Let $f:[0,1]\to \ell_\infty[0,1]$ be defined by $f(t)=\chi_{[0,t]}$. Is it true that $f$ is weakly continuous almost everywhere w.r.t. Lebesgue measure ? Here $\ell_\infty[0,1]$ represents the function space $L_\infty(\Omega,\Sigma,\mu)$ where $\Omega=[0,1], \Sigma=\mathcal{P}([0,1])$ and $\mu$ be the counting measure on $[0,1]$.
2 Answers
"Weakly continuous almost everywhere" means there is a negligible set of $t$'s outside which $t\mapsto x^*(f(t))$ is continuous for all $x^*\in\ell_\infty^*$, not the other way round. But $\delta_s(f(t))$ depends continuously on $t$ everywhere except when $t$ crosses $s$ : no common negligible set is suitable for all $x^*\in\ell_\infty^*$.
What Tanmoy Paul proves is: for all $x^*\in\ell_\infty^*$, $t\mapsto x^*(f(t))$ is continuous almost everywhere. That's true, but doesn't answer the question as stated.
This function is weakly continuous almost everywhere. One can show that $f$ is Riemann Int. over $[0,1]$. Now if the value of this integral is $x$, for some $x\in \ell_\infty[0,1]$, then for any $x^*\in \ell_\infty[0,1]^*$ $x^*(x)=\int_{[0,1]}x^*f(t)d\lambda(t)$. Since $f$ is Riemann Int. the integral in RHS is scalar valued Riemann Int. Thus $x^*f$ is Riemann Int which leads to that $x^*f$ is continuous almost everywhere.
-
$\begingroup$ Please clarify the Riemann Integrability of a vector valued function. Hope $\lambda$ is the Lebesgue measure. $\endgroup$ Commented Oct 18, 2015 at 4:19
-
3$\begingroup$ $\ell^{\infty}[0,1]$ as described by the OP looks to me like bounded functions on $[0,1]$, with sup norm. So $x\mapsto x(s)$ is in the dual, and $f$ is weakly discontinuous everywhere. $\endgroup$ Commented Oct 18, 2015 at 4:42
-
1$\begingroup$ @Christian Remling That's an answer, isn't it? (albeit to a trivial question). Why post answers as comments?? $\endgroup$ Commented Oct 18, 2015 at 10:13
-
1$\begingroup$ @Jean Duchon: I think people here prefer not to follow up elementary or trivial questions, in order to keep the topics within research level. Answering in comments is intended to stop the activity showing that the question is indeed trivial. $\endgroup$ Commented Oct 18, 2015 at 17:46