Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Is it possible to have a distribution $u\in \mathcal{D}'(\mathbb{R})$ such that the restriction of $u$ to $(0,\infty)$ is $\frac{1}{x}$ and the restriction of $u$ to $(-\infty,0)$ is $0$?

It seems to me that the answer should be 'no', but I couldn't figure how to prove it.

share|cite|improve this question

It is possible: define $u\in\mathcal D'(\mathbb R)$ by $u(\phi)=\int_0^\infty \frac 1 x (\phi(x)-\phi(0)e^{-x}) dx$.

Note that for any test function $\phi$, the function $x\mapsto \phi(x)-\phi(0)e^{-x}$ is smooth and sends zero to zero (and has compact support), which implies that $(\phi(x)-\phi(0)e^{-x})/x$ can be integrated on $(0,\infty)$. (Thanks Julián Aguirre for pointing out the problem with just integrating $\phi(x)/x$.)

$u$ is clearly linear. To show that $u$ is a distribution, consider a sequence $\phi_n\to 0$ in $\mathcal D(\mathbb R)$. The support of the functions $\phi_n$ is contained in some interval $[-K,K]$ (and we can assume $K\geq 1$). By integration by parts we have $$u(\phi_n)=\lim_{\epsilon\to 0^+}\int_\epsilon^K \frac 1 x (\phi_n(x)-\phi_n(0)e^{-x}) dx =-\lim_{\epsilon\to 0^+}\int_\epsilon^K \log(x) (\phi_n'(x)+\phi_n(0)e^{-x}) dx$$

Note $\int_0^K|\log x|dx\leq K\log K+1$. Thus $$|u(\phi_n)|\leq (K\log K+1) \max_{0\leq x\leq K}(|\phi_n'(x)|+|\phi_n(0)|)\to 0$$

share|cite|improve this answer
Note that $1/x$ is not locally integrable at $x=0$, so that $\phi(x)/x$ is not integrable if $\phi(0)\ne0$. – Julián Aguirre Mar 13 '13 at 17:28
@JuliánAguirre: thanks for pointing out the problem. – Colin McQuillan Mar 13 '13 at 17:36
Thanks, but I still have one silly question: why is $u$ a distribution? i.e. how can I find an appropriate bound to $|u(\phi)|$ given a compact set $K\subseteq \mathbb{R}$? – Dist Mar 13 '13 at 21:11
@Dist: I've added an argument bounding $|u(\phi)|$. – Colin McQuillan Mar 13 '13 at 23:40
I'm probably missing something silly again, but why does $u(\phi)=\int_0^\infty \frac{\phi(x)}{x}$ for $\phi$ with compact support? – Dist Mar 14 '13 at 15:10

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.