the holomorphic local inverse

When a holomorphic function can be undone near a point, the function that undoes it is not just some continuous patch — it is a holomorphic function in its own right. The holomorphic local inverse is exactly that undoing map: the function g defined on a neighborhood of w_0 = f(z_0) with g(f(z)) = z and f(g(w)) = w, and it inherits the full smoothness of the complex world.

Concretely, at a point z_0 where f is holomorphic with f'(z_0) not zero, the inverse function theorem hands you a small neighborhood V of w_0 and a map g : V to a neighborhood of z_0 satisfying f(g(w)) = w for all w in V. This g is holomorphic, with g'(w) = 1 / f'(g(w)). You can even read off its Taylor coefficients without inverting any formula explicitly: the Lagrange inversion technique expresses the coefficients of g around w_0 in terms of the derivatives of f at z_0. A useful way to picture g: it is the analytic continuation of the obvious power-series inverse, valid as long as you stay close enough that f remains one-to-one. The branches of the logarithm and the branches of fractional powers are exactly such local inverses of e^z and of power maps.

Where this matters: any time you want to solve f(z) = w for z holomorphically — defining inverse trigonometric functions, building conformal maps from their inverses, or running implicit-function arguments — the holomorphic local inverse is the rigorous object you lean on. The honest cautions track those of the inverse function theorem. The inverse exists only locally, on a neighborhood small enough that f is injective; trying to extend it around a critical point (where f' = 0) or around a branch point forces multivaluedness, which is exactly how the logarithm and roots acquire their branches and branch cuts. So 'the' local inverse is really one chosen branch, valid on a chosen neighborhood.