Let $\mathbb D$ denote the open unit disc in $\mathbb{C}$ . Let $Hol(\mathbb {D})$ denote the space of holomorphic functions on $\mathbb D$. The Hardy spaces on $\mathbb D$ are defined as follows.

$$H^p=\{f\in Hol(\mathbb {D}):\sup_{r<1}\int_{0}^{2\pi} |f(re^{i\theta}|^pd\theta<\infty\}\;\;\;(0<p<\infty)$$

$$H^\infty=\{f\in Hol(\mathbb {D}):\sup_{z\in\mathbb D}|f(z)|<\infty\}$$

Now let $w\in \mathbb {D}$ and $f\in H^p$ . Consider the new function $g$ on $\mathbb {D}$ defined as $$g(z)=\frac{f(z)-f(w)}{z-w}\;\;\;(z\in \mathbb {D})$$ It is said the $g\in H^p.$ Can anyone tell how? Does $g\in H^\infty$ also?

  • $\begingroup$ Good exercise I like it $\endgroup$ – Guy Fsone Oct 12 '17 at 14:58
  • $\begingroup$ Are you asking if $f\in H^\infty \implies g\in H^\infty?$ $\endgroup$ – zhw. Oct 12 '17 at 19:55
  • $\begingroup$ @zhw No. I am asking if $f\in H^p$ then does $g\in H^\infty$? $\endgroup$ – user429197 Oct 12 '17 at 21:40

Hint: By Analiticty,

$$f\in Hol(\Bbb D) \implies f(z) = \sum_{n =0}^{\infty}a_n(z-w)^n, ~~z\in D(w,\rho)$$ for som $\rho>$ such that $D(w,\rho)\subset \Bbb D.$

Remark $$a_n =\frac{f^{(n)}(w)}{n!}$$


  • For $z\in D(w,\rho)$

$$ g(z) = \frac{f(z) - f(w)}{z-w} = \sum_{n =1}^{\infty}a_n(z-w)^{n-1} ~~$$

$\bar{D}(w,\rho)\ni\mapsto g(z)$ is continuous with $~~g(w) = f'(w).$ This implies that, there exists some constant $C_\rho>0$ $$|g(z)| \le C_\rho,~~z\in \bar{D}(w,\rho)$$

  • For $z\in \Bbb D\setminus D(w,\rho)$

$$ |g(z)|\le \rho^{-1} (|f(z)|+|f(w)|).$$

  • Altogether give

$$ |g(z)|\le C_\rho+ \rho^{-1} (|f(z)|+|f(w)|)~~~~\forall ~~z\in\Bbb D.$$

Then it easy to see that, $g\in H^p$

Now if $g\in H^\infty \implies f\in H^\infty $ this will mean that $H^p\subset H^\infty$ which is not true!!

in fact

$$|f(z)|\le |f(z)-f(w)| +|f(w)| = |g(z)(z-w)|+|f(w)|\le 2\|g\|_\infty+|f(w)|. $$

Hence there exists $f\in H^p$ such that $g\not \in H^\infty$.

Better Rescaling on $g$ If you set $$g_\rho (z) = g(\rho z)~~~\text{for fixed $0<\rho<1$} $$ then $$g_\rho\in H^\infty$$

Since $\bar{\Bbb D}\subset D(0,\frac{1}{\rho}) $

$$g \in Hol(\Bbb D) \implies g_\rho\in Hol(D(0,\frac{1}{\rho})) \implies g_\rho ~~~\text{is bounded on $\Bbb D$}.$$

  • $\begingroup$ FsoneThank you for your answer. Can you tell if there is any way that I can scale $g$ so that it belongs to $H^\infty$? Or any factor which I multiply with $g $ and then it belongs to $H^\infty$? $\endgroup$ – user429197 Oct 13 '17 at 5:03
  • $\begingroup$ I think this question on rescaling is too broad and vague . since it depends on which kind of singularity that $f$ has and the type of resacaling you aim to do use. But any if you see the new edit this may be ok $\endgroup$ – Guy Fsone Oct 13 '17 at 7:43

The isolated sigularity at $w$ is removable, so $g$ is holomorphic on $\mathbb D.$ Here's the proof for $1\le p < \infty$: Choose $r_0 \in (|w|,1).$ Then for $r_0<r<1,$

$$\int_0^{2\pi} |g(re^{it})|^p\, dt \le \int_0^{2\pi} \frac{(|(f((re^{it})|+|f(w)|)^p}{(r_0-|w|)^p}\, dt$$ $$ \le \frac{1}{(r_0-|w|)^p}\int_0^{2\pi} 2^{p-1}(|f((re^{it})|^p+|f(w)|^p)\, dt,$$

which shows $g\in H^p.$ The proof for $0<p<1$ is very similar. Try your hand at proving $f\in H^\infty \implies g\in H^\infty.$


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.