Algebraic independence of Fredholm series (Q2773294)

The following Theorem 1 is the main result of the present paper. Suppose that for each integer \(d \geq 2\) all \(\sigma_{d,h}\) \((h=0,1,\ldots)\) belong to a finite set of non-zero algebraic numbers, and define \(f_d\) in the unit circle by the series \(\sum_{h\geq 0} \sigma_{d,h} z^{d^h}\). Then, for any \(\alpha\in\overline{\mathbb{Q}}\) with \(0 <|\alpha|< 1\), the values \(f_d(\alpha)\) \((d=2,3,\ldots)\) are algebraically independent. NEWLINENEWLINENEWLINETheorem 1 is deduced from Theorem 2 which, in turn, gives a sufficient criterion for the algebraic independence of the values of certain power series \((\ast)\): \(f_1^{(0)},\ldots,f_m^{(0)}\) in \(n\) complex variables at points \(\alpha = (\alpha_1,\ldots,\alpha_n)\in \overline{\mathbb{Q}}^n\) with \(0 <|\alpha_\nu|< 1\) for \(\nu = 1,\ldots,n\). The main hypothesis is that each function \(f_\mu^{(0)}\) from \((\ast)\) is the first term of an infinite sequence \((f_\mu^{(k)})_{k=0,1,\ldots}\) of power series such that \(f_\mu^{(0)}(z)-f_\mu^{(k)}(\Omega^{(k)} z)\) is a polynomial in \(z =(z_1,\ldots,z_n)\) ``under control''. Here \((\Omega^{(k)})_{k=0,1,\ldots}\) is a sequence of \({n\times n}\) matrices with non-negative integer entries, \(\Omega^{(0)} := E\), as they typically appear in Mahler's method. NEWLINENEWLINENEWLINEThe proof of Theorem 2 follows rather closely the proof of Theorem 3.3.2 of the author's monograph [Mahler functions and transcendence, Lect. Notes Math. 1631, Springer (1996; Zbl 0876.11034)]; this is now standard in the present subject. Worthwhile is the author's deduction of Theorem 1 from Theorem 2 where new reasonings due to \textit{D. W. Masser} [Q. J. Math., Oxf. (2) 50, 207-230 (1999; Zbl 0929.11021)] come into play.