4

I’m currently reading Alexer’s linear algebra done right. He proved in the book that $\dim \mathcal{L}(V,W) = \dim V \dim W$ holds if both $V$ and $W$ have finite dimension. I’m wondering if this identity still holds when one of $V$ and $W$ is infinite dimensional (dimension is defined as the cardinality of its Hamel bases).

Obviously the proof for finite dimensional cases does not work in the infinite dimensional case, since the standard basis that sends $v_i$ to $w_j$ and $v_k$ to $0$ $(\forall k \neq i)$where $i < \dim V$ and $j < \dim W$ (for all $i, j$) does not span $\dim \mathcal{L}(V,W)$, as linear combinations must be finite. After some searching I found this question which seems to be a counterexample to the identity, since the identity implies $\dim V^* = \dim V\dim F = \dim V$. But I still would like to know when exactly does this identity fail, like does it hold exactly for finite dimensional spaces or does it still hold for some infinite dimensional $V$ or $W$?

Jyrki Lahtonen
  • 140,891
Poscat
  • 175
  • 1
    First of all: have you looked around on this site? The question has probably been asked before. Please check this for a few minutes. Anyway: If V is infinite-dimensional and W is non-trivial, the equation breaks down. It holds when W is zero (clear) or if V is finite-dimensional: easy exercise using $\mathrm{hom}(K^n,W) \cong W^n$. – Martin Brandenburg Oct 30 '23 at 12:19
  • What search input should I use? Since the core of my question is the identity, I tried to use https://approch0.xyz to search for questions containing the identity but couldn't find one that completely and satisfyingly covers my question. – Poscat Oct 30 '23 at 14:55
  • https://math.stackexchange.com/questions/3160423 already covers most of your question, and I didn't need to search at all, it is displayed in the (AI-generated?) list of related questions to the right. – Martin Brandenburg Oct 30 '23 at 15:32
  • Searching for "Hom(V,W) infinite-dimensional" brings up https://math.stackexchange.com/questions/405584 where Tom's answer pretty much answers everything. – Martin Brandenburg Oct 30 '23 at 15:34
  • And https://math.stackexchange.com/questions/2648820 covers the special case $W=K$ (that you already found on MO, but it's here as well). – Martin Brandenburg Oct 30 '23 at 15:38

0 Answers0