Cumulants of the q-semicircular law, Tutte polynomials, and heaps

Abstract

The q-semicircular distribution is a probability law that interpolates between the Gaussian law and the semicircular law. There is a combinatorial interpretation of its moments in terms of matchings where q follows the number of crossings, whereas for the free cumulants one has to restrict the enumeration to connected matchings. The purpose of this article is to describe combinatorial properties of the classical cumulants. We show that like the free cumulants, they are obtained by an enumeration of connected matchings, the weight being now an evaluation of the Tutte polynomial of a so-called crossing graph. The case q=0 of these cumulants was studied by Lassalle using symmetric functions and hypergeometric series. We show that the underlying combinatorics is explained through the theory of heaps, which is Viennot's geometric interpretation of the Cartier-Foata monoid. This method also gives results for the classical cumulants of the free Poisson law.

14 pages

References in corpus (2)

• The multivariate Tutte polynomial (alias Potts model) for graphs and matroids
• The normal distribution is $⊞$-infinitely divisible

Cited by in corpus (5)

• Relations between cumulants in noncommutative probability
• Monotone, free, and boolean cumulants: a shuffle algebra approach
• Bernoulli variables, classical exclusion processes and free probability
• High temperature expansion of double scaled SYK
• Macdonald cumulants, $G$-inversion polynomials and $G$-parking functions