Let $(x_n)_n$ be a positive, disjoint, weakly null sequence in a Banach lattice $E$. If $(y_n)_n$ is a sequence such that $0\leq y_n\leq x_n$ for every $n\in \mathbb{N}$, we can garantee that $y_n$ is weakly null?


"Weakly null" means that $x_n$ converges to zero in the weak topology.

"disjoint" means that $\inf\{|x_n|,|x_m|\}=0 ~ \forall n\neq m$.

"positive" means that $x_n\geq 0$ for all $n\in\mathbb{N}$.

Thanks for any explanations.

  • $\begingroup$ I am not sure if I understand disjoint correctly. Wouldnt your definition be the same as $\inf_n \{|x_n|\}=0$? $\endgroup$ – supinf Mar 27 at 13:03
  • $\begingroup$ No. Disjoint is stronger. For example, the sequence $(x_n)_n:=(1,1,0,0,0,\cdots)$ in $\mathbb{R}$ haves $\inf_n|x_n|=0$, but $\inf\{|x_1|,|x_2|\}=1\neq 0$. $\endgroup$ – L26 Mar 27 at 13:24
  • $\begingroup$ Then your notation is incorrect. $\inf_{n\neq m}$ means you minimize over all $n,m$. Maybe you want to say is that $\inf \{|x_n|,|x_m|\}=0\forall n\neq m$? $\endgroup$ – supinf Mar 27 at 13:58
  • $\begingroup$ It's true. My notation is incorrect. Thanks! $\endgroup$ – L26 Mar 27 at 15:10

If $E^*$ allows for a decomposition of functionals into positive and negative parts then this can be proven. This is true at least if $E$ is a Hilbert lattice or equal to some $L^p$ with the natural ordering ($p\in [1,+\infty)$).

Let $f\in E^*$. Then there are $f^+,f^-$ with $f=f^++f^-$ with $f^+\ge0$, $f^-\le0$. Then it follows $0\le f^+(y_n) \le f^+(x_n)$ and $0\le -f^-(y_n) \le -f^-(x_n)$. This implies $y_n\rightharpoonup0$.

  • $\begingroup$ what is an example where this assumption is not true? $\endgroup$ – supinf Mar 27 at 14:31
  • $\begingroup$ Thanks! I think that your explanation holds for every Banach lattice, because the norm dual of a Banach lattice $E$ coincide with its order dual. And the order dual is a Dedekind complete Riesz space, so it is possible write $f\in E'$ as $f=f^++f^-$. $\endgroup$ – L26 Mar 27 at 15:26

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.