---
title: Local Equivalence Group in Mathematics
url: https://www.emergentmind.com/topics/local-equivalence-group
type: topic
---

# Local Equivalence Group in Mathematics

“Local equivalence group” does not denote a single universal object across mathematics. Rather, it is a family of constructions in which a global classification or comparison problem is reduced to data that are local in one of several precise senses: local homotopies and rational curves for algebraic groups, open subgroups or connected components for locally compact groups, local normalizers and Brauer quotients in modular representation theory, isotropy and moving frames in differential geometry, or localized chain-level packages in Floer- and Khovanov-theoretic concordance. A common pattern is that one defines an equivalence relation from restricted transformations, local deformations, or localized homological information, and then packages the resulting classes into a group or a functorial invariant. The phrase is therefore best understood contextually, with the specific ambient category determining both the “local” datum and the resulting equivalence structure.

## 1. Algebraic groups: \(R\)-equivalence, \(\mathbb{A}^1\)-equivalence, and sheaf-valued connected components

For algebraic groups, the data block uses “local equivalence group” for the object recording connected components of an algebraic group under equivalence generated either by rational curves or by naive affine-line homotopies. Let \(k\) be a field and \(G/k\) an algebraic group. For a field extension \(F/k\), \(R\)-equivalence on \(G(F)\) is generated by rational maps \(\mathbb{P}^1_F \dashrightarrow G_F\) defined at \(0,\infty\), while naive \(\mathbb{A}^1\)-equivalence is generated by morphisms \(\mathbb{A}^1_F \to G_F\). For a group scheme, both relations are compatible with multiplication, so \(G(F)/R\) and the \(\mathbb{A}^1\)-connected-component functor acquire group structures [1408.6627].

In Morel–Voevodsky \(\mathbb{A}^1\)-homotopy theory, the corresponding sheaf-valued invariant is
\[
\pi_0^{\mathbb{A}^1}(\mathcal{X}) := \pi_0\big(L_{\mathbb{A}^1}(\mathcal{X})\big),
\]
defined on the Nisnevich site of smooth \(k\)-schemes. For a scheme \(X\), one also has the naive sheaf
\[
S(X):=\pi_0(\mathrm{Sing}_*(X)),
\qquad
\mathrm{Sing}_*(X)_n=\underline{\mathrm{Hom}}(\Delta^n,X),
\]
with
\[
\Delta^n=\mathrm{Spec}\big(k[x_0,\dots,x_n]/(\sum_i x_i-1)\big).
\]
The local equivalence group functor is then the fieldwise assignment
\[
F\longmapsto G(F)/\!\sim_{\mathbb{A}^1},
\]
or equivalently \(F\mapsto \pi_0^{\mathbb{A}^1}(G)(F)\), followed by Nisnevich sheafification [1408.6627].

For anisotropic, semisimple, absolutely almost simple, simply connected groups in characteristic \(0\), Balwe–Sawant proved that \(R\)-equivalence and naive \(\mathbb{A}^1\)-equivalence coincide on every field extension:
\[
x\sim_R y \Longleftrightarrow x\sim_{\mathbb{A}^1} y,
\qquad
\forall\,F/k,\quad G(F)/R\cong \pi_0^{\mathbb{A}^1}(G)(F).
\]
This identifies the local equivalence group with the Nisnevich sheaf of \(\mathbb{A}^1\)-connected components in the anisotropic case and shows that \(\mathrm{Sing}_*(G)\) is not \(\mathbb{A}^1\)-local for such groups [1408.6627].

A related extension from fields to semilocal rings appears in the theory of reductive group schemes. Gille–Stavrova define \(R\)-equivalence for a group scheme \(G\) over a semilocal ring \(R\) by using homotopies over the localization \(R[t]_E\), where \(E\) is the multiplicative subset of polynomials invertible at \(0\) and \(1\). The local equivalence group is then the quotient \(G(R)/R\). For tori, this quotient is computed via flasque resolutions; for isotropic semisimple simply connected groups over regular semilocal domains containing a field, \(R\)-equivalence coincides with Karoubi–Villamayor \(\mathbb{A}^1\)-equivalence and, in rank at least \(2\), with the Whitehead quotient \(K_G(R)\) [2107.01950].

This algebraic-group usage is the most literal instance of a sheaf-theoretic “local equivalence group”: the invariant is defined on fields or semilocal rings, is functorial in the base, and packages path-like or rational-curve connectivity into a quotient group or sheaf of groups.

## 2. Locally compact abelian and Polish groups: open invertible subgroups, connected components, and Borel reducibility

In harmonic analysis on locally compact abelian groups, the phrase is used in a metric-algebraic sense. For an LCA group \(G\), let \(M(G)\) be the Banach algebra of bounded regular complex Borel measures under convolution, and let \(\mathrm{Inv}(M(G))\) be the open group of invertible measures. Hatori proved that if \(A_j\) is an open subgroup of \(M(G_j)^{-1}\) for \(j=1,2\), then any surjective isometry
\[
T:A_1\to A_2
\]
extends to a surjective either complex-linear or conjugate-linear isometry of the entire measure algebras, and after normalization becomes an algebra isomorphism. Consequently, the underlying LCA groups \(G_1\) and \(G_2\) are topologically isomorphic [1105.2657]. Here “local” refers to the fact that one works only with an open subgroup of the invertible group, not with the whole algebra.

The mechanism passes through the Gelfand transform, a clopen splitting of the maximal ideal space induced by a real-algebra isomorphism, and an idempotent
\[
J=\frac{-i\,S(i\delta_0)+\delta_0}{2}
\]
whose transform detects the “complex” and “conjugate” sectors. A contradiction argument using Pontryagin duality forces one of these sectors to vanish, so the normalized extension is globally either a complex-algebra isomorphism or a conjugate-algebra isomorphism, which is enough to recover the original group [1105.2657].

A different LCA-group usage appears in descriptive set theory. For a Polish group \(G\), one defines
\[
c(G)=\{x\in G^\omega:\exists g\in G\ \lim_{n\to\infty}x(n)=g\}
\]
and the right-coset equivalence relation
\[
E(G):\quad x\sim y \iff x^{-1}y\in c(G).
\]
When \(G\) and \(H\) are locally compact abelian Polish groups, Ding–Zheng show that
\[
E(G)\leq_B E(H)
\]
implies the existence of a continuous homomorphism
\[
S:G_0\to H_0
\]
whose kernel is non-archimedean; when \(G\) is compact and connected, the converse also holds [2209.12167]. In this setting, the “local equivalence group” is conceptual rather than formally named: the connected component \(G_0\) and the allowed non-archimedean kernels govern the Borel complexity of the local convergence relation on sequences.

An analogous rigidity theorem persists beyond the abelian case for locally compact TSI Polish groups with open identity component. If \(G\) is such a group and \(H\) is a nontrivial pro-Lie TSI Polish group, then
\[
E(G)\leq_B E(H)
\]
holds if and only if there exists a continuous homomorphism \(\phi:G_0\to H_0\) such that \(\ker(\phi)\) is non-archimedean and
\[
\phi\,\mathrm{Inn}_G(G_0)\subseteq \overline{\mathrm{Inn}_H(H_0)\,\phi}
\]
in the topology of pointwise convergence [2410.21851]. This refines the earlier abelian picture by showing that local inner-automorphism dynamics of the identity component are part of the relevant equivalence data.

These LCA and Polish-group variants share a structural principle with the algebraic-group case: local data, whether open neighborhoods in an invertible group or connected components modulo convergent-sequence noise, determine global equivalence.

## 3. Modular representation theory: local equivalence at normalizers, defect groups, and local Deligne–Lusztig data

In modular representation theory, “local equivalence” often means compatibility of a global equivalence with defect groups, Brauer pairs, or normalizers. Ruhstorfer’s equivariant Jordan decomposition provides one such instance for finite groups of Lie type. Starting from the Bonnafé–Rouquier equivalence attached to a semisimple \(\ell'\)-element \(s\in (G^*)^{F^*}\), one obtains a splendid Rickard or Morita equivalence between a block of \(G^F\) and a block of a suitable Levi subgroup or normalizer. The local version is formulated for Brauer correspondents at a common defect group \(D\), with local central idempotents
\[
B_Q := \mathrm{Tr}_{N_{G^F}(Q,b_Q)}^{N_{G^F}(Q)}(b_Q),
\qquad
C_Q := \mathrm{Tr}_{N_{L^F}(Q,c_Q)}^{N_{L^F}(Q)}(c_Q),
\]
and yields Morita equivalences between block algebras of the normalizers \(N_{G^F}(Q)\) and \(N_{L^F}(Q)\) [2010.04468].

In the equivariant setting, these local equivalences extend to crossed products by suitable automorphism groups \(A=\langle F_0,o\rangle\). Theorem 5.11 of the cited paper states that, for \(Q\) characteristic in the defect group and under the hypotheses of the global equivariant Morita theorem, a local Deligne–Lusztig bimodule extends to the relevant crossed product and induces a Morita equivalence
\[
A N_{G^F}(Q)\rtimes A\cdot B_Q
\;\text{--Morita}\simeq\;
A N_{L^F}(Q)\rtimes A\cdot C_Q
\]
[2010.04468]. Here “local” refers to passage from blocks of the ambient finite groups to blocks of the local normalizers attached to defect data.

A closely related but distinct use occurs in the theory of group graded basic Morita equivalences. Let \(A=OGb\) and \(A'=OG'b'\) be \(T\)-graded block extensions coming from normal subgroups. The global graded Morita equivalence descends, for each \(p\)-subgroup \(Q\leq P\), to a graded Morita equivalence between extended Brauer quotients. The grading group at \(Q\) is
\[
N^K(Q)/C_N(Q),
\]
where \(K=\mathrm{Aut}_T(Q)\), and the paper explicitly identifies this group with the fusion group of the local pointed algebra:
\[
F_A(Qa)\cong N^K(Qa)/C_N(Q)\cong F_{A'}(Q'a')
\]
[1607.02262]. In that context, the “local equivalence group” is the grading group controlling the local fusion and the graded Morita equivalence at the level of extended Brauer quotients.

Both representation-theoretic usages are local in Puig’s sense: one does not merely compare global categories, but requires the equivalence to descend to the local \(p\)-structure encoded by defect groups, Brauer functors, local points, and normalizers.

## 4. Geometry and analysis: moving frames, local equivalence pseudo-groups, and measure equivalence

In differential geometry, the phrase can refer to the pseudo-group of local transformations solving an equivalence problem for submanifolds. For a Lie pseudo-group \(G\) acting on the jet spaces of submanifolds, the local equivalence problem asks whether two submanifolds are related by a local element of \(G\). Within the equivariant moving-frame method, the local equivalence group at a jet is its isotropy group
\[
G_{z^{(\infty)}}=\{g^{(\infty)}\in G^{(\infty)}|_z:\ g^{(\infty)}\cdot z^{(\infty)}=z^{(\infty)}\}.
\]
When the prolonged action is free, this isotropy is trivial; when it is not free, Valiquette introduces partial moving frames on singular submanifold jets and shows that the equivalence map is determined only up to composition with isotropy at source and target [1304.1616].

The machinery is built from normalization equations, invariantization
\[
\iota(\Omega)=\widehat{\rho}^{\,*}\big[(\Omega)\big],
\]
and the recurrence formulas for invariant derivatives and Maurer–Cartan forms. In the regular case, coincidence of signature manifolds formed from a generating set of differential invariants is equivalent to local equivalence of submanifolds; in the singular case, the same criterion holds modulo the residual isotropy groups [1304.1616]. Here the “local equivalence group” is literally a local symmetry pseudo-group or its isotropy subgroup.

A different geometric-analytic usage is measure equivalence for locally compact second countable groups. For l.c.s.c. groups \(G\) and \(H\), a measure correspondence consists of a standard Borel \(G\times H\)-space with identifications by Haar measure and probability spaces on the quotient sides; if one can choose equivalent measures making the induced near-actions probability-measure preserving, then \(G\) and \(H\) are measure equivalent [1805.09063]. In the unimodular case, Koivisto–Kyed–Raum prove that measure equivalence is equivalent to stable orbit equivalence of cross-section equivalence relations of free ergodic pmp actions, and introduce uniform measure equivalence, which coincides with coarse equivalence for amenable unimodular l.c.s.c. groups [1703.08121].

In this setting the phrase “local equivalence group” is less a single object than an equivalence relation on groups themselves, but the local character is still evident: the comparison is carried by couplings, cross sections, orbit cocycles, and normalizers of compact neighborhoods rather than by an isomorphism of the groups.

## 5. Homological and concordance-theoretic local equivalence groups

In Floer- and Khovanov-theoretic applications, “local equivalence group” becomes a genuinely new algebraic target. Hendricks–Manolescu–Zemke introduced the local equivalence group \(\mathfrak{I}\) for involutive Heegaard Floer theory. Its objects are \(\iota\)-complexes \((C,\iota)\), where \(C\) is a finitely generated free \(\mathbb{F}[U]\)-complex whose localized homology is a single \(U\)-tower, and \(\iota\) is a grading-preserving chain homotopy involution. Two such complexes are locally equivalent if there exist \(\iota\)-equivariant chain maps in both directions that become quasi-isomorphisms after inverting \(U\). The set of local equivalence classes forms an abelian group \(\mathfrak{I}\) under tensor product, and there is a homomorphism
\[
h:\Theta^3_{\mathbb{Z}}\to \mathfrak{I},
\qquad
h(Y)=[(CF^-(Y),\iota)][-2].
\]
This construction is discussed in the 2025 paper on homology spheres of the trivial local equivalence class [2508.15384].

That paper emphasizes the distinction between local equivalence and homology cobordism. It exhibits infinite families of homology spheres whose images under \(h\) vanish in \(\mathfrak{I}\) but which are linearly independent in the rational homology cobordism group \(\Theta^3_{\mathbb{Q}}\), detected by the \(r_s\)-invariants from filtered instanton Floer homology [2508.15384]. The same examples include homology spheres that are trivial even in the local equivalence group of \(\operatorname{Pin}(2)\)-equivariant Seiberg–Witten Floer stable homotopy type.

A parallel construction appears in Khovanov theory. Dunfield–Lipshitz–Schütz define a Local Even–Odd triple, or LEO triple, consisting of an odd Khovanov complex \(C\), an equivariant even Bar–Natan complex \(D\), and a mod-\(2\) identification \(f\), subject to the condition that \(h^{-1}D\) is chain homotopy equivalent to a free rank-\(1\) \(\mathbb{Z}[h,h^{-1}]\)-module. Local maps are pairs of chain maps whose \(D\)-component becomes a homotopy equivalence after inverting \(h\) and which commute with the mod-\(2\) comparison up to homotopy. Local equivalence classes form an abelian group \(C_{LEO}\), and there is a concordance homomorphism from the smooth concordance group to \(C_{LEO}\) [2312.09114].

This Khovanov-theoretic local equivalence group supports refinements of Rasmussen’s \(s\)-invariant. The paper defines several invariants, including \(s_F^\pm\), Bockstein-based refinements \(r^{\beta_n}\) and \(s^{\beta_n}\), and the more comprehensive \(r_c,s_c\). For classes induced by knots, the vanishing of \(s_c\) on both a knot and its mirror completely detects triviality of its image in \(C_{LEO}\) [2312.09114]. A two-reduced variant even acquires a translation-invariant total order.

These constructions make the phrase unusually literal: the local equivalence group is a group of equivalence classes of chain-level packages modulo maps that are only required to be equivalences after localization.

## 6. Quantum information and related uses: local unitary groups, stabilizers, and symplectic moment maps

In quantum information, “local equivalence group” frequently means the group of local unitaries acting on a multipartite Hilbert space. For a system with subsystems of dimensions \(d_1,\dots,d_n\), the local unitary group is
\[
G=U(d_1)\times\cdots\times U(d_n),
\]
or \(SU(N)^L\) in the identical-subsystem case. It acts on pure states in projective space by tensor-product unitaries, and two pure states are locally equivalent precisely when they lie on the same \(G\)-orbit [1108.4134].

The symplectic-geometric formulation uses the Hamiltonian action of \(G\) on projective space with moment map
\[
\mu([ \psi ])(A)= i\langle \psi|A|\psi\rangle.
\]
For composite systems, \(\mu\) records the reduced density matrices. LU-equivalence then splits into two parts: first identifying the correct coadjoint orbit determined by the local spectra, and then analyzing the fiber of the moment map over a fixed value. In bipartite pure-state systems, equality of Schmidt spectra is sufficient because the fibers are contained in the corresponding local orbits; in multipartite systems, especially GHZ-type examples for \(L\ge 4\), the fibers can be larger than the local orbits, so equal one-particle spectra are not sufficient [1108.4134].

A stabilizer-based classification appears for symmetric multiqubit states. The relevant local unitary group is
\[
G=U(1)\times (SU(2))^n,
\]
and one studies the stabilizer subgroup
\[
\mathrm{Stab}(|\psi\rangle)
=
\{(e^{i\theta},g_1,\dots,g_n)\in U(1)\times (SU(2))^n:
(e^{i\theta}g_1^{(1)}\cdots g_n^{(n)})|\psi\rangle=|\psi\rangle\}.
\]
For symmetric states with positive-dimensional stabilizer, the classification is exhaustive: product states, generalized GHZ states, the two-qubit singlet, and Dicke states are characterized by their stabilizer Lie algebras. If the stabilizer is discrete, it is isomorphic to a finite subgroup of \(SO(3)\), identified with the rotational symmetry group of the Majorana constellation [1007.3920].

A plausible implication is that, in this quantum-information usage, the local equivalence group is not a quotient or sheaf but the acting symmetry group itself. Nevertheless, the recurring principle remains the same: a global equivalence problem is encoded by transformations restricted to act independently on local subsystems.

## 7. Comparative perspective

Across these domains, the phrase “local equivalence group” exhibits a stable formal pattern but not a fixed definition. In algebraic groups, it is a quotient such as \(G(F)/R\) or a Nisnevich sheaf like \(\pi_0^{\mathbb{A}^1}(G)\) [1408.6627; 2107.01950]. In LCA harmonic analysis, it is the global group recovered from the local metric structure of an open subgroup of invertible measures [1105.2657]. In descriptive set theory of Polish groups, it is the connected-component data and kernel structure controlling reducibility of sequence-coset equivalence relations [2209.12167; 2410.21851]. In modular representation theory, it is the local fusion or normalizer structure through which a global equivalence descends [2010.04468; 1607.02262]. In moving-frame geometry, it is the isotropy pseudo-group governing residual local symmetries [1304.1616]. In Floer and Khovanov theories, it becomes a new abelian group built from localized chain complexes [2508.15384; 2312.09114]. In quantum information, it is the local unitary group whose orbit structure defines local equivalence of states [1108.4134; 1007.3920].

What unifies these usages is not a shared formula but a shared method. One isolates transformations, homotopies, or algebraic data that are local in the relevant category, defines an equivalence relation from them, and then studies either the resulting quotient, the acting local symmetry group, or the functorial package of classes. The expression is therefore best treated as a context-dependent term of art rather than a single invariant with a universal definition.

Source: https://www.emergentmind.com/topics/local-equivalence-group