The two-dimensional ϕ4\phi^4 theory enjoys an exact strong-weak (Chang) self-duality relating its classically symmetric and classically broken descriptions, while the planar Ising model enjoys the Kramers-Wannier (KW) duality relating its disordered and ordered phases. Since ϕ4\phi^4 theory is the long-distance effective description of Z2\mathbb{Z}_2-symmetric lattice spin systems, it is natural to conjecture that Chang duality is the continuum limit of lattice KW duality, but the precise map between the two dualities has never been derived. The problem is to establish or refute this conjecture by constructing an explicit dictionary between the two dualities — at the level of couplings, renormalization scheme, operators, and phases — and to determine whether a strict finite-volume realization of Chang duality requires Z2\mathbb{Z}_2 gauge degrees of freedom, as KW duality does at finite volume. The question was posed as an open problem by Sberveglieri, Serone, and Spada (JHEP 02 (2021) 098).

Background

The Euclidean ϕ4\phi^4 theory in two dimensions with Lagrangian L=12(∂ϕ)2+12m2ϕ2+g Nm(ϕ4)\mathcal{L} = \frac{1}{2}(\partial\phi)^2 + \frac{1}{2}m^2\phi^2 + g\,\mathcal{N}_m(\phi^4) is super-renormalizable (the coupling gg has positive mass dimension, so only finitely many diagrams diverge), and its continuum limit can be defined nonperturbatively by normal-ordering the interaction with respect to the mass scale mm — the prescription Nm\mathcal{N}_m that removes self-contractions at a vertex. For m2>0m^2>0 the classical potential has a single Z2\mathbb{Z}_2-symmetric minimum, yet Chang (Phys. Rev. D 13 (1976) 2778) showed that the theory undergoes a second-order phase transition to a phase with spontaneously broken ϕ→−ϕ\phi \to -\phi symmetry as the dimensionless coupling x=g/m2x = g/m^2 increases (the critical coupling is xc≈2.75–3x_c \approx 2.75\text{--}3 in the normalization of Rychkov and Vitale). His argument relied on a strong-weak duality — now called Chang duality — stating that the theory with m2>0m^2>0 and coupling xx admits an equivalent dual description with negative classical squared mass, L′=12(∂ϕ)2−14M2ϕ2+g NM(ϕ4)\mathcal{L}' = \frac{1}{2}(\partial\phi)^2 - \frac{1}{4}M^2\phi^2 + g\,\mathcal{N}_M(\phi^4), whose dimensionless coupling X=g/M2X = g/M^2 is related to xx by

log⁡x−π3x=log⁡X+π6X,\log x - \frac{\pi}{3x} = \log X + \frac{\pi}{6X},

solved in terms of the Lambert WW function (the inverse of w↦weww \mapsto w e^w). Dual branches exist only for x≥x∗=π/(3W(2/e))≈2.26x \geq x^* = \pi/(3W(2/e)) \approx 2.26; on the lower branch X→0X \to 0 as x→∞x \to \infty, so the strongly coupled direct description maps to a weakly coupled dual description quantized around a Z2\mathbb{Z}_2-breaking vacuum. Chang duality has been verified numerically at finite volume by Hamiltonian truncation (a method that diagonalizes the truncated Hamiltonian of the theory quantized on a circle; Rychkov and Vitale, Phys. Rev. D 93 (2016) 065014), proven to hold to all orders in perturbation theory and verified by Borel resummation of high-order series in the infinite-volume continuum theory (Serone, Spada, and Villadoro, JHEP 05 (2019) 047), and reinterpreted by Sberveglieri, Serone, and Spada (JHEP 02 (2021) 098) as an exact analytic change of renormalization scheme — a one-parameter (κ\kappa) family of which Chang's original relation is the κ=0\kappa = 0 member in d=2d=2 (the d=3d=3 analogue is Magruder duality).

On the lattice side, the ferromagnetic Ising model on a planar graph admits the Kramers-Wannier (KW) duality (Kramers and Wannier, Phys. Rev. 60 (1941) 252): an exact transformation of the partition function that maps the high-temperature (disordered) phase to the low-temperature (ordered) phase of a dual model on the dual graph, exchanging spin variables for disorder variables living on faces; on the square lattice the dual coupling obeys tanh⁡(βJ)=e−2β∗J∗\tanh(\beta J) = e^{-2\beta^* J^*}, and the self-dual point sinh⁡(2βJ)=1\sinh(2\beta J) = 1 coincides exactly with the critical point. At finite volume on a torus, the KW transformation does not map the spin model strictly to itself: it produces a Z2\mathbb{Z}_2 lattice gauge theory (spins on links with a plaquette constraint and topological sectors), so a proper finite-volume realization of the duality requires Z2\mathbb{Z}_2 gauge degrees of freedom. In the continuum, the corresponding structure is the gauging of the global Z2\mathbb{Z}_2 symmetry: starting from the critical point, the 2d Ising conformal field theory (CFT), gauging Z2\mathbb{Z}_2 implements KW duality as a non-invertible topological duality defect line (a topological operator whose fusion with itself does not obey group multiplication) that exchanges the order operator σ\sigma with the disorder operator μ\mu. The lattice origin of this continuum duality line has been made precise through lattice duality defects satisfying topological commutation relations (Aasen, Mong, and Fendley, J. Phys. A 49 (2016) 354001), and the gauging construction generalizes to higher dimensions (Choi, Cordova, Hsin, Lam, and Shao, Phys. Rev. D 105 (2022) 125016); the continuum limit of the lattice KW defect continues to be actively studied.

The bridge between the two pictures is the physical observation that the 2d ϕ4\phi^4 theory is the long-distance (Landau-Ginzburg) effective description of Z2\mathbb{Z}_2-symmetric lattice spin systems near criticality, so KW duality should persist in the continuum limit, where it is defined starting from the Ising CFT point. Sberveglieri, Serone, and Spada noted that it is then natural to conjecture that continuum KW duality is closely related to Chang duality — both exchange the symmetric/disordered phase with the broken/ordered phase of the same universality class — but that the precise map between the two dualities is unknown. The identification is not naive: KW duality is an exact symmetry of a discrete-spin lattice model whose self-dual point is the critical point, whereas Chang duality is a scheme-dependent statement about the interacting continuum ϕ4\phi^4 theory whose dual branches merge below the critical coupling; and no exact self-duality of the continuous-spin lattice ϕ4\phi^4 model is known that could serve as an intermediate rung.

Problem Statement

Determine the precise relation between the strong-weak (Chang) duality of the two-dimensional continuum ϕ4\phi^4 theory and the Kramers-Wannier (KW) duality of the planar Ising model, as posed by Sberveglieri, Serone, and Spada (JHEP 02 (2021) 098). Concretely: (1) establish or refute the conjecture that Chang duality is the continuum limit of lattice KW duality, by constructing an explicit map between the two dualities. This includes identifying which lattice Z2\mathbb{Z}_2 spin system and which KW transformation correspond to the continuum ϕ4\phi^4 theory and its Chang transformation; giving the dictionary between the lattice dual-coupling relation and the continuum coupling relation log⁡x−π/(3x)=log⁡X+π/(6X)\log x - \pi/(3x) = \log X + \pi/(6X) (or its scheme-generalized form, in which case specifying which renormalization scheme the lattice limit selects is part of the answer); and specifying how the ordered/disordered lattice phases and the order/disorder operators map onto the broken/symmetric continuum phases and their field (and kink) operators. (2) Determine whether a strict finite-volume realization of Chang duality requires introducing Z2\mathbb{Z}_2 gauge degrees of freedom, as the finite-volume realization of KW duality on a torus does — and if it does, construct that finite-volume definition; if it does not, exhibit a finite-volume duality realization without gauge fields. The conjecture in (1) may equally be refuted by showing that the two dualities are not related as continuum limit and descendant, provided the disagreement is demonstrated by a concrete, checkable property of the two dualities in the continuum limit; such a refutation settles (1) but does not by itself resolve the independent finite-volume deliverable (2), which must still be established whether (1) is established or refuted.

References

  1. G. Sberveglieri, M. Serone, G. Spada, "Self-Dualities and Renormalization Dependence of the Phase Diagram in 3d O(N)O(N) Vector Models", JHEP 02 (2021) 098, DOI 10.1007/JHEP02(2021)098, arXiv:2010.09737, https://arxiv.org/abs/2010.09737
  2. S.-J. Chang, "Existence of a second-order phase transition in a two-dimensional ϕ4\phi^4 field theory", Phys. Rev. D 13 (1976) 2778, DOI 10.1103/PhysRevD.13.2778, https://doi.org/10.1103/PhysRevD.13.2778
  3. S. Rychkov, L. G. Vitale, "Hamiltonian truncation study of the ϕ4\phi^4 theory in two dimensions. II. The Z2\mathbb{Z}_2-broken phase and the Chang duality", Phys. Rev. D 93 (2016) 065014, DOI 10.1103/PhysRevD.93.065014, arXiv:1512.00493, https://arxiv.org/abs/1512.00493
  4. M. Serone, G. Spada, G. Villadoro, "λϕ4\lambda\phi^4 theory — Part II. The broken phase beyond NNNN(NNNN)LO", JHEP 05 (2019) 047, DOI 10.1007/JHEP05(2019)047, arXiv:1901.05023, https://arxiv.org/abs/1901.05023
  5. H. A. Kramers, G. H. Wannier, "Statistics of the Two-Dimensional Ferromagnet. Part I", Phys. Rev. 60 (1941) 252, DOI 10.1103/PhysRev.60.252, https://doi.org/10.1103/PhysRev.60.252
  6. P. Romatschke, "On self-dualities for scalar ϕ4\phi^4 theory", arXiv:2602.18286 (2026), https://arxiv.org/abs/2602.18286
  7. P. Romatschke, U. Romatschke, "Testing scalar field self-dualities in d=2d=2 using a variational method", Phys. Rev. D 114 (2026) 014521, DOI 10.1103/25zd-4r9z, arXiv:2604.15988, https://arxiv.org/abs/2604.15988
  8. Y. Liang, "Physical reduced states and continuum characters of the lattice Kramers-Wannier defect", arXiv:2607.01137 (2026), https://arxiv.org/abs/2607.01137
  9. D. Aasen, R. S. K. Mong, P. Fendley, "Topological defects on the lattice: I. The Ising model", J. Phys. A 49 (2016) 354001, DOI 10.1088/1751-8113/49/35/354001, arXiv:1601.07185, https://arxiv.org/abs/1601.07185
  10. Y. Choi, C. Córdova, P.-S. Hsin, H. T. Lam, S.-H. Shao, "Noninvertible duality defects in 3+13+1 dimensions", Phys. Rev. D 105 (2022) 125016, DOI 10.1103/PhysRevD.105.125016, arXiv:2111.01139, https://arxiv.org/abs/2111.01139