A-localized states for clock models on trees and their extremal decomposition into glassy states

Christof Kuelske; Niklas Schubert

REVIEW 3 major objections 5 minor 47 references

A-localized states for clock models on trees and their extremal decomposition into glassy states

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read At sufficiently low temperature, the A-localized Gibbs states of $\mathbb{Z}_q$-clock models on regular trees are not pure: their extremal decomposition is atomless and supported on uncountably many inhomogeneous pure states.

desk verdict Solid low-temperature extension of the free-state glassy decomposition to A-localized clock states; the flaws are fixable, and the Peierls-bound worry about Lemma 4 comes from a misread exponent. read the letter →

arxiv 2411.10271 v1 pith:74RZYQES submitted 2024-11-15 math.PR

Christof Kuelske , Niklas Schubert This is my paper · ORCID

classification math.PR MSC 60K3582B2082B44

keywords Gibbsmeasuresontreestree-indexedMarkovchainsclockmodelsextremaldecompositionlocalizationreconstructionglassystatesPottsmodel

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes that, at sufficiently low temperature, the A-localized Gibbs states of $\mathbb{Z}_q$-valued clock models on regular trees are not extremal, even though their single-site marginals look sharply ordered: the mass concentrates on a finite subset $A \subset \mathbb{Z}_q$ of spin values. The central claim is that for any localization set with $|A| \geq 2$, such a state $\mu_A$ is a genuine convex mixture of uncountably many spatially inhomogeneous pure states, with an extremal decomposition measure that has no atoms. This matters because it shows tree geometry makes ordered states thermodynamically unstable in a strong sense: low-temperature pure phases are hidden, boundary-condition-dependent states, so a measurement of $\mu_A$ represents uncertainty over a continuous family of phases rather than a single equilibrium state. The same conclusion is new even for the Potts model, where previously only partial extremality statements were available for A-localized states. The proof works for all ferromagnetic nearest-neighbor clock potentials satisfying a mild cost bound (the $u, U, d$-condition), with no FKG or symmetry assumptions.

What carries the argument

The argument runs on a good site/bad site decomposition adapted to A-localization. Edges where the configuration changes spin value or touches $A^c$ are called A-irregular; a bad event is the presence of a contour whose boundary has too high a density of A-irregular edges. Lemma 1 is the load-bearing estimate: for $\mu_A$-almost every boundary condition $\omega$, the probability that the $\pi$-kernel at the root disagrees with $\omega$ is bounded by the indicator of the bad event plus a small error $\epsilon_1(\beta)$, while bad events themselves decay exponentially in contour size with rate $\lambda(\beta) \to \infty$. To prove this, the model is coarse-grained to a two-letter state space $\{A, A^c\}$, which destroys the Markov chain property except in the Potts case, and the coarse-grained sums are bounded by a non-stochastic matrix $M$ whose entries absorb the transition probabilities together with exponential weights; an induction called propagation of smallness over subgraphs splits the $e^{-\beta u}$ factors among the $d$ children of each vertex and yields the exponential decay. The reconstruction consequences then follow the branch-overlap route: single-site reconstruction gives a positive Edwards-Anderson parameter, and multi-site reconstruction on suitably thinned branches shows the thinned branch-overlap $\varphi^{\omega}$ concentrates near $1$ under $\pi(\cdot|\omega)$ but near $\sum_a \pi_A(a)^2$ under $\pi(\cdot|\omega')$, so typical kernels are mutually singular almost surely.

What would settle it

Direct numerical test on the Potts model ($q \geq 3$) on the binary tree, where explicit formulas exist: sample boundary conditions $\omega, \omega'$ independently from $\mu_A$ for a two-point localization set $A$ at large $\beta$, and estimate the distribution of the thinned branch overlap $\varphi^{\omega}$ under the conditional kernels $\pi(\cdot|\omega)$ and $\pi(\cdot|\omega')$. If the two distributions fail to separate — that is, if $\pi(\varphi^{\omega}|\omega)$ does not remain close to $1$ while $\pi(\varphi^{\omega}|\omega')$ stays below $\sum_a \pi_A(a)^2$ for most pairs — the multi-site reconstruction of Theorem 5 fails. A cheaper check: directly estimate $\mu_A(B_A(\gamma))$ for contours of increasing size at couplings near the threshold $(d^2+1)u = dU$; the bound (3.4) predicts exponential decay in $|\gamma|$ with rate $\lambda(\beta)$, and its failure would break the Peierls step.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is Theorem 2: for a $\mathbb{Z}_q$-valued nearest-neighbor clock model on a $d$-regular tree whose potential satisfies the $u, U, d$-bounds, for any $A \subset \mathbb{Z}_q$ with $|A| \geq 2$ and all $\beta$ large enough, the A-localized Gibbs state $\mu_A$ is not extremal, and its extremal decomposition measure $\alpha_{\mu_A}$ is supported on uncountably many inhomogeneous states, with $\alpha_{\mu_A}(\{\nu\}) = 0$ for every extremal $\nu$. Equivalently, $\mu_A = \int \nu\, \alpha_{\mu_A}(d\nu)$ over pure states, and the mixing measure is continuous. This is obtained through a restricted reconstruction statement: only spin values $a \in A$ can be recovered from the boundary at infinity (Theorem 3), which forces the Edwards-Anderson parameter to be strictly positive (Theorem 4); then a multi-site reconstruction on thinned branches (Theorem 5) shows that typical pairs of boundary conditions produce mutually singular pure states (Theorem 6), which is exactly the atomless property of the decomposition. The extremal states are described as low-temperature perturbations around mostly flat almost ground states with spin values in $A$.

Load-bearing premise

The argument assumes as its starting point that the A-localized state from reference [1] exists with the stated quantitative bounds — single-site mass essentially all on $A$, transitions out of $A$ exponentially rare at low temperature — and builds every subsequent estimate on those bounds; if those bounds failed or were unavailable for a given clock model, the atomless decomposition result would not follow from this paper's proof.

Editorial extensions

If this is right

If the main theorem is correct, every A-localized state with $|A| \geq 2$ at low temperature is a mixture of uncountably many inhomogeneous pure states, so no single boundary condition at infinity can account for its local statistics.
The restricted reconstruction bound (Theorem 3) quantifies the information flow: initial spins in $A$ stay recoverable at infinity with probability close to 1, while spins outside $A$ are lost, a sharp asymmetry inside one Gibbs state.
The Edwards-Anderson parameter is strictly positive at low temperature (Theorem 4), giving a concrete order parameter for the non-extremality of $\mu_A$.
Typical extremal measures are almost surely mutually singular (Theorem 6), so the decomposition is genuinely continuous rather than a countable sum of pure states.
The result is new already for Potts-model A-localized states, where previously only partial extremality statements were known.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: by analogy with the Potts case in reference [27], $\mu_A$ may be extremal in an intermediate temperature window below the uniqueness threshold, with the atomless glassy decomposition appearing only at low temperature; the authors leave this open.
Editorial inference: the one-step-jump condition $(d^2+1)u > dU$ could likely be relaxed to multi-step returns from $A^c$ into $A$, which would extend the result to models with slower relaxation in spin space, such as p-SOS-type models.
Editorial inference: a testable quantitative prediction is that the information loss rate for initial spins in $A^c$ is governed by the second eigenvalue of the transition matrix restricted to $A$, while recovery of spins in $A$ is governed by the exponentially small bad-event probability.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The new thing here is real: the paper takes the free-state extremal decomposition machinery from [8] and makes it work for A-localized clock states, including Potts A-localized states, where previously only existence and non-convex-combination results were known. The A-irregular edges and the A/Ac coarse-graining are genuine technical additions, and they let the authors prove the Peierls-type bound (Lemma 1) that the whole argument hangs on. Sections 3 and 4.1 are written carefully, and the restricted single-site reconstruction plus the almost-sure singularity of the π-kernels are valuable results in their own right. I did not find a circularity problem: the paper leans on [1] and [8], but those are external results with independent proofs, and the theorem here is not baked into the inputs.

The soft spots are real but minor. The proof of Theorem 4 uses the condition (1 - C2 e^{-βu})^{-1} < |A|-1, which is impossible for |A|=2; the theorem includes |A|=2, so this case needs a slightly sharper bound on the single-site marginal or a separate argument. Lemma 5 says beta 'small enough' where the argument needs beta large; that is a typo-level fix. Theorem 2 states A⊂Zq with |A|≥2 but Lemma 1 and all the reconstruction theorems require A⊊Zq; the statement should either exclude A=Zq or point out that the A=Zq case is covered by [8].

About the stress-test concern: I checked the exponent in Lemma 4 carefully. The second rate in c is u/d, not 1/d — the induction step splits e^{-βu} uniformly over the d children, which is exactly why the u/d rate appears in Remark 2 and in the g(1) term. So the counterexample d=2, u=1/2, U=3/4 gives c = min{0.5, 0.25} = 0.25 and the induction closes. If the published formula literally typesets 1/d, that is a typo that should be fixed, but as a mathematical objection the stress-test does not land.

The reader's CONDITIONAL verdict is about right. The main theorem is very likely correct, the proof strategy is sound, and the issues I found do not threaten the central argument. This paper deserves a serious referee: it extends known results in a nontrivial way and introduces tools that will be used for other localized states. My recommendation is to send it to review, with the referee asked to verify the |A|=2 case in Theorem 4 and to clean up the statement typos. I would cite it if I work on tree Gibbs measures, and I would bring it to a reading group on disordered systems or tree-indexed Markov chains.

Referee Report

3 major / 5 minor

Summary. The paper studies Z_q-valued ferromagnetic clock models on d-regular trees with nearest-neighbor potentials satisfying the u,U,d-bounds (2.2). Building on the construction of homogeneous tree-indexed Markov chain Gibbs states μ_A localized on a subset A⊂Z_q from [1], it claims that for |A|≥2 and sufficiently large inverse temperature, μ_A is not extremal and its extremal decomposition measure is atomless, supported on uncountably many spatially inhomogeneous pure states. The proof introduces a new notion of A-irregular edges and a coarse-grained state space, proving a Peierls-type bound with errors (Lemma 1), then derives restricted single-site reconstruction (Theorem 3), non-extremality via the Edwards-Anderson parameter (Theorem 4), multi-site reconstruction on thinned branches (Theorem 5), and almost-sure singularity of the π-kernels (Theorem 6), which yields the atomless decomposition.

Significance. If the proof is correct, this is a valuable extension of the recent atomless decomposition results for free states of Potts and clock models to the non-symmetric A-localized states, including the Potts model as a special case. The proposed A-irregular edge decomposition and the local coarse-graining method are genuinely new technical tools that may be useful in other tree models. The paper also makes the interesting observation that reconstruction holds only for spin values in the localization set A, not for all states. The explicit exponential bounds and the clear statement of the u,U,d-regime are strengths. However, the current version contains a load-bearing technical gap in the main Peierls estimate and a separate gap in the proof of non-extremality for |A|=2, both of which need to be repaired before the main claims are established.

major comments (3)

[§3.2, Lemma 4 and Lemma 1(b)] The constant c in Lemma 4 (and in the definition of λ(β) in Lemma 1) is specified as c = min{ (d^2+1)/d · u − U, 1/d }. However, the induction step combining equations (3.23) and (3.24) requires c to be no larger than each of (d^2+1)/d · u − U, (d−2+1/d)·u, and u/d. The stated value can violate the last two requirements. For example, with d=2, u=1/2, U=3/4, which satisfies (2.2) because 5u = 2.5 > 2U = 1.5, the stated c equals min{0.5, 0.5} = 0.5, while u/d = 0.25 and (d−2+1/d)u = 0.25. Consequently f(β) = (2C3+C1)e^{−0.5β} decays faster than the terms C1 e^{−0.25β+t} and C3 e^{−0.25β+t} appearing in the prefactor, and the closing inequality P ≤ f(β)e^t + 1 cannot hold for large β. Thus the induction in Lemma 4, and hence the Peierls-type bound (3.4) of Lemma 1(b), is not proved for all potentials satisfying (2.2). This is load-bearing for Theorems 2, 3, 5 and 6. A likely repair is to replace the second entry 1/d by u/d and to re-check the constant c̃ in Lemma 1, but as written the proof has a real gap.
[§4.1, proof of Theorem 4] The proof of Theorem 4 chooses β large enough such that (1 − C2 e^{−βu})^{-1} < |A| − 1. For |A| = 2 this condition is impossible, because the left-hand side is strictly larger than 1 while |A|−1 = 1. Since Theorem 2 explicitly includes |A| = 2, the argument as written does not cover this case. The final bound (4.3) may still be true, for instance by using the spin-space symmetry that the authors state holds for |A|=2, but that argument is not given. This gap directly affects the proof of part i) of Theorem 2.
[§4.2, Lemmas 6–8] The multi-site reconstruction theorem (Theorem 5) and the almost-sure singularity theorem (Theorem 6) rely on Lemmas 6, 7 and 8, but the proofs of these lemmas are only sketched and refer to [8] for the main ideas. In particular, Lemma 8's covariance bound mixes the exponential decay e^{−c1 λ(β)} with the Perron-Frobenius decay |λ2(PA)|^{c2/6}. In the low-temperature A-localized regime the second eigenvalue of PA may be close to 1, so the resulting rate may not be small uniformly in β. Since these lemmas are essential for the atomless-decomposition claim, the authors should provide a complete and self-contained proof, or at least a precise adaptation of [8] with explicit estimates for the A-localized case.

minor comments (5)

[§3.3, Lemma 5] The statement 'If β is small enough, h possesses a unique minimizer' should read 'If β is large enough'. The minimizer formula (3.26) gives t* ≥ 0 only when f(β) is sufficiently small, which happens for large β, not small β.
[§4.2.1, Lemma 8] The statement 'For small enough β' should read 'For large enough β', since the covariance decay uses λ(β) → ∞ as β → ∞.
[§3.2, definition of c in Lemma 4 and Lemma 1] The expression c := min{ (d^2+1)/d · u − U, 1/d } is dimensionally inconsistent: the second argument should presumably be u/d (an energy), not 1/d. This is closely related to Major Comment 1 and should be corrected in both Lemma 1 and Lemma 4.
[§4.1, proof of Theorem 4] The sentence 'Choosing β large enough such that (1−C2e^{−βu})^{-1} < (|A|−1)' is also problematic for |A|=2, as noted in Major Comment 2; even for |A|≥3 the proof should explain how this condition is compatible with (iv) of Proposition 1.
[Throughout] The notation 1/d in the definition of c and the phrase 'small enough β' in two lemmas suggest typographical errors that should be fixed during revision; these typos do not affect the overall strategy but are confusing to the reader.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main theorem is derived from new Peierls-type estimates, not built into the inputs.

full rationale

The derivation chain is not circular. The paper's main result (Theorem 2) is the non-extremality and atomless extremal decomposition of the A-localized states mu_A. These conclusions are not assumed in any input. The existence and quantitative localization bounds for mu_A are imported from Theorem 1, restated from [1] (Theorem 3.1), and Proposition 1 is proved in Appendix B from that theorem. Although [1] shares an author with the present paper, it is a published theorem with an independent proof whose assumptions (strong coupling, u, U, d-bounds) do not include extremality or the atomless property of the decomposition measure. The proof of Theorem 2 then rests on new technical content: the A-irregular edge notion, the coarse-graining Lemma 2, the propagation Lemma 4, and the Peierls-type bound Lemma 1. These are proved inside the paper. The multi-site reconstruction argument adapts the strategy of [8], but [8] concerns free states and is used as a proof template, not as a black-box theorem that already contains the conclusion. No fitted parameter is renamed as a prediction, no normalization is defined in terms of the target quantity, and no uniqueness claim is imported from the authors' prior work to force the choice of states. The skeptical concern about the constant c in Lemma 4 exceeding the decay rate needed to close the induction is an internal estimate-correctness issue, not a circularity: it does not identify an input that is equivalent to the output by construction. Overall, the central claim has independent mathematical content and the load-bearing dependencies are external, checkable results.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The model has no fitted free parameters; constants C1 through C5 are explicit bounds. The central claim rests on external theorems: existence and localization bounds of mu_A from [1], and the textbook Gibbs theory of extremal decompositions. No new physical entities are introduced.

assumptions (6)

domain assumption Existence and quantitative concentration of mu_A from Theorem 1 of [1] (Theorem 3.1 there).
The current paper does not rederive the construction of mu_A; all estimates in Sections 3 and 4 start from the bounds (2.9) and Proposition 1.
domain assumption Ferromagnetic nearest-neighbor clock potentials with discrete rotational symmetry and u,U,d-bounds (2.2), including (d^2+1)u > dU.
This modeling assumption ensures the one-step jump from the complement of A back to A is sufficiently likely; Lemma 4's propagation estimates depend on it.
domain assumption A is a proper subset of Z_q with |A| >= 2.
The main theorems require the localization set to be a proper subset of the state space; Lemma 1 and Theorems 3-6 assume this, and Theorem 2's statement should carry it explicitly.
standard math General Gibbs theory: DLR equations, extremal decomposition via pi-kernels, tail triviality of extremals.
Invoked in Appendix A and Section 2.1; these are textbook results (Georgii, Friedli-Velenik), not proved here.
domain assumption Tree-indexed Markov chain representation of mu_A with strictly positive, aperiodic transition matrix P_A.
Used for the coarse-grained expectation (3.6) and for exponential convergence in Lemma 8; strict positivity follows from Proposition 1.
standard math Exponential decay of correlations along sufficiently thinned branches under extremal Gibbs measures and under mu_A.
Used in Lemmas 6-8 via Borel-Cantelli, Chebyshev, and Perron-Frobenius arguments; standard but not fully proved in this paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A-localized states for clock models on trees and their extremal decomposition into glassy states." pith.science (2026). https://pith.science/paper/74RZYQES

@misc{pith2026241110271,
  author       = {Pith},
  title        = {Pith review of: A-localized states for clock models on trees and their extremal decomposition into glassy states},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/74RZYQES}},
  note         = {Machine review of arXiv:2411.10271}
}

abstract

We consider $\mathbb{Z}_q$-valued clock models on a regular tree, for general classes of ferromagnetic nearest neighbor interactions which have a discrete rotational symmetry. It has been proved recently that, at strong enough coupling, families of homogeneous Markov chain Gibbs states $\mu_A$ coexist whose single-site marginals concentrate on $A\subset \mathbb{Z}_q$, and which are not convex combinations of each other [AbHeKuMa24]. In this note, we aim at a description of the extremal decomposition of $\mu_A$ for $|A|\geq 2$ into all extremal Gibbs measures, which may be spatially inhomogeneous. First, we show that in regimes of very strong coupling, $\mu_A$ is not extremal. Moreover, $\mu_A$ possesses a single-site reconstruction property which holds for spin values sent from the origin to infinity, when these initial values are chosen from $A$. As our main result, we show that $\mu_A$ decomposes into uncountably many extremal inhomogeneous states. The proof is based on multi-site reconstruction which allows to derive concentration properties of branch overlaps. Our method is based on a new good site/bad site decomposition adapted to the $A$-localization property, together with a coarse graining argument in local state space.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 25 canonical work pages

[8]

To appear in Ann

Coquille, L., K¨ ulske, C., Le Ny, A.: Non-Atomicity of th e extremal decomposition of the free state for ﬁnite-spin models on Cayley trees. To appear in Ann. Inst . H. Poincar´ e Probab. Statist.. Preprint https://arxiv.org/abs/2310.11101 (2024)

arXiv 2024
[1]

Abbondandolo, A., Henning, F., K¨ ulske, C., Majer, P.: I nﬁnite-Volume States with Irreducible Localization Sets for Gradient Models on Trees. J. Stat. Phy s. 191, Paper No. 63 (2024). https://doi.org/10.1007/s10955-024-03278-9

work page doi:10.1007/s10955-024-03278-9 2024
[2]

Arguin, L.-P., Chatterjee, S.: Random overlap structur es: properties and applications to spin glasses. Probab. Theory Relat. Fields 156, 375-413 (2013). https://doi.org/10.1007/s00440-012-0431-6

work page doi:10.1007/s00440-012-0431-6 2013
[3]

Arguin, L.-P., Damron, M., Newman, C.M., Stein, D.L.: Un iqueness of ground states for short-range spin glasses in the half-plane. Comm. Math. Phys. 300, 641-657 (2010). https://doi.org/10.1007/s00220-010-1130-8

doi:10.1007/s00220-010-1130-8 2010
[4]

A Mathematical Perspective (Cambridge Series in Statistical and Probabilistic Mathematics), vol

Bovier, A.: Statistical Mechanics of Disordered System s. A Mathematical Perspective (Cambridge Series in Statistical and Probabilistic Mathematics), vol. 18, pp. x iv+312. Cambridge University Press, Cambridge (2006). https://doi.org/10.1017/CBO9780511616808

doi:10.1017/cbo9780511616808 2006
[5]

Br´ emaud, P.: Markov chains—Gibbs ﬁelds, Monte Carlo si mulation and queues (Texts in Applied Mathematics), vol. 31, pp. xvi+557, Second editi on. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-45982-6

doi:10.1007/978-3-030-45982-6 2020
[6]

Brito, A.F., Redinz, J.A., Plascak, J.A.: Two-dimensio nal XY and clock models stud- ied via the dynamics generated by rough surfaces. Phys. Rev. E 81, 031130 (2010). https://doi.org/10.1103/PhysRevE.81.031130

work page doi:10.1103/physreve.81.031130 2010
[7]

Caputo, P., Lubetzky, E., Martinelli, F., Sly, A., Tonin elli, F.L.: Dynamics of (2 + 1)-dimensional SOS surfaces above a wall: Slow mixing induced by entropic repul sion. Ann. Probab. 42(4), 1516– 1589 (2014). https://doi.org/10.1214/13-AOP836

work page doi:10.1214/13-aop836 2014

Show all 47 references

[9]

Coquille, L., K¨ ulske, C., Le Ny, A.: Extremal inhomogen eous Gibbs states for SOS- models and ﬁnite-spin models on trees. J. Stat. Phys. 190, Paper No. 71, 26 (2023). https://doi.org/10.1007/s10955-023-03081-y

2023 doi
[10]

Cotar, C., Deuschel, J.-D., M¨ uller, S.: Strict Convex ity of the Free Energy for a Class of Non-Convex Gradient Models. Comm. Math. Phys. 286, 359–376 (2009). https://doi.org/10.1007/s00220-008-0659-2

2009 doi
[11]

Cotar, C., K¨ ulske, C.: Uniqueness of gradient Gibbs me asures with disorder. Probab. Theory Relat. Fields 162, 587–635 (2015). https://doi.org/10.1007/s00440-014-0580-x

2015 doi
[12]

Dario, P., Harel, M., Peled, R.: Random-ﬁeld random sur faces. Probab. Theory Relat. Fields 184, 91–158 (2023). https://doi.org/10.1007/s00440-022-01179-0

2023 doi
[13]

Dominguez, T., Mourrat, J.-C.: Statistical mechanics of mean-ﬁeld disordered systems—a Hamilton- Jacobi approach (Zurich Lectures in Advanced Mathematics) , pp. vi+361. EMS Press, Berlin (2024). https://doi.org/10.4171/zlam/32

2024 doi
[14]

van Enter, A.C.D., K¨ ulske, C.: Nonexistence of Random Gradient Gibbs Measures in Continuous Interface Models in d = 2. Ann. Appl. Probab. 18(1), 109–119 (2008). https://doi.org/10.1214/07-AAP446

2008 doi
[15]

The Annals of Applied Probability 10(2), 410–433 (2000)

Evans, W., Kenyon, C., Peres, Y., Schulman, L.J.: Broad casting on trees and the Ising model. The Annals of Applied Probability 10(2), 410–433 (2000). https://doi.org/10.1214/aoap/1019487349

2000
[16]

A concrete mathematical introduction, pp

Friedli, S., Velenik, Y.: Statistical mechanics of lat tice systems. A concrete mathematical introduction, pp. xix+622. Cambridge University Press, Cambridge (2018). https://doi.org/10.1017/9781316882603

2018 doi
[17]

Fr¨ ohlich, J., Spencer, T.: The Kosterlitz-Thouless t ransition in two-dimensional abelian spin systems and the Coulomb gas. Comm. Math. Phys. 81(4), pp. 527–602 (1981). http://projecteuclid.org/euclid.cmp/1103920388

1981
[18]

Funaki, T., Spohn, H.: Motion by Mean Curvature from the Ginzburg-Landau Interface Model. Comm. Math. Phys. 185, 1–36 (1997). https://doi.org/10.1007/s002200050080

1997 doi
[19]

Gandolfo, D., Maes, C., Ruiz, J., Shlosman, S.: Glassy s tates: the free Ising model on a tree. J. Stat. Phys. 180, 227–237 (2020). https://doi.org/10.1007/s10955-019-02382-5

2020 doi
[20]

Gandolfo, D., Ruiz, J., Shlosman, S.: A manifold of pure Gibbs states of the Ising model on a Cayley tree. J. Stat. Phys. 148, 999–1005 (2012). https://doi.org/10.1007/s10955-012-0574-y

2012 doi
[21]

Ganikhodjaev, N.N.: On Pure Phases of the Ferromagneti c Potts Model on the Bethe lattice. Dokl. Acad. Nauk. Rep. Uzb. 6, 4–7 (1992)

1992
[22]

Georgii, H.-O.: Gibbs measures and phase transitions ( de Gruyter Studies in Mathematics), Second edition, vol. 9, pp. xiv+545. W alter de Gruyter & Co., Berlin (2011). https://doi.org/10.1515/9783110250329

2011 doi
[23]

Preprint https://hal.science/hal-04653617v2 (2024) 31 Extremal decomposition of A-localized states

Goswami, A., Kumar, R., Gope, M., Sahoo, S.: Phase trans itions in q-state clock model. Preprint https://hal.science/hal-04653617v2 (2024) 31 Extremal decomposition of A-localized states

2024
[24]

In: Longo, R., Mathematical physics in mathematics and physics

Guerra, F.: Sum rules for the free energy in the mean ﬁeld spin glass model. In: Longo, R., Mathematical physics in mathematics and physics. Quantum a nd Operator Algebraic Aspects (Siena, 2000), Fields Inst. Commun. Vol. 30, pp. 161–170, Amer. Math . Soc., Providence, RI (200...

2001 doi
[25]

Guerra, F.: Broken replica symmetry bounds in the mean ﬁ eld spin glass model. Comm. Math. Phys. 233, 1–12 (2003). https://doi.org/10.1007/s00220-002-0773-5

2003 doi
[26]

Krzakala, F., M´ ezard, M., Sausset, F., Sun, Y.F., Zdeb orov´ a, L.: Statistical-physics-based reconstruction in compressed sensing. Phys. Rev. X 2, 021005 (2012). https://doi.org/10.1103/PhysRevX.2.021005

2012 doi
[27]

Random Struct

K¨ ulske, C., Rozikov, U.A.: Fuzzy transformations and extremality of Gibbs measures for the Potts model on a Cayley tree. Random Struct. Alg. 50, 636–678 (2017). https://doi.org/10.1002/rsa.20671

2017 doi
[28]

Lammers, P., Toninelli, F.: Height function localisat ion on trees. Combin. Probab. Comput. 33(1), 50–64 (2024). https://doi.org/10.1017/s0963548323000329

2024 doi
[29]

Le Ny, A.: Introduction to (generalized) Gibbs measure s (Ensaios Matem´ aticos [Mathematical Surveys]), vol. 15, pp. ii+126. Sociedade Brasileira de Matem´ atica, R io de Janeiro (2008). ISBN: 978-85-85818-38-8

2008
[30]

Maes, C., Shlosman, S.: Rotating states in driven clock - and XY-models. J. Stat. Phys. 144, 1238–1246 (2011). https://doi.org/10.1007/s10955-011-0325-5

2011 doi
[31]

M´ ezard, M., Parisi, G., Virasoro, M.A.: Spin glass the ory and beyond (W orld Scientiﬁc Lecture Notes in Physics), vol. 9, pp. xiv+461. W orld Scientiﬁc Publ ishing Co., Inc., Teaneck, NJ (1987). https://doi.org/10.1142/0271

1987 doi
[32]

The Annals of Applied Probability 11(1), 285–300 (2001)

Mossel, E.: Reconstruction on Trees: Beating the Secon d Eigenvalue. The Annals of Applied Probability 11(1), 285–300 (2001). https://doi.org/10.1214/aoap/998926994

2001
[33]

Nuclear Physics B 854, 780–814 (2012)

Ortiz, G., Cobanera, E., Nussinov, Z.: Dualities and th e phase diagram of the p-clock model. Nuclear Physics B 854, 780–814 (2012). https://doi.org/10.1016/j.nuclphysb.2011.09.012

2012 doi
[34]

Journal of Physics A: Mathematical and General 13, L115 (1980)

Parisi, G.: A sequence of approximated solutions to the S-K model for spin glasses. Journal of Physics A: Mathematical and General 13, L115 (1980). https://doi.org/10.1088/0305-4470/13/4/009

1980 doi
[35]

Parisi, G.: Field theory, disorder and simulations (W o rld Scientiﬁc Lecture Notes in Physics), vol. 49, pp. vi+503. W orld Scientiﬁc Publishing C o., Inc., River Edge, NJ (1992). https://doi.org/10.1142/9789812799050

1992 doi
[36]

Scholten, P.D., Irakliotis, L.J.: Critical behavior o f the q-state clock model in three dimensions. Phys. Rev. B 48, 1291–1294 (1993). https://doi.org/10.1103/PhysRevB.48.1291

1993 doi
[37]

Soci´ et´ e math´ ematique de France (2005)

Sheﬃeld, S.: Random surfaces (Ast´ erisque 304). Soci´ et´ e math´ ematique de France (2005). http://www.numdam.org/item/AST_2005__304__R1_0/

2005
[38]

xviii+317

Stein, D.L., Newman, C.M.: Spin glasses and complexity (Primers in Complex Systems), pp. xviii+317. Princeton University Press, Princeton, NJ (2013). https://doi.org/10.1515/9781400845637

2013 doi
[39]

Talagrand, M.: The Parisi formula. Ann. of Math. 163, 221–263 (2006). https://doi.org/10.4007/annals.2006.163.221

2006 doi
[40]

Volum e I: Basic examples (Ergebnisse der Mathematik und ihrer Grenzgebiete

Talagrand, M.: Mean ﬁeld models for spin glasses. Volum e I: Basic examples (Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Math ematics]...

2011 doi
[41]

Volum e II: Advanced replica-symmetry and low temper- ature (Ergebnisse der Mathematik und ihrer Grenzgebiete

Talagrand, M.: Mean ﬁeld models for spin glasses. Volum e II: Advanced replica-symmetry and low temper- ature (Ergebnisse der Mathematik und ihrer Grenzgebiete. 3 . Folge. A Series of Modern Surveys in Mathe- matics [Results in Mathematics and Related Areas. 3rd Serie s. A Ser...

2011 doi
[42]

Todoroki, N., Ueno, Y., Miyashita, S.: Ordered phase an d phase transitions in the three-dimensional general- ized six-state clock model. Phys. Rev. B 66, 214405 (2002). https://doi.org/10.1103/PhysRevB.66.214405

2002 doi
[43]

Tomita, Y., Okabe, Y.: Probability-changing cluster a lgorithm for two-dimensional XY and clock models. Phys. Rev. B 65, 184405 (2002). https://doi.org/10.1103/PhysRevB.65.184405

2002 doi
[44]

Ueno, Y.: Description of ordering and phase transition s in terms of local connectivity: Proof of a novel type of percolated state in the general clock model. J . Statist. Phys. 80, 841–873 (1995). https://doi.org/10.1007/BF02178558

1995 doi
[45]

Ueno, Y., Kasono, K.: Incompletely ordered phases and p hase transitions in the three-dimensional general clock model. Phys. Rev. B 48, 16471 (1993). https://doi.org/10.1103/PhysRevB.48.16471

1993 doi
[46]

Velenik, Y.: Localization and delocalization of rando m interfaces. Probab. Surv. 3, 112–169 (2006). https://doi.org/10.1214/154957806000000050

2006 doi
[47]

Zachary, S.: Countable state space Markov random ﬁelds and Markov chains on trees. Ann. Probab. 11(4), 894–903 (1983). https://doi.org/10.1214/aop/1176993439 32

1983

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [8]

To appear in Ann

Coquille, L., K¨ ulske, C., Le Ny, A.: Non-Atomicity of th e extremal decomposition of the free state for ﬁnite-spin models on Cayley trees. To appear in Ann. Inst . H. Poincar´ e Probab. Statist.. Preprint https://arxiv.org/abs/2310.11101 (2024)

arXiv 2024

[2] [1]

Abbondandolo, A., Henning, F., K¨ ulske, C., Majer, P.: I nﬁnite-Volume States with Irreducible Localization Sets for Gradient Models on Trees. J. Stat. Phy s. 191, Paper No. 63 (2024). https://doi.org/10.1007/s10955-024-03278-9

work page doi:10.1007/s10955-024-03278-9 2024

[3] [2]

Arguin, L.-P., Chatterjee, S.: Random overlap structur es: properties and applications to spin glasses. Probab. Theory Relat. Fields 156, 375-413 (2013). https://doi.org/10.1007/s00440-012-0431-6

work page doi:10.1007/s00440-012-0431-6 2013

[4] [3]

Arguin, L.-P., Damron, M., Newman, C.M., Stein, D.L.: Un iqueness of ground states for short-range spin glasses in the half-plane. Comm. Math. Phys. 300, 641-657 (2010). https://doi.org/10.1007/s00220-010-1130-8

doi:10.1007/s00220-010-1130-8 2010

[5] [4]

A Mathematical Perspective (Cambridge Series in Statistical and Probabilistic Mathematics), vol

Bovier, A.: Statistical Mechanics of Disordered System s. A Mathematical Perspective (Cambridge Series in Statistical and Probabilistic Mathematics), vol. 18, pp. x iv+312. Cambridge University Press, Cambridge (2006). https://doi.org/10.1017/CBO9780511616808

doi:10.1017/cbo9780511616808 2006

[6] [5]

Br´ emaud, P.: Markov chains—Gibbs ﬁelds, Monte Carlo si mulation and queues (Texts in Applied Mathematics), vol. 31, pp. xvi+557, Second editi on. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-45982-6

doi:10.1007/978-3-030-45982-6 2020

[7] [6]

Brito, A.F., Redinz, J.A., Plascak, J.A.: Two-dimensio nal XY and clock models stud- ied via the dynamics generated by rough surfaces. Phys. Rev. E 81, 031130 (2010). https://doi.org/10.1103/PhysRevE.81.031130

work page doi:10.1103/physreve.81.031130 2010

[8] [7]

Caputo, P., Lubetzky, E., Martinelli, F., Sly, A., Tonin elli, F.L.: Dynamics of (2 + 1)-dimensional SOS surfaces above a wall: Slow mixing induced by entropic repul sion. Ann. Probab. 42(4), 1516– 1589 (2014). https://doi.org/10.1214/13-AOP836

work page doi:10.1214/13-aop836 2014