Dirichlet convolution¶

A binary operation on arithmetic functions defined by summing f(d)g(n/d) over the positive divisors d of n.

Version: v2 · 2026-09-19 · History
Domain-specific #: 4196
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomain: Multiplicative Number Theory → Mathematics

Core Idea¶

The operation is associative and commutative over a commutative coefficient ring, has the delta function as identity, preserves multiplicativity and turns multiplication of Dirichlet series into convolution of coefficients. Every factorization n=ab contributes the product f(a)g(b); regrouping triples proves associativity, and recursive cancellation constructs inverses for functions whose value at one is invertible. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

How would you explain it like I'm…

All-the-Ways-to-Multiply Mix

Dirichlet convolution is a way to mix two lists of numbers. For a number like 6, you find every way to make it by multiplying two numbers — 1×6, 2×3, 3×2, 6×1. For each way, you multiply the matching entries from the two lists, then add all those answers up.

Mixing Rules Over Factor Pairs

An arithmetic function is a rule that gives a number for each counting number 1, 2, 3 and so on. Dirichlet convolution combines two such rules, f and g, into a new one. To find the new rule's value at a number n, you list every way to write n as a times b, compute f(a) times g(b) for each way, and add them all up. For n = 6 that means the pairs (1,6), (2,3), (3,2) and (6,1). This operation behaves nicely: the order doesn't matter, and there's a special rule that acts like 'times 1' and changes nothing.

Divisor-Sum Convolution

Dirichlet convolution combines two arithmetic functions f and g (functions on positive integers) into f * g, where (f * g)(n) is the sum of f(a)g(b) over all pairs of positive integers with ab = n, equivalently the sum over divisors d of n of f(d)g(n/d). Because every n has finitely many divisors, the sum is finite. It is associative and commutative (when the values come from a commutative ring), and its identity is the delta function that is 1 at n = 1 and 0 elsewhere. A function has a convolution inverse exactly when its value at 1 is invertible; the inverse is built step by step by cancellation. For example, the inverse of the constant function 1 is the Möbius function, which gives Möbius inversion. If f and g are multiplicative, so is f * g, and multiplying Dirichlet series corresponds to convolving their coefficients.

Dirichlet convolution of arithmetic functions f, g: ℕ⁺ → R is (f * g)(n) = Σ_{ab=n} f(a)g(b) = Σ_{d|n} f(d)g(n/d), a finite sum over positive divisors. Over a commutative coefficient ring it is associative (by regrouping factorizations n = abc) and commutative, with identity ε, where ε(1) = 1 and ε(n) = 0 for n > 1, making the arithmetic functions a commutative ring. A function f is invertible if and only if f(1) is a unit, with the inverse constructed recursively from (f * f⁻¹)(n) = 0 for n > 1. The Möbius function μ is the inverse of the constant function 1, which is the content of Möbius inversion. The convolution of multiplicative functions is multiplicative. Formally, and where the series converge absolutely, the product of Dirichlet series Σ f(n)n⁻ˢ and Σ g(n)n⁻ˢ is the Dirichlet series of f * g.

Scope of Application¶

Dirichlet convolution belongs to multiplicative number theory and is useful where the analyst can specify the typed multiplicative number theory carrier, defining objects and relations, parameters, conventions, evidence, boundary cases, and comparison targets, then evaluate the arithmetic-function domain and coefficient ring, positive-divisor convention, convolution formula, finite sum, identity function, associativity and commutativity assumptions, invertibility criterion, multiplicative-function closure, Möbius inverse and Dirichlet-series convergence qualifications are explicit. The scope is broad within that domain but bounded by the need for the arithmetic-function domain and coefficient ring, positive-divisor convention, convolution formula, finite sum, identity function, associativity and commutativity assumptions, invertibility criterion, multiplicative-function closure, Möbius inverse and Dirichlet-series convergence qualifications are explicit.

Clarity¶

The abstraction clarifies a crowded vocabulary by making the arithmetic-function domain and coefficient ring, positive-divisor convention, convolution formula, finite sum, identity function, associativity and commutativity assumptions, invertibility criterion, multiplicative-function closure, Möbius inverse and Dirichlet-series convergence qualifications are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test.

Manages Complexity¶

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Dirichlet convolution. Dirichlet convolution compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning¶

Identify the carrier. State what the elements, states, objects, or observations are: the typed multiplicative number theory carrier, defining objects and relations, parameters, conventions, evidence, boundary cases, and comparison targets. Reject examples whose alleged carrier belongs to a different problem.
Lock the constitutive rule. Express the arithmetic-function domain and coefficient ring, positive-divisor convention, convolution formula, finite sum, identity function, associativity and commutativity assumptions, invertibility criterion, multiplicative-function closure, Möbius inverse and Dirichlet-series convergence qualifications are explicit independently of one notation or implementation.

Knowledge Transfer¶

Knowledge transfers strongly among subfields of multiplicative number theory because they reuse the typed multiplicative number theory carrier, defining objects and relations, parameters, conventions, evidence, boundary cases, and comparison targets, Every factorization n=ab contributes the product f(a)g(b); regrouping triples proves associativity, and recursive cancellation constructs inverses for functions whose value at one is invertible., and type the carrier, state every parameter and convention in the definition, test that the arithmetic-function domain and coefficient ring, positive-divisor convention, convolution formula, finite sum, identity function, associativity and commutativity assumptions, invertibility criterion, multiplicative-function closure, Möbius inverse and Dirichlet-series convergence qualifications are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions¶

Current abstraction Dirichlet convolution Domain-specific

Parents (1) — outgoing typed relations

Dirichlet convolution is a kind of Convolution Prime

It remains its own entry because its identity is fixed by the arithmetic-function domain and coefficient ring, positive-divisor convention, convolution formula, finite sum, identity function, associativity and commutativity assumptions, invertibility criterion, multiplicative-function closure, Möbius inverse and Dirichlet-series convergence qualifications.

Hierarchy path (1) — routes to 1 parentless root

Dirichlet convolution → is a kind of Convolution → is a kind of Function (Mapping)

Neighborhood in Abstraction Space¶

Dirichlet convolution sits in a crowded region of the domain-specific corpus (12^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Integer Classifications & Arithmetic Functions (45 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25