Jordan canonical forms of Riordan arrays (Q2038845)

A Riordan array is an infinite lower triangular matrix \(R=[r_{nk}]\) that is identified with a pair of formal power series \((d(t),h(t))\) where \(r_{nk}\) is the coefficient of \(t^n\) in \(d(t)(h(t))^n\). The Jordan form of \(R\) is a triangular matrix \(J_R=X^{-1}RX\) that is a direct sum of possibly infinite Jordan cells (Jordan blocks). This note is devoted to proving two results. Precisely, Theorem 1 characterises when the Jordan form \(J_R\) is itself a Riordan array. Theorem 2 characterises when the transition matrix \(X\) is a Riordan array.