[Paper Review] The classical umbral calculus, and the flow of a Drinfeld module

Dong Quan Ngoc Nguyen|arXiv (Cornell University)|May 9, 2014

Stochastic processes and financial applications6 references3 citations

TL;DR

This paper recasts flows in finite characteristic—central to David Goss's additive harmonic analysis in function fields—using the classical umbral calculus, providing a more conceptual and generalized framework. By modeling sequences of functions as evaluations of umbrae under a linear functional, the authors derive generalized flow formulas without relying on differential Fourier transforms, and establish duality between flow maps, confirming Goss's duality as natural and extending his framework to broader settings including Drinfeld modules.

ABSTRACT

David Goss developed a very general Fourier transform in additive harmonic analysis in the function field setting. In order to introduce the Fourier transform for continuous characteristic $p$ valued functions on $\mathbb{Z}_p$, Goss introduced and studied an analogue of flows in finite characteristic. In this paper, we use another approach to study flows in finite characteristic. We recast the notion of a flow in the language of the classical umbral calculus, which allows to generalize the formula for flows first proved by Goss to a more general setting. We study duality between flows using the classical umbral calculus, and show that the duality notion introduced by Goss seems a natural one. We also formulate a question of Goss about the exact relationship between two flows of a Drinfeld module in the language of the classical umbral calculus, and give a partial answer to it.

Motivation & Objective

• To generalize Goss's flow formulas in finite characteristic beyond additive functions using umbral calculus.
• To provide a conceptual framework for flows in function fields by interpreting formal substitutions as umbral evaluations.
• To clarify the duality between flow maps in Goss's theory using umbral algebraic structures.
• To address Goss's open question on the relationship between two distinct flows of a Drinfeld module using umbral techniques.

Proposed method

• Introduce an umbral map that encodes sequences of functions as evaluations of powers of an umbra under a linear functional.
• Define a flow map via the umbral exponential operator, generalizing the classical Taylor expansion to non-Archimedean fields.
• Use the classical umbral calculus to interpret Goss's formal power series as evaluations in a topological umbral algebra.
• Incorporate non-Archimedean topology into the umbral framework to extend the domain of the evaluation map to bounded power series.
• Establish duality between flow maps by exploiting the duality in the umbral algebra and the pairing between umbrae and power series.
• Apply the framework to Drinfeld modules, recovering both the naive and twisted flows as special cases of the generalized umbral flow.

Experimental results

Research questions

• RQ1 How can the notion of a flow in finite characteristic be reinterpreted using the classical umbral calculus to reduce formality and increase conceptual clarity?
• RQ2 What is the precise algebraic structure underlying the duality between flow maps in Goss's theory, and is it naturally explained via umbral duality?
• RQ3 Can the relationship between the naive flow and the twisted flow of a Drinfeld module be fully characterized using umbral techniques?
• RQ4 To what extent can Goss's flow formula be generalized beyond sequences arising from additive functions?
• RQ5 Is there a canonical umbral structure that captures the behavior of Bernoulli–Carlitz numbers in function fields, analogous to the role of Bernoulli numbers in classical umbral calculus?

Key findings

• The generalized flow formula holds over any complete non-Archimedean field, not just those arising from additive functions, broadening the scope of Goss's original result.
• The duality between flow maps is naturally explained through the duality in the umbral algebra, confirming that Goss's duality is algebraically canonical.
• The naive flow of a Drinfeld module is recovered as a special case of the umbral flow when the umbra corresponds to the exponential of the module.
• The twisted flow is shown to arise from a different umbral structure, and its relationship to the naive flow is clarified via the evaluation of symmetric functions in umbrae.
• The paper provides a partial answer to Goss's question by expressing the two flows in terms of distinct umbral maps whose evaluations yield the respective flow operators.
• The umbral framework allows for a unified treatment of flows in finite characteristic, replacing the need for differential Fourier transforms on measures.