The τ2\tau_2-model — the second element of the fusion hierarchy of the six-vertex Yang–Baxter algebra evaluated in cyclic representations at qN=1q^N=1 — is the vehicle through which the chiral Potts model is solved, and its spectrum (eigenvalues and eigenstates) is completely characterized by Sklyanin's separation of variables even for the most general, site-dependent, non-self-adjoint cyclic representations. What is missing is the operator-level Baxter Q-operator formulation in that generality. Baxter's generalized Q-operator, constructed by gauge transformations from generalized discrete dilogarithm functions, has never been proven to form a commuting one-parameter family or to commute with the τ2\tau_2-transfer matrix; consequently the coincidence of τ2\tau_2-eigenstates with Q-eigenstates — equivalently, when QQ acts polynomially in the spectral variable, the completeness of the associated Bethe ansatz type equations off the self-adjoint subvariety, and on the chiral Potts curves the reverse inclusion between τ2\tau_2 and chiral Potts eigenspaces — remains unproven for general representations. The problem asks for a proof, or an explicit finite counterexample, of the commuting-family, commutation, and eigenstate-coincidence properties for general cyclic representations, which would either establish Baxter's proposed 'generalized chiral Potts model' as a genuine commuting integrable family with complete Bethe equations, or refute it.

Background

A large class of two-dimensional exactly solved lattice models is organized by the Yang–Baxter algebra: an associative algebra generated by the entries of an L-matrix satisfying the Yang–Baxter (star-triangle) commutation relations, whose transfer matrices commute and can be diagonalized exactly. The six-vertex Yang–Baxter algebra admits, at roots of unity qN=1q^N=1, a family of NN-dimensional cyclic representations: here two generators XX, YY of the algebra act as N×NN \times N matrices obeying XN=YN=1X^N = Y^N = \mathbf{1} and XY=q2YXXY = q^2 YX. The τ2\tau_2-model of Bazhanov and Stroganov is the integrable spin chain obtained by taking the transfer matrix τ2(λ)\tau_2(\lambda) that is the second element of the fusion hierarchy of this algebra — the family of transfer matrices generated by fusion, i.e. by composing L-matrices in larger auxiliary representations before taking the trace (hence the name) — in such cyclic representations; the quantum space is a tensor product of cyclic representations with site-dependent complex parameters, each site specified by a point in C3\mathbb{C}^3, and τ2(λ)\tau_2(\lambda) is generically neither self-adjoint nor normal. The chiral Potts model is the ZNZ_N-symmetric generalization of the Ising model: its Boltzmann weights satisfy the star-triangle relation but depend on rapidities uniformized by algebraic curves of genus greater than one (for N>2N>2), so they lack the difference property of vertex models. The two models are linked on a special subfamily of cyclic τ2\tau_2-representations, those parametrized by points on the chiral Potts curves: there the Baxter Q-operator of the τ2\tau_2-model coincides with the transfer matrix of the (inhomogeneous) chiral Potts model, which is how the chiral Potts spectrum is accessed through τ2\tau_2. A Baxter Q-operator is an auxiliary transfer matrix Q(λ)Q(\lambda) satisfying, together with τ2(λ)\tau_2(\lambda), a functional (TQ) relation of the form τ2(λ) Q(λ)=a(λ) Q(λq)+d(λ) Q(λq−1)\tau_2(\lambda)\,Q(\lambda) = a(\lambda)\,Q(\lambda q) + d(\lambda)\,Q(\lambda q^{-1}); when QQ acts polynomially in the spectral variable, its zeros obey Bethe ansatz type equations, and 'completeness' means that the solutions of these equations are in bijection with the full finite-dimensional spectrum. Sklyanin's separation of variables (SOV) is a method that represents the spectral problem of a transfer matrix in a basis of 'quantum separate variables' zz, in which the transfer matrix acts by multiplication and the eigenvalue function t(λ)t(\lambda) is characterized directly as a solution of a functional equation; it does not by itself produce a Q-operator, and the passage from the SOV characterization to Bethe equations requires one. Grosjean and Niccoli (2012) used SOV to characterize the complete τ2\tau_2-spectrum for the most general cyclic representations, proved simplicity of that spectrum, and proved polynomiality of the Baxter solutions and completeness of the Bethe ansatz type equations for a class of self-adjoint representations — a subvariety of parameter space on which τ2(λ)\tau_2(\lambda) is self-adjoint (and, on the chiral Potts curves, the chiral Potts transfer matrix is normal), so that simultaneous diagonalization follows. Earlier, Baxter (2004) had extended the eigenvalue analysis to completely general cyclic representations by constructing a 'generalized' Q-operator through gauge transformations built from generalized discrete dilogarithm functions, in effect proposing a 'generalized chiral Potts model' transfer matrix. However — as Grosjean and Niccoli stress when reproducing Baxter's construction in their Appendix B (Remark 8) — for these general non-self-adjoint representations it was never proven that the generalized Q-operators form a commuting family [Q(λ),Q(μ)]=0[Q(\lambda),Q(\mu)] = 0 for all λ,μ\lambda,\mu, nor that they commute with τ2(λ)\tau_2(\lambda); only the eigenvalue-level content is controlled, since the Baxter equations of the generalized Q-operator and of the SOV-induced Q-operator coincide up to a gauge transformation. In the introduction's words, the missing commutativity 'leaves the definition of generalized chP-model only formal'. The same gap reappears on the chiral Potts curves as Section 5.3 of the same paper: the chiral Potts transfer matrix TλchPT^{\mathrm{chP}}_\lambda and τ2(λ)\tau_2(\lambda) are each commuting one-parameter families, but their mutual commutativity is proven only at coincident spectral parameters, so the reverse eigenspace inclusion (every τ2\tau_2-eigenstate is a TλchPT^{\mathrm{chP}}_\lambda-eigenstate) and simultaneous diagonalizability are proven only on the self-adjoint subvariety where TλchPT^{\mathrm{chP}}_\lambda is normal.

Problem Statement

For the τ2\tau_2-model on a finite chain with general cyclic representations of the six-vertex Yang–Baxter algebra at qN=1q^N=1 — site-dependent parameters in C3\mathbb{C}^3, not restricted to the self-adjoint subvariety nor to the chiral Potts algebraic curves — settle the operator-level status of Baxter's generalized Q-operator construction:

  1. Prove, or disprove by an explicit counterexample, that the generalized Baxter Q-operator defined through the gauge-transformation kernel built from generalized discrete dilogarithm functions satisfies [Q(λ),Q(μ)]=0[Q(\lambda),Q(\mu)] = 0 for all spectral parameter values λ,μ\lambda,\mu, and [τ2(λ),Q(μ)]=0[\tau_2(\lambda),Q(\mu)] = 0, so that it defines a commuting integrable family (Baxter's proposed 'generalized chiral Potts model') on the full space of cyclic representations.

  2. Prove, or disprove, that in these general representations the eigenstates of τ2(λ)\tau_2(\lambda) coincide with the eigenstates of the generalized Q-operator — the direction of inclusion not implied by the Baxter equation — so that, whenever QQ acts polynomially in the spectral variable, the associated Bethe ansatz type equations are complete off the self-adjoint subvariety.

  3. The same question on the chiral Potts curves, where the relevant Q-operator is the inhomogeneous chiral Potts transfer matrix TλchPT^{\mathrm{chP}}_{\lambda}: prove, or disprove, the reverse eigenspace inclusion (every τ2\tau_2-eigenstate is a TλchPT^{\mathrm{chP}}_{\lambda}-eigenstate) and the simultaneous diagonalizability of τ2(λ)\tau_2(\lambda) and TλchPT^{\mathrm{chP}}_{\lambda} for general parameter choices in the chiral Potts representation space RNchPR_N^{\mathrm{chP}}, currently established only on the self-adjoint subvariety RNchP,s−adjR_N^{\mathrm{chP},\mathrm{s-adj}} where TλchPT^{\mathrm{chP}}_{\lambda} is normal.

An admissible answer is either a theorem covering the general cyclic representations as stated (or a precisely delimited classification of the representations for which the properties hold and fail, with boundary examples on both sides), or an explicit finite counterexample — a concrete cyclic representation with all parameters, the explicit τ2(λ)\tau_2(\lambda) and generalized Q(λ)Q(\lambda) matrices or kernels, and exact computations demonstrating the failure of the specific property being refuted. All three enumerated items are mandatory: a complete answer must settle each of them, either by establishing the property or by showing, through a counterexample, a classification, or boundary examples, where it fails. The independently refutable sub-claims are (a) [Q(λ),Q(μ)]=0[Q(\lambda),Q(\mu)] = 0, (b) [τ2(λ),Q(μ)]=0[\tau_2(\lambda),Q(\mu)] = 0, (c) the coincidence of τ2\tau_2- and QQ-eigenstates, and (d) the on-curve reverse eigenspace inclusion together with the simultaneous diagonalizability of τ2(λ)\tau_2(\lambda) and TλchPT^{\mathrm{chP}}_\lambda. An explicit counterexample to any one of (a)–(d) resolves the problem on the negative side, since it disproves the corresponding statement and therefore the claimed commuting-family and complete-Bethe structure as a whole; a proof of some of these sub-claims does not discharge the remaining ones. Item 3 is the restriction of items 1–2 to the chiral Potts representation space RNchPR_N^{\mathrm{chP}}, where the generalized Q-operator coincides with TλchPT^{\mathrm{chP}}_\lambda, and is therefore subsumed by items 1–2: a proof of items 1–2 settles item 3 for chiral Potts parameters, while a counterexample on the chiral Potts curves refutes items 1–2 as well. Results restricted to the self-adjoint subvariety, to alternating-rapidity specializations, to homogeneous on-curve quantum spaces, or to numerical evidence for small chains do not settle the problem as stated.

References

  1. N. Grosjean, G. Niccoli, 'The tau_2-model and the chiral Potts model revisited: completeness of Bethe equations from Sklyanin's SOV method', J. Stat. Mech. (2012) P11005. DOI 10.1088/1742-5468/2012/11/P11005, arXiv:1205.4614, https://arxiv.org/abs/1205.4614
  2. R. J. Baxter, 'Transfer Matrix Functional Relations for the Generalized tau_2 (t_q) Model', J. Stat. Phys. 117 (2004) 1–25. DOI 10.1023/B:JOSS.0000044062.64287.b9, https://doi.org/10.1023/B:JOSS.0000044062.64287.b9
  3. V. V. Bazhanov, Yu. G. Stroganov, 'Chiral Potts model as a descendant of the six-vertex model', J. Stat. Phys. 59 (1990) 799–817. DOI 10.1007/BF01025851, https://doi.org/10.1007/BF01025851
  4. N. Grosjean, J.-M. Maillet, G. Niccoli, 'On the form factors of local operators in the lattice sine-Gordon model', J. Stat. Mech. (2012) P10006. DOI 10.1088/1742-5468/2012/10/P10006, arXiv:1204.6307, https://arxiv.org/abs/1204.6307
  5. N. Grosjean, J.-M. Maillet, G. Niccoli, 'On the Form Factors of Local Operators in the Bazhanov–Stroganov and Chiral Potts Models', Ann. Henri Poincaré 16 (2015) 1103–1153. DOI 10.1007/s00023-014-0358-9, arXiv:1309.4701, https://arxiv.org/abs/1309.4701
  6. X. Xu, J. Cao, S. Cui, W.-L. Yang, K. Shi, Y. Wang, 'Off-diagonal Bethe Ansatz solution of the tau_2-model', JHEP 09 (2015) 212. DOI 10.1007/JHEP09(2015)212, arXiv:1507.03367, https://arxiv.org/abs/1507.03367
  7. R. Weston, 'Cyclic representations of Uq(sl^2)U_q(\hat{\mathfrak{sl}}_2) and its Borel subalgebras at roots of unity and Q-operators', SciPost Phys. Core 9, 016 (2026). DOI 10.21468/SciPostPhysCore.9.1.016, arXiv:2412.14811, https://arxiv.org/abs/2412.14811
  8. S.-s. Roan, 'Bethe equation of tau^(2)-model and eigenvalues of a finite-size transfer matrix of a chiral Potts model with alternating rapidities', J. Stat. Mech. (2008) P10001. DOI 10.1088/1742-5468/2008/10/P10001, arXiv:0805.1585, https://arxiv.org/abs/0805.1585
  9. H. Au-Yang, J. H. H. Perk, 'CSOS models descending from chiral Potts models: Degeneracy of the eigenspace and loop algebra', arXiv:1511.08523 (2015), https://arxiv.org/abs/1511.08523