---
title: Module of Universal Norms
url: https://www.emergentmind.com/topics/module-of-universal-norms
type: topic
---

# Module of Universal Norms

Searching arXiv for recent and foundational papers on the arithmetic notion of the module of universal norms.
The module of universal norms is, in arithmetic Iwasawa theory, the subgroup or cohomological module consisting of elements that are norms from every finite layer of an infinite \( \mathbb Z_p \)- or \( \mathbb Z_\ell \)-extension. For a totally real number field \(K\) with cyclotomic \( \mathbb Z_\ell \)-extension \(K_\infty=\bigcup_n K_n\), Jaulent defines the group of universal norms by
\[
N_K:=\bigcap_{n\ge 0} N_{K_n/K}(E_{K_n})\subseteq E_K,
\]
where \(E_K:=\varprojlim_n E_{K_n}\) is the projective limit of the \( \ell \)-adic unit groups, often called “the group of logarithmic units of \(K\),” and \(T_K:=\varprojlim_n Cl(K_n)\) is the classical Iwasawa module of \( \ell \)-classes [1904.07014]. In later work, the same norm-coherence principle is recast for \(p\)-adic Galois representations and for higher-rank exterior-bidual constructions, yielding closely related modules defined by intersections of corestriction images or by Iwasawa-cohomological descent [2010.02292; 2007.07454].

## 1. Cyclotomic definition and \(\Lambda\)-module framework

Fix a prime \( \ell \) and a totally real number field \(K\) of degree \(d=[K:\mathbb Q]\). Let \(K_\infty=\bigcup_n K_n\) be its cyclotomic \( \mathbb Z_\ell \)-extension, with
\[
\Gamma:=\operatorname{Gal}(K_\infty/K)\simeq \mathbb Z_\ell,\qquad \Lambda=\mathbb Z_\ell[[\Gamma]]\simeq \mathbb Z_\ell[[T]]
\]
after choosing a topological generator \( \gamma \) of \( \Gamma \) via \( \gamma=1+T \) [1904.07014]. For each finite layer \(K_n\), the \( \ell \)-adic completion \(E_{K_n}\) of the unit group, and more generally the \(S\)-unit group when \(S\) contains the primes above \( \ell \), forms an inverse system under the norm maps. The intersection
\[
\bigcap_n N_{K_n/K}(E_{K_n,S})
\]
is independent of \(S\) and equals \(N_K\) [1904.07014].

This definition isolates the largest \(\Lambda\)-submodule of \(E_K\) consisting of elements that are norms from each finite layer \(K_n\). Since \(N_K\) is a closed subgroup of the compact \(\Lambda\)-module \(E_K\), it is itself a compact \(\Lambda\)-module [1904.07014]. Under Leopoldt’s conjecture, \(T_K\) is a finitely generated torsion \(\Lambda\)-module, while \(E_K\) is \(\Lambda\)-free of rank
\[
r=r_1+r_2-1=d-1
\]
with no torsion except the roots of unity [1904.07014]. In this setting, the universal-norm module is the unit-theoretic counterpart to the Iwasawa module of classes.

## 2. Fixed points, co-invariants, and Greenberg’s conjecture

A key structural relation is the pair of isomorphisms
\[
T_K^\Gamma \simeq E_K/N_K,
\qquad
T_K/(\Gamma-1)T_K \simeq Cl^{\log}_K,
\]
where \(Cl^{\log}_K\) is the \( \ell \)-group of logarithmic ideal-classes of \(K\) [1904.07014]. These formulas identify the quotient \(E_K/N_K\) with the \(\Gamma\)-fixed submodule of the Iwasawa class-group module and the co-invariants of \(T_K\) with the logarithmic class group. The quotient \(E_K/N_K\) is therefore finite exactly when \(T_K^\Gamma\) is finite.

Greenberg’s conjecture for \(K\) and \(\ell\) asserts that the classical invariants \(\lambda(T_K)\) and \(\mu(T_K)\) vanish, equivalently that \(T_K\) is a pseudo-null \(\Lambda\)-module, or again that the orders of the finite \( \ell \)-class groups \(Cl(K_n)\) remain bounded as \(n\to\infty\) [1904.07014]. Under Leopoldt’s conjecture for \(K\) at \(\ell\), Jaulent proves that Greenberg’s conjecture holds if and only if
\[
(E_K:N_K)=|Cl^{\log}_K|.
\]
The proof proceeds by comparing the exact cohomology sequences for \(\Gamma\) on \(E_K\) and the co-invariant sequence on \(T_K\), obtaining
\[
|T_K^\Gamma|=|E_K/N_K|,
\qquad
|T_K/(\Gamma-1)T_K|=|Cl^{\log}_K|
\]
and then identifying the two finite quantities under Greenberg’s condition [1904.07014].

In the abelian real case \(K/\mathbb Q\) with Galois group \(A\) of order prime to \(\ell\), \(E_K\) decomposes over \(\mathbb Z_\ell[A]\) into its irreducible \( \ell \)-adic character components. Under Greenberg, each universal-norm component \(N_K^\varphi\subseteq E_K^\varphi\) is the unique \(\mathbb Z_\ell[A]\)-submodule of index \(|Cl^{\log}_K|_\varphi\) [1904.07014]. This gives a characterwise formulation of the same index phenomenon.

## 3. Capitulation, circular units, and special cases

The universal-norm quotient also admits a capitulation interpretation. Assuming Leopoldt in \(K\) and Gross–Kuz’min in \(K_\infty\), Jaulent shows that for large \(n\),
\[
(E_K:N_K)=|Cap_{K_n/K}|,
\]
where
\[
Cap_{K_n/K}=\ker\!\bigl[Cl^{\log}_K\to Cl^{\log}_{K_n}\bigr]
\]
is the subgroup of classes that capitulate logarithmically in \(K_n\) [1904.07014]. In particular,
\[
(E_K:N_K)\le |Cl^{\log}_K|,
\]
with equality if and only if Greenberg holds. The universal-norm index is thus bounded by the logarithmic class-group order and becomes equal to it precisely in the conjectural pseudo-null regime.

Jaulent also compares universal norms with circular units. For an odd prime \(\ell\) and \(K=\mathbb Q(\zeta_{\ell^n})^+\), the maximal real subfield of \(\mathbb Q(\zeta_{\ell^n})\), one has
\[
K \text{ satisfies Greenberg } \Longleftrightarrow N_K=C_K,
\]
where \(C_K\) is the circular unit group; equivalently,
\[
N_{K_n/K}(E_{K_n})=C_K
\]
for all large \(n\) [1904.07014]. This places universal norms in direct relation with the classical analytic unit theory of cyclotomic fields.

A contrasting phenomenon occurs when \(\ell\) splits completely in a real abelian \(K\). In that case, the intersection of Sinnott’s circular units with the logarithmic units is just the roots of unity \(A_K\), and
\[
N_K\simeq A_K\oplus \mathbb Z_\ell^{[K:\mathbb Q]},
\qquad
E_K/N_K\simeq A_K,
\]
independently of any conjecture [1904.07014]. At the semi-local level, one decomposes each local \( \ell \)-adic completion \(R_{K_v}\) as a product of its ordinary units and logarithmic units, and the global injection \(E_K\hookrightarrow \prod R_{K_v}\) identifies \(N_K\) inside \(E_K\) up to a finite defect that vanishes under Leopoldt [1904.07014].

Jaulent also proves a criterion of non-triviality: for \(K\) totally real abelian of degree prime to \(\ell\), under Leopoldt and Gross–Kuz’min in \(N=K(\zeta_\ell)\), the quotient \(E_K/N_K\) can be trivial only if three canonical radicals coincide in \(K\): the logarithmic Kummer radical, the initial radical of \( \mathbb Z_\ell \)-extensions, and Tate’s universal kernel [1904.07014].

## 4. \(p\)-units and the first layers of \(\mathbb Z_p\)-extensions

In the cyclotomic \( \mathbb Z_p \)-extension of a number field \(F\) containing \( \mu_p \), the analogous object is the inverse limit of \(p\)-units. For
\[
F_n=F(\zeta_{p^n}),\qquad F_\infty=\bigcup_{n\ge 0}F_n,\qquad \Gamma=\operatorname{Gal}(F_\infty/F)\simeq \mathbb Z_p,
\]
and
\[
U_n=\mathcal O_{F_n}[1/p]^*\otimes \mathbb Z_p,
\]
the norm maps make \( \{U_n\}_n \) into an inverse system, and one sets
\[
U_\infty=\varprojlim_n U_n,
\]
described as the \(\Lambda\)-module of “universal norms” in the \(p\)-unit tower [1303.0176].

Its structure is explicit: \(U_\infty\) contains as its maximal \(\Lambda\)-torsion submodule exactly the copy of \( \mu_{p^\infty}(F_\infty) \), and after factoring out this torsion one obtains a free \(\Lambda\)-module of rank
\[
r=r_1+r_2-1
\]
through an exact sequence
\[
0\to \mu_{p^\infty}(F_\infty)\to U_\infty\to \Lambda^r\to 0
\]
[1303.0176]. This is the \(p\)-unit analogue of the freeness phenomenon for logarithmic units.

The same paper relates \(U_\infty\) to the Kummer radical
\[
A_F\subseteq F^\times/(F^\times)^p,
\]
consisting of those classes \(\bar a\) such that the Kummer extension \(F(\sqrt[p]{a})\) sits inside some \( \mathbb Z_p \)-extension of \(F\). Under Leopoldt’s conjecture for \(F\), the torsion module \( \mu_{p^\infty}(F_\infty) \) is canonically Pontrjagin-dual to \(A_F\), equivalently
\[
A_F \simeq \operatorname{Hom}_\Lambda\!\bigl(U_\infty^{\mathrm{tors}},\,\mathbb Q_p/\mathbb Z_p\bigr)
\]
[1303.0176]. The module of universal norms thereby controls the first layers of all \( \mathbb Z_p \)-extensions of \(F\).

For \(F=\mathbb Q(\zeta_p)\) with \(p\) an odd regular prime, one has
\[
U_\infty\simeq \mu_{p^\infty}(F_\infty)\oplus \Lambda^r
\]
with \(r=(p-3)/2\), so the only nonzero \(p\)-torsion of \(U_\infty\) is \( \mu_p(F)=\langle \zeta_p\rangle \), and
\[
A_F\simeq \mathbb Z/p
\]
generated by the class of \( \zeta_p \) in \(F^\times/(F^\times)^p\) [1303.0176].

## 5. \(p\)-adic representations and the Fargues–Fontaine curve

For a finite extension \(K/\mathbb Q_p\), a \(p\)-adic representation \(V\in \operatorname{Rep}_{\mathbb Q_p}(G_K)\) with \(G_K\)-stable lattice \(T\), and an algebraic extension \(L/K\), the universal-norm module is defined cohomologically by
\[
H^1_{\mathrm{Iw},g}(L/K,T)
:=
\bigcap_{K'\subset L\ \mathrm{finite}}
\operatorname{cor}_{K'/K} H^1_g(K',T),
\]
equivalently as the image of the natural projection \(H^1_{\mathrm{Iw}}(L/K,T)\to H^1(K,T)\) [2010.02292]. This reframes the universal-norm condition inside Bloch–Kato local conditions and Iwasawa cohomology.

The central theorem concerns de Rham representations with Hodge–Tate weights \(\ge 0\). If \(V_0\subset V\) is the unique maximal subrepresentation whose Hodge–Tate weights are \(\ge 1\), with quotient \(V'=V/V_0\), then for perfectoid \(\widehat L\) the restriction map \(H^1(L,V/T)\to H^1(L,V'/T')\) is surjective and
\[
0\to H^1_e(L,V/T)\to H^1(L,V/T)\to H^1(L,V'/T')\to 0
\]
is exact [2010.02292]. Equivalently,
\[
H^1_e(L,V/T)=\ker\!\bigl(H^1(L,V/T)\to H^1(L,V'/T')\bigr).
\]

The proof uses the classification of vector bundles over the Fargues–Fontaine curve. The argument identifies \(E_+(V/T)\) with global sections of a truncated vector bundle over the curve, applies the Harder–Narasimhan theory, and exploits the vanishing of \(H^1(L,E_+(V_0))\) when \(\widehat L\) is perfectoid [2010.02292]. In the case of an abelian variety \(A/K\), where \(V=V_p(A)\) has Hodge–Tate weights \(\{0,1\}\), this recovers the Coates–Greenberg statement that the image of the Kummer map
\[
\kappa_L:A(L)\otimes \mathbb Q_p/\mathbb Z_p \to H^1(L,A[p^\infty])
\]
coincides with the image of the weight-\(1\) subrepresentation \(V_0\) [2010.02292].

## 6. Higher-rank universal norms and basic norm-coherent sequences

Higher-rank Iwasawa theory generalizes universal norms from single cohomology classes to equivariant exterior biduals. For a finite abelian extension \(L/K\), a \( \mathbb Z_p \)-extension \(L_\infty/K\), coefficient ring \(R\), and \(R\)-free \(p\)-adic Galois representation \(T\), one defines for each finite layer \(L_n\) and each rank \(r\ge 1\)
\[
UN_n^r(T):=\bigcap_{m\ge n}\operatorname{im}\!\bigl(\operatorname{cores}_{m,n}^r\bigr)
\subseteq {}^r_{R[\operatorname{Gal}(L_n/K)]}H^1_\Sigma(L_n,T),
\]
together with the inverse-limit module of rank-\(r\) norm-coherent sequences
\[
NS^r(T):=\varprojlim_n {}^r_{R[\operatorname{Gal}(L_n/K)]}H^1_\Sigma(L_n,T)
\]
[2007.07454].

Under mild hypotheses and for \(1\le r\le r_T\), \(NS^r(T)\) is a finitely generated free \(\Lambda\)-module of rank
\[
[L:K]\binom{r_T}{r},
\]
and descent induces an isomorphism
\[
NS^r(T)\otimes_\Lambda R[\operatorname{Gal}(L_n/K)]
\xrightarrow{\ \simeq\ }
UN_n^r(T)
\]
[2007.07454]. In particular each \(UN_n^r(T)\) is free of the same rank over the finite-layer group ring. There is also a short exact sequence
\[
0\to UN_n^r(T)\to {}^r_{R[G_n]}UN_n^1(T)\to
\bigl(Ext^1_\Lambda({}^r_\Lambda H^1_{\mathrm{Iw}},\Lambda)\bigr)^{\Gamma^n}\to 0
\]
with finite cokernel [2007.07454].

At the basic rank \(r=r_T\), determinant formalism produces a rank-one \(\Lambda\)-submodule
\[
NS^b(T)\subseteq NS^{r_T}(T),
\]
the module of basic norm-coherent sequences, and a perfect \(\Lambda\)-bilinear pairing between \(NS^{r_T}/NS^b\) and a Fitting-ideal module built from \(H^2_{\Sigma,\mathrm{Iw}}\) [2007.07454]. In the cyclotomic case \(K=\mathbb Q\), \(T=\mathbb Z_p(1)\), the basic module agrees characterwise with the module generated by Rubin–Stark or cyclotomic elements, and the quotient \(NS^1/NS^b\) refines the classical Iwasawa Main Conjecture [2007.07454].

## 7. Terminological scope across disciplines

The arithmetic literature uses “module of universal norms” for norm-coherent subgroups of units, cohomology classes, or exterior biduals. The phrase “universal norms,” however, also appears in unrelated arXiv contexts. One example is the ontology network ValueNet, in which FOLK and That’s All Folks are described as a “single ontology of norms” uniting top-down value theories, bottom-up folk values, and linguistic or factual triggers [2303.00632]. Another is the “Normative Module,” an architectural component of a generative agent that maintains weights over candidate institutions and transforms action values by expected sanction costs in order to support multi-agent cooperation [2405.19328].

This suggests a terminological ambiguity across research areas. In arithmetic, “universal norms” denotes elements surviving all norm maps in a tower; in the ontology and multi-agent literature, the phrase designates resources or modules concerned with social and institutional norm recognition rather than field norms or Galois cohomology [2303.00632; 2405.19328]. For mathematical usage, the decisive markers are the cyclotomic tower, the Iwasawa algebra \( \Lambda \), and constructions such as \(N_K\), \(H^1_{\mathrm{Iw},g}(L/K,T)\), and \(UN_n^r(T)\).

Source: https://www.emergentmind.com/topics/module-of-universal-norms