Polynomials with no zeros on a face of the bidisk (Q265507)

This paper is devoted to the extension of the results proved in a previous work [the first and the second author, J. Eur. Math. Soc. (JEMS) 16, No. 9, 1849--1880 (2014; Zbl 1303.42010)], where the authors have given a complete characterization of the positive trigonometric polynomials \(Q(\theta,\varphi)\) on the bicircle. Let \(\mathbb D\) be the unit disk, \(\mathbb D=\{z\in\mathbb C:\, | z| <1\}\). Let us consider the measures on \(\mathbb T^2=(\partial\mathbb D)\times (\partial\mathbb D)\) of the form NEWLINE\[NEWLINE\frac{| dz| \, | dw| }{(2\pi)^2\, | p(z,w)| ^2}NEWLINE\]NEWLINE where \(p\) is a polynomial. These are called Bernstein-Szegő measures.NEWLINENEWLINEWith ``the split-shift orthogonality condition'' one denotes the existence of a special orthogonal decomposition of spaces of polynomials. Instead, the ``matrix condition'' involves certain natural truncated shift operators, and it is related to an earlier characterization due to Jeffrey S. Geronimo and Plamen Iliev. These are two different characterizations which use a Hilbert space structure naturally associated to the trigonometric polynomial. In Section 6, the authors prove that the matrix condition is equivalent to the split-shift condition.NEWLINENEWLINEIn Section 4, the authors show that a Bernstein-Szegő measure with \(p(z,w)\neq 0\) for \(| z| =1\), \(| w| \leq 1\) implies the split-shift condition using a decomposition of \(p(z,0)\) into stable and unstable factors. In Section 5 one shows that the split-shift condition implies the existence of a Bernstein-Szegő measure. In Section 9, the authors give necessary and sufficient conditions in terms of moments when a bivariate Borel measure supported on the bicircle is a Bernstein-Szegő measure with \(p\) nonzero for \(| z| =1\), \(| w| \leq 1\).NEWLINENEWLINEFinally, Section 10 is devoted, among others, to give a different proof of the sums of squares formula (Theorem 2.4), obtaining a determinantal representation which generalizes one of the main theorems in \textit{J. Agler} and \textit{J. E. McCarthy}'s paper [Acta Math. 194, No. 2, 133--153 (2005; Zbl 1102.32005); correction ibid. 205, No. 1, 1--3 (2010)].