2 DZMITRY DUDKO AND MIKHAIL LYUBICH
1. Introduction
Local dynamics near a neutral fixed point, and a closely related dynamical theory of circle homeo-
morphisms, is a classical story going back to Poincar´e, Fatou, and Julia. It followed up in the next two
decades with breakthroughs by Denjoy (1932) and Siegel (1942) on the linearization of circle diffeomor-
phisms and local maps, respectively. In the 1950-60s, the Kolmogorov–Arnold–Moser (KAM) theory
emerged resulting in the reenvision of the near-rotation phenomenon in mathematics and physics. In
the mid-1970s, Renormalization Ideas were introduced into Dynamics by Feigenbaum, Coullet and
Tresser and led, in particular, to numerous conjectures in Low-Dimensional Dynamics including the
nature of the KAM small divisor problem.
The theory of analytic circle diffeomorphisms and local theory for neutral holomorphic germs and
quadratic polynomials received an essentially complete treatment in the second half of the last century
in the work by Arnold (in the KAM framework), Bruno, Herman, Yoccoz, and Perez-Marco. About
at the same time (1980–90s) a global and semi-local theory for neutral Siegel quadratic polynomials
fθ:z7→ e2πiθ z+z2with rotation numbers θof bounded type was designed on the basis of the Douady-
Ghys surgery. Renormalization Theory of Siegel maps, also initiated by physicists, was mathematically
designed by McMullen in [McM1] in the mid-1990s. In the late-2010s, the hyperbolicity of Siegel
Renormalization of bounded type was established in the framework of Pacman Renormalization [DLS]
providing tools to study near-Siegel maps [DL1].
A remarkable progress in understanding Near-Neutral Complex Dynamics came in the 2000s, when
Inou and Shishikura established uniform a priori bounds for quadratic polynomials fθwith rotation
numbers of high type (near-parabolic perturbative regime). This theory found numerous applica-
tions, from constructing examples of Julia sets of positive area by Buff-Cheritat (2000s, [BC]) and
Avila-Lyubich (2010s, [AL2])) to a complete description, for high type rotation numbers, of the topo-
logical structure of the Mother Hedgehogs that capture the semi-local dynamics of neutral quadratic
polynomials (Shishikura-Yang, Cheraghi (2010s)). See §1.3 and §1.5 for a more detailed historical
account.
In this paper, we lay down a foundation for the renormalization theory based upon almost-invariant
objects, called pseudo-Siegel disks, that “hide” various geometric irregularities; compare with Item b
Z
in §1.5. In the almost invariant framework, we will establish uniform “pseudo-Siegel” a priori bounds
for neutral quadratic polynomials fθwith artbitrary rotation numbers and show that the postcritical
set of every such fθis “small”: it lies within the compact Mother Hedgehog Hθso that the restriction
fθ:Hθýis a homeomorphism. Pseudo-Siegel bounds can be transferred into other forms of a priori
bounds [DL2, DLL] making the theory more complete and compatible with the previous developments;
see §1.2 for the summary.
Ideas of the current paper helped to confirm the MLC at the classical Feigenbaum parameter [DL4]
– historically, the most important parameter for the development of the Dynamical Renormalization
Theory; see Remark 1.8. Applications to McMullen’s problem on Sierpinski hyperbolic components
and to the construction of Herman curves have already appeared in [DLu, Lim]. More applications of
the theory developed in the current paper are underway; see §1.4.
Our proof is based upon analysis of degenerating Siegel disks of bounded type. The Near-Degenerate
Regime and its principles, in the quadratic-like renormalization context, were originally designed by
Jeremy Kahn [K], with a key analytic tool, the Covering Lemma, appeared in [KL1]. They serve
as an entry point for our paper. One of the major subtleties of our situation is that Siegel disks
of bounded type do not have uniformly bounded qc-geometry since they may develop long fjords in
all scale. (Otherwise, Cremer and even parabolic points would not have existed.) To deal with this