Beyond the Mandelbrot set, an intro to holomorphic dynamics

Holomorphic dynamics turns the Mandelbrot set from a one-off curiosity into a recurring pattern: iterating complex-analytic functions produces stable behaviors (fixed points or cycles) separated by fractal boundaries where nearby starting points diverge into radically different outcomes. The core setup is simple—take a complex function that has a complex derivative (polynomials, exponentials, trig functions, and many rational functions qualify), then repeatedly apply it to an initial value. The orbit either settles into a cycle, converges to a limit point, escapes to infinity, or behaves chaotically; and when the chaotic cases are visualized, the dividing lines often form intricate fractals.

The discussion connects this general framework to two landmark constructions. Newton’s method, iterated in the complex plane, yields “Newton fractals” by coloring each starting seed according to which root it converges to. In parallel, the classic Mandelbrot set arises from a different choice: fix the starting value z = 0 while varying a parameter c in the quadratic map z ↦ z^2 + c. Points c for which the orbit stays bounded form the Mandelbrot set, while those that escape are colored by how quickly they diverge. The key contrast is structural: Mandelbrot images scan parameter space (changing the function), whereas Newton fractals scan initial-value space (changing the seed).

To build a theory rather than just pictures, the transcript emphasizes fixed points and—crucially—stability. A fixed point satisfies f(z) = z, and whether nearby points are drawn in or pushed away is determined by the derivative at that point. If |f′(z*)| < 1 the fixed point is attracting; if |f′(z*)| > 1 it is repelling; and for Newton’s method, roots of the polynomial become super-attracting because the derivative at those points is 0. The same logic extends to periodic orbits: two-cycles, n-cycles, and the fact that periodic points proliferate rapidly because solving f^n(z) = z becomes a high-degree polynomial problem.

A practical numerical question then appears: can Newton’s method get trapped in an attracting cycle that is not a root? Yes. For a specific cubic, Newton iteration can converge to a genuine attracting two-cycle, meaning some initial guesses never reach any root. Visualizing which cubics have such “bad” attracting cycles leads to a striking universality. A theorem attributed to Fatou implies that if a rational map has an attracting cycle, at least one critical point of an iterate must lie in that cycle. In the cubic Newton setting, this translates into a test using only the midpoint (the average) of the three roots: if that midpoint falls into an attracting cycle, then the polynomial has one. Scanning the third root λ while holding two roots fixed produces a parameter-space diagram whose zoomed-in black region matches the Mandelbrot set’s iconic cardioid-and-bulb geometry. The Mandelbrot shape thus emerges as a general feature of parameter spaces for Newton-type dynamics, not just the quadratic family.

Finally, the transcript explains why fractal boundaries are unavoidable. Boundaries between different basins have a “multicolor” property: arbitrarily small neighborhoods around a Julia set point eventually contain points with every available limiting behavior. That forces boundaries to be rough rather than smooth. The Fatou set consists of points that fall into predictable stable behavior; the Julia set is the complement, where nearby points expand apart and behave chaotically. A Julia-set “stuff-goes-everywhere” principle—linked to Montel’s theorem—states that iterating a neighborhood of a Julia point eventually spreads across the complex plane (with at most two exceptions). This supplies a logical bridge from chaos to fractals: maximal instability at the Julia set boundary is what generates the intricate, self-similar-looking shapes seen in these dynamical pictures.