MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

My question comes from trying to understand a technical step in this paper by Bourgain.

Let $R,L$ be positive integers and let $f(x)=\sum_{|n|\leq RL}a_ne^{2\pi inx}$ be a trigonometric polynomial. Assume $f(x)>0$. Let $F_L(x)=\sum_{|n|\leq L}\frac{L-|n|}Le^{2\pi inx}$ be the Fejer kernel. Define (as usual) the convolution $$(f*F_L)(x)=\int_0^1f(t)F_L(x-t)dt=\sum_{|n|\leq L}a_n\frac{L-|n|}Le^{2\pi inx}$$ Does it follow (and how) that $$f(x)\leq10R(f*F_L)(x)$$?

In the paper we have a concrete $f(x)$, so maybe this is not true in general. There we have $$a_n=1-\cos\left(2\pi\frac{RL-|n|}N\right)$$ where $N$ is a large positive integer, (at least bigger than $4RL$).

share|cite|improve this question
Don't we have a long list of MO questions that begin with "My question comes from trying to understand a technical step in this paper by Bourgain"? – Bill Johnson Feb 14 '13 at 2:43
Can you provide the links, Bill? – fedja Feb 14 '13 at 6:27
Here is one:… – user21162 Feb 14 '13 at 6:46
Yeah, but that one was due to a misreading and the author had the privilege to get an explanation from Terry Tao himself, so I wouldn't complain too much if I were in his shoes (LOL). Of course, as stated in the post, the statement was just patently false. You can put almost all energy to any given sufficiently large annulus $R<|\xi|<2R$ you want (just take your favorite bounded compactly supported function $f$ and move the main bulk of $\widehat f$ anywhere you want by adjusting the phase). – fedja Feb 14 '13 at 11:23
up vote 13 down vote accepted

That is true for all non-negative trigonometric polynomials, though not entirely obvious unless you are a Fourier analyst yourself. To see it, just note that the convolution with $K_{RL}=2F_{2RL}-F_{RL}$ recovers $f$ faithfully and $F_{RL}\le RF_L$. Of course, to Jean such things are as obvious as $2\times 2+1=5$ (he writes $10$ instead of $5$ just out of the traditional analyst's habit to have a 100% security margin in the constants), but I agree that it may be a bit perplexing for poor mortals like you and me. Joe Diestel just told me at a beer party tonight that the most common phrase in Bourgain's early papers was "By standard techniques we conclude from here that". :)

share|cite|improve this answer
I was trying to be humorous with my other comment, fedja, but Joe is correct about Jean's early papers. You had an easier time reading his papers in harmonic analysis as by then he put in a few more details for us mortals. – Bill Johnson Feb 14 '13 at 16:16
Thanks! It still took me some time to work with these hints but I got it now. Just to make sure I understood it, I think we actually get $$f=K_{RL}*f\leq2F_{2RL}*f\leq4RF_l*f$$ so with a $4$ and not $5$, right? It's good to know that I'm not alone having a hard time understanding Bourgain's papers... – Joel Moreira Feb 14 '13 at 18:53
@Joel Yes, it is 4, but the other traditional analyst's habit is to make sure that an occasional wrong sign in an identity does not invalidate the argument :). @Bill Yeah, my first exposure to Jean's writing was reading his opus on the existence of $\Lambda(p)$ sets for $2<p<4$ in French (surprisingly, that one was written well and it took me just a couple of months to figure out what was going on there :)). – fedja Feb 14 '13 at 19:00
@Fedja, perhaps you were reading Queffélec's (more detailed) french exposition, as the original paper is in english :) (although, maybe there was some earlier preprint in circulation?) – Mark Lewko Feb 14 '13 at 23:11

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.