Divided differences and multivariate holomorphic calculus

Abstract

We review the multivariate holomorphic functional calculus for tuples in a commutative Banach algebra and establish a simple “naïve” extension to commuting tuples in a general Banach algebra. The approach is naïve in the sense that the naïvely defined joint spectrum maybe too big. The advantage of the approach is that the functional calculus then is given by a simple concrete formula from which all its continuity properties can easily be derived.
We apply this framework to multivariate functions arising as divided differences of a univariate function. This provides a rich set of examples to which our naïve calculus applies. Foremost, we offer a natural and straightforward proof of the Connes–Moscovici Rearrangement Lemma in the context of the multivariate holomorphic functional calculus. Secondly, we show that the Daletski–Krein type noncommutative Taylor expansion is a natural consequence of our calculus. Also Magnus’ Theorem which gives a nonlinear differential equation for the $\log$ of the solutions to a linear matrix ODE follows naturally and easily from our calculus. Finally, we collect various combinatorial related formulas.