the holomorphic inverse function theorem

If a holomorphic function has nonzero derivative at a point, can you undo it nearby — solve f(z) = w for z, smoothly and uniquely? Yes, and beautifully: the inverse you get is itself holomorphic. The holomorphic inverse function theorem says that a single nonzero derivative guarantees a local two-sided inverse that is just as nice as the original function. No loss of smoothness, no surprises.

The statement: if f is holomorphic near z_0 and f'(z_0) is not zero, then there is a neighborhood of z_0 on which f is one-to-one (injective), its image is an open neighborhood of w_0 = f(z_0), and the local inverse map g = f^(-1) is holomorphic there. Moreover the derivative of the inverse obeys the familiar rule g'(w_0) = 1 / f'(z_0), and more generally g'(w) = 1 / f'(g(w)). The reason f is locally injective is the local mapping theorem with m = 1: when f'(z_0) is not zero, f looks locally like a conformal stretch-rotate, which is one-to-one. That the inverse is holomorphic (not merely continuous) is the genuinely complex bonus: once you know the inverse is differentiable in the real sense, the chain rule forces it to satisfy the Cauchy-Riemann equations, so it is holomorphic.

This is the local cornerstone of building inverse functions in complex analysis — local logarithms (inverting e^z where its derivative e^z is never zero), local n-th roots, local inverses of conformal maps. It is what makes f a local biholomorphism at every regular point. Two honest cautions. First, it is purely local: f'(z_0) not zero gives injectivity only in some small neighborhood, never global injectivity by itself (e^z has nonzero derivative everywhere yet is far from one-to-one, being 2 pi i-periodic). Second, the converse needs care: a holomorphic injective function does have nonvanishing derivative, but the everyday inverse theorem is the one-directional 'nonzero derivative implies local inverse.'