Laurent Bartholdi, Ivailo Hartarsky, Ivan Mitrofanov
Consider the following plaquette model from statistical physics: a lamp lies at every vertex of the triangular lattice and a switch lies at every even vertex of the (bipartite) dual hexagonal lattice. Each switch toggles the three lamps on its face. The energy of a configuration is the number of ON lamps. For the Glauber dynamics associated with the Gibbs measure defined by this Hamiltonian at any inverse temperature , we show that, in any dimension , the infinite volume relaxation time satisfies for some . Our result entails that the Gibbs measure is unique. The scaling was conjectured by Newman and Moore in 1999 and matches the behaviour of supercritical rooted kinetically constrained models such as the East model, thus recovering fragile glass phenomenology in the absence of kinetic constraints. More precisely, we show that, on a torus of side length , when and , we have . Quite surprisingly, however, we also prove that, on non-periodic finite domains of size for large , we have the much larger asymptotics . The main ingredients of the proofs are new results in extremal and enumerative combinatorics and rely on renormalisation ideas for the dynamics and its groundstates also known as the Ledrappier subshift. We note consequences of our results to geometric group theory (more precisely to the complexity of the word problem for the Baumslag finitely presented group) and to ergodic theory.
This paper resolves a 25-year-old conjecture by Newman and Moore (1999) concerning the super-Arrhenius relaxation time of the triangular plaquette model (TPM), a statistical mechanics model with three-body interactions on the triangular lattice. The main results establish that in any dimension , the infinite-volume relaxation time satisfies , confirming the predicted scaling. The paper also proves sharp asymptotics on dyadic tori of side (when ), and discovers a surprising phenomenon: on non-periodic finite domains, the relaxation time is exponentially larger (), suggesting a remarkable boundary-condition-dependent speedup in larger volumes.
The paper is technically formidable, combining tools from at least four distinct mathematical areas:
1. Extremal combinatorics (Section 2.1): A renormalization-based proof that moving a single ON lamp out of a ball of radius requires at least simultaneously ON lamps. The inductive argument through nested simplices and the classification of plaquette-inverted plaquette intersections is clean and sharp (tight up to additive constants in ).
2. Entangled configurations (Section 2.2): Construction of admissible configurations requiring additional ON lamps to annihilate, using a novel metric based on "long vectors" (edges of large plaquettes) and an exponential decay lemma (Lemma 2.17) proved by induction on dimension with careful parity-class arguments.
3. Enumerative combinatorics (Section 2.3): A recursive bound on the generating function of cycles (groundstate configurations) via a dimension-reduction scheme, yielding polynomial decay estimates that control partition function ratios.
4. Probabilistic arguments (Section 3): Standard but carefully executed Glauber dynamics techniques — spectral gap bounds, canonical paths, path coupling with block dynamics — are combined with the combinatorial inputs.
The proofs are self-contained, carefully structured, and appear correct. The paper handles the delicate interplay between periodic and non-periodic boundary conditions with precision. The upper bound proof via path coupling at scale (Section 3.4) is particularly elegant, using the cycle generating function estimates to establish that boundary conditions have negligible influence above a critical scale.
Statistical physics and glassy dynamics: This is arguably the first rigorous confirmation of fragile glass phenomenology () in a model without artificially imposed kinetic constraints. Previous results of this type required kinetically constrained models (East model, etc.) where constraints are imposed ad hoc. The TPM achieves the same super-Arrhenius behavior purely from energetic considerations, providing a more physically motivated explanation.
Geometric group theory: The connection to Baumslag's metabelian group is genuinely novel. Corollary 4.1 gives a lower bound on the -filling length function, and Corollary 4.2 produces words requiring polynomially many extra generators to reduce — both are new results about the word problem complexity.
Ergodic theory: Theorem 5.2 resolves a question from Arenas-Carmona, Berend, and Bergelson (2008) by showing the function (minimal number of large plaquettes to represent admissible sets) grows superlinearly: .
Combinatorics: The classification of small admissible configurations (Lemma 2.16), the metric and its properties, and the recursive cycle-counting estimates are independently interesting combinatorial results.
The paper addresses a long-standing open problem at the intersection of mathematical physics and probability. The recent surge of interest in kinetically constrained models (with mathematical breakthroughs by Martinelli, Morris, Toninelli, and others in the 2010s-2020s) makes this particularly timely. The fact that plaquette models can reproduce the same universality class behavior without kinetic constraints addresses a fundamental criticism of the KCM paradigm. The connections to quantum plaquette models (currently active in physics) add further relevance.
This is an outstanding paper that solves a well-known conjecture through a creative synthesis of combinatorial, algebraic, and probabilistic techniques. The discovery of boundary-condition-sensitive relaxation regimes and the cross-disciplinary applications significantly amplify its impact beyond the immediate problem.
Generated Jun 16, 2026
Paper 2 likely has higher impact: it delivers an optimal, dimension-free sparsification theorem for Gaussian-process suprema with tight ε-dependence, improving a very recent (2026) result by an exponential factor. Such a broadly applicable tool can propagate across probability, convex geometry, high-dimensional statistics, learning theory, property testing, and optimization, with clear downstream algorithmic/complexity consequences. Paper 1 is deep and novel for glassy dynamics and links to combinatorics/group/ergodic theory, but its immediate applications are narrower and its impact more specialized within mathematical physics/dynamics.
Paper 2 likely has higher impact: it establishes conformally covariant scaling limits for interfaces and loop ensembles in the near-critical planar RFIM, connecting rigorously to (massive) SLE/CLE theory and highlighting a sharp absolute-continuity vs singularity dichotomy. This advances a central, timely area (2D critical/near-critical phenomena, random media) with broad relevance across probability, mathematical physics, and complex analysis. Paper 1 is strong and novel (super-Arrhenius relaxation, combinatorics, links to dynamics and group theory) but is more specialized and less broadly foundational than new RFIM scaling-limit results.
Paper 2 solves a long-standing 1999 conjecture by Newman and Moore regarding the relaxation time scaling in a plaquette model. Furthermore, it demonstrates significantly broader scientific impact by connecting statistical physics with extremal combinatorics, geometric group theory, and ergodic theory. Paper 1 offers a solid theoretical contribution to stochastic analysis on Wasserstein spaces, but its scope and potential for cross-disciplinary applications are narrower compared to the multi-field relevance and breakthrough nature of Paper 2.
Paper 2 has broader scientific impact across multiple fields: statistical physics, combinatorics, geometric group theory, and ergodic theory. It resolves a 1999 conjecture by Newman and Moore about super-Arrhenius relaxation, connects to glass physics phenomenology, and discovers surprising boundary-condition-dependent behavior. The interdisciplinary nature—linking plaquette models to the Baumslag group's word problem and the Ledrappier subshift—gives it exceptional breadth. Paper 1, while technically strong with sharp estimates for FEP freezing times, is more narrowly focused within interacting particle systems.
Paper 2 resolves a 25-year-old conjecture (Newman and Moore, 1999) in statistical physics and introduces novel techniques linking extremal combinatorics with renormalization. Its results have profound, diverse implications that extend beyond statistical physics into geometric group theory and ergodic theory, suggesting a broader and deeper scientific impact compared to the incremental, albeit useful, improvements in concentration inequalities presented in Paper 1.
Paper 2 has broader scientific impact across multiple fields: statistical physics, combinatorics, geometric group theory, and ergodic theory. It resolves a 1999 conjecture by Newman and Moore about super-Arrhenius relaxation, connects fragile glass phenomenology to plaquette models without kinetic constraints, and discovers surprising boundary-condition dependence. The interdisciplinary nature (extremal combinatorics, renormalization, Baumslag groups, Ledrappier subshift) and the resolution of a long-standing conjecture give it wider reach. Paper 1, while technically strong, addresses a more specialized question in population genetics with a partial (perturbative) answer.
Paper 1 solves a long-standing 1999 conjecture by Newman and Moore regarding scaling in statistical physics. It demonstrates exceptional breadth of impact by bridging statistical physics (fragile glass phenomenology) with extremal combinatorics, geometric group theory, and ergodic theory. While Paper 2 provides a valuable mathematical framework for infinite-dimensional stochastic systems (like Cahn-Hilliard), Paper 1's resolution of a 25-year-old open problem and its surprising findings on non-periodic domains give it a higher degree of novelty and wider cross-disciplinary relevance, likely leading to greater overall scientific impact.
Paper 2 resolves a conjecture from 1999 (Newman-Moore) about super-Arrhenius relaxation in plaquette models, bridging statistical physics, combinatorics, geometric group theory, and ergodic theory. It establishes surprising dimension-dependent and boundary-condition-dependent behavior with broad implications for glass physics and fragile glass phenomenology. The breadth of connections across multiple fields (statistical mechanics, combinatorics, group theory, ergodic theory), resolution of a longstanding conjecture, and the surprising discovery of dramatically different behavior on periodic vs. non-periodic domains give it higher impact than Paper 1, which provides rigorous but more incremental foundations for dynamical low-rank approximations of SDEs.
Paper 1 introduces a highly novel algebraic framework to resolve the Group Spencer conjecture and significantly advance the Matrix Spencer conjecture. By bridging C*-algebras and discrepancy theory, it provides a methodological breakthrough with profound implications for theoretical computer science, functional analysis, and algorithmic design. While Paper 2 solves an important 1999 statistical physics conjecture with impressive mathematical breadth, Paper 1's resolution of central structural conjectures in modern discrepancy theory gives it a higher potential impact across foundational mathematics and computer science.
Paper 2 resolves a 1999 conjecture by Newman and Moore about super-Arrhenius relaxation in plaquette models, establishes surprising dimension-dependent and boundary-condition-dependent behavior, and connects to multiple fields including statistical physics, combinatorics, geometric group theory, and ergodic theory. Its breadth of impact across disparate mathematical and physical disciplines, resolution of a longstanding conjecture, and discovery of unexpected phenomena (dramatic dependence on boundary conditions) give it higher potential impact than Paper 1, which, while technically strong, represents an incremental (though significant) extension of existing information-theoretic sensitivity bounds for SDEs.