It was taken from Complex Analysis (Princeton Lectures in Analysis, No. 2) by Elias M. Stein and Rami Shakarchi page 92.

This is the rough idea of the proof of open mapping theorem:

Theorem: If $f$ is non-constant holomorphic function in a region $\Omega$, then $f$ is open.

Let $w_0$ belong to the image of $f$, say $w_0=f(z_0)$. We must prove that all points $w$ near $w_0$ also belongs to the image of f. Define $g(z) = f(z) - w$. Choose $\delta \gt 0 $ s.t. the disc $|z-z_0| \lt \delta$ is contained the the domain $\Omega$. Then at the end it proves that $g$ has a zero inside the circle $|z-z_0| = \delta$.

Why is it the end of the proof? Can someone elaborate a little bit more?

And this is the rough idea of the proof of maximum modulus principle:

Theorem: If $f$ is non-constant holomorphic function in a region $\Omega$, then $f$ cannot attain a maximum in $\Omega$.

Suppose that $f$ did attain a maximum at $z_0$. Since $f$ is holomorphic it is an open mapping, and therefore, if $D \subset \Omega $ is a small disc centered at $z_0$, its image $f(D)$ is open and contains $f(z_0)$. This proves that there are points in $z \in D$ such that $|f(z)| \gt |f(z_0)|$, a contradiction.

How comes there are points in $z \in D$ such that $|f(z)| \gt |f(z_0)|$? Did I miss some basic stuffs?

  • 1
    $\begingroup$ First, you want to prove the image is open. So take $w_0$ and prove all points near $w_0$ are in the image. Second, you have an open subset of $\mathbb{C}$ containing $f(z_0)$, so clearly there are points with higher absolute value than $f(z_0)$ (imagine a ball) $\endgroup$ Aug 26, 2017 at 16:57

1 Answer 1


If $g$ has a zero in the disc $D=\{z:|z-z_0|<\delta\}$, then there is $z\in D$ with $f(z)=w$. This means that if $z$ is "near" to $w_0$, say in an $\varepsilon$-neighbourhood $U=\{w:|w-w_0|<\varepsilon\}$ then $w\in f(D)$. This means that $U\subseteq f(D)\subseteq f(\Omega)$. Therefore $f(\Omega)$ is a neighbourhood of each of its points, and so is open.

For the last question, split into cases $f(z_0)=0$ and $f(z_0)\ne0$. For the first case, you only need that $f$ takes nonzero values near $z_0$. In the second case you have $f(z_0)(1+\varepsilon)\in f(D)$ for small positive $\varepsilon$.

  • $\begingroup$ "In the second case you have $f(z_0)(1+\varepsilon)\in f(D)$", does that mean I just have to take the first order approximation of $f$ and consider its absolute value? $\endgroup$ Aug 26, 2017 at 17:13
  • 2
    $\begingroup$ @mathshungry I don't think that's what I meant.... Note that this number is $f(z_0)+\epsilon f(z_0)$ and for small enough $\epsilon$ this is in the image of $\Omega$ under $f$. $\endgroup$ Aug 26, 2017 at 17:22

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.