---
title: Triple Quadratic Residue Symbols
url: https://www.emergentmind.com/topics/triple-quadratic-residue-symbols
type: topic
---

# Triple Quadratic Residue Symbols

Triple quadratic residue symbols are third-order arithmetic invariants attached, in the classical setting, to triples of distinct rational primes \(p_1,p_2,p_3\) satisfying \(p_i\equiv 1 \pmod 4\) and \(\left(\frac{p_i}{p_j}\right)=1\) for all \(i\neq j\). Classically the symbol is Rédei’s symbol \([p_1,p_2,p_3]\in\{\pm1\}\), defined by the decomposition law of \(p_3\) in a dihedral extension of degree \(8\) determined by \(p_1\) and \(p_2\). In contemporary treatments it appears equally as a decomposition symbol, a cohomological secondary invariant, a mod \(2\) Milnor invariant, and a triple Massey-product evaluation [2407.02063].

## 1. Classical definition over \(\mathbf Q\)

For distinct primes \(p_1,p_2\) with
\[
p_1\equiv p_2\equiv 1\pmod 4,\qquad \left(\frac{p_1}{p_2}\right)=1,
\]
there exist integers \(x,y,z\) such that
\[
x^2-p_1y^2-p_2z^2=0,\qquad (x,y,z)=1,\qquad y\equiv 0\pmod 2,\qquad x-y\equiv 1\pmod 4.
\]
With
\[
\alpha=x+y\sqrt{p_1},
\]
one forms
\[
R^{(2)}_{\{p_1,p_2\}}=\mathbf Q(\sqrt{p_1},\sqrt{p_2},\sqrt{\alpha}).
\]
This is the unique Galois extension of \(\mathbf Q\), determined by \(\{p_1,p_2\}\), whose Galois group is the dihedral group of order \(8\), unramified outside \(p_1,p_2\), and \(\infty\) [1906.00627].

If a third prime \(p_3\) also satisfies
\[
p_3\equiv 1\pmod 4,\qquad \left(\frac{p_i}{p_j}\right)=1\quad (1\le i\ne j\le 3),
\]
then \(p_3\) is unramified in \(R^{(2)}_{\{p_1,p_2\}}/\mathbf Q\). Let \(K^{(2)}_{\{p_1,p_2\}}=\mathbf Q(\sqrt{p_1},\sqrt{p_2})\), and let \(\tilde{\mathfrak p}\) be a prime of \(K^{(2)}\) over \(p_3\). The classical triple quadratic residue symbol is
\[
[p_1,p_2,p_3]
=
\frac{\left(\frac{R^{(2)}/K^{(2)}}{\tilde{\mathfrak p}}\right)(\sqrt{\alpha})}{\sqrt{\alpha}}
\in \{\pm1\}.
\]
Equivalently, it records whether \(p_3\) splits completely in the degree-\(8\) dihedral extension attached to \(p_1,p_2\): it is \(1\) in the completely split case and \(-1\) otherwise [1906.00627].

This construction already exhibits the basic structural feature of triple quadratic residue symbols: they are not pairwise residue symbols in disguise, but decomposition invariants in a nonabelian extension.

## 2. Cohomological formulation and arithmetic Chern–Simons theory

A recent reformulation embeds the quadratic case into a general mod \(n\) theory. Let \(F\) be a number field containing \(\mu_n\), let \(G=\mathrm{Gal}(\overline F/F)\), and let
\[
\chi_i:G\to \mathbf Z/n\mathbf Z,\qquad i=1,2,3,
\]
be continuous Kummer characters. Under the standing hypotheses that the ramification sets \(S_1,S_2,S_3\) are disjoint, each \(\chi_i\) is tame locally, and
\[
[\operatorname{loc}_v^2(\chi_i\cup\chi_j)]=0
\quad\text{in }H^2(G_v,\mathbf Z/n\mathbf Z)
\]
for all non-archimedean \(v\) and all \(i\neq j\), one defines a triple symbol
\[
[\chi_1,\chi_2,\chi_3]_n\in \mu_n
\]
by first choosing a global trivialization \(d\eta=\chi_1\cup\chi_2\cup\chi_3\), then local unramified trivializations \(\eta_v\), and finally setting
\[
\mathfrak d(\chi_1,\chi_2,\chi_3)
=
\sum_{v\in S}\operatorname{inv}_v\bigl(\operatorname{loc}_v^2(\eta)-\eta_v\bigr)\in \mathbf Z/n\mathbf Z,
\]
\[
[\chi_1,\chi_2,\chi_3]_n=\zeta^{\mathfrak d(\chi_1,\chi_2,\chi_3)}.
\]
The symbol is independent of the choices of \(\eta\), \(\eta_v\), and the auxiliary set \(S\), and it is alternating:
\[
[\chi_1,\chi_2,\chi_3]_n
=
[\chi_{\sigma(1)},\chi_{\sigma(2)},\chi_{\sigma(3)}]_n^{|\sigma|}
\qquad (\sigma\in S_3).
\]
The construction is explicitly motivated by arithmetic Chern–Simons theory: the symbol measures the discrepancy between a global trivialization of a degree-\(3\) class and local unramified trivializations [2407.02063].

In the quadratic specialization
\[
n=2,\qquad F=\mathbf Q,\qquad \mu_2=\{\pm1\},
\]
take distinct primes \(p_1,p_2,p_3\) with
\[
p_i\equiv 1\pmod 4,\qquad \left(\frac{p_i}{p_j}\right)=1 \quad (1\le i<j\le 3),
\]
and let \(\chi_i\) be the quadratic Kummer characters defined by
\[
(-1)^{\chi_i(g)}=\frac{g(\sqrt{p_i})}{\sqrt{p_i}}.
\]
Then \([\chi_1,\chi_2,\chi_3]_2\) is defined and
\[
[\chi_1,\chi_2,\chi_3]_2=[p_1,p_2,p_3].
\]
In this sense the classical Rédei symbol is recovered exactly as a cohomological secondary invariant [2407.02063].

## 3. Arithmetic-topological interpretations

In Ihara theory, triple quadratic residue symbols are expressed by mod \(2\) triple Milnor invariants of Frobenius elements. For a suitable punctured projective line \(X=\mathbf P^1\setminus A\), Ihara’s Galois representation acts on the pro-\(2\) fundamental group, and the preferred pro-\(2\) longitudes admit Magnus expansions whose coefficients are mod \(2\) Milnor invariants. In the quadratic setting one has the central formula
\[
[p_1,p_2,p_3]=(-1)^{\mu_2(\sigma_{p_3};123)},
\]
and hence
\[
\mu_2(123)=\mu_2(\sigma_{p_3};123).
\]
The mechanism is a mod \(2\) Heisenberg covering whose monodromy along the third longitude isolates the coefficient of \([x_1,x_2]\), so that the triple symbol becomes the arithmetic analogue of a mod \(2\) triple linking invariant [1906.00627].

The same quadratic symbol can also be written in terms of the second \(2\)-adic Galois polylogarithm. With
\[
x^2-p_1y^2-p_2w^2=0,\qquad
z=p_1(y/x)^2,
\]
and \(\widetilde\sigma_{p_3}\) an extension of Frobenius above \(p_3\), one has
\[
[(p_1),(p_2),(p_3)]_2
=
(-1)^{p_x(\widetilde\sigma_{p_3})-\operatorname{li}^{(2)}_2(z)(\widetilde\sigma_{p_3})}.
\]
The functional equation for \(\ell\)-adic Galois polylogarithms then implies the reciprocity law
\[
[p_1,p_2,p_3]_2\,[p_2,p_1,p_3]_2=1,
\]
which in the quadratic case is the usual symmetry in the first two entries, since the symbol takes values in \(\{\pm1\}\) [1909.06705].

These two viewpoints are compatible rather than competing. The first emphasizes nonabelian Galois actions and Heisenberg monodromy; the second expresses the same third-order phenomenon through the mod \(2\) value of the second Galois polylogarithm.

## 4. Massey products and generalizations to real quadratic fields

The triple symbol admits a cohomological interpretation via Massey products. In the quadratic examples where the pairwise cup products vanish, one may choose \(1\)-cochains \(c_{12},c_{23}\) with
\[
dc_{12}=\chi_1\cup\chi_2,\qquad dc_{23}=\chi_2\cup\chi_3,
\]
and define
\[
\langle \chi_1,\chi_2,\chi_3\rangle
=
[c_{12}\cup\chi_3+\chi_1\cup c_{23}].
\]
For \(F=\mathbf Q\), \(n=2\), and \(v=p_3\), the local invariant of this triple Massey product satisfies
\[
\mathfrak d(\chi_1,\chi_2,\chi_3)=\langle \chi_1,\chi_2,\chi_3\rangle_{p_3},
\]
so that
\[
[p_1,p_2,p_3]=(-1)^{\langle \chi_1,\chi_2,\chi_3\rangle_{p_3}}.
\]
Thus the triple quadratic residue symbol is also the local invariant of a triple Massey product [2407.02063].

A genuine generalization beyond \(\mathbf Q\) has now been constructed for real quadratic fields. Let \(k\) be a real quadratic field with trivial narrow class group, let \(\varepsilon\) be its fundamental unit, and let
\[
S=\{\mathfrak p_1,\mathfrak p_2,\mathfrak p_3\}
\]
be distinct finite primes of \(k\), none above \(2\), with \(\mathrm N\mathfrak p_i\equiv 1\pmod 4\). Under the conditions
\[
\left(\frac{\pi_i}{\pi_j}\right)=1 \quad (1\le i\ne j\le 3),\qquad
\left(\frac{\varepsilon}{\mathfrak p_i}\right)=1 \quad (i=1,2,3),
\]
the symbol is defined by
\[
[\mathfrak p_1,\mathfrak p_2,\mathfrak p_3]=(-1)^{\mu_2(123)}.
\]
It detects complete decomposition of \(\mathfrak p_3\) in a unique Rédei type \(D_8\)-extension \(K/k\) attached to \(\{\mathfrak p_1,\mathfrak p_2\}\):
\[
[\mathfrak p_1,\mathfrak p_2,\mathfrak p_3]=1
\iff
\mathfrak p_3 \text{ is completely decomposed in }K/k.
\]
The same paper gives the Massey-product formula
\[
[\mathfrak p_1,\mathfrak p_2,\mathfrak p_3]
=
(-1)^{\langle \chi_1,\chi_2,\chi_3\rangle(\delta_3)}.
\]
In this real quadratic setting, symmetry is proved only in the first two variables:
\[
[\mathfrak p_1,\mathfrak p_2,\mathfrak p_3]
=
[\mathfrak p_2,\mathfrak p_1,\mathfrak p_3],
\]
and no full three-variable symmetry is claimed [2509.00667].

## 5. Position in a higher-order hierarchy

Triple quadratic residue symbols sit in a broader hierarchy of nilpotent decomposition symbols. One formulation presents the progression
\[
\left(\frac{p_1}{p_2}\right),\qquad [p_1,p_2,p_3],\qquad [p_1,p_2,p_3,p_4]
\]
as successive invariants associated respectively with
\[
N_2(\mathbb F_2)=\mathbf Z/2\mathbf Z,\qquad
N_3(\mathbb F_2)=D_8,\qquad
N_4(\mathbb F_2).
\]
In this hierarchy the Rédei symbol is the \(3\)-fold case and satisfies
\[
[p_1,p_2,p_3]=(-1)^{\mu_2(123)},
\]
while a \(4\)-th multiple residue symbol satisfies
\[
[p_1,p_2,p_3,p_4]=(-1)^{\mu_2(1234)}.
\]
This places triple quadratic residue symbols within a general theory of arithmetic Milnor invariants and nilpotent extensions, rather than treating them as isolated artifacts of genus theory [1311.1289].

Adjacent work also shows that genuinely triple-dependent quadratic phenomena need not always appear under the name “triple quadratic residue symbol.” In the degree-\(8\) field
\[
L^+=\mathbf Q(\sqrt2,\sqrt{pq},\sqrt{ps}),
\]
a residual bit
\[
\delta(p,q,s)\in\{0,1\}
\]
decides which of two squareclasses gives the final unit generator. The standard pairwise residue datum
\[
\mathcal D(p,q,s)=\left(p\bmod 8,\ q\bmod 8,\ s\bmod 8,\ \left(\frac{q}{p}\right),\ \left(\frac{s}{p}\right),\ \left(\frac{q}{s}\right)\right)
\]
does not determine \(\delta(p,q,s)\): the paper gives explicit triples
\[
(7,19,3),\qquad (7,3,59)
\]
with the same \(\mathcal D\) but opposite values of \(\delta\). The paper does not define a standard object called a triple quadratic residue symbol, but it isolates a triple-dependent residual squareclass invariant that is invisible to pairwise Legendre data alone [2605.13888].

## 6. Scope, reciprocity, and common distinctions

A persistent source of confusion is the phrase “triple quadratic.” In the Rédei–Milnor–Massey tradition it means a ternary, third-order symbol attached to three primes or three Kummer characters. It does not mean repeated \(2\)-power residuosity. In particular, the rational \(2^k\)-th power residue symbols
\[
\left(\frac{a}{p}\right)_{2^k}
\]
form a different hierarchy: \(k=1\) is quadratic, \(k=2\) is quartic, and \(k=3\) is octic. That theory concerns iterated \(2\)-power residuosity of a single element modulo a prime or composite modulus, not a ternary residue symbol of Rédei type [1604.07791].

Another common misconception is that the triple quadratic residue symbol is defined unconditionally once three primes are given. Every formulation in the modern literature imposes pairwise compatibility conditions. In the classical rational case these are the conditions
\[
p_i\equiv 1\pmod 4,\qquad \left(\frac{p_i}{p_j}\right)=1.
\]
In the cohomological theory they become the disjointness, tameness, and local vanishing conditions on pairwise cup products. This is the arithmetic analogue of the topological fact that a triple linking invariant is defined only after the relevant pairwise linking numbers vanish [2407.02063].

A further distinction concerns symmetry. Over \(\mathbf Q\), the Rédei symbol is fully symmetric in \(p_1,p_2,p_3\); in the cohomological formulation this follows from alternation together with the fact that inversion is trivial in \(\{\pm1\}\). Over real quadratic fields, by contrast, only symmetry in the first two variables is presently established [1311.1289].

Taken together, these developments show that triple quadratic residue symbols are best understood as third-order arithmetic invariants with several equivalent realizations: decomposition laws in \(D_8\)-extensions, mod \(2\) Milnor invariants, local evaluations of triple Massey products, and global–local defects arising from cohomological trivializations. The classical symbol over \(\mathbf Q\) remains the model case, but the modern theory has made its higher-order nature precise and transportable to broader arithmetic settings.

Source: https://www.emergentmind.com/topics/triple-quadratic-residue-symbols