# Schwartz space inequality

Let $g$ be a function in the Schwartz space $\mathscr S (\mathbb R)$. Show that for any $l \ge 0$, we have $\sup_x |x|^l |g(x-y)|\le A_l (1+|y|)^l$ by considering separately the cases $|x|\le 2|y|$ and $|x|\ge 2 |y|$.

The Schwartz space is defined as the set $\mathscr S (\mathbb R)$ of all indefinitely differentiable functions $f:\mathbb R\to \mathbb R$ such that $\sup_{x\in\mathbb R} |x|^k |f^{(l)}(x)|<\infty$ for all $k, l \ge 0$.

(This was used in a proof in Elias Stein's book on Fourier Analysis and is not a homework problem. The book just didn't go through this particular step.)

-
Should not it be Schwartz? – Giuseppe Jun 1 '11 at 10:54
$$(1+|x-y|)^{-1} \le \frac {1+|y|}{1+|x|}$$ – Piero D'Ancona Jun 1 '11 at 11:04
@Giuseppe: indeed it should (and I've corrected it). This is the opposite of what I often tell my students is the most common spelling mistake in mathematics. – Mark Meckes Jun 1 '11 at 14:26
@Piero D'Ancona: I'm mostly having trouble with finding where to use the two cases $|x|\ge 2|y|$ and $|x|\le 2|y|$. Any help? – Feynmaniac Jun 2 '11 at 6:48
@Feynmaniac: About an approach which doesn't separate the two cases $|x|\ge 2|y|$ and $|x|\le 2|y|$: couldn't it be possible the following proof? 1)use the triangle inequality $|x|\le |y|+|x-y|$, 2)apply the binomial formula to $(|y|+|x-y|)^l$, 3)apply the hypothesis on $g$, and finally $|y|^k\le(1+|y|)^l$, for any $y\in\mathbb{R}$ and $0\le k\le l$. – Giuseppe Jun 3 '11 at 7:40

We have $$\vert x\vert^l\vert g(x-y)\vert\le {(\vert x-y\vert+\vert y\vert)}^l\vert g(x-y)\vert \le (1+\vert y\vert)^l(1+\vert x-y\vert)^l\vert g(x-y)\vert$$ so that $$\vert x\vert^l\vert g(x-y)\vert\le (1+\vert y\vert)^l\underbrace{\sup_{z\in\mathbb R} (1+\vert z\vert)^l\vert g(z)\vert}_{A_l}$$ which is the sought inequality, where $A_l$ in a semi-norm of $g$ in the Schwartz space. Note that no differentiability property for $g$ is necessary, we have used only fast decay of the function itself.