7
$\begingroup$

This is from Gamelin's book on Complex Analysis.

Problem: Show that if $f(z)$ is continuous on a domain $D$, and if $f(z)^8$ is analytic on $D$, then $f(z)$ is analytic on $D$. (I assume the intention is that $D$ is nice, i.e. open and connected)

I'm not exactly sure how to approach this. I'm guessing it has something to do with zeroes of $f$. For example, around non-zeroes, $f$ is analytic since $z^{1/8}$ is nonzero in a neighborhood. Or something along those lines. The details are eluding me.

Any help would be appreciated!

|cite|improve this question
$\endgroup$
  • 1
    $\begingroup$ It's true that you can prove a function analytic by proving that every point has a neighborhood in which it's analytic. So you can immediately reduce to the case of a small disk around a zero of $f$. Can you then show that the order of the zero of $f(z)^8$ there must be a multiple of $8$? $\endgroup$ – Greg Martin Jun 11 '14 at 23:15
  • 2
    $\begingroup$ Does this work? Say $z_0$ is a zero of $f$. We have $f(z)^8 = (z-z_0)^n g(z)$ for some $n$, where $g$ is analytic. Then $f = (z-z_0)^{n/8} g^{1/8}$, where $g^{1/8}$ is analytic (since $z_0$ isn't a zero of $g$). Since $f$ is continuous, the limit as $z\to z_0$ of both sides must be $0$, but this isn't true if $n/8$ isn't an integer. $\endgroup$ – Sameer Kailasa Jun 11 '14 at 23:23
8
$\begingroup$

Suppose $f(z) \neq 0$. Then $${{f(z+h)^8 - f(z)^8}\over{h}} = {{f(z+h) - f(z)}\over{h}}\left(f(z+h)^7 + f(z+h)^6f(z) + \dots + f(z)^7\right).$$ Using continuity, we see the expression in parentheses is nonzero for small $h$ and has limit $8f(z)^7$ as $h \to 0$. Dividing by this expression and taking limits, we get $${{\left(f(z)^8\right)'}\over{8f(z)^7}} = f'(z).$$ If $f(z)$ has some zeros in the domain, then as $f(z)^8$ is holomorphic so its zeros are isolated as well as those of $f(z)$, since zeroes of $f(z)$ and $f(z)^8$ are the same set. Now $f(z)^8$ is locally bounded at those points so $f(z)$ is also locally bounded at those points. Applying removable singularity, we can conclude that $f(z)$ is also holomorphic at all zeros of $f(z)$.

|cite|improve this answer
$\endgroup$
  • $\begingroup$ So in the last paragraph, you know that $f(z)$ is holomorphic on $D\setminus f^{-1}(0)$, and so now you just have to show it's also holomorphic at its zeroes? Can't we directly say that $f$ is continuous, therefore it's bounded at those points? $\endgroup$ – Jack M Jun 13 '14 at 8:19

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.