Arithmetic Functions and Geometry

Andrew Kobin · 2025

Read Arithmetic Functions and Geometry by Andrew Kobin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-13328.txt.

Introduction

Andrew Kobin revisits classic arithmetic functions in a 2025 expository note. The functions are Euler's totient, the divisor sum functions, and Dedekind's psi function. Kobin places them in a unifying algebraic framework that highlights their links to geometry. The framework builds on prior work with zeta functions and Möbius inversion. The goal is to give clear context for similar constructions in the future. The note also makes an original observation about Dedekind's psi function.

Catalog record

CREATORAndrew Kobin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-13328

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Arithmetic Functions and Geometry Andrew Kobin Abstract In this expository note, we revisit several classical arithmetic functions - namely Euler’s totient function, the divisor sum functions and Dedekind’s ψ \psi -function - within a unifying algebraic framework that highlights their connections to geometry. This framework builds on prior work involving zeta functions and Möbius inversion. While our main goal is to provide a clear context for similar constructions in the future, we also make an original observation regarding Dedekind’s ψ \psi -function. 1 Introduction In a recent article [ Kob ] , we gave a description of various zeta functions in number theory and algebraic geometry using a more abstract algebraic framework. This framework generalizes some classical notions in the theory of arithmetic functions, including the Möbius inversion principle (also see [ Koc ] for part of the story). In this note, we give a similar treatment to several other arithmetic functions, including the totient function (Section 2 ), the divisor sum functions (Section 3 ) and Dedekind’s \psi -function (Section 4 ). A key theme is that these functions have certain geometric origins, which we will begin discussing in Section 1.3 . Many of the ideas present in this article are well-known to experts and we make no claims of originality, with one exception in Section 4 . Instead, our aim is to tell the story of these arithmetic functions through a modern, often geometric lens and motivate future study of similar functions (see Section 5 ). 1.1 Generalities on arithmetic functions Classically, an arithmetic function is any function f:\mathbb{N}\rightarrow\mathbb{C} , possibly with some conditions such as weak multiplicativity that we will not adopt here. To every such f , we can assign a Dirichlet series F(s)=\sum_{n=1}^{\infty}\frac{f(n)}{n^{s}},\qquad s\in\mathbb{C}, which encodes the arithmetic of f complex-analytically. For example, the arithmetic function \zeta:\mathbb{N}\rightarrow\mathbb{C} defined by \zeta(n)=1 for all n\in\mathbb{N} corresponds to the Riemann zeta function \zeta_{\mathbb{Q}}(s) . One thing that makes the set of arithmetic functions interesting is that it admits the structure of a commutative \mathbb{C} -algebra by defining (f*g)(n)=\sum_{ij=n}f(i)g(j), called the Dirichlet convolution of f and g . More intuitively, f*g is defined by ordinary multiplication of Dirichlet series: \left(\sum_{n=1}^{\infty}\frac{f(n)}{n^{s}}\right)\left(\sum_{n=1}^{\infty}\frac{g(n)}{n^{s}}\right)=\sum_{n=1}^{\infty}\frac{(f*g)(n)}{n^{s}}. 1.2 Abstract arithmetic functions The \mathbb{C} -algebra of classical arithmetic functions is merely a special case of the more general notion of an incidence algebra , as surveyed in [ Kob , Koc ] . In the most general version of the story, an algebra of “objective arithmetic functions” can be attached to any locally finite decomposition space (or locally finite 2 -Segal space). For concreteness, we restrict the discussion here to locally finite monoids and their vector space incidence algebras. Fix a field k . For a monoid M , the incidence algebra I(M) is defined to be the k -vector space of functions f:M\rightarrow k equipped with a convolution product * defined by (f*g)(x)=\sum_{ab=x}f(a)g(b) for all x\in M . When M is a locally finite monoid 1 1 1 M is locally finite if for each x\in M there are only finitely many a,b\in M with ab=x . , this gives I(M) the structure of an associative, unital k -algebra, with unit \delta(x)=\delta_{x,1} , where 1 is the identity element of M . Each locally finite monoid M admits a zeta function \zeta\in I(M) defined by \zeta(x)=1 for all x\in M . Möbius inversion asserts that \zeta is invertible in I(M) , with inverse given by \mu(x)=\begin{cases}1,&\text{if }x=1\\ -\sum_{ab=x}\mu(a),&\text{if }x\not=1.\end{cases} The usual application of Möbius inversion is to invert a relation between two functions f,g\in I(M) , e.g. convolving both sides of f=g*\zeta with \mu yields f*\mu=g , or: f(x)=\sum_{ab=x}g(a)\quad\text{implies}\quad g(x)=\sum_{ab=x}f(a)\mu(b). (1) The goal of this short note is to highlight several pairs of arithmetic functions (f,g) satisfying formula ( 1 ) and extend them to a more general arithmetic geometric context. 1.3 Arithmetic functions coming from geometry Many arithmetic functions “come from geometry”, a phrase which we will not give precise meaning to. Instead, we will explore the geometric flavor of several examples, even when this connection is not typically mentioned in the literature. To begin with, the functions \zeta and \mu that form the basis for Möbius inversion have well-known geometric interpretations. The Riemann zeta function \zeta_{\mathbb{Q}}(s) can be viewed as the zeta function \zeta_{\operatorname{Spec}\mathbb{Z}}(s) attached to the scheme \operatorname{Spec}\mathbb{Z} and more generally, Dedekind zeta functions of number fields and Hasse–Weil zeta functions of varieties can be interpreted as zeta functions for schemes over \operatorname{Spec}\mathbb{Z} . This often provides a geometric explanation for relations between these different functions; cf. [ Kob ] . In fact, these zeta functions are mere shadows 2 2 2 In technical terms, they are decategorifications of generating functions whose coefficients lie in some Grothendieck ring. of an abstract zeta function, such as Kapranov’s motivic zeta function Z_{mot}(X,t)=1+\sum_{n=1}^{\infty}[\operatorname{Sym}^{n}X]t^{n}, which lies in the power series ring 1+tK_{0}({\tt Sch}_{X})[[t]] , where K_{0}({\tt Sch}_{X}) is the Grothendieck ring of the category of X -schemes. Here, \operatorname{Sym}^{n}X denotes the n th symmetric power of X , parametrizing unordered n -tuples of points in X . Choosing X appropriately and applying a motivic measure to the coefficients of Z_{mot}(X,t) recovers the more familiar zeta functions in the previous paragraph. Like \zeta , the classical Möbius function \mu:\mathbb{N}\rightarrow\mathbb{C} is an element of the incidence algebra I(\mathbb{N}^{\times}) , where \mathbb{N}^{\times} is the multiplicative monoid of natural numbers. It too is a shadow of a motivic generating function, namely M_{mot}(X,t)=1+\sum_{\mu}(-1)^{||\mu||}[\operatorname{Sym}^{\mu}X]t^{|\mu|}=1+\sum_{n=1}^{\infty}\sum_{\mu\vdash n}(-1)^{||\mu||}[\operatorname{Sym}^{\mu}X]t^{n} where the sum is over all partitions \mu=(\mu_{1},\ldots,\mu_{k}) of natural numbers n\in\mathbb{N} , with |\mu|:=\sum_{i=1}^{k}\mu_{i}=n , ||\mu|| equal to the number of distinct elements of \mu and \operatorname{Sym}^{\mu}X:=\coprod_{i=1}^{k}\operatorname{Sym}^{\mu_{i}}X . This can be rewritten [ VW , Prop. 3.7] M_{mot}(X,t)=1+\sum_{\mu}(-1)^{||\mu||}[\operatorname{Conf}^{\mu}X]t^{|\mu|}, where \operatorname{Conf}^{\mu}X\subseteq\operatorname{Sym}^{\mu}X is a configuration space , i.e. the open subscheme parametrizing distinct tuples of points. For a lengthier discussion of generating functions involving configuration spaces, see [ VW , Sec. 5] . Remark 1.1 . As an amusing sidenote, the other famous mathematical object named for August Ferdinand Möbius, the Möbius strip, is also a configuration space: \operatorname{Conf}^{2}(S^{1}) . Another example worth mentioning here is the partition function p(n) , which counts the number of ways of writing n as a sum of positive integers. On the geometric side, the values of p(n) are encoded in the Hilbert schemes \operatorname{Hilb}^{n}\mathbb{A}^{2} , parametrizing collections of n unordered points 3 3 3 To be more accurate, while \operatorname{Sym}^{n}\mathbb{A}^{2} parametrizes n unordered points in the plane, \operatorname{Hilb}^{n}\mathbb{A}^{2} parametrizes such points together with an infinitesimal structure when points collide . In fancier terms, the natural map \operatorname{Hilb}^{n}\mathbb{A}^{2}\rightarrow\operatorname{Sym}^{n}\mathbb{A}^{2} is a desingularization. in \mathbb{A}^{2} . Specifically, p(n)=\chi(\operatorname{Hilb}^{n}\mathbb{A}^{2}) , where \chi denotes the Euler characteristic. We can thus view the partition generating function \sum_{n=0}^{\infty}p(n)t^{n} , originally studied by Euler, as a shadow of the motivic generating function P_{mot}(\mathbb{A}^{2},t)=1+\sum_{n=1}^{\infty}[\operatorname{Hilb}^{2}\mathbb{A}^{2}]t^{n}. For a version of this relation when \mathbb{A}^{2} is replaced by a smooth surface X , see [ Göt ] . 1.4 Overview For the rest of this article, we will highlight more arithmetic functions and discuss their possible geometric origins. We summarize things in the following table. Function Symbol Formula Geometric interpretation zeta \zeta \zeta(n)=1 symmetric powers (Sec. 1.3 ) \begin{matrix}\text{M\"{o}bius}\\ \text{(Sec.~\ref{sec:fromgeom})}\end{matrix} \mu \mu(n)=\begin{cases}1,&\text{if }n=1\\ -\sum_{ab=n}\mu(a),&\text{if }n\not=1\end{cases} configuration spaces partition p p(n)=\#\{\lambda\vdash n\} Hilbert schemes (Sec.

Read the complete text