I have some problems reading Pointwise convergence of Fourier series by Fefferman https://www.jstor.org/stable/1970917

When I proceed to Lemma 2, Chapter 6, I could not verify either of the following:

"Trivial estimates show that $|T_{p'}T_p^*f(x)|\leq \delta^{10}/|I'^*|\int_{E(p)|f(y)|dy}$, if $\mathrm{distance}(\omega,\omega')>\delta^{-\epsilon/2}|\omega|$, "

"$T_{p'}T_p^*f(x)=0$ if $I\not\subseteq I'^*$."

For the second, I could only show that this is the case if $I^*\cap I'^*=\varnothing$, but I think it just needs some modification (any suggestions?) to be true.

However, for the first one, the only place that involves $\delta$ is the distance condition. However, I simply do not know how to use that in the trivial estimate. I tried using the decay of the Fourier transform, but it turns out I could only obtain a term involving $\delta^{\epsilon/2}$.

Thank you for spending time reading my question. If you happened to have read the paper before, could you help me understand what he meant?


1 Answer 1


I have solved the problem by reading the following master's thesis:


The idea is that we could iterate the process for $\sim \epsilon^{-1}$ times, so as to bootstrap the bounds on the right hand side from the $\delta^{\frac \epsilon 2}$ to $\delta^{10}$. The implicit constant here may become incredibly large, but it does no harm, since in Lemma 2, we tolerate a loss of $\eta$ on the power of $\delta$.

However, there seems also a problem in the thesis when I proceed to Chapter 6, Lemma 5.

Below is a link to another problem I raised.

Fefferman's article: Pointwise convergence of Fourier series, II

Also, I would like to correct a typo in the statement: the $\delta^{\frac 1 2-\eta}$ should be replaced by $\delta^{\frac 1 4-\eta}$.


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.