Skip to main content
3 of 4
added top-level tag; http://meta.mathoverflow.net/questions/1457/why-are-mo-tags-formatted-as-they-are
Martin Sleziak
  • 4.7k
  • 4
  • 35
  • 40

A question on ultraproducts of $L_{p}(\mu)$-spaces

Let $I$ be a set, $\mathcal{U}$ be an ultrafilter on $I$ and $1\leq p<\infty$. Let $X_{i}=L_{p}(\mu_{i})(i\in I)$, where $\mu_{i}$ is a probability measure for each $i\in I$. Relying on standard results from Banach lattice theory, we can prove that the ultraproduct $(X_{i})_{\mathcal{U}}$ of $(X_{i})_{i\in I}$ is isometrically isomorphic to $L_{p}(\mu)$ for some measure $\mu$.

Question 1: Can the above measure $\mu$ be taken to be finite or even a probability measure?

Question 2: Is there a direct construction of the measure $\mu$ without relying on Banach lattice theory?

Thank you!

Dongyang Chen
  • 3.3k
  • 1
  • 13
  • 16