Terminology Clarification: ``Natural Topology on $\mathscr{L}\big(V,V'\big)$''

73 Views Asked by At

I'm reading Karoubi's $K$-Theory, and am a bit uncertain of what he's calling a natural topology:

Let $X$ be a topological space, and let $V,V'$ be vector spaces over a field $k$. Let $X\times V,X\times V'$ be trivial vector bundles, and let $\mathscr{L}\big(V,V'\big)$ denote the space of linear functions $g:V\rightarrow V'$.

Karoubi gives a theorem that a certain map $\overline{g}:X\rightarrow \mathscr{L}\big(V,V'\big)$ is continuous with respect to the natural topology on $\mathscr{L}\big(V,V'\big)$.

Does he mean the compact-open topology? If so, it was never required that $V$ be a topological vector space, or even that $k$ be a topological field. I'm aware that the compact-open topology can be fashioned so long as there is a topology on $k$, but this was never required. Is he making these assumptions implicitly, or is there an alternative topology that he is referring to?

1

There are 1 best solutions below

2
On BEST ANSWER

As Karoubi states at the very start of the first chapter, $k$ is restricted to be either $\mathbb{R}$ or $\mathbb{C}$. Note that the natural topology on $\mathscr{L}(V,V')$ is just its unique topological vector space topology (the unique topology that makes any linear isomorphism $k^N\to\mathscr{L}(V,V')$ a homeomorphism; note that in this context $V$ and $V'$ are assumed to be finite-dimensional). This turns out to coincide with the compact-open topology but that is a nontrivial fact and usually the topology is not thought of as the compact-open topology.