MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $B=\int_{Y}^{\oplus}B_yd_y$ be the direct integral decomposition of vNa $B$ into factors and if $P=\int_{Y}^{\oplus} p_ydy$ is a projection in $B$ and $p_y$ is equivalent to a projection $q_y$ in $B_y$ for all or a.e. $y\in Y$. Is there a good choice of $q_y$ up to equivalence such that $Q=\int_{Y}^{\oplus} q_ydy$ makes sense and $P$ and $Q$ are equivalent in $B$?

share|cite|improve this question
I don't think the question makes sense as is. Why not choose $q_y = p_y$? – Jesse Peterson Oct 17 '11 at 21:40
@Jesse: I presume that when the OP says “good choice of $q_y$ up to equivalence” he means the choice of a representative of the equivalence class of a section with respect to sets of measure 0. – Dmitri Pavlov Oct 18 '11 at 7:47
That's certainly one interpretation. I wonder if the OP would mind clarifying what is meant by "good", "choice" (since the $q_y$ are already defined), and "up to equivalence"? – Jesse Peterson Oct 18 '11 at 12:46
For each $y\in Y$ we can find a $q_y$ which is equivalent to $p_y$ (we don't know whether $q_y$ equals $p_y$ or not, let's say they are different), but $y\mapsto q_y$ is not necessarily measurable in the sense that $Q=\int_Y^{\oplus}q_ydy$ makes sense. So I want to find $q'_y$ which is equivalent to $q_y$ in $B_y$ and $y\mapsto q'_y$ is measurable. I need a kind of measurable selection theorem. Thanks for your attention :) – m07kl Oct 18 '11 at 15:18
@m07kl: So what is wrong with choosing $q_y' = p_y$? – Jesse Peterson Oct 18 '11 at 15:35
up vote 3 down vote accepted

Suppose B is a trivial bundle whose fibers are type I2 factors and p is a constant section of B corresponding to some projection with 1-dimensional image. Projections with 1-dimensional image in a type I2 factor can be identified with angles, i.e., elements of iR/πiZ. If q is given by some non-measurable section of B, then there is no way to choose its equivalence class in such a way that Q exists.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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