scientific article; zbMATH DE number 1003120
The 1990s saw a great deal of work being done in the area of quadratic polynomials and unimodal maps of the interval. There were deep results about the structure of the Mandelbrot set, local connectivity of Julia sets, dense hyperbolicity in the logistic family, measure-theoretic attractors for unimodal maps and convergence of renormalization. Impressive as all this was, the larger mathematical public has not generally been able to follow this progress closely. One reason is the great complexity of the proofs. Even the key ideas have not been widely understood. The paper at hand is designed to serve a very useful purpose. Its objective is to present the main concepts in a non-technical way understandable to non-specialists. The paper indeed gives a broad and lively account of the recent research in the field and can be read by mathematicians of various specialities. However, this needed and approachable paper is flawed on several counts. To begin with the least serious, there is evidence of haste in misprints and occasional convoluted sentences. A glaring example of the latter is found in paragraph 4.4, where one reads that ``Quantifying in ingenious way the ideas outlined in par. 3, Yoccoz has transferred the last result to the parameter plane. This is about as clear and graceful as saying that Newton quantified the ideas of gravity to transfer the law of falling apples to the sky. Worse, the paper presents ideas of the field in a one-sided way. In fact, conformal dynamics of the 90ss was the interest of many people with many sources of inspiration. To force all this into the framework of ``renormalization does not contribute to the intellectual depth of the survey. The word ``renormalization in one-dimensional dynamics used to mean the Feigenbaum-Tresser process for unimodal maps, later extended to circle mappings. The author gives a wider meaning to this term in Section 2.2. This extension of the concept allows him to present a much wider scope of ideas as if they flowed from renormalization. This is a personal point of view and not widely shared. For me, the most of important source of ideas was the quest for induced hyperbolicity. Loosing this from the sight, the author overlooks an attractive connection with works in higher dimensions based on the same idea. Other researchers brought yet different viewpoints which did not receive attention in the paper. Objectionable as his choice may be, the author surely had the license to use any organizing concept he liked. The worst flaw of the paper is where he extends his license to put a spin on the history of subject, often simply ignoring references to relevant papers. It is necessary to compile a list of such omissions here. Note that citations with letters are from the paper, while those with numbers, missing there, are explained at the bottom of this review. \(\bullet\) Theorem 4.6 for real polynomials follows, with stated consequences like Corollary 4.7 and Corollary 4.11, follows from [2], a preprint of which was circulated in July of 1993. This approach has also been continued to an independent proof of Theorem 4.12, see [3]. \(\bullet\) Lemma 4.13 is due to Keller and Nowicki, see [5]. \(\bullet\) On the level of detail with which the proof of Theorem 5.6 was given, there is no difference between the arguments of [L4] and [Sw]. By the way, Theorem 4.10 is referenced, which must be a misprint since only Lemma 4.10 is found in the paper. It should be noted that a fact equivalent to Lemma 4.10 appears in the second version of Świątek's proof, circulated in July of 1993, thus still prior to [L4]. \(\bullet\) The discussion in paragraph 7.3 ignores the paper of Hofbauer and Keller, see [4]. That paper provided a setting for the random walk argument in terms of the so-called Hofbauer tower construction and singled out the Fibonacci map as the most likely example with a wild attractor. It is likely that for the subsequent work of Keller this article was a more important inspiration than computer experiments made by the author and Sutherland. Also, the proof of the existence of wild attractors due to Bruin, Keller, Nowicki and Van Strien appeared as a preprint in 1993 and more recently in [1]. \(\bullet\) A stronger version of Theorem 6.3 is due to Nowicki and Van Strien, see [6]. In summary, the paper should be a useful reading for non-specialists once they have been cautioned not to take it to be the whole story. For specialists, it is also strongly recommended as an outspoken voice in the controversy. References. [1] \textit{H. Bruin}, \textit{G. Keller}, \textit{T. Nowicki}, and \textit{S. Van Strien}, Ann. Math., II. Ser. 143, No. 1, 97-130 (1996; Zbl 0848.58016) [2] \textit{J. Graczyk} and \textit{G. Świątek}, Ann. Sci. Éc. Norm. Super., IV. Ser. 29, No. 4, 399-482 (1996) [3] \textit{J. Graczyk} and \textit{G. Świątek}, `Polynomial-like property for real quadratic polynomials', to appear in Topology Proceedings [4] \textit{F. Hofbauer} and \textit{G. Keller}, Ann. Inst. Henri Poincaré, Phys. Théor. 53, No. 4, 413-425 (1990; Zbl 0721.58018). [5] \textit{G. Keller} and \textit{T. Nowicki}, Ergodic Theory Dyn. Syst. 15, No. 1, 99-120 (1995; Zbl 0853.58072). [6] \textit{S. Van Strien} and \textit{T. Nowicki}, `Polynomial maps with a Julia set of positive Lebesgue measure: Fibonacci maps', preprint IMS Stony Brook (1994). For the entire collection see [Zbl 0833.00016].
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4335075)
