If we have a compact subset $ A $ of a topological vector space, is it necessary for $ A $ to be bounded? If not, can we put any simple conditions on the topological vector space to make boundedness a necessary condition for compactness?

By boundedness, I mean the following definition: We say that a subset $ A $ of a topological vector space $ V $ is bounded iff for every open neighbourhood U of $ 0 $, there is a scalar $ x $ such that for every $ y > x $, $ A \subset yU $.


Any compact metric space is bounded as the product $A\times A$ is compact and the metric is a real-valued function on $A\times A$, which must have a maximum and a minimum. For another possible intepretation of bounded, the distance function from the origin is a real valued function on $A$ and therefore has a maximum and a minimum.

For your edit, note that for any open neighborhood $U$ of $0$, the neighborhoods $yU$ for $y\in \mathbb{R}$ cover $A$. Thus $A$ is covered by finitely many such neighborhoods by compactness. In particular there is a $x$ such that $A\subseteq xU$, and therefore $A\subseteq yU$ for all $y\geq x$.

  • $\begingroup$ I've edited my question to clarify what I mean by boundedness. This is a notion independent of the metric on the vector space, even if such a metric exists. $\endgroup$ – Ormi Mar 22 '15 at 16:22
  • $\begingroup$ @Ormi Please see my edit. $\endgroup$ – Matt Samuel Mar 22 '15 at 16:25
  • 1
    $\begingroup$ @MattSamuel don't you need to argue that $U$ may be assumed balanced w.l.o.g.? Otherwise, how do you deduce $A \subseteq yU$ from $A \subseteq xU$? $\endgroup$ – Tom Hirschowitz Aug 14 '18 at 20:43

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.