This problem arose while I was trying to solve exercise 6.41 in Folland's real analysis.

Suppose $1<p\leq \infty$ and let $q$ be the conjugate exponent of $p$ (i.e. $p^{-1}+q^{-1}=1$). If $T$ is a bounded operator in $L^p$ such that $$ \int (Tf) g = \int f(Tg) \qquad \forall f,g\in L^P\cap L^q $$ then can one show that $T$ is linear?

I was able to show that for any functions $f_1, f_2$ on $L^p\cap L^q$ and any $g\in L^p \cap L^q$ we have
$$ \int (T[f_1+f_2]) g = \int (Tf_1 + Tf_2)g $$ I am not sure how to proceed from here. Any hints are appreciated.

My ultimate goal is to extend $T$ to a bounded operator on $L^r$ where $r$ is some number between $p$ and $q$. If there is another approach which does not require to show that $T$ is linear that would also solve my problem as well.

Thanks in advance.


It seems to me like you are pretty much done. Indeed, if you can show that $$\int (T[f_1 + \alpha f_2])g = \int (Tf_1 + \alpha Tf_2)g$$ for all $g \in L^p \cap L^q$, then by boundedness of $T$ and density of $L^p \cap L^q$ in $L^q$, you have $$\int (T[f_1 + \alpha f_2])g = \int (Tf_1 + \alpha Tf_2)g, \,\,\, \forall g \in L^q.$$ But then $h \in L^p$ defined by $h = T[f_1 + \alpha f_2] - Tf_1 - \alpha Tf_2$ satisfies $$\int hg = 0, \,\,\, \forall g \in L^q$$ and so $$\| h \|_p = \sup_{g \in L^q, \|g \|_q \le 1} \left \lvert \int hg \right\rvert = 0$$ which shows that $h = 0$ and thus $$T[f_1 + \alpha f_2] = Tf_1 + \alpha Tf_2,$$ so $T$ is linear.

  • $\begingroup$ So I have to assume that $T$ is a bounded operator from $L^p$ to $L^p$? (and not just from $L^p$). Otherwise, I don't know how to go from $g\in L^q\cap L^p$ to $g\in L^q$ $\endgroup$ – Quoka Mar 6 '18 at 20:34
  • $\begingroup$ I'm not exactly sure I understand what you're asking. I thought you had already assumed that $T$ was bounded from $L^p \to L^p$; otherwise what does "$T$ is a bounded operator in $L^p$" mean? $\endgroup$ – User8128 Mar 6 '18 at 20:41
  • $\begingroup$ The author never specified. I think this assumption is indeed required, I was wondering if maybe I was missing something.Thanks! $\endgroup$ – Quoka Mar 6 '18 at 20:45

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.