I am trying to prove Liouville's theorem: An entire bounded function is constant. I'm trying to use the Mean Value Theorem from my textbook.

MVT: $\space\space $If $f$ is analytic in $D$ and $a \in D$, then $f(a)$ is equal to the mean value of $f$ taken around the boundary of any disk centered at $a$ and contained in $D$. That is,

$f(a)$ = $\frac{1}{2\pi} $$\int_{0}^{2\pi} f(a + re^{i\theta}) d\theta$ when $D(a; r) \subset D$


Assumtion: Suppose that $f(z)$ is analytic and NOT constant on a circle $C$.

Clearly $f(z)$ is bounded, giving us the necessary assumption (hold your horses for the entire part).

$$ (*) \space \space \space \space \space \space\space \space \space|f(z)| \leq{\frac{1}{2\pi}\int_{0}^{2\pi}|f(z + re^{i\theta})| d\theta} \leq{\max{_\theta(|f(z + re^{i\theta})}}|)$$ $(**)$ Since $f(z)$ is bounded, $\exists{}$ $z$, s.t.equality is achieved in $(*)$

The only way equality holds, is if $f$ is constant throughout the circle. So f is constant throughout the circle

Let $f$ be entire. Since entire functions are by definition holomorphic and analytic on their domain, our statement $(**)$ still holds since $f$ is still bounded by assumption.

So $f$ is constant. Contradiction. $QED$

  • $\begingroup$ Clarification: for the entire part of the theorem to work, take a circle but let $r$ => $\infty$ $\endgroup$
    – lf2225
    Dec 17 '18 at 4:47

It is not at all clear why boundedness of $f$ must imply equality is achieved in (*).

  • $\begingroup$ Since $f$ is bounded, $f(z)$ is less or equal to the right hand side of $(*)$ for some $z$. Same holds for absolute value of each side. Do you disagree? $\endgroup$
    – lf2225
    Dec 17 '18 at 4:40
  • $\begingroup$ No, you just prove that a bounded function is bounded. Surprise. -- Note that the circle moves with $z$, so that your claim that "equality must hold" because of some compactness argument fails because for a fixed circle your theorem only makes a claim on one point $z$, the center of that circle. $\endgroup$ Dec 17 '18 at 17:40
  • $\begingroup$ What if you use the MVT directly without the inequality and take r goes to infinity. I'm trying to not use another point $|f(w)|$ (and that the difference between the two approaches 0 as r => $\infty$) $\endgroup$
    – lf2225
    Dec 17 '18 at 19:39

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.