Some properties of generalized hypergeometric coefficients (Q919501)

It is possible to generalize the theory of hypergeometric series to several variables by introducing partitions of integers as the index set and using some classical theory of symmetric functions. This paper defines \({}_ 2F_ 1(a,b;c;z)\) where a, b, c are complex parameters and \(z=(z_ 1,z_ 2,...,z_ n)\in {\mathbb{C}}^ n\), for some \(n=1,2,3... \). This is a sum over partitions \((\mu_ 1,\mu_ 2,...,\mu_ n)\) with n parts (that is \(\mu_ 1\geq \mu_ 2\geq \mu_ 3\geq...\geq \mu_ n\geq 0\) and each \(\mu_ i\in {\mathbb{Z}}_+)\). The authors define a ``generalized hypergeometric coefficient'' depending on a, b, c, and \(\mu\). The series is then the sum of these coefficients multiplying the Schur functions (usually denoted \(s_{\mu})\) in z. There is some discussion of related series, e.g. \({}_ 1F_ 0(a;z)=\prod^{n}_{j=1}(1-z_ j)^{-a}\) and how they can be used to establish summation formulas for the generalized coefficients. Those are analogues of the Chu-Vandermonde and the Saalschütz sums. Integral formulas of Euler type for the \({}_ 2F_ 1\) functions are discussed. The application to analyzing tensor products of representations of SU(3) is presented in the context of the multiplication of two hypergeometric series.