Arithmetic Triple Milnor Invariants

• Arithmetic triple Milnor invariants are arithmetic analogues of topological linking invariants that quantify higher-order interactions among primes, valuations, or Galois elements.
• They arise from explicit constructions using Ihara theory, Magnus expansions, and unitriangular representations, recovering classical symbols such as Rédei’s and cubic residue symbols.
• Their cohomological interpretation via Massey products and unipotent representations bridges number theory and topology, offering insights into decomposition laws in arithmetic topology.

Searching arXiv for recent and foundational papers on arithmetic triple Milnor invariants, Milnor triple linking, and arithmetic topology. Arithmetic triple Milnor invariants are arithmetic analogues of Milnor invariants from link theory, developed within arithmetic topology and Ihara theory to measure higher-order interactions among primes, valuations, orderings, or Galois elements in a way parallel to triple linking phenomena for three-component links. In the topological model, Milnor’s triple linking number $\mu$ for an ordered, oriented three-component link is an integer when the pairwise linking numbers vanish and is otherwise well-defined modulo $\gcd(p,q,r)$, where $p,q,r$ are the pairwise linking numbers (0901.1612). In the arithmetic model, mod $l$ or mod $q$ Milnor invariants are defined from Galois actions, Magnus expansions, and suitable nilpotent or unitriangular Galois extensions; in the triple case they recover classical residue symbols such as Rédei’s triple symbol and triple cubic residue symbols, and they admit interpretations via Massey products in Galois cohomology (Hirano et al., 2019, Efrat, 2024, Amano et al., 2014).

1. Topological prototype: Milnor’s triple invariant and its arithmetic pattern

Milnor classified three-component links in the $3$-sphere up to link homotopy by the pairwise linking numbers $p,q,r$ together with the residue class of one further integer $\mu$, the triple linking number, which is well-defined modulo $\gcd(p,q,r)$ (0901.1612). Thus, when all pairwise linking numbers vanish, $\mu$ is an ordinary integer; the Borromean rings have all pairwise linking numbers zero and $\gcd(p,q,r)$0 (0901.1612).

This modular behavior is the first arithmetic feature of the invariant. The same paper associates to a three-component link $\gcd(p,q,r)$1 a geometrically natural characteristic map

$\gcd(p,q,r)$2

such that link homotopies of $\gcd(p,q,r)$3 become homotopies of $\gcd(p,q,r)$4, and shows that the pairwise linking numbers are equal to the degrees of the restrictions of $\gcd(p,q,r)$5 to the coordinate $\gcd(p,q,r)$6-tori (0901.1612). Pontryagin’s homotopy classification of maps $\gcd(p,q,r)$7 uses the same primary data $\gcd(p,q,r)$8 together with an “ambiguous Hopf invariant” $\gcd(p,q,r)$9, well-defined modulo $p,q,r$0, and the relation

$p,q,r$1

holds in the sense stated in the paper (0901.1612). When $p,q,r$2, both $p,q,r$3 and $p,q,r$4 are integers, and Whitehead’s integral formula can be adapted to give an explicit integral formula for $p,q,r$5 through the characteristic map and the scalar Laplacian on the $p,q,r$6-torus (0901.1612, DeTurck et al., 2011).

This topological picture supplies the template for the arithmetic theory. A plausible implication is that the indeterminacy of triple Milnor data in topology—integer-valued only after vanishing of lower-order information, and otherwise defined modulo a gcd—motivates the mod $p,q,r$7 and mod $p,q,r$8 formulations used in arithmetic topology.

2. Ihara theory and mod $p,q,r$9 Milnor invariants of Galois elements

In Ihara theory, arithmetic analogues of Milnor invariants are introduced for Galois elements via Ihara’s Galois representation on the pro-$l$0 fundamental group of a punctured projective line (Hirano et al., 2019). The representation is a continuous homomorphism

$l$1

where $l$2 is the pro-$l$3 geometric fundamental group of $l$4 (Hirano et al., 2019). For $l$5, the $l$6-th pro-$l$7 longitude $l$8 of a Galois element $l$9 records how $q$0 acts around the puncture $q$1 (Hirano et al., 2019).

The pro-$q$2 Magnus expansion

$q$3

provides the coefficients from which mod $q$4 Milnor invariants are extracted (Hirano et al., 2019). For a multi-index $q$5, one defines

$q$6

with $q$7 (Hirano et al., 2019). In this framework, triple Milnor invariants are the length-$q$8 instances of this Magnus-coefficient construction.

The same paper shows that triple quadratic residue symbols of primes in the rational number field and triple cubic residue symbols of primes in the Eisenstein number field are expressed by mod $q$9 and mod $3$0 triple Milnor invariants of Frobenius elements (Hirano et al., 2019). It also introduces the dilogarithmic mod $3$1 Heisenberg ramified covering $3$2 of $3$3, regarded there as a higher analog of the dilogarithmic function for the gerbe associated to the mod $3$4 Heisenberg group, and studies monodromy transformations of certain functions on $3$5 along the pro-$3$6 longitudes of Frobenius elements for $3$7 (Hirano et al., 2019).

This establishes the principal arithmetic mechanism: Galois actions produce longitudes, longitudes produce Magnus coefficients, and triple Magnus coefficients encode triple residue phenomena.

3. Residue symbols, Heisenberg extensions, and the triple arithmetic invariant

The arithmetic triple Milnor invariant appears concretely in the classical mod $3$8 and mod $3$9 cases as higher residue symbols. The paper on linking invariants for valuations and orderings states that Morishita’s mod-$p,q,r$0 arithmetic Milnor invariants provide a decomposition law for primes in canonical Galois extensions of $p,q,r$1 with unitriangular Galois groups, and contain the Legendre and Rédei symbols as special cases (Efrat, 2024). It further explains that Morishita proposed mod-$p,q,r$2 arithmetic Milnor invariants for number fields containing the $p,q,r$3-th roots of unity and satisfying certain class field theory assumptions (Efrat, 2024).

For a field $p,q,r$4 containing the $p,q,r$5-th roots of unity, one considers a Galois group $p,q,r$6 with restricted ramification and a canonical surjective homomorphism

$p,q,r$7

where $p,q,r$8 is the unitriangular $p,q,r$9 group (Efrat, 2024). If $\mu$0, the arithmetic Milnor invariant or linking invariant at $\mu$1 is defined by

$\mu$2

and it vanishes exactly when the corresponding valuation or ordering splits completely in the associated $\mu$3-extension (Efrat, 2024). In the triple case mod $\mu$4, this recovers the arithmetic analogue of Rédei’s symbol (Efrat, 2024).

The Eisenstein-field case is developed explicitly in the study of mod $\mu$5 triple Milnor invariants and triple cubic residue symbols (Amano et al., 2014). For primes $\mu$6 in the set

$\mu$7

with a suitable auxiliary prime $\mu$8, the Galois group of the maximal pro-$\mu$9 extension unramified outside $\gcd(p,q,r)$0 admits a link-group presentation, and Magnus expansions define the mod $\gcd(p,q,r)$1 Milnor invariants $\gcd(p,q,r)$2 (Amano et al., 2014). For $\gcd(p,q,r)$3, $\gcd(p,q,r)$4 is the mod $\gcd(p,q,r)$5 Magnus coefficient of $\gcd(p,q,r)$6 in $\gcd(p,q,r)$7 (Amano et al., 2014).

Under the vanishing of all lower-length Milnor numbers $\gcd(p,q,r)$8, the triple invariant is well-defined for the ordered triple $\gcd(p,q,r)$9, and the triple cubic residue symbol is defined by

$\mu$0

This symbol describes the decomposition law of $\mu$1 in a mod $\mu$2 Heisenberg extension of degree $\mu$3 over $\mu$4 (Amano et al., 2014). The extension is a Rédei type Heisenberg extension $\mu$5 with Galois group $\mu$6, and it is constructed explicitly in the form

$\mu$7

from solutions of a norm equation (Amano et al., 2014). The triple cubic residue symbol generalizes both the cubic residue symbol and Rédei’s triple symbol (Amano et al., 2014).

4. Magnus coefficients, unipotent representations, and cohomological interpretation

A central structural fact is that arithmetic triple Milnor invariants can be read as Magnus coefficients. In the extension of Morishita’s theory from number fields to general fields, the Magnus homomorphism

$\mu$8

is used for the free pro-$\mu$9 product $\gcd(p,q,r)$00 (Efrat, 2024). If

$\gcd(p,q,r)$01

then the Milnor $\gcd(p,q,r)$02-tuple invariant is the coefficient $\gcd(p,q,r)$03 associated with the upper-right matrix entry in the unitriangular representation (Efrat, 2024). Corollary 5.2 in that paper states that the linking invariant satisfies

$\gcd(p,q,r)$04

for $\gcd(p,q,r)$05 and a suitable lift $\gcd(p,q,r)$06 (Efrat, 2024). Thus arithmetic triple Milnor invariants correspond to length-$\gcd(p,q,r)$07 Magnus coefficients.

The cohomological interpretation uses Massey products. The same paper relates the linking invariant to Massey product elements in Galois cohomology, and states that higher arithmetic Milnor invariants generalize the relation between the Legendre symbol and cup products (Efrat, 2024). For cohomology classes $\gcd(p,q,r)$08, the $\gcd(p,q,r)$09-fold Massey product

$\gcd(p,q,r)$10

contains the pullbacks along homomorphisms $\gcd(p,q,r)$11 arising from the arithmetic construction (Efrat, 2024). In the triple case, this gives the cohomological realization of the arithmetic triple Milnor invariant.

The Eisenstein-field paper states the triple Massey product formulation explicitly: if $\gcd(p,q,r)$12 are the cohomology classes corresponding to the three distinguished primes, then

$\gcd(p,q,r)$13

where $\gcd(p,q,r)$14 is the triple Massey product and $\gcd(p,q,r)$15 is dual to the relation at $\gcd(p,q,r)$16 (Amano et al., 2014). The paper emphasizes that this generalizes the cup-product interpretation of classical residue symbols to a higher-order cohomological operation (Amano et al., 2014).

This cohomological picture clarifies why unitriangular Galois groups are natural: they are the algebraic receptacles for defining systems of Massey products, and the $\gcd(p,q,r)$17-entry records the highest-order interaction.

5. Generalization from number fields to valuations and orderings

Arithmetic triple Milnor invariants are no longer confined to primes of number fields. The paper “Linking Invariants for Valuations and Orderings on Fields” extends Morishita’s theory from the number field context to general fields by introducing a linking invariant for discrete valuations and orderings (Efrat, 2024). Here discrete valuations replace finite primes, and orderings enter when $\gcd(p,q,r)$18, using Kummer theory for valuations and Becker–Artin–Schreier theory for orderings (Efrat, 2024).

In this setting, one fixes independent discrete valuations $\gcd(p,q,r)$19 and, for $\gcd(p,q,r)$20, orderings $\gcd(p,q,r)$21, together with closed subgroups $\gcd(p,q,r)$22 and $\gcd(p,q,r)$23 of the maximal pro-$\gcd(p,q,r)$24 Galois group $\gcd(p,q,r)$25 (Efrat, 2024). After passing to the quotient unramified outside the chosen set, one assumes that the resulting group is minimally generated by the images of the $\gcd(p,q,r)$26, and then constructs the globalization

$\gcd(p,q,r)$27

compatible with the local data (Efrat, 2024).

The linking homomorphism

$\gcd(p,q,r)$28

generalizes the classical bilinear linking number to the arithmetic setting (Efrat, 2024). The triple arithmetic Milnor invariant is then the appropriate higher unipotent entry, defined by the same projection procedure that produces the mod-$\gcd(p,q,r)$29 linking invariant (Efrat, 2024). The paper presents the following correspondence: the Legendre symbol is the $\gcd(p,q,r)$30 Milnor invariant; Rédei’s symbol is the $\gcd(p,q,r)$31 or triple arithmetic Milnor invariant; and higher linking invariants correspond to higher Milnor $\gcd(p,q,r)$32-invariants, higher Magnus coefficients, and higher Massey products (Efrat, 2024).

This suggests a broad arithmetic-topological dictionary in which valuations and orderings play the role of link components, restricted ramification replaces geometric complement data, and unitriangular extensions replace nilpotent link-group quotients.

6. Variants, lifts, and related developments

Several later developments show how triple Milnor invariants can be lifted, reexpressed, or computed in forms that illuminate their arithmetic structure, although these works are primarily topological rather than arithmetic in the strict number-theoretic sense.

For three-component links, integer-valued invariants $\gcd(p,q,r)$33 were defined as concordance and weak-cobordism invariants that lift certain Milnor invariants (Davis et al., 21 May 2026). The construction introduces an invariant $\gcd(p,q,r)$34, a three-component analogue of the Kojima–Yamasaki $\gcd(p,q,r)$35-invariant, and proves

$\gcd(p,q,r)$36

where $\gcd(p,q,r)$37 is generated by lower-order invariants (Davis et al., 21 May 2026). This is a topological lifting result, but it mirrors the arithmetic theme that classical Milnor invariants often appear only modulo indeterminacy, while more elaborate structures can furnish integer-valued lifts.

For algebraically split three-component links, Heegaard Floer homology gives formulas for the square of the triple Milnor invariant, including

$\gcd(p,q,r)$38

and for Brunnian L-space links the simplification

$\gcd(p,q,r)$39

holds (Gorsky et al., 2020). That work also states that $\gcd(p,q,r)$40 detects the Borromean rings among Brunnian L-space three-component links (Gorsky et al., 2020). These are not arithmetic invariants of fields, but they reinforce the role of the triple Milnor invariant as a rigid higher-order quantity.

Other topological formulations include configuration-space integral realizations for string links (Koytcheff, 2012, Koytcheff et al., 2015), Gauss-diagram expressions and non-torsion-valued variations (Ito et al., 2021, Intawong et al., 2022), and chord-diagram or multiple-crossing formulas (Hirata, 5 Oct 2025, Okuhara et al., 28 Mar 2025). A plausible implication is that these explicit models may inform computational approaches to arithmetic analogues, because the arithmetic theory already relies on Magnus coefficients, longitudes, and nilpotent presentations closely paralleling the combinatorial structures used in topology.

7. Conceptual significance and scope

Arithmetic triple Milnor invariants sit at the intersection of knot theory, Galois theory, and cohomology. The topological source invariant detects higher-order linking not seen by pairwise linking numbers, and in the three-component case it is the additional datum completing Milnor’s link-homotopy classification together with $\gcd(p,q,r)$41 (0901.1612). The arithmetic analogue transfers this pattern to primes, valuations, and orderings through pro-$\gcd(p,q,r)$42 or pro-$\gcd(p,q,r)$43 fundamental groups, unitriangular Galois representations, and Magnus expansions (Hirano et al., 2019, Efrat, 2024).

The concrete arithmetic meaning is decomposition law. In Morishita-type theories, the invariant describes whether a prime, valuation, or ordering splits completely in a canonical nilpotent extension with unitriangular Galois group; in the mod $\gcd(p,q,r)$44 case it contains the Legendre and Rédei symbols, and in the mod $\gcd(p,q,r)$45 Eisenstein case it becomes the triple cubic residue symbol governing decomposition in a mod $\gcd(p,q,r)$46 Heisenberg extension of degree $\gcd(p,q,r)$47 (Efrat, 2024, Amano et al., 2014). The cohomological meaning is higher reciprocity: cup products control pairwise residue data, whereas triple Massey products control triple Milnor data (Efrat, 2024, Amano et al., 2014).

The resulting theory is therefore not merely an analogy of vocabulary. It consists of explicit invariants defined by Magnus coefficients, realized by unipotent Galois extensions, and interpreted by Massey products. This suggests that “arithmetic triple Milnor invariant” names a precise higher-order reciprocity invariant: a mod-$\gcd(p,q,r)$48 or mod-$\gcd(p,q,r)$49 obstruction to complete splitting, modeled on Milnor’s triple linking number and expressed in the language of Galois actions, Heisenberg-type coverings, and higher cohomology operations (Hirano et al., 2019, Efrat, 2024, Amano et al., 2014).