On the dichotomy of Perron numbers and beta-conjugates (Q1022353)

Let \(\beta > 1\) be an algebraic integer, \(d_\beta(1) = 0.t_1t_2\dots\) denote the \(\beta\)-expansion of \(1\), and \(f_\beta(z) = -1 + \sum_{i \geq 1} t_i z^i\). If the latter is not a polynomial then the radius of convergence of the power series is \(1\). The beta-conjugates of \(\beta\) are the reciprocals of the zeros of \(f_\beta(z)\) other than \(\beta\) and its Galois conjugates. The dichotomy referred to in the title is whether \(f_\beta(z)\) is a rational function, or whether it is an analytic function inside the unit disk with the unit circle as its natural boundary, these being the only two possibilities according to a well-known theorem of Szegö. The author is interested in the angular distribution of the beta-conjugates relative to the Galois conjugates. He surveys results and conjectures from the literature that bear on this question, for Pisot, Salem and Perron numbers that are Parry numbers (i.e. the first case of the dichotomy). (Parry numbers were called beta-numbers by Parry, but the new designation is sensible and welcome). The examples are well chosen and interesting. No definitive general answers are obtained, but some interesting conjectures are proposed.