arxiv: 2605.12708 · v1 · submitted 2026-05-12 · 🧮 math.PR · math-ph· math.MP

Recognition: unknown

Equilibrium Biphasicity and Non-Binary Pathwise Confinement in Stochastic Ising Models

Jean-Gabriel Attali

Authors on Pith no claims yet

Pith reviewed 2026-05-14 19:42 UTC · model grok-4.3

classification 🧮 math.PR math-phmath.MP

keywords Ising modelGlauber dynamicspathwise confinementGibbs phasesmagnetizationstochastic dynamicsequilibrium states

0 comments

The pith

In low-temperature 2D Ising models with Glauber dynamics, some trajectories remain confined to zero magnetization yet their averages converge to the equal mixture of the two pure phases.

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

The paper constructs initial configurations for the two-dimensional Ising model at low temperature whose Glauber dynamics keep the magnetization converging to zero. These paths lie in a set R that has measure zero under the pure phases but is pathwise absorbing for these starts. Despite the confinement, the long-run time averages of the process still converge to one-half the sum of the plus and minus phase measures. This shows that the pathwise behavior of the dynamics is more complex than the classification of equilibrium Gibbs states would suggest, without introducing a new equilibrium phase.

Core claim

The central claim is that there exists a dense family of initial configurations whose trajectories under Glauber dynamics are confined to the centered sector C0 where magnetization Mn(σ) goes to 0, which is contained in R = Omega minus (P+ union P-), yet the Cesaro averages converge to 1/2 (μ+ + μ-). This establishes that the pathwise absorbing structure is richer than the Gibbs phase classification.

What carries the argument

The centered sector C0 = {σ : Mn(σ) → 0}, which carries the pathwise confinement while allowing Cesaro averages to mix the two phases.

If this is right

The set R, though null under pure phases, supports dense pathwise confined subsets.
Pathwise confinement to neutral magnetization does not create a third equilibrium state.
Cesaro averages can still recover the mixture of extremal states even from confined paths.
The absorbing geometry of the dynamics exceeds the static Gibbs classification.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

This suggests that long-time behavior in spin systems can have transient-like confinement without altering equilibrium measures.
Similar constructions might apply to other models with multiple phases to reveal hidden pathwise structures.
Testing numerically for small lattices could show if such confined trajectories persist.

Load-bearing premise

The assumption that at low temperatures the two pure phases exhaust all extremal equilibrium states for the two-dimensional Ising model under Glauber dynamics.

What would settle it

Finding an initial configuration in the constructed dense family where the trajectory escapes C0 or where the Cesaro average fails to converge to the mixture of the two phases.

read the original abstract

For the low-temperature two-dimensional Ising model, the two pure Gibbs phases exhaust the extremal equilibrium states, but not the pathwise absorbing structure of the Glauber dynamics. Let \[ P^\pm=\{\sigma:M_n(\sigma)\to \pm m_\beta\},\qquad R=\Omega\setminus(P^+\cup P^-). \] We show that \(R\) is null under both pure phases but contains a dense pathwise confined subset. More precisely, we construct a dense family of initial configurations whose trajectories are confined to the centered sector \[ C_0=\{\sigma:M_n(\sigma)\to0\}\subset R. \] Nevertheless, the corresponding Cesaro averages converge to \(\frac12(\mu^++\mu^-)\). Thus the pathwise absorbing geometry is richer than the Gibbs-phase classification, without creating a third Gibbs phase.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The abstract's construction of dense zero-magnetization trajectories whose Cesaro averages converge to the phase mixture contains an internal contradiction that the paper does not resolve.

read the letter

The paper shows that for low-temperature 2D Ising Glauber dynamics the set R outside the two pure magnetization limits is null under the equilibrium measures yet contains a dense subset of initial conditions whose paths stay in configurations with magnetization tending to zero. The Cesaro occupation measures for those paths are claimed to converge to the symmetric mixture of the two pure phases. If the construction works this would separate pathwise confinement from the Gibbs classification in a way not already in the literature cited in the abstract. The low-temperature assumption and the exhaustion of extremal states by the two pure phases are standard, so the setup itself is not the issue. The central claim, however, runs into a direct problem. The mixture puts full mass on configurations whose magnetization is near plus or minus the spontaneous value. Once the trajectory is eventually trapped in small-magnetization states, the long-run occupation measure is supported only on small-magnetization configurations and cannot converge to a measure that charges only the large-magnetization ones. The portmanteau argument in the stress-test note makes this incompatibility explicit, and the abstract supplies no detail that would evade it. Without the full proofs it is impossible to see whether the construction somehow avoids the support mismatch or whether the statement of the result is simply imprecise. This work is aimed at researchers who care about metastability and the gap between equilibrium phases and pathwise dynamics. A reader already working in that area might want to examine the construction once the contradiction is cleared up. I would not send the manuscript to peer review in its present form; the authors need to reconcile the pathwise confinement with the measure-theoretic properties of the mixture before the claim can be taken seriously.

Referee Report

1 major / 0 minor

Summary. The manuscript constructs a dense family of initial configurations for the low-temperature two-dimensional Ising model with Glauber dynamics such that the trajectories remain confined to the centered sector C0 = {σ : Mn(σ) → 0} ⊂ R, while the associated Cesàro averages (empirical occupation measures) converge to ½(μ⁺ + μ⁻). This is presented as showing that the pathwise absorbing geometry is richer than the classification into the two pure Gibbs phases, without creating a third equilibrium state.

Significance. If the central construction were valid, it would demonstrate a non-trivial separation between pathwise confinement and equilibrium phase classification in stochastic Ising dynamics, potentially informing studies of metastability and long-time behavior beyond the extremal Gibbs measures.

major comments (1)

[Abstract] Abstract: The stated pathwise confinement to C0 (Mn(σ_t) → 0) is incompatible with convergence of the occupation measures to ν = ½(μ⁺ + μ⁻). Confinement implies that for all t > T0, |M(σ_t)| < m_β/2, so for T ≫ T0 the measure (1/T)∫_0^T δ_σt dt is supported on {|M| < m_β/2}. Yet ν({|M| > m_β/2}) = 1. Assuming the weak topology used for the convergence makes {|M| > m_β/2} open, the Portmanteau theorem yields that the limit measure cannot place full mass on this set, contradicting the claimed convergence. This internal inconsistency is load-bearing for the main claim.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and for identifying this apparent inconsistency between pathwise confinement and convergence of occupation measures. We maintain that the construction is consistent once the discontinuity of the magnetization functional in the product topology is taken into account. We respond to the single major comment below.

read point-by-point responses

Referee: [Abstract] Abstract: The stated pathwise confinement to C0 (Mn(σ_t) → 0) is incompatible with convergence of the occupation measures to ν = ½(μ⁺ + μ⁻). Confinement implies that for all t > T0, |M(σ_t)| < m_β/2, so for T ≫ T0 the measure (1/T)∫_0^T δ_σt dt is supported on {|M| < m_β/2}. Yet ν({|M| > m_β/2}) = 1. Assuming the weak topology used for the convergence makes {|M| > m_β/2} open, the Portmanteau theorem yields that the limit measure cannot place full mass on this set, contradicting the claimed convergence. This internal inconsistency is load-bearing for the main claim.

Authors: We respectfully disagree. The magnetization functional M(σ), given by the limit of spatial averages Mn(σ) over boxes of side length n centered at the origin, is discontinuous with respect to the product topology on the configuration space Ω = {-1,1}^{ℤ²}. Consequently, the set {|M| > m_β/2} is not open. Explicitly, one can construct a sequence σ^k converging to the all-plus configuration (where |M| = m_β) in the product topology, yet with M(σ^k) = 0 for each k: let σ^k agree with the all-plus configuration inside the box of radius k and arrange the exterior so that the spatial averages converge to zero (possible since the inner box contributes o(1) to the limit). Thus the pure-phase configurations lie in the closure of {|M| < m_β/2}. The empirical occupation measures for large T are supported inside {|M| < m_β/2}, but their weak limits may charge the closure. Portmanteau therefore does not produce a contradiction, as the hypothesis that {|M| > m_β/2} is open fails. The claimed convergence to ½(μ⁺ + μ⁻) remains compatible with confinement to C0. revision: no

Circularity Check

0 steps flagged

No significant circularity; direct construction of confined trajectories

full rationale

The paper presents a mathematical construction of a dense family of initial configurations for the low-temperature 2D Ising model with Glauber dynamics. These trajectories remain confined to the centered sector C0 where Mn(σ)→0, while the associated Cesàro averages converge to ½(μ⁺ + μ⁻). No load-bearing step in the abstract or described derivation reduces the claimed result to its own inputs by definition, renames a fitted parameter as a prediction, or relies on a self-citation chain whose content is unverified outside the paper. The result is offered as an existence statement that enriches the pathwise geometry beyond the Gibbs-phase classification, without any self-referential loop or ansatz smuggled via prior work by the same authors. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Relies on standard background facts about the 2D Ising model at low temperature; no free parameters or invented entities introduced in the abstract.

axioms (1)

domain assumption The two pure Gibbs phases exhaust the extremal equilibrium states for the low-temperature 2D Ising model.
Invoked directly in the first sentence of the abstract as the equilibrium classification.

pith-pipeline@v0.9.0 · 5446 in / 1166 out tokens · 48565 ms · 2026-05-14T19:42:38.888793+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

Harris , title =

Theodore E. Harris , title =. Advances in Mathematics , volume =

work page
[2]

, title =

Liggett, Thomas M. , title =

work page
[3]

2011 , series =

Hans-Otto Georgii , title =. 2011 , series =

work page 2011
[4]

Sacha Friedli and Yvan Velenik , title =

work page
[5]

Ethier and Thomas G

Stewart N. Ethier and Thomas G. Kurtz , title =. 1986 , series =

work page 1986
[6]

Communications in Mathematical Physics , volume =

Michael Aizenman , title =. Communications in Mathematical Physics , volume =

work page
[7]

Random Fields, Vol

Yasunari Higuchi , title =. Random Fields, Vol. II , series =

work page
[8]

Lectures on Probability Theory , series =

Durrett, Richard , title =. Lectures on Probability Theory , series =. 1995 , doi =

work page 1995
[9]

, title =

Kechris, Alexander S. , title =. 1995 , doi =

work page 1995
[10]

Walters, Peter , title =

work page
[11]

Substitutions in Dynamics, Arithmetics and Combinatorics , series =

work page
[12]

Substitution Dynamical Systems---Spectral Analysis , series =

Queff. Substitution Dynamical Systems---Spectral Analysis , series =

work page
[13]

Colloquium Mathematicum , volume=

Topological zero-one laws , author=. Colloquium Mathematicum , volume=. 1978 , organization=

work page 1978
[14]

and Stroock, Daniel W

Holley, Richard A. and Stroock, Daniel W. , title =. Communications in Mathematical Physics , volume =

work page
[15]

Inventiones Mathematicae , volume =

Lindenstrauss, Elon , title =. Inventiones Mathematicae , volume =

work page
[16]

Rosenblatt, Murray , title =

work page
[17]

2017 , doi =

Generic points of invariant measures for an amenable residually finite group action with the weak specification property , journal =. 2017 , doi =

work page 2017
[18]

arXiv preprint , year =

Attali, Jean-Gabriel , title =. arXiv preprint , year =. XXXX.XXXXX , archivePrefix =

work page