4

$ \mathbb{P}^{1}\left(\mathbb{C}\right) :=\mathbb{C}\cup{\infty} $

(There are some different definitions, the one that I know is the stereographic projection and defining the image of the north pole under the projection as $\infty $ in $ \mathbb{C} $.

Define also $ \text{ord}_{z=a}\left(f\right) $ as the following:

  1. if $ f $ has a zero of multiplicity $m\geq1 $ at $ a $, then $ \text{ord}_{z=a}\left(f\right)= m $.

  2. if $ a $ is a pole of multiplicity $m\geq 1 $ for $ f $, then $ \text{ord}_{z=a}\left(f\right) = -m $

  3. if $ f $ defined well\has a removable singularity at $ a $ and $f(a) \neq 0 $ then $ \text{ord}_{z=a}\left(f\right) =0 $.

Prove that the following sum has finitely many nonzero terms and that $ \sum_{z\in\mathbb{P}^{1}\left(\mathbb{C}\right)}\text{ord}_{z}\left(f\right)=0 $

Where we define $ \text{ord}_{z=\infty}\left(f\right)=\text{ord}_{\omega=0}\left(g\right) $

Where $ g $ is the nonzero meromorphic function defined by $ g\left(\omega\right):=f\left(\frac{1}{\omega}\right) $.

Im have absolutely no intuition for the proof, I cant even understand why it is correct. For example what about the function $ \frac{1}{z}+\frac{1}{z-1} $ ?

It has a pole of order $ 1 $ in $0 $ and in $ 1 $ so that together it summed to $-2 $, and the order at $\infty $ is the order of $ z+\frac{z}{1-z} $ at $0 $, which is $1$. So it seems like the sum is not $0 $.

Any help would be appreciated.

  • 1
    $\frac{1}{z}+\frac{1}{z-1}$ also has a zero of order $1$ in $\frac{1}{2}$. – Z. Zhu Aug 13 '21 at 01:09
  • @GeorgeBrown What about $ \frac{1}{\sin\pi z} $ ? – DirichletIsaPartyPooper Aug 13 '21 at 06:50
  • 1
    @euler0.0: Are you aware that the meromorphic functions on $\Bbb P^1(\Bbb C)$ are exactly the rational functions, i.e. quotients of two polynomials? – Martin R Aug 13 '21 at 06:51
  • Think of factorization of $f(z) = (z-a_1)^{d_1}\dots (z-a_k)^{d_k}$, for $f \in \mathbb{C}(z)$ a rational function of $z$. Each linear factor $(z-a)$ increases 1 for $\operatorname{ord}{a} f$ and decrease $1$ for $\operatorname{ord}\infty f$, and vice versa for $(z-a)^{-1}$. I think this would help you to have some intuitions. – dust05 Aug 13 '21 at 06:53
  • Also you think of $\sin(z)$ having essential singularity at $z = \infty$ with this manner. – dust05 Aug 13 '21 at 06:53
  • @MartinR The definition that I know is holomorphic except for maybe polar points – DirichletIsaPartyPooper Aug 13 '21 at 06:57
  • 2
    @euler0.0: That is correct. But on the (compact) extended complex plane that are exactly the rational functions. The reason is that there can be at most finitely many poles, and if you subtract the principal parts at those poles then you have a holomorphic and bounded function, which is constant. – Or see this: https://math.stackexchange.com/q/2004311/42969. – Martin R Aug 13 '21 at 07:00

1 Answers1

2

The meromorphic functions $f:\Bbb P^1(\Bbb C) \to \Bbb P^1(\Bbb C)$ are exactly the rational functions, see for example

In order for $\sum_{z \in\Bbb P^1(\Bbb C)} \operatorname{ord}_z(f)$ to make sense we must assume that $f$ is not identically zero and not identically $\infty$. So we have $f(z) = p(z)/q(z)$ where $p, q$ are polynomials, both not the zero polynomials. Let $m = \deg(p)$ and $n = \deg(q)$ be the degrees of the polynomials.

If $m = n$ then $f$ has $m$ zeros and $m$ poles, counted with multiplicity, and $\lim_{z \to \infty} f(z) = a \ne 0, \infty$. In this case, $$ \sum_{z \in\Bbb P^1(\Bbb C)} \operatorname{ord}_z(f) = m - m = 0 \, . $$

If $m > n$ then $f$ has $m$ zeros and $n$ poles, and $f(1/z)$ has a pole of order $m-n$ at zero. In this case $$ \sum_{z \in\Bbb P^1(\Bbb C)} \operatorname{ord}_z(f) = m - n - (m-n) = 0 \, . $$

The case $m < n$ works similarly.

Example: $$ f(z) = \frac{1}{z} + \frac{1}{z-1} = \frac{2z-1}{z(z-1)} $$ has two simple poles and one simple zero in $\Bbb C$. Also $$ f\left( \frac 1 z\right) = \frac{z(z-2)}{z-1} $$ has a simple zero at $z=0$, so that $\operatorname{ord}_\infty(f) = 1$.


Remark: The same is true for meromorphic functions on arbitrary compact Riemann surfaces, see for example

Martin R
  • 113,040
  • Thanks, a question if I may, why can we say that if $ f $ is meromorphic then it has finitely many zeroes in any compact set? I can understand that it is true for homorphic function, because then we can use the uniqueness theorem and claim that otherwise it would be constant zero, but when a function is meromorphic, the accumulation point of the zeroes may not be included in its holomorphic region – DirichletIsaPartyPooper Aug 13 '21 at 08:45
  • @euler0.0: The zeros of a meromorphic function cannot accumulate at a pole. – Martin R Aug 13 '21 at 09:15
  • Oh, I see, since the function strive to $\infty $ around pole. But why cant the function has infinitly many poles? – DirichletIsaPartyPooper Aug 13 '21 at 09:17
  • 1
    @euler0.0: The poles are by definition of a meromorphic function isolated in the domain of the function. – Martin R Aug 13 '21 at 09:19