1
$\begingroup$

This is a follow-up question on this. Let $A$ be a von Neumann algebra and $P$ be its projection lattice.

For $p,s,q \in P$, let us define $ p \perp q \mid s \iff ps^\perp q = 0$ where $s^\perp = 1-s$.

Let $P_0 \subset P$ be a complete sublattice of $P$. Is the following true over $P_0$ (i.e., for any $p,q,r,s \in P_0$): \begin{align*} r \perp p \mid (q \vee s), \quad r \perp q \mid (p \vee s) \implies r \perp (p\vee q) \mid s\;? \end{align*}

$\endgroup$
  • 1
    $\begingroup$ What conditions are you having in mind? You must formulate your question more clearly, because, being a mathematician, I would say the conditions are precisely those that you gave. $\endgroup$ – user1688 Feb 3 '16 at 7:32
  • $\begingroup$ @Anton, not sure what you mean. Let me ask this: Is the statement true for every complete sublattice $P_0$, or one needs to assume something more about $P_0$? $\endgroup$ – passerby51 Feb 3 '16 at 14:35
  • $\begingroup$ @Anton, OK, I understand what you mean. I just reworded it as a yes/no question. But my guess is the answer is no. I was hoping that there is a simpler condition that one can verify for $P_0$, from which mine follows, but perhaps not. In other words, I was hoping that someone recognize a connection with some other more well-understood condition. Otherwise, I will delete the question. $\endgroup$ – passerby51 Feb 3 '16 at 14:44
2
$\begingroup$

I think it is false - let $r=1$ and take any distinct rank $1$ projections $p,q,s\in M_2$ so $p\vee q=p\vee s=q\vee s=1$.

$\endgroup$
  • 2
    $\begingroup$ Or just $p = q = r = 1$ and $s = 0$. $\endgroup$ – Nik Weaver Feb 3 '16 at 16:36
  • $\begingroup$ @Tristan, how can $r=1$ and a rank 1 projection at the same time? $\endgroup$ – passerby51 Feb 3 '16 at 16:38
  • $\begingroup$ Yep, that'll do it! $\endgroup$ – Tristan Bice Feb 3 '16 at 16:39
  • $\begingroup$ Oops, I meant $s$ for the second $r$. In any case, Nik's answer is even simpler. $\endgroup$ – Tristan Bice Feb 3 '16 at 16:42
  • $\begingroup$ @Nik and Tristan, OK... fair enough. Maybe I should have mentioned distinct elements $p,q,r,s$, but Tristan's example seem to cover that. Though not sure if I understand that example correctly. Isn't $M_2$ 4-dimensional? Then, why $p \vee q = 1$ for example? $\endgroup$ – passerby51 Feb 3 '16 at 16:48

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.