I'm doing some exercises, but there is something in one of the questions I don't quite understand, so I'm really hoping someone might be able to clarify. The question is as follows: Let $\mathcal H$ be a Hilbert Space with inner product $\lt\cdot,\cdot\gt$, and corresponding norm $\|\cdot\|$. Furthermore there is the set $B=\{x\in\mathcal H |\|x\|\le1\}$.
a) Show, that B is closed and bounded$^1$ regarding (the metric corresponding to) $\|x\|$.
b) Show, that if $\mathcal H$ doesn't have a finite dimension, then B isn't compact regarding (the metric corresponding to) $\|x\|$. (Hint: Look at a countable orthonormal system $\{e_n|n\in\mathbb N\}$ in $\mathcal H$.
c) Give an example of a Hilbert Space $\mathcal H$, where B isn't compact.

What I don't understand is, if I have to show in a) that B is compact, how can it suddenly be not-compact in b)? Where do we assume that $\mathcal H$ has a finite dimension in a)? And finally, in c) - haven't we just, in b) given an example - the example being all Hilbert Spaces which does not have a finite dimension?

$^1$ This is the same as being compact, right?


Closed and bounded does not imply compact - but compact implies closed and bounded. (I've made the same mistake many times.)

a) Of course if you could prove that $B$ is compact then you'd achieve this...but this claim is false!

b) So up to here is consistent. You don't assume $\mathcal{H}$ has finite dimension in $A$, since again closed and bounded is not the same as compact.

c) Sure, that's a valid example. Maybe you want to try considering the question for finite dimensions?

  • 1
    $\begingroup$ a) If B was always compact then you'd run into problems with compact operators, because the identity is never a compact operator. $\endgroup$ – Gyu Eun Lee Dec 10 '12 at 4:22

a) No, this is not in general true. For example, if you give $\mathbb{Z}$ the discrete metric, the whole space is closed and bounded but not compact. I think you can actually solve part a) by using the continuity of $\|\cdot\|$ and the fact that you're ball is the preimage of $[0,1]$.

b) You should try as hard as you can. If you need to, you can peek here Is any closed ball non-compact in an infinite dimensional space?.

c) Yes, you are correct. Take your pick, $L^2$ for example.

  • $\begingroup$ @JonasMeyer You are of course correct--my head must have been elsewhere (I was thinking of Banach spaces). $\endgroup$ – Alex Youcis Dec 10 '12 at 4:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.