I have been struggling on the following problem.

Suppose $f$ is an entire analytic function such that $|f(z)|>1$ if $|z|>1$. Show that $f$ is a polynomial.

My idea is as followed: all zeros of $|f(z)|$ lie inside $|z|\leq 1$. Applying Argument Principle, we can show that number of zeros of $f$ is bounded. So we can assume


where g is entire analytic without any zeros. Then I would like to apply Liouville's Theorem: the point is that it isn't too clear to me why $|\dfrac{1}{g(z)}|$ is a bounded function.

  • 2
    $\begingroup$ Hint: if $f$ is entire, then either $f$ is a polynomial, or $f$ has an essential singularity at $\infty$. $\endgroup$ May 20 '12 at 0:34

If the Taylor series about 0 does not terminate, $f(1/z)$ has an essential singularity at $0$ (why?)

Then from the Casorati–Weierstrass theorem (have you learnt this?) you know $f(1/z)$ cannot be bounded from below near 0. Contradiction!

  • $\begingroup$ This is great. I didn't approach in this way because the above exercise was taken from the chapter on "Argument Principle, Rouche's Thm and Liouville's Thm". $\endgroup$
    – iloveinna
    May 20 '12 at 0:53
  • $\begingroup$ @Montez, can you explain the line "$f(z)$" cannot be bounded from below near 0" $\endgroup$
    – Deepak
    Dec 21 '12 at 20:40
  • 1
    $\begingroup$ The Casorati-Weierstrass theorem says that if $p$ is an essential singularity of an analytic function $g$, the image of a deleted neighbourhood of $p$ is dense in the complex plane. If it's dense, it takes values arbitrarily close to $0$. $\endgroup$ Dec 23 '12 at 7:05
  • 1
    $\begingroup$ @Deepak I didn't say $f(z)$ is bounded from below near $0$. I said $f(1/z)$ bounded from below near zero, which is the same as $f(z)$ bounded from below away from $0$, which is given, which contradicts the Casorati-Weierstrass theorem if the $f(1/z)$ has an essential singularity at zero. $\endgroup$
    – Spook
    Dec 23 '12 at 16:33

The function $1/f$ is bounded at infinity, hence has a fake singularity there and can be holomorphically extended through infinity.
Hence $1/f$ is actually rational ($\iff$ meromorphic on the whole Riemann sphere) and so also $f$ is rational.
But since $f$ does not have any pole at finite distance (since it is an entire function) it is a polynomial (and has a pole at infinity if it isn't a constant).


The number of zeros of $f$ can be finite as the set of zeros of $f$ will have a accumulation point and $f$ will be equivalent to $0$ (not possible) then $$g=\frac{f}{(z-a)\cdots(z-x)}$$ where $a$ to $x$ are the finite number of zeros of $f$ then $g$ is analytic consider $h=1/g$ see that $h(z)$ is never zero as we have deleted the zeros of $f$ and also $f$ tends to infinity as $z$ tends to infinity (by the given condition) which implies $$|h(z)| \le a+|z|^n,$$ where $n$ is the number of zeros of $f$, but by fundamental theorem of algebra we get $h(z)=c$ (constant) thus $$f=\frac{(z-a)\cdots(z-x)}{k}.$$


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.