Fluctuation spectrum of the broken replica theory in finite dimensions: infrared divergences and the lower critical dimension of replica symmetry breaking in spin glasses

In his 1980 Physics Reports note on the broken replica (replica-symmetry-breaking) theory of spin glasses, Parisi listed among the remaining problems of the theory 'the computations of the fluctuations induced corrections, of the Goldstone modes and of the lower critical dimension'. This statement reconstructs that question in modern terms: the completion of the Hessian (Gaussian fluctuation) eigenvalue computation about the Parisi saddle point beyond the restricted subspace of hierarchical $L^{2}(0,1)$ perturbations, whose known eigenvalues accumulate at zero, and the determination of the nature of the infrared divergences and of the lower critical dimension of the replica-symmetry-broken phase. The mean-field fluctuation analysis was developed (replicon marginal stability; replica-Fourier diagonalization of the Hessian), and the two companion issues treated alongside it in this reconstruction were substantially resolved by later work: non-equilibrium properties (remanence, aging) by dynamical mean-field theory (Sompolinsky; Cugliandolo-Kurchan), and the quantitative connection between the Parisi free-energy equation and the TAP formalism by the rigorous generalized-TAP program. The surviving open core is the finite-dimensional part: the fluctuation spectrum of the replica-symmetry-broken phase in spatial dimension $d$, the infrared behavior of its massless replicon-like sector, and the existence and value of the lower critical dimension of replica-symmetry-breaking criticality. This core remains genuinely open and actively contested, with opposing studies appearing in 2024 and 2025.

Background

Spin glasses are disordered magnets in which randomly placed, frustrated (impossible to satisfy simultaneously) interactions freeze the spins into a state with no conventional long-range order. The Edwards-Anderson (EA) model places Ising spins on a lattice with quenched (frozen) random nearest-neighbor couplings; the Sherrington-Kirkpatrick (SK) model is its infinite-range, mean-field version. To average the free energy over the disorder, one introduces $n$ identical copies ('replicas') of the system, computes with integer $n$, and continues analytically to $n \to 0$ — the replica trick. The order parameter is the overlap matrix $Q_{\alpha\beta}$ between replicas $\alpha$ and $\beta$. The replica-symmetric (RS) ansatz, in which all off-diagonal elements are equal, describes the high-temperature phase, but de Almeida and Thouless showed it is unstable below a transition line in the temperature-magnetic-field plane (the AT line): the transverse 'replicon' eigenvalue of the Hessian (the matrix of second functional derivatives, governing Gaussian fluctuations) about the RS saddle becomes negative. Parisi's replica symmetry breaking (RSB) scheme repairs this: the stable saddle has a hierarchical structure in which, in the $n\to 0$ limit, the order parameter becomes a monotone function $q(x)$ on an internal coordinate $x\in[0,1]$ — the embedding of the Hilbert space $L^{2}(0,1)$ into the space of replica matrices — with $q(1)=q_{\mathrm{EA}}$, the Edwards-Anderson order parameter (and $q(0)=0$ in zero magnetic field); the equilibrium states are organized ultrametrically (any three states have two of their pairwise overlaps equal). The Thouless-Anderson-Palmer (TAP) equations are an alternative, local-magnetization-based mean-field description of the same model. Fluctuation corrections to the saddle-point (mean-field) thermodynamics are controlled by the spectrum of the Hessian of the replicated effective Hamiltonian at the RSB saddle. Thouless, de Almeida and Kosterlitz computed the eigenvalues within the restricted subspace of hierarchical perturbations — functions on $[0,1]$ — and found the spectrum accumulating at zero: the RSB phase is only marginally stable, with a divergent nonlinear susceptibility. Parisi's 1980 Physics Reports note listed among the remaining problems of the broken replica theory 'the computations of the fluctuations induced corrections, of the Goldstone modes and of the lower critical dimension' — in the terms of this reconstruction, the completion of this eigenvalue computation beyond the restricted subspace, and its finite-dimensional consequences: the replicated Landau-Ginzburg functional (obtained by adding a spatial gradient term and describing the short-range EA model) has upper critical dimension 6 — the dimension above which mean-field exponents hold and below which fluctuations dominate — and the infrared (long-wavelength) behavior of its massless replicon-like modes should decide the lower critical dimension, the spatial dimension below which the RSB ordered phase and its transition disappear. Bray and Roberts showed already in 1980 that perturbative renormalization-group (RG) analysis (the systematic integration of short-wavelength fluctuations to obtain an effective long-distance theory) fails to find a stable fixed point for the AT transition below $d=6$. The competing droplet picture — in which only two pure states exist and interfaces between them cost a free energy growing with system size — predicts no AT line and no RSB phase at all in finite dimension. Whether and where RSB survives in finite dimension has remained the central open controversy of spin-glass theory, pursued since through multiloop expansions of the RSB field theory, nonperturbative instanton arguments (localized saddle-point field configurations) and real-space arguments, large-scale Monte Carlo simulations, and one-dimensional power-law chains and hierarchical-lattice proxy models (self-similar lattices whose effective dimension can be tuned) that interpolate in effective dimension.

Problem Statement

Determine the infrared behavior of the replica-symmetry-broken phase of the replicated spin-glass effective Hamiltonian in spatial dimension $d$: whether the infrared divergences of its massless replicon-like sector are harmless or destroy the phase, whether the de Almeida-Thouless transition in a field (and the RSB ordered phase it bounds) exists below the upper critical dimension $d=6$, and what the lower critical dimension of replica-symmetry-breaking criticality is, i.e. the dimension below which it disappears. The decisive requirement is a controlled analysis of the fluctuation spectrum of the Parisi broken-replica saddle point in spatial dimension $d$ — the Hessian (Gaussian fluctuation operator) of the replicated effective Hamiltonian — determining the fate of the infrared divergences of its massless replicon-like sector and, thereby, whether the RSB ordered phase and its de Almeida-Thouless transition survive below $d=6$ and, if so, the value of the lower critical dimension. The starting point is the restricted subspace of hierarchical $L^{2}(0,1)$ perturbations, whose known eigenvalues accumulate at zero (Thouless, de Almeida and Kosterlitz, 1980), together with the replica-Fourier diagonalization program (Crisanti and De Dominicis, 2015) that treats the hierarchical sector systematically; extending the mean-field Hessian diagonalization beyond that subspace is the historical motivation for the question — among the items Parisi flagged in 1980 — rather than a gating deliverable, and is required only insofar as the finite-dimensional conclusion needs. This is the surviving open core of the question posed in Parisi's 1980 Physics Reports note. The two companion issues — the computation of non-equilibrium statistical properties (remanence, aging) within the broken replica theory, and a quantitative derivation connecting the Parisi free-energy equation (its stochastic-differential-equation interpretation) to the TAP approach — have been substantially resolved by later work and are excluded from this problem.

References

1. G. Parisi (1980). 'Mean field theory for spin glasses'. Physics Reports 67(1), 25-28. DOI: 10.1016/0370-1573(80)90075-7. URL: https://doi.org/10.1016/0370-1573(80)90075-7
2. S. F. Edwards and P. W. Anderson (1975). 'Theory of spin glasses'. Journal of Physics F: Metal Physics 5, 965-974. DOI: 10.1088/0305-4608/5/5/017. URL: https://doi.org/10.1088/0305-4608/5/5/017
3. D. Sherrington and S. Kirkpatrick (1975). 'Solvable Model of a Spin-Glass'. Physical Review Letters 35, 1792-1796. DOI: 10.1103/PhysRevLett.35.1792. URL: https://doi.org/10.1103/PhysRevLett.35.1792
4. D. J. Thouless, P. W. Anderson and R. G. Palmer (1977). 'Solution of "Solvable model of a spin glass"'. Philosophical Magazine 35, 593-601. DOI: 10.1080/14786437708235992. URL: https://doi.org/10.1080/14786437708235992
5. J. R. L. de Almeida and D. J. Thouless (1978). 'Stability of the Sherrington-Kirkpatrick solution of a spin glass model'. Journal of Physics A: Mathematical and General 11, 983-990. DOI: 10.1088/0305-4470/11/5/028. URL: https://doi.org/10.1088/0305-4470/11/5/028
6. G. Parisi (1979). 'Infinite Number of Order Parameters for Spin-Glasses'. Physical Review Letters 43, 1754-1756. DOI: 10.1103/PhysRevLett.43.1754. URL: https://doi.org/10.1103/PhysRevLett.43.1754
7. D. J. Thouless, J. R. L. de Almeida and J. M. Kosterlitz (1980). 'Stability and susceptibility in Parisi's solution of a spin glass model'. Journal of Physics C: Solid State Physics 13, 3271-3280. DOI: 10.1088/0022-3719/13/17/017. URL: https://doi.org/10.1088/0022-3719/13/17/017
8. A. J. Bray and S. A. Roberts (1980). 'Renormalisation-group approach to the spin glass transition in finite magnetic fields'. Journal of Physics C: Solid State Physics 13, 5405-5412. DOI: 10.1088/0022-3719/13/29/019. URL: https://doi.org/10.1088/0022-3719/13/29/019
9. H. Sompolinsky (1981). 'Time-Dependent Order Parameters in Spin-Glasses'. Physical Review Letters 47, 935-938. DOI: 10.1103/PhysRevLett.47.935. URL: https://doi.org/10.1103/PhysRevLett.47.935
10. C. De Dominicis, M. Gabay and B. Duplantier (1982). 'A Parisi equation for Sompolinsky's solution of the SK model'. Journal of Physics A: Mathematical and General 15, L47-L49. DOI: 10.1088/0305-4470/15/1/010. URL: https://doi.org/10.1088/0305-4470/15/1/010
11. K. Binder and A. P. Young (1986). 'Spin glasses: Experimental facts, theoretical concepts, and open questions'. Reviews of Modern Physics 58, 801-976. DOI: 10.1103/RevModPhys.58.801. URL: https://doi.org/10.1103/RevModPhys.58.801
12. L. F. Cugliandolo and J. Kurchan (1993). 'Analytical solution of the off-equilibrium dynamics of a long-range spin-glass model'. Physical Review Letters 71, 173-176. DOI: 10.1103/PhysRevLett.71.173. URL: https://doi.org/10.1103/PhysRevLett.71.173
13. E. Bolthausen (2013). 'An Iterative Construction of Solutions of the TAP Equations for the Sherrington-Kirkpatrick Model'. Communications in Mathematical Physics 325, 333-366. DOI: 10.1007/s00220-013-1862-3. URL: https://doi.org/10.1007/s00220-013-1862-3
14. L. Leuzzi, G. Parisi, F. Ricci-Tersenghi and J. J. Ruiz-Lorenzo (2008). 'Dilute One-Dimensional Spin Glasses with Power Law Decaying Interactions'. Physical Review Letters 101, 107203. DOI: 10.1103/PhysRevLett.101.107203; arXiv:0801.4855. URL: https://doi.org/10.1103/PhysRevLett.101.107203
15. G. Parisi and T. Temesvári (2012). 'Replica symmetry breaking in and around six dimensions'. Nuclear Physics B 858, 293-316. DOI: 10.1016/j.nuclphysb.2012.01.014; arXiv:1111.3313. URL: https://arxiv.org/abs/1111.3313
16. T. Temesvári (2014). 'The perturbative structure of spin glass field theory'. Nuclear Physics B 880, 528-551. DOI: 10.1016/j.nuclphysb.2014.01.020; arXiv:1312.7310. URL: https://doi.org/10.1016/j.nuclphysb.2014.01.020
17. A. Crisanti and C. De Dominicis (2015). 'Replica Fourier Transform: Properties and applications'. Nuclear Physics B 891, 73-105. DOI: 10.1016/j.nuclphysb.2014.12.002; arXiv:1410.4023. URL: https://doi.org/10.1016/j.nuclphysb.2014.12.002
18. M. C. Angelini and G. Biroli (2015). 'Spin Glass in a Field: a New Zero-Temperature Fixed Point in Finite Dimensions'. Physical Review Letters 114, 095701. DOI: 10.1103/PhysRevLett.114.095701; arXiv:1409.1011. URL: https://arxiv.org/abs/1409.1011
19. P. Charbonneau and S. Yaida (2017). 'Nontrivial Critical Fixed Point for Replica-Symmetry-Breaking Transitions'. Physical Review Letters 118, 215701. DOI: 10.1103/PhysRevLett.118.215701; arXiv:1607.04217. URL: https://arxiv.org/abs/1607.04217
20. M. Baity-Jesi, E. Calore, A. Cruz, L. A. Fernandez, J. M. Gil-Narvión, A. Gordillo-Guerrero, D. Iñiguez, A. Maiorano, E. Marinari, V. Martin-Mayor, J. Monforte-García, A. Muñoz Sudupe, D. Navarro, G. Parisi, S. Perez-Gaviro, F. Ricci-Tersenghi, J. J. Ruiz-Lorenzo, S. F. Schifano, B. Seoane, A. Tarancón, R. Tripiccione and D. Yllanes (2017). 'A statics-dynamics equivalence through the fluctuation-dissipation ratio provides a window into the spin-glass phase from nonequilibrium measurements'. Proceedings of the National Academy of Sciences 114(8), 1838-1843. DOI: 10.1073/pnas.1621242114. URL: https://doi.org/10.1073/pnas.1621242114
21. E. Bolthausen (2018). 'A Morita type proof of the replica-symmetric formula for SK'. arXiv:1809.07972; published in 'Statistical Mechanics of Classical and Disordered Systems', Springer Proceedings in Mathematics & Statistics (2019), pp. 63-93. DOI: 10.1007/978-3-030-29077-1_4. URL: https://arxiv.org/abs/1809.07972
22. B. Atalay and A. N. Berker (2018). 'Lower lower-critical spin-glass dimension from quenched mixed-spatial-dimensional spin glasses'. Physical Review E 98, 042125. DOI: 10.1103/PhysRevE.98.042125; arXiv:1809.06029. URL: https://doi.org/10.1103/PhysRevE.98.042125
23. M. A. Moore (2019). 'Why replica symmetry breaking does not occur below six dimensions in Ising spin glasses'. arXiv:1902.07099. URL: https://arxiv.org/abs/1902.07099
24. A. Auffinger and A. Jagannath (2019). 'Thouless-Anderson-Palmer equations for generic p-spin glasses'. Annals of Probability 47(4), 2230-2256. DOI: 10.1214/18-AOP1307; arXiv:1612.06359. URL: https://doi.org/10.1214/18-AOP1307
25. W.-K. Chen, D. Panchenko and E. Subag (2023). 'Generalized TAP Free Energy'. Communications on Pure and Applied Mathematics 76(7), 1329-1415. DOI: 10.1002/cpa.22040; arXiv:1812.05066. URL: https://doi.org/10.1002/cpa.22040
26. A. Adhikari, C. Brennecke, P. von Soosten and H.-T. Yau (2021). 'Dynamical Approach to the TAP Equations for the Sherrington-Kirkpatrick Model'. Journal of Statistical Physics 183(3), 35. DOI: 10.1007/s10955-021-02773-7; arXiv:2102.10178. URL: https://doi.org/10.1007/s10955-021-02773-7
27. E. Subag (2023). 'The free energy of spherical pure p-spin models: computation from the TAP approach'. Probability Theory and Related Fields 186, 715-734. DOI: 10.1007/s00440-023-01200-0; arXiv:2101.04352. URL: https://doi.org/10.1007/s00440-023-01200-0
28. M. C. Angelini, C. Lucibello, G. Parisi, G. Perrupato, F. Ricci-Tersenghi and T. Rizzo (2022). 'Unexpected upper critical dimension for spin glass models in a field predicted by the loop expansion around the Bethe solution at zero temperature'. Physical Review Letters 128, 075702. DOI: 10.1103/PhysRevLett.128.075702; arXiv:2103.17080. URL: https://arxiv.org/abs/2103.17080
29. W.-K. Chen and S. Tang (2024). 'On the TAP Equations via the Cavity Approach in the Generic Mixed p-Spin Models'. Communications in Mathematical Physics 405(4), 87. DOI: 10.1007/s00220-024-04971-2; arXiv:2207.05203. URL: https://doi.org/10.1007/s00220-024-04971-2
30. J.-C. Mourrat (2023). 'Un-inverting the Parisi formula'. arXiv:2308.10715; published as Annales de l'Institut Henri Poincaré, Probabilités et Statistiques 61(4) (2025). DOI: 10.1214/24-aihp1487. URL: https://arxiv.org/abs/2308.10715
31. M. Aguilar-Janita, V. Martin-Mayor, J. Moreno-Gordo and J. J. Ruiz-Lorenzo (2024). 'Evidence of a second-order phase transition in the six-dimensional Ising spin glass in a field'. Physical Review E 109, 055302. DOI: 10.1103/PhysRevE.109.055302; arXiv:2306.00569. URL: https://arxiv.org/abs/2306.00569
32. B. Vedula, M. A. Moore and A. Sharma (2024). 'Evidence that the de Almeida-Thouless transition disappears below six dimensions'. Physical Review E 110, 054131. DOI: 10.1103/PhysRevE.110.054131; arXiv:2402.03711. URL: https://doi.org/10.1103/PhysRevE.110.054131
33. M. Aguilar-Janita, V. Martin-Mayor, J. Moreno-Gordo and J. J. Ruiz-Lorenzo (2025). 'Evidence of de Almeida-Thouless line below six dimensions'. Journal of Statistical Mechanics: Theory and Experiment 2025(11), 113301. DOI: 10.1088/1742-5468/ae1a4a; arXiv:2510.14446. URL: https://doi.org/10.1088/1742-5468/ae1a4a
34. P. Charbonneau (2022). 'From the replica trick to the replica symmetry breaking technique'. IAMP News Bulletin 2022(October), 5-25; arXiv:2211.01802. URL: https://arxiv.org/abs/2211.01802