Let $E, F$ be locally convex $\mathbb{K}$ vector spaces and $T: E\to F$ linear and continuous. Let $T': F'\to E'$ be defined through $T'(g)=g\circ T$

$T': F'\to E'$ is weak-$\ast$-continuous (hence continuous when $F'$ and $E'$ are provided with the weak-$\ast$ topology)

This question is related to $T: E\to F$ weak continuous, $E, F$ locally convex and unfortunatly I do not know where to start with the proof. I struggle with the definition of the weak topology and the weak-$\ast$-topology.

Can you help me, or give me a hint. Thanks in advance.


Assume $g_{\delta}\rightarrow g$ in $w^{\ast}$ of $F'$, and let $x\in E$, then $T(x)\in F$ and hence $|T'(g_{\delta})(x)-T'(g)(x)|=|g_{\delta}\circ T(x)-g\circ T(x)|=|g_{\delta}(T(x))-g(T(x))|\rightarrow 0$.

  • $\begingroup$ This might be a dumb question, but why is it enough to show this? And why holds $|g_\delta(T(x))-g(T(x))|\to 0$ $\endgroup$ – Cornman Nov 28 '17 at 18:31
  • 1
    $\begingroup$ The point is to show that the seminorm $p_{x}(T'(g_{\delta})-T'(g))\rightarrow 0$ (actually it is sufficient to show at the zero, that is, $g=0$), here $p_{x}(\eta)=|\eta(x)|$, $\eta\in E'$. Why that holds? Because $g_{\delta}$ converges to $g$ $w^{\ast}$ of $F'$, that means, for any $y\in F$, $p_{y}(g_{\delta}-g)=|g_{\delta}(y)-g(y)|\rightarrow 0$. $\endgroup$ – user284331 Nov 28 '17 at 18:33
  • $\begingroup$ Is $p_x$ a special seminorm, or just a seminorm? I could also assume $g_\delta\to 0$ in $w^\ast$, since it is sufficient to show it at zero. Hence $|T'(g_\delta(x)-T'(0)(x)|=|g_\delta(T(x))-0|\to 0$ since $g_\delta\to 0$ $\endgroup$ – Cornman Nov 28 '17 at 18:43
  • $\begingroup$ Yes, as I said it suffices to do for $0$. The weak star topology is generated by the bunch of seminorms $p_{x}(\eta)=|\eta(x)|$, and the convergence must satisfy all these seminorms, that is, for any fixed $x$, one must have $p_{x}(\eta_{\delta})=|\eta_{\delta}(x)|\rightarrow 0$. $\endgroup$ – user284331 Nov 28 '17 at 18:45
  • $\begingroup$ Do you have a hint on the linked question too? I would love to figue that out mostly on my own, but I am stuck. :( $\endgroup$ – Cornman Nov 28 '17 at 18:49

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.