Persistent Symmetry Groups

• Persistent symmetry groups are structures that capture surviving symmetries under continuous changes of parameters such as time, temperature, or scale in both physical and algebraic systems.
• They enable classification of dynamical invariances, symmetry-protected phenomena, and symmetry-breaking patterns using categorical, combinatorial, and computational methods.
• Applications range from modeling unitary Fermi gases and Floquet systems to analyzing spin textures, using tools like symmetry barcodes and polybarcodes for measurable invariants.

Searching arXiv for recent and foundational papers on “persistent symmetry groups” and closely related formulations. Searching arXiv for exact paper identifiers and titles mentioned in the source block to ground the article’s citations. Persistent symmetry groups are symmetry structures that remain operative under evolution in time, temperature, coupling space, word length, or other parameters; in a more formal categorical setting, they are the groups of symmetries that survive along a parametrized family of finite configurations. Across current research, the expression covers exact dynamical symmetries such as $SO(2,1)$ in a unitary Fermi gas, spatio-temporal point groups of the form $(R|\tau)$, hidden or restored spin symmetries such as the persistent spin helix, and persistence formalisms that record the birth, death, and reappearance of symmetries by barcodes and related invariants (Liu et al., 11 Aug 2025, Padmanabhan et al., 2017, Sun et al., 2024, Sasaki et al., 2014).

1. Conceptual scope

In current usage, persistent symmetry groups do not denote a single universal construction. Rather, they organize several closely related phenomena. In one line of work, persistence is literal: a symmetry group is tracked along a parameterized family, and one asks which symmetry elements survive from one parameter value to another. In another, persistence refers to exact dynamical invariance under continuous evolution, as in Floquet systems or many-body breathing modes. In a third, it refers to symmetry protection of physical observables against perturbations that would generically destroy them, as in persistent spin textures or persistent spin helices. A further usage concerns persistent symmetry breaking, where a broken-symmetry phase survives to arbitrarily high temperature in scale-invariant nonlocal models (Liu et al., 11 Aug 2025, Padmanabhan et al., 2017, Kilic et al., 2024, Chai et al., 2021, Chai et al., 2021).

Setting	Symmetry structure	Persistence variable
Parametrized configurations	$\mathrm{Sym}(X)$ and its persistent images	Parameter $t$
Spatio-temporal systems	$(R|\tau)$ point groups	Time translation
Unitary Fermi gas	$SO(2,1)$ dynamical symmetry	Real time
Spin-orbit systems	PSH or PST symmetry	Gate, momentum, crystal symmetry
Thermal field theory	Broken $O(N)$, $\mathbb Z_2$, or $O(m)$ sectors	Temperature
Infinite words	$G_n(w)\le S_n$	Factor length $(R|\tau)$0

This breadth is not merely terminological. It indicates that the persistence problem can be posed at several levels: for group elements, for representations, for broken-symmetry patterns, and for symmetry-constrained observables. A recurring theme is that the full symmetry expected from a naive ambient space is often reduced to a smaller but rigid surviving group, and that this surviving group can be classified, measured, or tracked.

2. Formal persistence for parametrized configurations

The most explicit general theory treats a finite configuration $(R|\tau)$1 in a metric space by its geometric symmetry group

$(R|\tau)$2

A persistent $(R|\tau)$3-configuration is then a functor

$(R|\tau)$4

where $(R|\tau)$5 is the category of $(R|\tau)$6-point subsets together with morphisms induced by ambient homeomorphisms. For a morphism $(R|\tau)$7, conjugation does not automatically send geometric symmetries of $(R|\tau)$8 to geometric symmetries of $(R|\tau)$9. The surviving subgroup is therefore restricted to those $\mathrm{Sym}(X)$0 for which $\mathrm{Sym}(X)$1, and the persistent symmetry group is defined from the image of this induced homomorphism (Liu et al., 11 Aug 2025).

This obstruction to ordinary functoriality is the reason the theory is formulated in a span bicategory rather than as a direct functor into $\mathrm{Sym}(X)$2. The assignment from configurations to symmetry data becomes a pseudofunctor into $\mathrm{Sym}(X)$3, with a span

$\mathrm{Sym}(X)$4

attached to each morphism. The resulting categorical picture is substantial: it permits a notion of persistent symmetry group for nonabelian objects, a symmetry-type bicategory built from conjugacy classes of closed subgroups, and a representation-theoretic extension in which persistence representations of persistence groups generalize the classical decomposition theorem of persistence modules (Liu et al., 11 Aug 2025).

Two invariants play the role ordinarily played by barcodes. A symmetry bar is a maximal interval on which a specific symmetry persists; the symmetry barcode records birth, death, and persistence. A polybar records the full set of parameter values at which a given ambient isometry is an actual symmetry, thereby allowing disappearance and later reappearance. For finite-type polybarcodes, the left expansion distance equals the interleaving distance, and both are bounded by an indexed interleaving distance. The same framework introduces degree of symmetry, symmetry entropy, and symmetry defect, and shows that $\mathrm{Sym}(X)$5-approximate symmetries in Euclidean space behave like compact finite approximate subgroups of $\mathrm{Sym}(X)$6 through the inclusion

$\mathrm{Sym}(X)$7

(Liu et al., 11 Aug 2025).

3. Dynamical and spatio-temporal persistence

A distinct but related tradition studies symmetry of motion rather than symmetry of static configurations. In spatio-temporal point groups, an element

$\mathrm{Sym}(X)$8

acts on a space-time point $\mathrm{Sym}(X)$9 by

$t$0

These groups are subgroups of $t$1 and classify systems invariant only after a spatial operation accompanied by a time translation. The construction can be obtained either from subgroup data with isomorphic factor groups or from one-dimensional irreducible representations of an ordinary point group, each irrep determining which spatial operations acquire nontrivial time shifts. A complete listing is given for the 32 crystallographic point groups, together with formulas for non-crystallographic cases (Padmanabhan et al., 2017).

The paradigmatic example is the classical harmonic oscillator, for which $t$2 is a symmetry. In Floquet settings the same formalism yields selection rules; for linearly polarized high-harmonic generation, the symmetry $t$3 implies that only odd harmonics survive. This use of persistence is exact and kinematic: the symmetry holds throughout the motion, but only in combined space-time form (Padmanabhan et al., 2017).

A many-body realization appears in the 3D unitary Fermi gas trapped in an isotropic harmonic potential. There the Hamiltonian admits ladder-like operators $t$4 satisfying the $t$5 Lie algebra, and repeated application of $t$6 generates a conformal tower with level spacing $t$7. The same symmetry enforces the exact breathing equation for the mean-square radius and fixes the breathing frequency at

$t$8

Experimentally the mode oscillates at twice the trapping frequency even for large excitation amplitudes, with damping-to-frequency ratio as small as $t$9. The oscillation frequency and damping rate remain nearly constant across densities and temperatures, whereas the analogous 2D setting is degraded by the quantum anomaly. In this setting, persistence is symmetry-protected nonequilibrium dynamics rather than merely recurrent structure (Sun et al., 2024).

4. Spin, crystal, and projective symmetry

In semiconductor quantum wells, the persistent spin helix is realized when the Rashba and Dresselhaus coefficients satisfy

$(R|\tau)$0

At this point a spin $(R|\tau)$1-type symmetry is recovered: the effective spin-orbit field becomes effectively uniaxial, spin rotation invariance is restored in a hidden form, and Dyakonov–Perel relaxation is strongly suppressed for the helical mode. A fitting-parameter-free transport method based on weak-localization anisotropy determines $(R|\tau)$2 from the direction of the effective field, and gate tuning yields $(R|\tau)$3 near $(R|\tau)$4 V, essentially the persistent spin helix condition (Sasaki et al., 2014).

A common misconception is that persistent spin textures require this Rashba–Dresselhaus fine tuning. Bulk non-centrosymmetric crystals with non-symmorphic symmetry provide a different mechanism. In orthorhombic systems with glide reflections and screw rotations, little-group constraints at high-symmetry points force transverse spin components to vanish. Around $(R|\tau)$5, only $(R|\tau)$6 survives; around $(R|\tau)$7, only $(R|\tau)$8 survives. The effective low-energy Hamiltonian near $(R|\tau)$9 reduces within a doublet to a form with a single allowed spin component, and BiInO$SO(2,1)$0 in space group $SO(2,1)$1 is identified as a representative material (Tao et al., 2018).

This symmetry-enforced view is generalized to all 230 crystallographic space groups by a little-group and corepresentation analysis of nonmagnetic type-II magnetic space groups. The principal result is that every nonmagnetic noncentrosymmetric bulk crystal except $SO(2,1)$2 has symmetry-protected persistent spin texture somewhere in the Brillouin zone. Type I PST occurs in nondegenerate bands, while type II PST occurs in degenerate bands and is naturally tied to nonsymmorphic symmetry at Brillouin-zone boundaries. Rotation or screw axes pin the spin along the axis, and mirror or glide planes pin it normal to the plane (Kilic et al., 2024).

The superconducting analogue is the fermion Projective Symmetry Group. In a paired state, fermion parity remains exact even though charge $SO(2,1)$3 is broken, so the correct symmetry object for Bogoliubov quasiparticles is a central extension

$SO(2,1)$4

The pairing symmetry determines phases $SO(2,1)$5, these phases modify the normal-state symmetry action projectively, and the resulting PSG governs the symmetry and topology of fermionic excitations. For the 32 crystalline point groups, the correspondence between pairing symmetry and fermion PSG is explicitly enumerated when the normal and superconducting states share the same spin rotational symmetry (Yang et al., 2023).

5. Persistent symmetry breaking

Persistence need not mean that a symmetry remains unbroken. In thermal field theory it can also mean that a broken-symmetry pattern survives to arbitrarily high temperature. One UV-complete nonlocal conformal model contains an $SO(2,1)$6 vector $SO(2,1)$7 and a singlet $SO(2,1)$8, with global symmetry

$SO(2,1)$9

Conformal perturbation theory yields two IR fixed points for $O(N)$0, one with negative mixed coupling $O(N)$1. On that branch, the thermal mass $O(N)$2 becomes negative for $O(N)$3, so the finite-temperature minimum has $O(N)$4, giving the breaking pattern

$O(N)$5

At $O(N)$6, the fixed points lie on a conformal manifold $O(N)$7, $O(N)$8, and finite temperature deforms a moduli space of vacua without restoring the discrete symmetry (Chai et al., 2021).

The continuous-symmetry generalization replaces $O(N)$9 by two long-range vector multiplets transforming under $\mathbb Z_2$0 and $\mathbb Z_2$1. Weakly coupled IR fixed points with $\mathbb Z_2$2 drive one thermal mass negative and produce

$\mathbb Z_2$3

at any nonzero temperature. The construction works for $\mathbb Z_2$4, including $\mathbb Z_2$5, with existence criteria summarized as $\mathbb Z_2$6 for suitable parameters, and already $\mathbb Z_2$7 in the $\mathbb Z_2$8 case (Chai et al., 2021).

These results are sometimes read as a contradiction of the Coleman–Hohenberg–Mermin–Wagner theorem. The papers instead present them as bypasses of that theorem’s locality and short-range assumptions: the models are nonlocal or long-range, so the standard infrared obstruction does not apply (Chai et al., 2021, Chai et al., 2021).

6. Algebraic, combinatorial, and constructive manifestations

Persistent symmetry also appears in algebraic and combinatorial settings where the relevant variable is not time but combinatorial size, group action, or exact invariance class. For pfaffians of symmetric matrices, the symmetry group of the pfaffian polynomial is precisely the dihedral group $\mathbb Z_2$9. When $O(m)$0, the pfaffian reduces to a constant multiple of the cyclic product $O(m)$1; when $O(m)$2, it becomes $O(m)$3. In both families, the surviving symmetry is the rigid cycle symmetry of a $O(m)$4-gon rather than the full symmetric group (Dzhumadil'daev, 2022).

For Boolean functions, symmetry groups coincide with automorphism groups of hypergraphs, or relation groups. Among simple permutation groups, the relation-group property is almost downward closed: with one exceptional family, if a simple permutation group is a relation group, then every subgroup is a relation group. The obstructions are sharply identified, notably small parallel multiples of alternating groups together with $O(m)$5 and $O(m)$6 (Grech et al., 2019).

For infinite words, each length $O(m)$7 carries a subgroup $O(m)$8 consisting of permutations that preserve the factor language of length $O(m)$9. Any single subgroup of $G_n(w)\le S_n$0 can occur as $G_n(w)\le S_n$1 for some word, but the full sequence $G_n(w)\le S_n$2 is highly constrained. Persistent cycles or adjacent transpositions appearing at infinitely many lengths force universality and hence $G_n(w)\le S_n$3 for all $G_n(w)\le S_n$4. By contrast, Sturmian, Arnoux–Rauzy, and Thue–Morse words eventually retain only identity and reversal, whereas period-doubling, paperfolding, and some Toeplitz words exhibit recursively large or exponentially growing symmetry groups (Luchinin et al., 2021).

A constructive computational direction appears in planar symmetric pattern generation. There the problem is to impose exact symmetry from any of the 17 planar crystallographic groups while preserving continuity of a 2D representation. The solution embeds a target planar group $G_n(w)\le S_n$5 into an affine reflection supergroup $G_n(w)\le S_n$6 and represents a $G_n(w)\le S_n$7-invariant field as a sum of $G_n(w)\le S_n$8-invariant coefficient fields against fixed basis functions. This yields exact symmetry control compatible with continuous optimization, with applications to pattern design, paper-cutting design, stylized topology design, and material design (Lin et al., 1 Jun 2026).

A related but distinct persistence program studies homological invariants of groups themselves rather than symmetry groups: persistent homology of $G_n(w)\le S_n$9-groups packages the homology of successive quotients from standard group-theoretic series into persistence matrices and barcodes, providing strong invariants for groups of order at most $(R|\tau)$00 (Ellis et al., 2010). This suggests that persistence methods and symmetry-group methods are increasingly intertwined, even when their primary invariants are different.