1
$\begingroup$

Let $H$ be a Hilbert space and $B(H)$ denotes the space of all continuous linear operators on $H$. I am looking for a class/example of bounded linear functionals $B(H)\to \mathbb C$ which cannot be reduced to the type $$T\mapsto\sum_{i=1}^\infty k_i\left<Tx_i,y_i\right>$$ for some fixed $x_i$'s and $y_i$'s in $H$ and $k_i$'s in $\mathbb C$. Do we have any explicit examples?

$\endgroup$

1 Answer 1

9
$\begingroup$

So, you are asking about non-normal functionals on $B(H)$. This is very similar to the question of what are the functionals on $\ell_\infty$ that are not in $\ell_1$?

Fix an ultrafilter $U$ on $\mathbb N$ and let $(e_n)$ be an ONB for $H$. Define

$$\langle \phi, T\rangle = \lim_{n,U} \langle Te_n, e_n\rangle\quad (T\in B(H)).$$

Then this functional is not normal.

$\endgroup$
5
  • $\begingroup$ Thank you for answering. Can you suggest some literature where I could find these results along with proofs? $\endgroup$
    – NewB
    Sep 29, 2020 at 22:14
  • 2
    $\begingroup$ @NewB, I'd go for Takesaki, Theory of operator algebras, vol. 1. $\endgroup$ Sep 30, 2020 at 6:06
  • $\begingroup$ As far as I could understand, you gave a recipe for linear functionals on B(H) which are not coming through the natural map $J:H^*\otimes^\gamma H\to B(H)^*$, since projective tensor product $H^*\otimes^\gamma H$ is nothing but trace class operators, the predual of $B(H)$. Does this work for any Banach space projective tensor product? I mean, can we get linear functionals on $(X\otimes^\gamma Y)^{*}$ which are not coming through J map, through the same recipe? $\endgroup$
    – NewB
    Sep 30, 2020 at 15:47
  • $\begingroup$ Also, do you mean non-principle ultrafilter? Because for principle ultrafilter generated by $k$, the functional defined above is $T\mapsto \left<Te_k,e_k\right>$, which is again normal. Am I missing here something? $\endgroup$
    – NewB
    Oct 1, 2020 at 20:24
  • $\begingroup$ @NewB, of course non-principal. $\endgroup$ Oct 2, 2020 at 7:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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