2
$\begingroup$

Let $V$ be a complex vector space and let $\{0\} \subset V_1 \subset V_2 \subset \dots \subset V$ be an increasing sequences of subspaces of $V$, whose union is $V$. Suppose that each $V_n$ is a locally convex Hausdorff topological vector space and that every inclusion $V_{n} \subset V_{n+ 1}$ is continuous. We can then endow $V$ with two topologies:

  1. The finest locally convex vector space topology, making all inclusions $V_n \subset V$ continuous (which can be shown to exist).
  2. The finest topology on $V$ which makes makes all inclusion maps $V_n \subset V$ continuous (a general, point-set topology construction that works for all topological spaces).

Both of these topologies have universal properties characterizing continuous maps out of $V$: In the first case, this universal property holds within the category of locally convex spaces, while in the second, it holds in the category of all topological spaces.

My questions are:

  1. Is there an example in which the second topology does even yield continuous vector space operations on $V$?
  2. Does there exist an example in which the sceond topology does yield continuous vector spaces operations on $V$ but the two topologies are nonetheless different?
$\endgroup$
1
$\begingroup$

I think there would be problems with the "second" topology, as topological vector spaces do need to be Hausdorff and have continuous operations... I have not thought about that more general issue, but I have thought about the contrast between colimits and coproducts in the category of locally convex TVS's, and colimits and coproducts in the larger category of not-necessarily-locally-convex TVS's. (As expected, colimits are quotients of coproducts.)

Arbitrary locally convex TVS colimits exist, but uncountable not-necessarily-locally-convex TVS coproducts (and corresponding colimits) _do_not_exist_. E.g., an uncountable coproduct of copies of \C does not exist as a TVS. :)

One proof of the latter uses the non-local-convexity of the spaces $L^p$ with $0<p<1$ (defined slightly differently from the case of $p\ge 1$).

(This is http://www.math.umn.edu/~garrett/m/fun/uncountable_coproducts.pdf)

$\endgroup$

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.