The paper studies certain families of polynomials that generalize classical objects in combinatorics called Narayana polynomials. These polynomials appear in counting problems related to paths, trees, and other discrete structures, and they come in two flavors traditionally labeled type A and type B. The authors introduce a continuous parameter that interpolates between these two types, and then they add a second layer of generalization by introducing a parameter called q, which deforms the polynomials in a structured algebraic way. The starting point for building these deformations is a well-known tool from the theory of orthogonal polynomials called the Rogers (or continuous q-ultraspherical) functional, which provides a systematic way to extract moments from a weighted integral.
The central question the paper addresses is whether these new q-deformed polynomial families are "totally positive," meaning that certain infinite matrices built from their coefficients have all non-negative minors. Total positivity is a strong and desirable property with connections to probability, combinatorics, and analysis. The authors prove this for both families they construct, using different techniques for each. For the first family, they use moment representations tied to discrete probability measures. For the second, they use a factorization method involving objects called q-Bessel functions and a classical theorem due to Edrei about infinite products. Together these results extend and unify theorems previously known only for the pure type A and type B cases.
Beyond proving total positivity, the authors also investigate the relationship between their two q-deformed families. They show that an explicit change of basis, expressed as a triangular transformation, links the two families to each other and explains why they both reduce to the same classical polynomials when q equals 1. They also verify a weaker positivity condition on the polynomial coefficients themselves, up to second order in q, for the boundary cases of the interpolation. The overall effect is a coherent framework that unifies a collection of previously separate results about Narayana polynomials into a single two-parameter theory.