I am reviewing some complex analysis and I have just gotten to the portion on analytic continuation. My question is about the proof of the following theorem:

Theorem. Suppose $f$ is a holomorphic function in a region $\Omega$ that vanishes on a sequence of distinct points with a limit point in $\Omega$. The $f$ is identically zero.

Proof. Suppose $z_0 \in \Omega$ is a limit point for the sequence $\{w_k\}$ and $f(w_k) = 0$. Chose a disc $D \subset \Omega$ centered on $z_0$, and consider the power series expansion of $f$ in $D$ $$f(z) = \sum a_n(z-z_0)^n$$ If $f$ is not identically 0, there exists a smallest $m$ such that $a_m\neq 0$. Then, $f$ can be rewritten as $$f(z) = a_m(z-z_0)^m(1+g(z-z_0))$$ where $g(z-z_0) \to 0$ as $z \to z_0$. Taking $z = w_k$, we obtain a contradiction since $a_m(w_k-z_0)^m \neq 0$ and $g(w_k-z_0) \neq 0$, but $f(w_k) = 0$ ... $\blacksquare$

The proof concludes the argument by using the connectedness of $\Omega$. However, my question is why must $1+g(w_k-z_0) \neq 0$.

  • $\begingroup$ I would have just written $f(z) = (z-z_0)^mh(z),$ where $h$ is continuous and $h(z_0)\ne 0.$ $\endgroup$ – zhw. Jul 8 '17 at 16:13
  • $\begingroup$ That is an alternative of course. I was just going through the proof in Stein and Shakarchi's Complex Analysis book where it was written this way. $\endgroup$ – Jonathan Davidson Jul 8 '17 at 20:16

I think there is a typo here, it should be $1+g(w_k-z_0)\neq 0$. This is true because $g(z-z_0)\to 0$ as $z\to z_0$.

Then we have $$0 = f(w_k) = a_m(w_k-z_0)^m(1+g(w_k-z_0))$$ for sufficiently large $k$, all the terms on the right are not zero, so a contradiction.

  • $\begingroup$ Oh, so if i choose $w_k$ close enough to $z_0$, that term should be really close to 1 by continuity of $g$. Thanks!!! $\endgroup$ – Jonathan Davidson Jul 8 '17 at 5:02

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.