Skip to content

Divisor Function

The multiplicative arithmetic-function family \(\sigma_z(n)=\sum_{d\mid n}d^z\), including divisor count and divisor sum as distinguished cases.

Version
v1 · 2026-08-30 · History
Domain-specific #
1702
Origin domain
number theory
Subdomain
arithmetic functions

Core Idea

The divisor functions are the arithmetic-function family

\[ \sigma_z(n)=\sum_{d\mid n}d^z, \]

where \(n\) is a positive integer, the sum ranges over its positive divisors, and \(z\) is real or complex. NIST's Digital Library of Mathematical Functions uses this definition and identifies \(\sigma_0(n)\) with the divisor-counting function and \(\sigma_1(n)\) with the sum of divisors.[1] When the phrase “the divisor function” appears without a parameter, it often means \(d(n)=\tau(n)=\sigma_0(n)\); notation must therefore be declared.

The family converts the divisor lattice of one integer into a weighted aggregate. Its most powerful property is multiplicativity: if \(\gcd(m,n)=1\), then \(\sigma_z(mn)=\sigma_z(m)\sigma_z(n)\). Hence prime factorization turns an apparently global divisor enumeration into independent prime-power calculations. If \(n=\prod_i p_i^{a_i}\), then

\[ \sigma_z(n)=\prod_i\left(1+p_i^z+p_i^{2z}+\cdots+p_i^{a_i z}\right), \qquad \tau(n)=\prod_i(a_i+1). \]

This divisor-sum, prime-power, and multiplicative reconstruction package is the stable abstraction. Apostol treats arithmetical functions and Dirichlet multiplication as foundational machinery of analytic number theory.[2]

Structural Signature

Recognition roles: a positive integer \(n\) — its complete set of positive divisors — a declared exponent \(z\) — the weight \(d^z\) — finite aggregation over \(d\mid n\) — coprime multiplicativity — prime-power factorization — specializations \(\sigma_0\) and \(\sigma_1\).

  • Positive-integer input. The standard family is defined on \(n\ge1\).
  • Complete divisor domain. Both \(1\) and \(n\) are included; “proper divisors” defines the related aliquot sum instead.
  • Power weight. The exponent controls whether divisors are counted, summed, reciprocally weighted, or otherwise transformed.
  • One output per input. The result is an arithmetic function, not the set of divisors itself.
  • Multiplicativity, not complete multiplicativity. Products split only for coprime inputs; generally \(\sigma_z(p^{a+b})\ne\sigma_z(p^a)\sigma_z(p^b)\).
  • Prime-local reconstruction. Values on prime powers determine the function on all positive integers.
  • Analytic transform. In its convergence region, the Dirichlet series factors as \(\sum_{n\ge1}\sigma_z(n)n^{-s}=\zeta(s)\zeta(s-z)\).[3]

Recognition test: identify the divisor set and weights, verify the value on a prime power, and test multiplicativity on coprime inputs. A function summing over residues, prime divisors only, or proper divisors only is a neighbor, not the same family.

What It Is Not

  • It is not the divisor set. The function aggregates divisors into a number.
  • It is not a divisor summatory function such as \(\sum_{n\le x}\tau(n)\), which aggregates divisor-function values across inputs.
  • It is not the aliquot sum \(s(n)=\sigma_1(n)-n\), which excludes \(n\) itself.
  • It is not Euler's totient \(\varphi(n)\), which counts residues coprime to \(n\).
  • It is not Pillai's arithmetical function \(\sum_{k=1}^n\gcd(k,n)\). Pillai's function can be regrouped over divisors but uses totient-weighted gcd classes.
  • It is not the generalized \(k\)-fold divisor function \(d_k(n)\) counting ordered \(k\)-factorizations, except that \(d_2(n)=\tau(n)\).
  • It is not a generic function of a divisor. The complete divisor sum and power weighting are identity-bearing.

Scope of Application

In elementary number theory, \(\tau(n)\) counts factors, \(\sigma_1(n)\) classifies perfect, deficient, and abundant numbers through comparison with \(2n\), and \(s(n)\) generates aliquot sequences. In multiplicative number theory, the prime-power product supplies exact evaluation and Euler products. In analytic number theory, summatory behavior leads to the Dirichlet divisor problem; DLMF records

\[ \sum_{n\le x}d(n)=x\log x+(2\gamma-1)x+O(\sqrt{x}) \]

as the classical elementary error bound, while explicitly separating the unresolved optimal exponent question.[4]

Divisor functions also appear in zeta-function identities and modular-form coefficient formulas. Those applications use the same arithmetic function; they do not broaden the definition. The family can take complex \(z\), but applied claims about order, positivity, or inequalities must state parameter restrictions.

Clarity

The parameterized notation resolves recurring ambiguity. “Number of divisors” is \(\sigma_0\), “sum of divisors” is \(\sigma_1\), and “sum of reciprocal divisors” is \(\sigma_{-1}\). Writing \(\sigma_z\) prevents a proof about one specialization from being silently applied to another.

Prime factorization also clarifies mechanism. The value is not mysterious global behavior: every divisor chooses an exponent \(0\le j_i\le a_i\) independently at each prime. Summing \(d^z\) therefore becomes a product of finite geometric sums. Evidence fails to identify the function when a text says only “divisor statistic” without specifying whether all positive divisors and which weights are used.

Manages Complexity

Naively, evaluating \(\sigma_z(n)\) suggests scanning integers up to \(n\). Once the factorization \(n=\prod p_i^{a_i}\) is known, the product formula reduces evaluation to \(\sum_i(a_i+1)\)-scale local terms rather than enumerating all candidates. The compression is structural: a divisor corresponds bijectively to one choice of exponent at each prime.

The abstraction keeps the factorization, exponent parameter, inclusion convention, and arithmetic domain explicit. It discards the ordering of divisors because addition is commutative. Computational cost is not magically eliminated—integer factorization may be hard—but proofs and exact formulas can reason from a supplied factorization.

Abstract Reasoning

Multiplicativity: when \(\gcd(m,n)=1\), every divisor of \(mn\) factors uniquely as \(ab\) with \(a\mid m,b\mid n\), so \((ab)^z=a^zb^z\) and the double sum factors. Prime-power evaluation: divisors of \(p^a\) are exactly \(1,p,\ldots,p^a\). Dirichlet convolution: if \(\operatorname{id}_z(n)=n^z\), then \(\sigma_z=\mathbf1*\operatorname{id}_z\), explaining the product of zeta functions in the Dirichlet series. Bounds and classification: \(\sigma_1(n)\) compares total divisor mass with \(2n\); \(\tau(n)\) records the size of the divisor lattice without weights.

These deductions depend on positive divisors and unique factorization in the integers. Extending the symbol to ideals, polynomials, or other monoids requires a new divisor notion and is not automatic.

Knowledge Transfer

Within number theory, the full mechanism transfers literally among exact computation, multiplicative-function proofs, Euler products, average orders, perfect-number questions, and modular forms. The prime-local factorization and divisor aggregation remain unchanged.

Outside this domain, Aggregation and factorwise decomposition are portable parent structures, but the named divisor function is not. In algebraic number theory, analogous ideal-divisor sums are genuine extensions with different objects and norms; in data analysis, “divisor function” has no general transferable meaning.

Examples

The integer 12

The positive divisors of \(12\) are \(1,2,3,4,6,12\). Therefore

\[ \tau(12)=\sigma_0(12)=6, \qquad \sigma_1(12)=1+2+3+4+6+12=28. \]

Since \(12=2^2\cdot3\), the product formulas give \(\tau(12)=(2+1)(1+1)=6\) and \(\sigma_1(12)=(1+2+4)(1+3)=7\cdot4=28\). The two methods map the complete divisor set, power weight, aggregation, and prime-local reconstruction roles.

Coprime reconstruction

Take \(8=2^3\) and \(9=3^2\). They are coprime, so

\[ \sigma_1(72)=\sigma_1(8)\sigma_1(9) =(1+2+4+8)(1+3+9)=15\cdot13=195. \]

Direct factorization \(72=2^3 3^2\) gives the same product. By contrast, \(\sigma_1(4)=7\) but \(\sigma_1(2)^2=9\); this negative case demonstrates why multiplicative does not mean completely multiplicative.

Structural Tensions

T1: One familiar “divisor function” versus a parameterized family. Common usage reserves \(d(n)\) for counting, while \(\sigma_z\) names the wider power-sum family. Diagnostic: Has the exponent and notation been stated before using an identity?

T2: Global divisor enumeration versus prime-local product. The definition sums over all divisors, while the efficient structure factors across primes. Diagnostic: Is coprimality or a prime factorization available to justify the product step?

T3: Exact values versus aggregate asymptotics. Multiplicativity computes one input from its factors; analytic number theory studies sums over many inputs where exceptional factorization patterns matter. Diagnostic: Is the claim pointwise, average-order, normal-order, or summatory?

T4: Autonomous arithmetic family versus generic Aggregation. Aggregation explains many-to-one summation, but not divisibility, power weights, multiplicativity, prime-power formulas, or zeta identities. Diagnostic: Can the proposed reduction distinguish \(\sigma_z\) from aliquot, totient, and Pillai functions without restating the divisor-sum contract?

Structural–Framed Character

Divisor functions are mathematically structural. Their definition and identities do not depend on institutions or evaluation. Yet the term is domain-specific because positive-integer divisibility, arithmetic functions, Dirichlet convolution, and Euler products are indispensable.

Notation is convention-framed: \(d,\tau,\sigma_0\) may name the divisor count, and \(\sigma\) often means \(\sigma_1\). This human variation affects communication but not the underlying family once parameters are fixed.

Structural Core vs. Domain Accent

The portable core is weighted aggregation over a finite relation-defined set, followed by factorwise reconstruction when the relation decomposes. The domain accent is positive-integer divisibility, prime factorization, power weights, arithmetic multiplicativity, and Dirichlet series.

The candidate does not clear the prime bar; outside number theory the exact vocabulary and mechanism do not recur across unrelated substrates. It clears the domain-specific bar through a stable family identity, multiple specializations, diagnostics, formulas, and autonomous consequences.

Divisor Function presupposes prime:aggregation: each value collapses a complete weighted divisor set to one number. The proposed relation is compositional rather than subsumptive because an arithmetic function is not itself the generic act of aggregation.

domain_specific:natural_number supplies the input carrier but is not the closest mechanism. domain_specific:pillai_s_arithmetical_function is a sibling arithmetic function whose divisor-class formula uses different weights. prime:partition is declined because prime-exponent decomposition aids evaluation but does not define the divisor sum.

Relationships to Other Abstractions

Local relationship map for Divisor FunctionParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Divisor FunctionDOMAINPrime abstraction: Aggregation — presupposesAggregationPRIME

Current abstraction Divisor Function Domain-specific

Parents (1) — more general patterns this builds on

  • Divisor Function presupposes Aggregation Prime

    Divisor Function presupposes prime:aggregation: each value collapses a complete weighted divisor set to one number.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Divisor Function sits in a sparse region of the domain-specific corpus (69th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Computational Number Theory & Enumeration (13 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

  • Divisor summatory function: accumulates \(\tau(n)\) over \(n\le x\), rather than divisors within one \(n\).
  • Aliquot sum: excludes \(n\) itself and equals \(\sigma_1(n)-n\).
  • Euler totient: counts coprime residues, not divisors.
  • Pillai's arithmetical function: sums gcd values over residues and only later regroups by divisors.
  • Generalized divisor function \(d_k\): counts ordered \(k\)-factorizations; only \(d_2=\tau\).
  • Prime omega functions: count distinct or repeated prime factors, not all divisors.
  • Divisibility sequence: a sequence whose terms divide later terms under an index relation, not an arithmetic divisor sum.

References

[1] NIST Digital Library of Mathematical Functions, §27.2 “Functions,” especially Eq. 27.2.10, https://dlmf.nist.gov/27.2. registry ↩

[2] Tom M. Apostol, Introduction to Analytic Number Theory (Springer, 1976), chapters on arithmetical functions and Dirichlet multiplication, https://doi.org/10.1007/978-1-4757-5579-4. registry ↩

[3] NIST Digital Library of Mathematical Functions, §27.4 “Euler Products and Dirichlet Series,” especially Eq. 27.4.11, https://dlmf.nist.gov/27.4. registry ↩

[4] NIST Digital Library of Mathematical Functions, §27.11 “Asymptotic Formulas: Partial Sums,” especially Eq. 27.11.2, https://dlmf.nist.gov/27.11. registry ↩