3
$\begingroup$

Let f and g be two holomorphic functions on a domain $\Omega$. Suppose that $\frac{f}{g}$ is always finite (while g can be zero at some points). Is it true that then $\frac{f}{g}$ is holomorphic? Actually I do not know any example of 2 such functions, except of course the trivial case $f=cg$ where c is a constant. It will be thus very helpful if someone can provide nontrivial examples. Thanks in advance!

Edit: finally, this is standard thing about removable singularities. Nothing more needed! Thanks!

$\endgroup$
8
  • $\begingroup$ An easy example: $f(z)=z^{n+1}$, $g(z)=z^n$ $\endgroup$
    – Salmorejo
    Jul 29, 2014 at 9:22
  • $\begingroup$ You might like to look at this Wikipedia page. $\endgroup$
    – user149874
    Jul 29, 2014 at 9:24
  • $\begingroup$ @Adolfo Thanks for your comment! About that easy example, well I'm looking for examples different from that kind. To produce examples of that kind, we take h holomorphic and we write h= (hf)/f :). I want to see more nontrival examples, where f and g don't have easily seen common factors. $\endgroup$ Jul 29, 2014 at 9:30
  • $\begingroup$ How about $g=e^z$? There are entire functions. $\endgroup$ Jul 29, 2014 at 9:32
  • $\begingroup$ @Jez Thanks. I see your point, so it seems that the answer is affirmative. $\endgroup$ Jul 29, 2014 at 9:33

1 Answer 1

1
$\begingroup$

Answer: yes. I am going to formalize where you said "$f/g$ is always finite" to mean: "for any $P \in \Omega$, if $g$ has a zero at $P$, then $f$ has a zero of at least the same order."

Recall that we define the order of a zero to be smallest $n \geq 0$ such that $f^{(n)}(P) \neq 0$ but $f^{(n-1)}(P) = 0$. That this exists (when the function is not identically the zero function) is a consequence of holomorphicity, which is obvious if you already know that holomorphic functions are analytic.

Given that $f$ and $g$ are holomorphic, this is the only hypothesis needed to ensure that $f/g$ admits a holomorphic continuation to a domain including $P$. Moreover, this condition is necessary. This is seen easily by Taylor expanding at $P$.

In particular, if a continuous extension exists, it is automatically holomorphic.

$\endgroup$
2
  • $\begingroup$ thanks, I get it. Being a physics major, sometimes I forget to formalize things properly. Another naive question, what about real functions (smooth but not analytic for example)? From your answer, it seems to me that holomorphicity=analyticity is a key ingredient. $\endgroup$ Jul 29, 2014 at 10:50
  • $\begingroup$ i don't know how to formalize the question for such a function, e.g. consider a bump function; what is its order of vanishing at the origin? $\endgroup$
    – hunter
    Jul 29, 2014 at 11:27

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.