trivial metric-induced topology to be weaker than the quotient one? stronger? – Yemon Choi Sep 9 '11 at 20:10
Crossposted to MSE:… where it has received answers. I therefore vote to close here – Yemon Choi Sep 28 '11 at 21:40

Thank you for your answers, and please excuse my lack of proficiency in the subject. With that been said, I am still confused how there could be such choice in topology. For example, how can there be a weaker-than-the-quotient metrizable topology, wouldn't this make the quotient topology Hausdorff? Further, is there a metrizable topology for any cardinal number space which is not the trivial one or the topology of all subsets?

What about the existence of a norm on the space, regardless of the quotient topology? Is it constructable?

share|cite|improve this answer
Please add your comments to the original question, rather than adding them as an "answer" – Yemon Choi Sep 9 '11 at 23:22
It is hard to answer your further questions without you making more precise what you mean by "a metric on $\ell^1/\Phi$". You seem to be assuming that such a metric is related in some way to the given quotient topology, but you do not say so explicitly. – Yemon Choi Sep 9 '11 at 23:23
Thank you again for your response. I now realize that the original question was very poorly defined. What I was trying to do originally was to see if a seminorm exists on \ell_1 that vanishes on \Phi and therefore gives rise to a norm on the \ell_1/\Phi. Such a norm cannot be continuous with respect to the \ell_1 norm, but the point was to see if it would be possible to explicitly construct a norm on the quotient space(and not use an axiom of choice type of argument). Asking about metrizability was pointless. To the moderators, I apologize for posting a comment earlier as an answer. – Ivan Sep 10 '11 at 0:44
Ivan: thanks for clarifying. I suggest that you edit your original question, and add this more precise version to it - people don't always read through the comments. – Yemon Choi Sep 10 '11 at 1:21
Ivan, you may need to register an account and have it merged with the unregistered one you used post the question. Just flag your post for moderator attention when you've registered. – j.c. Sep 29 '11 at 3:58

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.