Robnik and Berry (1986) showed that a quantum system whose naive time-reversal symmetry is broken can nevertheless display Gaussian-orthogonal-ensemble energy-level statistics whenever some other anti-unitary symmetry is present ("false time-reversal violation"), and they classified which anti-unitary symmetries force a real Hamiltonian representation and hence orthogonal-ensemble statistics. The algebraic side of this program is essentially complete: Dyson's threefold way, the Altland-Zirnbauer tenfold extension, and later constructions (including GSE statistics in spinless time-reversal-invariant graphs) settle which anti-unitary symmetries admit a real Hamiltonian representation, in any spatial dimension, and experiments have confirmed the logic beyond planar billiards. The dynamical side remains open: the symmetry-based prediction is guaranteed only for almost-all (generic) matrices, with no practical criterion deciding whether an individual, possibly non-generic system obeys the symmetry-predicted universality class, and the semiclassical justification of the classification is established only for scaling systems whose classical dynamics does not depend on energy. The problem is to determine this validity domain: (a) a checkable criterion for when an individual system - including time-reversal-invariant billiards of constant width that exhibit GUE-typed spectra despite "almost chaotic" dynamics - follows or violates the symmetry-based prediction, and (b) a semiclassical justification of the classification for non-scaling, energy-dependent classical dynamics, including the possibility of an energy-dependent symmetry classification.

Background

When the highly excited part of the spectrum of a bounded quantum system (a particle in a hard-walled planar enclosure, a "billiard"; an atom; a quantum dot) is examined after unfolding - rescaling the levels so that the mean spacing is one - the statistics of the spacings are, for classically chaotic systems, described by random matrix theory (RMT): the fluctuations match one of Wigner and Dyson's Gaussian ensembles of large random matrices. The Gaussian orthogonal ensemble (GOE) consists of real symmetric matrices and yields linear level repulsion, P(s)∼sP(s)\sim s at small spacing ss; the Gaussian unitary ensemble (GUE) consists of complex Hermitian matrices with quadratic repulsion P(s)∼s2P(s)\sim s^2; the Gaussian symplectic ensemble (GSE) consists of quaternion self-dual matrices with P(s)∼s4P(s)\sim s^4 and Kramers degeneracy (pairwise degeneracy forced by an anti-unitary symmetry whose square is −1-1).

Which ensemble appears is fixed by symmetry. An anti-unitary operator AA is one that is antilinear, A(α∣ψ⟩)=α∗A∣ψ⟩A(\alpha|\psi\rangle)=\alpha^{*}A|\psi\rangle, and unitary. For a spinless particle the time-reversal operator TT reduces to complex conjugation; invariance under TT makes the Hamiltonian real in a suitable basis and gives GOE statistics, while a magnetic field breaks TT and gives GUE statistics. Robnik and Berry (1986) sharpened this picture: a system with neither TT nor any geometric (point-group) symmetry can still show GOE rather than GUE statistics whenever it possesses some other anti-unitary symmetry AA (whose classical limit is an anticanonical transformation, one preserving Hamilton's equations but not necessarily the canonical one-form), so that the naive expectation of GUE statistics is a "false time-reversal violation". Their central example: for a particle in a magnetic field B\mathbf{B} with scalar potential VV, the combination TSxTS_x of time reversal with the mirror reflection SxS_x is a symmetry whenever BB and VV are invariant under SxS_x; they verified numerically for an Aharonov-Bohm billiard - a planar hard-walled enclosure pierced by an infinitesimally thin magnetic flux line, which breaks TT locally but can preserve TSxTS_x - that the spectral fluctuations are GOE as predicted. For anti-unitary operators of the two classes they analyze (with A2=1A^2=1, and related operators with U^2=1\hat U^2=1), they construct an AA-adapted basis in which the Hamiltonian matrix is real, which is the mechanism behind the orthogonal-type prediction.

The classification carries an important caveat, and the caveat is the seed of the open problem. The argument that symmetry fixes the universality class holds for almost all matrices of the relevant Gaussian ensemble, in the spirit of the ensemble-ergodicity results for the Gaussian ensembles; it provides no criterion for deciding whether a particular, individual system is generic. On the algebraic side this gap has since been closed: Dyson's threefold way (1962) classifies anti-unitary symmetries by the type of their co-representations (the representations of a group extended by an anti-unitary element) as real, complex, or pseudo-real (quaternionic); only the real type admits a basis in which the Hamiltonian is real, the pseudo-real type instead giving GSE-type statistics with Kramers degeneracy - so anti-unitary symmetries that do not guarantee a real Hamiltonian representation do exist. Altland and Zirnbauer (1997) extended the classification to ten symmetry classes in mesoscopic normal-superconducting hybrid structures; the standard reference compilation of these facts is the review of Guhr, Müller-Groeling and Weidenmüller. Akila and Gutkin (2019) realized GSE statistics even in spinless, time-reversal-invariant quantum graphs (finite metric graphs with scattering vertices) possessing only a C4C_4 rotational symmetry, by resolving the spectrum into symmetry sectors associated with irreducible representations (the elementary, indecomposable blocks of the symmetry group's action on quantum states), and experiments of Aßmann et al. (2016) on Rydberg excitons in Cu2_2O confirmed in a three-dimensional solid that quadratic level repulsion sets in only when all anti-unitary symmetries are broken - extending the false-time-reversal logic beyond planar billiards.

What remains open is dynamical rather than algebraic. Gutkin (2007) exhibited time-reversal-invariant convex billiards of constant width whose classical motion is "almost" chaotic and whose spectra nevertheless show GUE-typed statistics, because the dynamics is non-ergodic in the momentum directions of phase space; he argued this behavior is generic within the class of constant-width billiards. This demonstrates that symmetry information alone cannot decide the level statistics of an individual system. Robnik's subsequent program (Robnik 1992 on Aharonov-Bohm billiards; Orel and Robnik 2026 on a C3C_3-symmetric billiard) keeps encountering deviations of individual systems from random-matrix predictions and resolves them case by case with high-precision numerics. Meanwhile the semiclassical derivations of the symmetry dictionary - Joyner, Müller and Sieber (2012) showed by trace-formula methods (a trace formula rewrites the quantum spectral density as a sum over classical periodic orbits, and the diagonal approximation keeps only the pairs of orbits with identical or mutually time-reversed actions), extended to all orders for the two-point correlations, that real irreducible representations give GOE statistics and complex ones GUE inside each symmetry sector - assume a fixed symmetry structure and effectively scaling systems, that is, billiards whose classical dynamics is independent of energy. For general systems the classical dynamics changes with energy, the semiclassical-limit arguments do not transfer directly, and the symmetry classification itself may become energy-dependent.

Problem Statement

Determine the validity domain of the anti-unitary symmetry classification of energy-level statistics for individual quantum systems with predominantly chaotic classical dynamics, in the setting of Robnik and Berry (1986). Two linked questions constitute the open core:

(a) Non-generic individual systems. Formulate a practical criterion - checkable from the classical dynamics or from finite spectral data of a given individual system, and not merely an almost-all (measure-zero-exception) genericity argument over the Gaussian ensembles - that decides whether a system possessing a specified anti-unitary symmetry follows the symmetry-predicted universality class of its unfolded level statistics (GOE-type when the symmetry admits a basis in which the Hamiltonian is real, GSE-type when the symmetry's co-representation is pseudo-real (quaternionic) and forces Kramers degeneracy, and GUE-type when the co-representation is complex, including when no anti-unitary symmetry constrains the Hamiltonian) in the semiclassical limit. The criterion must account for systems of the kind exhibited by Gutkin (2007): time-reversal-invariant, almost chaotic convex billiards of constant width whose non-ergodicity in the momentum directions of phase space produces GUE-typed statistics, and it should state under what dynamical conditions such deviations can and cannot occur.

(b) Non-scaling, energy-dependent dynamics. Justify the semiclassical limit of the anti-unitary symmetry classification for systems that are not scaling, that is, whose classical dynamics depends on energy, where the trace-formula arguments used for billiards do not transfer directly; and determine whether and when an energy-dependent symmetry classification - different universality classes in different energy windows, as the relevant classical symmetries change - must replace a single global classification.

The criterion of (a) and the semiclassical justification of (b) must also cover, and be consistent with, the known confirming and exceptional examples (Aharonov-Bohm billiards with TSxTS_x symmetry; constant-width billiards; C3C_3- and C4C_4-symmetric billiards and quantum graphs; Rydberg excitons).

References

  1. M. Robnik and M. V. Berry, "False time-reversal violation and energy level statistics: the role of anti-unitary symmetry", Journal of Physics A: Mathematical and General 19, 669-682 (1986). DOI: 10.1088/0305-4470/19/5/020, https://doi.org/10.1088/0305-4470/19/5/020
  2. F. J. Dyson, "The Threefold Way. Algebraic Structure of Symmetry Groups and Ensembles in Quantum Mechanics", Journal of Mathematical Physics 3, 1199-1215 (1962). DOI: 10.1063/1.1703863, https://doi.org/10.1063/1.1703863
  3. A. Altland and M. R. Zirnbauer, "Nonstandard symmetry classes in mesoscopic normal-superconducting hybrid structures", Physical Review B 55, 1142-1161 (1997). DOI: 10.1103/PhysRevB.55.1142, https://doi.org/10.1103/PhysRevB.55.1142
  4. T. Guhr, A. Müller-Groeling and H. A. Weidenmüller, "Random-matrix theories in quantum physics: common concepts", Physics Reports 299, 189-425 (1998). DOI: 10.1016/S0370-1573(97)00088-4, https://doi.org/10.1016/S0370-1573(97)00088-4 (preprint: arXiv:cond-mat/9707301, https://arxiv.org/abs/cond-mat/9707301)
  5. B. Gutkin, "Dynamical 'breaking' of time reversal symmetry", Journal of Physics A: Mathematical and Theoretical 40, F761-F769 (2007). DOI: 10.1088/1751-8113/40/31/F02, https://doi.org/10.1088/1751-8113/40/31/F02
  6. M. Akila and B. Gutkin, "GSE spectra in uni-directional quantum systems", Journal of Physics A: Mathematical and Theoretical 52, 235201 (2019). DOI: 10.1088/1751-8121/ab1956, https://doi.org/10.1088/1751-8121/ab1956
  7. C. H. Joyner, S. Müller and M. Sieber, "Semiclassical approach to discrete symmetries in quantum chaos", Journal of Physics A: Mathematical and Theoretical 45, 205102 (2012). DOI: 10.1088/1751-8113/45/20/205102, https://doi.org/10.1088/1751-8113/45/20/205102 (preprint: arXiv:1202.4998, https://arxiv.org/abs/1202.4998)
  8. M. Aßmann, J. Thewes, D. Fröhlich and M. Bayer, "Quantum chaos and breaking of all anti-unitary symmetries in Rydberg excitons", Nature Materials 15, 741-745 (2016). DOI: 10.1038/nmat4622, https://doi.org/10.1038/nmat4622
  9. M. Robnik, "Improved statistics of energy levels for Aharonov-Bohm chaotic billiards", Journal of Physics A: Mathematical and General 25, 1399-1403 (1992). DOI: 10.1088/0305-4470/25/5/038, https://doi.org/10.1088/0305-4470/25/5/038
  10. M. Orel and M. Robnik, "Spectral statistics and localization properties of a C3C_3-symmetric billiard" (2026). arXiv:2603.03460, https://arxiv.org/abs/2603.03460
  11. M. Bonkhoff, G. Gurney, F. Theel, P. Schmelcher, N. L. Harshman and T. Posske, "Symmetry and integrability in the anyon-Hubbard model" (2026). arXiv:2605.28956, https://arxiv.org/abs/2605.28956