I am reading Lindenstrauss' Classical Banach Spaces II and I am having trouble with the following characterization of integrals. First a couple of preliminary definitions:

Let $(\Omega, \Sigma, \mu)$ a complete $\sigma$-finite space. A Banach space $X$ consisting of equivalence classes, modulo equality almost everywhere, of locally integrable real valued functions on $\Sigma$ is called a Köhte function space if the following conditions hold:

  1. If $\vert f (\omega) \vert \le \vert g(\omega) \vert $ a.e. on $\Omega$ with $f$ measurable and $g \in X$, then $f \in X$ and $\Vert > f \Vert \le \Vert g \Vert$.
  2. For every $\sigma \in \Sigma$ with $\mu ( \sigma ) < \infty$ the characteristic function $\chi_\sigma$ of $\sigma$ belongs to X.

Also we have that:

Every measurable function $g$ on $\Omega$ such that $gf \in L_1(\mu)$ for every $f \in X$ defines an element $x_g^* \in X^*$ by $x_g^*(f) = \int_\Omega > f(\omega) g(\omega) d\mu$. Any functional of the form $x_g^*$ is called an integral.

Then the book states (this is the part that I haven't been able to work around):

It is an immediate consequence of the Radon-Nikodym theorem that a functional $x^* \in X^*$ is an integral is and only if, for every sequence $\{ f_n \} \in X$ with $f_n ( \omega) \downarrow 0$ a.e. we have that $\vert x^* \vert (f_n) \to 0$.

I don't understand how exactly this follows from the Radon-Nikodym theorem. Could anyone clarify it to me?

  • $\begingroup$ When you say $gf\in L^1$ you mean that $gf\in L^1$ for all $f\in X$, right? $\endgroup$ – David C. Ullrich Apr 28 '16 at 15:08
  • $\begingroup$ @DavidC.Ullrich Yes, I forgot to add that. I have edited the post to reflect it. $\endgroup$ – A. A. Apr 28 '16 at 15:09
  • $\begingroup$ Also what does $|x^*|(f)$ mean? Is that $|x^*(f)|$ or something else? $\endgroup$ – David C. Ullrich Apr 28 '16 at 15:09
  • $\begingroup$ The reason I suspect it may be something else is that if $\nu$ is a complex measure then $|\nu|(E)\ne|\nu(E)|$. Also if it just meant $|x^*(f)|$ it's hard to see why he'd include the absolute value sign when he says $|x^*|(f_n)\to0$. So: What is $|x^*|(f)$? $\endgroup$ – David C. Ullrich Apr 28 '16 at 15:21

I haven't worked out the details - also I suspect we still haven't been given all the relevant definitions. But in case it helps, here's how the argument "must" go in outline, if it's by R-N:

Somehow we reduce to the case $\mu(\Omega)<\infty$. Define a complex measure $\nu$ by $$\nu(E)=x^*(\chi_E).$$Detail: Something shows somehow that $\nu$ is in fact countably additive. Now $$x^*\phi=\int\phi\,d\nu$$for integrable simple functions $\phi$. Detail: Somehow, maybe simple functions are dense, this extends to $\phi\in X$.

And now somehow that last condition shows somehow that $\nu<<\mu$.


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.