Continued fractions, discrete groups and complex dynamics (Q1851480)

This essay is directed to the proposition that much research currently being carried out in the theory of continued fractions might with profit and propriety be placed in a more general setting. Detailed treatment of this theme is given to the analytic theory of continued fractions with regard to Möbius maps in hyperbolic space. Some new results are given -- for example there are extensions of the Stern-Stolz theorem for complex numbers (if \(\{\sum |b(i)|\mid i\geq 1\} \) converges, then \(1/[b(1)+ 1/[b(2)+ \cdots ]]\) diverges by oscillation) -- but they serve merely to prop the narrative. The presentation is pleasantly pontifical. It is not invidious to suggest that future ventures indicated by the author will experience a common fate. As has been remarked by the reviewer [The work of E. B. Christoffel on the theory of continued fractions, in E. B. Christoffel Festschrift (Birkhäuser, Basel) (1980)] the theory of continued fractions often functions as a breakthrough device, to be discarded after initial utility. Perhaps the clearest evidence supporting this assertion is Khinchin's metric theory of continued fractions [\textit{A. Ya. Khinchin}, Continued fractions (The University of Chicago Press, Chicago and London) (1964; Zbl 0117.28601)]: the continuum is now investigated by other methods. A further example may well be provided by \textit{L. Jacobsen} [Math. Scand. 60, 129-147 (1987; Zbl 0611.40002)] on the simultaneous convergence and simultaneous divergence of neighbouring continued fractions, which may well lead to a general theory of iterative processes in which the original continued fraction example plays but an insignificant role. To proceed further, in compex variable theory the partial sums of the power series expansion of a linear fractional function \(C(z)\) are polynomials and may be expressed in the form \(S(i|z)= C(z)+ B(z)(z/\alpha)^i\) (\(i\geq 0\)) (\(C\) and \(B\) are both linear fractional): there is a simple continued fraction algorithm for reconstructing \(C\) from three successive \(S(i)\). The \(S(i)\) lie on a spiral: the algorithm may be represented as a method, in which computational steps have been replaced by projections and inversions, for locating the centre of a spiral from three successive points on it. This all extends to higher order spirals (a second order spiral is formed from one first order spiral whose centre moves on another, and so on) to \(n\) complex dimensions and also to points in a plane not having Desargue's property (the latter is treated by the reviewer in [Math. Balk. 5, 299-318 (1975; Zbl 0381.17001)] but becomes a not very important contribution to the theory of non-Desarguean geometry). The author gives an extensive bibliography but does not mention one important work which may well be of interest to the general reader - the delightful monograph in Romanian [\textit{G. Sudan}, Geometrizarea fracţiilor continue [Geometrization of continued fractions] (Editura Tehnica, Bucureşti) (1959; Zbl 0090.26302)]. It deals with many topics (e.g. Ford's theory of rational approximation to irrationals) touched on by the author and seems to be quite unknown among research workers in the field.