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

Let S be a set of integers and denote the characteristic function of S as $\chi_{S}(n)$. Define an operator on the space of trig functions by the relation $\hat{Tf}(n) = \chi_{S}(n) \hat{f}(n)$. Here $\hat{f}(n)$ is the n-th Fourier coefficient of f.

For $p\geq 2$ we'll call S a $L^p$ multiplier set (or just $L^p$ multiplier) if there is an inequality of the form $\Vert Tf\Vert\_{p} \leq c \Vert f\Vert\_p$. If this inequality holds for some p but fails for $p+\epsilon$ for every $\epsilon>0$, we'll say that S is a strict $L^p$ multiplier.

Note that every set is a $L^2$ multiplier and if S is a $L^p$ multiplier for some p then it is a $L^q$ multiplier for $2 \leq q \leq p$. Moreover, it follows from a result of Zygmund that almost every (in the obvious sense) set is a strict $L^2$ multiplier. (I also think you can prove this via Khintchine's inequality, but I haven't checked this argument.)

Do strict $L^p$ multiplier sets exist for every $p>2$? Note that this is similar to the $\Lambda(p)$ problem, however, I don't see how to transform a strict $\Lambda(p)$ set into a strict $L^p$ multiplier set.

share|cite|improve this question
up vote 6 down vote accepted

Let M_p be the class of L^p muliplier sets, as considered in this equation. It is known:

  1. This is an algebra of sets, but not a sigma-algebra. [Not sure of the reference.]

  2. The inclusion M_p \subsetneq M_q is strict for 2\le p

MR1728363 (2001g:42013) Mockenhaupt, Gerd(5-NSW-SM); Ricker, Werner J.(5-NSW-SM) Idempotent multipliers for $L^p(\bf R)$.

share|cite|improve this answer
Thanks! It turns out that the proof follows easily from Bourgain's \Lambda(p) set construction. Let E be a strict \Lambda(p) set. Clearly E is a is a L^{p} multiplier, since ||Tf||_{p} << ||f||_{2} << ||f||_{p}. Conversely, if E was a L^{p+\epsilon} multiplier, we can show it must be a \Labmda(p+\epsilon') set by interpolating the estimates ||Tf||_{p} << ||f||_{2} with ||Tf||_{p+\epsilon} << ||f||_{p+\epsilon}. It was this last step that I was missing. – Mark Lewko Oct 31 '09 at 3:25

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.