I want to derive $$\pi \cot(\pi z) = \sum_{-\infty}^{\infty}\frac{1}{z-n} + \frac{1}{n}$$ without taking derivatives.
I know through Mittag Leffler that $$\pi \cot(\pi z) = g(z) +\sum_{-\infty}^{\infty}\frac{1}{z-n} + \frac{1}{n}$$ for some entire analytic $g(z)$ but I'm having a hard time showing that $g(z) = 0$. I do not want to take the integral of $\frac{\pi^2}{\sin^2(\pi z)}$.
I have deduced that $g(0) = 0$, $g$ is odd, and $g$ is periodic with period $1$.
If I could deduce that $g$ was bounded even in a strip of length $1$ I wold be finished because of Louiville Theorem. To this end, I can show that $\pi \cot(\pi z)$ is bounded towards infinity but I was not able to show that the sum could be bounded.
I don't know if that is Cauchy's method, but one way to establish $g\equiv 0$ without differentiating is to use a functional equation:
$$2\pi \cot (2\pi z) = \pi \frac{\cos^2(\pi z) - \sin^2(\pi z)}{\sin (\pi z)\cos (\pi z)} = \pi \cot (\pi z) + \pi\cot \left(\pi \left(z+\tfrac{1}{2}\right)\right),$$
and with
$$s(z) = \frac{1}{z} + \sum_{n\neq 0} \left(\frac{1}{z-n} + \frac{1}{n}\right) = \lim_{N\to\infty} \sum_{n=-N}^N \frac{1}{z-n}$$
we have
$$\begin{align} s(z) + s\left(z+\tfrac{1}{2}\right) &= \lim_{N\to\infty} \left(\sum_{n=-N}^N \frac{1}{z-n} + \sum_{n=-N}^N\frac{1}{z+\frac{1}{2}-n}\right)\\ &= 2\lim_{N\to\infty} \left(\sum_{n=-N}^N\frac{1}{2z-2n} + \sum_{n=-N}^N \frac{1}{2z+1-n}\right)\\ &= 2\lim_{N\to\infty} \sum_{k=-(2N+1)}^{2N} \frac{1}{2z-k}\\ &= 2s(2z), \end{align}$$
so the entire function $g$ also satisfies the functional equation
$$g(z) + g\left(z+\tfrac{1}{2}\right) = 2g(2z).\tag{$\ast$}$$
But a continuous $1$-periodic function $f$ satisfying the functional equation $(\ast)$ is constant on $\mathbb{R}$. Separately considering the real and imaginary part of $f$, we may assume that $f$ is real-valued (and our $g$ is real-valued on $\mathbb{R}$). By continuity and periodicity, there is an $x_0 \in [0,1]$ with $f(x) \leqslant f(x_0)$ for all $x\in \mathbb{R}$. Let $x_k = 2^{-k}x_0$ for $k\in\mathbb{N}$. Then
$$2 f(x_0) = 2f(x_k) = 2f(2x_{k+1}) = f(x_{k+1}) + f\left(x_{k+1}+\tfrac{1}{2}\right)$$
shows inductively that $f(x_k) = f(x_0)$ for all $k$. Hence, by continuity, $f(0) = \max \{ f(x) : x\in\mathbb{R}\}$. The analogous argument shows $f(0) = \min \{ f(x) : x\in \mathbb{R}\}$, so $f \equiv f(0)$ is constant.
Since $g$ is holomorphic, it follows that $g$ is constant on $\mathbb{C}$, hence $g(z) \equiv g(0) = 0$.