Before I jump in, just want to quickly note that this post exists already, I am using the same book, and I had a very similar confusion which led me to trying to prove the theorem in a slightly different way. My question is whether or not the following proof is sufficient to establish the theorem -- the proof itself differs from the one in both Stein & Shakarchi and Rudin (see the linked post for the two book titles if it's unclear).

For ease, I'll restate the theorem:

$\textbf{Thm:}$ Let $f(z)$ be a holomorphic function defined in a region $\Omega$. Let $w_k$ denote a zero of $f$ in $\Omega$ and let the sequence $W=\{w_k\}$ be a sequence of zeros which converge to a limit point in $\Omega$. Then $f$ is identically zero.

$\textbf{Pf:}$ Let $z_0$ be the limit point of the sequence $W$. An existing theorem proves that since $f$ is holomorphic, it is also analytic with some positive radius of convergence when expanded at $z_0$. If $f$ is not identically zero, then there is at least one term in its expansion about $z_0$ which is not zero. By a separate previous theorem, we also know that the terms for the expansion at the derivatives of $f$ at $z_0$. Now, by the definition of convergence:


(This is the step in logic that I am unsure of): Since every of the $f(w_k)$ are zero, it follows that $f(z_0)$ is likewise zero. Hence the leading term in the expansion at $z_0$ must be proportional to the first derivative of $f$ or a higher derivative.

Next, consider the function $f: \mathbb{C} \to \mathbb{R}$ defined by $z \to |f(z)|$. Since this function has zeros at the same points in the region as $f$, we know that $|f(z)|$ has a maximum value on the interval $(|w_k|,|w_{k+1}|)$. Write this value as $w'_k$. Since the sequence $\{w_k\}$ converges to $z_0$, we have $|w_k - w_{k+1}| \to 0$ as $k \to \infty$. It follows then that $|w'_k - w'_{k+1}|$ also tends to $0$ as $k$ tends to infinity, where $w'_{k+1}$ is the value at which $|f|$ is max on $(|w_{k+1}|,|w_{k+2}|)$. But then this set also forms a sequence which converges to $z_0$ in the limit. But then by the same argument, $f'(z_0)$ must also be zero. By induction, every derivative of $f$ at $z_0$ must be zero and hence the function is identically zero in the neighborhood of $z_0$.

Then you can use the connectedness of the region to prove it is identically zero everywhere in $\Omega$.

$\textbf{Comments:}$ Now, I just want to mention the few spots at which I'm not 100% certain this follows:

  1. Does $f(z_0)=0$ per the argument I made?
  2. Does $\{w'_k\}$ also converge to $z_0$?
  3. Does it make sense to talk about $|f(z)|$ being defined on $\Omega$ as well?

Thanks in advance for any insight offered.


1) follows from the fact that any holomorphic function is continuous. 2) is false. Your $w_k'$ is a positive number and it cannot converge to $z_0$ unless $z_0$ is a non-negative real number. If $z_0=i$, for example, then 2) fails. 3) is true. Surely $|f(z)|$ is well defined continuous function on $\Omega$.

  • $\begingroup$ Could I say that the argument implies there is a complex number whose absolute value is between $|w_k|$ and $|w_{k+1}|$ which is a zero of $f'$? I.e. there is certainly a zero of $f'$ in the set $D_{|w_k|}(z_0) - D_{w_{k+1}}(z_0)$? $\endgroup$ – Ryan S Oct 2 '18 at 15:16

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.