🔢 mathematics

Universality of two-dimensional Markovian holonomy fields

This paper establishes a universality theorem demonstrating that a broad class of two-dimensional lattice gauge theories on compact surfaces, driven by conjugation-invariant Lévy processes on compact Lie groups, converge to a unique continuum Markovian holonomy process, thereby unifying models like Yang-Mills, Villain, Wilson, and Manton under a single gauge-theoretic invariance principle.

Original authors: Thibaut Lemoine, Elias Nohra

Published 2026-08-18

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Thibaut Lemoine, Elias Nohra

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world where the fundamental forces of nature are not just invisible fields, but are instead woven from the very shape of space itself. In the realm of theoretical physics, scientists study how particles move and interact by looking at the geometry of the universe. On a flat, two-dimensional surface, like a sheet of paper, there is a powerful mathematical tool called a gauge theory. It describes how a particle's internal state changes as it travels along a path. If the particle travels in a circle and returns to its starting point, the change it undergoes is called a holonomy. This concept is crucial because it helps physicists understand the behavior of forces like electromagnetism and the strong nuclear force, which hold atoms together. For decades, researchers have tried to understand what happens to these forces when they are smoothed out from a jagged, pixelated grid into a perfectly continuous surface. The question has been whether the specific rules used to build the grid matter in the end, or if all different grids eventually melt down into the same smooth reality.

A team of mathematicians has now answered this question with a definitive proof, showing that for a vast class of these theories, the microscopic details of the grid do not matter. They demonstrated that no matter how you construct the discrete grid or which specific rules you use to calculate the energy of the loops on that grid, as long as the rules follow a certain basic pattern, the system will always converge to the same continuous behavior. This result is a universality theorem, a concept familiar to those who study how random walks turn into smooth diffusion, but here it is applied to the complex geometry of gauge fields. The researchers proved that the only thing that survives the transition from a jagged lattice to a smooth surface is a single, fundamental signature of the rules used. Everything else—the specific shape of the grid, the size of the squares, or the exact formula used for the energy—fades away, leaving behind a universal process that depends only on that one signature.

To understand what the team actually did, one must picture the surface they are working on. They started with a compact, curved surface, like a sphere or a doughnut, and covered it with a mesh of polygons, similar to a geodesic dome. On the edges of these polygons, they placed variables that represent the state of a particle as it moves. The rules of the game, known as the action, dictate how these variables interact when they form a closed loop around a polygon. In the past, physicists had to choose a specific set of rules, such as the Wilson action or the heat-kernel action, to simulate these systems on a computer. The worry was that choosing one set of rules might lead to a different final result than choosing another. The new work shows that this worry is unfounded for a broad category of theories. The researchers took a very general set of rules, which they call scaling-admissible, and showed that as the mesh becomes infinitely fine—meaning the polygons get smaller and smaller—the behavior of the loops on the mesh becomes indistinguishable from a specific, smooth mathematical process known as a Markovian holonomy field.

The mechanism behind this discovery is a clever separation of concerns. The team developed a method to break down the complex calculation of the entire system into two distinct parts. One part depends entirely on the specific rules chosen for the grid and the size of the polygons. The other part depends only on the shape of the surface and how the loops are tangled together. As the grid gets finer, the part that depends on the specific rules simplifies dramatically. It turns out that the complex, messy details of the rules collapse into a single, simple number that describes how the system behaves at the tiniest scales. This number is the only thing that remains relevant in the final, smooth limit. The part of the calculation that depends on the shape of the surface and the loops remains constant, unaffected by the refinement of the grid. Because these two parts are separated, the researchers could prove that the final result is determined solely by that single number, regardless of the initial complexity of the rules.

This finding is significant because it unifies several different approaches that physicists have used for decades. For example, the standard model of particle physics often uses the Wilson action, while other mathematical approaches use the heat-kernel action. These are mathematically distinct ways of defining the energy of the system. The paper proves that both of these approaches, and many others, belong to the same universality class. They all lead to the same continuous theory of the Yang-Mills holonomy process, which is the mathematical description of the force fields in our universe. The result also extends beyond the standard Brownian motion-like behavior. It includes more exotic types of processes, such as those involving sudden jumps, showing that the framework is robust enough to handle a wide variety of physical scenarios.

The proof relies on a rigorous construction of the grid sequences. The researchers had to ensure that as they made the grid finer, they were not accidentally changing the topology of the loops or the surface in a way that would skew the results. They introduced a concept of "stable approximations," where the grid is refined in a way that preserves the essential shape and connectivity of the loops. By proving that such stable sequences always exist and that the convergence holds for any such sequence, they eliminated the possibility that the result was an artifact of a specific way of drawing the grid. The work is a complete mathematical proof, not a simulation or a suggestion. It establishes with certainty that the continuum limit is universal for this broad class of theories.

In the end, the paper provides a deep insight into the nature of physical laws. It suggests that the macroscopic world we observe is remarkably robust against the microscopic details of how it is constructed. Just as a photograph looks the same whether it is printed on high-resolution paper or a low-resolution screen, the fundamental behavior of these gauge theories remains the same regardless of the specific lattice used to model them. The researchers have shown that the universe, at least in this mathematical sense, cares only about the most essential features of the rules, filtering out the noise of the microscopic construction to reveal a single, elegant truth. This universality gives physicists confidence that their models of the fundamental forces are capturing something real and enduring, independent of the specific mathematical tools used to describe them.

Technical Summary: Universality of Two-Dimensional Markovian Holonomy Fields

Problem Statement
The paper addresses the scaling limit problem for two-dimensional lattice gauge theories on compact surfaces. While it is well-established that every Lévy process on a compact Lie group arises as the scaling limit of a suitable class of discrete random walks, the gauge-theoretic analogue remained less understood. Specifically, the authors investigate whether discrete lattice gauge theories, defined by general central plaquette actions (which are generally not projectively consistent under subdivision), converge to the continuous holonomy processes induced by Markovian holonomy fields.

The central challenge lies in the fact that generic lattice actions depend on the specific topological map (the choice of graph and face areas), whereas the continuous theory (specifically the Markovian holonomy fields constructed by Lévy) is defined by a coherent family of measures satisfying covariance and cutting-gluing axioms. The authors aim to prove that this dependence on the microscopic lattice structure disappears in the continuum limit, identifying a "universality class" for each admissible Lévy symbol.

Methodology
The proof strategy relies on a separation of spectral (action-dependent) and topological (action-independent) data via state-sum formulas. The methodology proceeds in three main stages:

State-Sum Factorization: The authors derive exact formulas for the partition function and Wilson loop integrals on an arbitrary topological map embedded in a surface. These formulas decompose the expectation values into a product of two distinct components:
- Spectral Coefficients: Products of Fourier coefficients (character expansion coefficients) of the plaquette weights, which depend on the specific action and the areas of the faces.
- Topological Coefficients: Integrals over the gauge group determined solely by the topology of the surface, the marked ribbon type of the loop configuration, and the boundary conditions. Crucially, these coefficients are invariant under graph refinement once the ribbon type is fixed.
Scaling-Admissibility and Spectral Limits: The authors introduce the notion of "scaling-admissible" families of plaquette weights. A family is scaling-admissible for a Lévy symbol $\psi$ if its Fourier coefficients $a_t(\lambda)$ satisfy $a_t(\lambda) = e^{-t\psi(\lambda)} + o_\lambda(t)$ as $t \to 0$ , subject to a uniform spectral summability condition. This condition ensures that the product of Fourier coefficients over a sequence of refining faces converges to the exponential of the total area times the Lévy symbol, effectively recovering the semigroup property in the limit.
Geometric Approximation and Regular Convergence: To handle the mismatch between the index sets of discrete processes (loops on a graph) and continuous processes (rectifiable loops on a surface), the authors define:
- Regular loop families: Finite families of loops whose union forms a minimal embedded ribbon graph.
- Stable graph approximations: Sequences of topological maps that approximate a regular loop family while preserving its marked ambient ribbon type and the areas of the complementary components.
- Universally stable graph sequences: A single sequence of graphs that admits stable approximations for all finite regular loop families simultaneously. The existence of such sequences is proven using PL topology and general-position arguments.

Key Contributions

Universality Theorem: The paper proves that for any compact connected Lie group and any admissible conjugation-invariant Lévy process, the continuum scaling limit of a broad class of lattice gauge theories depends only on the infinitesimal Fourier symbol of the plaquette action.
State-Sum Formulas: The authors provide rigorous proofs for state-sum expansions of partition functions and Wilson loop integrals for general central actions (not just the heat-kernel action). These formulas separate the action-dependent spectral data from the topological data, a structural result of independent interest.
Extension to Non-Brownian Fields: The framework generalizes beyond the standard Yang-Mills (Brownian) case. It encompasses non-Brownian Markovian holonomy fields, such as those generated by subordinated Brownian motions or stable-type processes, provided the plaquette action is scaling-admissible for the corresponding Lévy symbol.
Geometric Framework: The paper formalizes the notion of "regular convergence" for holonomy processes, resolving the technical difficulty of comparing processes indexed by different sets of loops via universally stable graph sequences.

Main Results

Theorem 1.1 (Universality of Holonomy Processes): Let $\Sigma$ be a compact connected oriented Riemannian surface and $G$ a compact connected Lie group. If $(G_n)$ is a universally stable sequence of embedded topological maps and $(Q_t)$ is a family of plaquette weights scaling-admissible for a Lévy symbol $\psi$ , then the associated lattice holonomy processes converge regularly to the holonomy process $H_\psi$ induced by the Markovian holonomy field associated with $\psi$ .
Theorem 1.3 & 1.4 (State-Sum Formulas): The partition function and Wilson loop integrals admit absolutely convergent expansions where the dependence on the action and face areas is isolated in products of Fourier coefficients, while the dependence on the surface topology and loop configuration is isolated in topological coefficients.
Proposition 4.10: The existence of universally stable graph sequences is established for any compact smooth surface, ensuring that the scaling limit can be defined consistently for all regular loop families.

Significance and Claims
The authors position this work as a gauge-theoretic analogue of invariance principles for random walks and Lévy processes. The primary significance claimed is the identification of universality classes for two-dimensional lattice gauge theories parametrized by admissible Lévy symbols.

The paper claims that:

Microscopically different lattice theories (e.g., Wilson, Manton, and heat-kernel actions) share the same continuum limit at the level of holonomy processes if they share the same infinitesimal Lévy symbol.
The framework unifies the treatment of Brownian (Yang-Mills) and genuinely non-Brownian holonomy fields.
The state-sum factorization provides a mechanism for universality: the topological coefficients are invariant under refinement, while the spectral coefficients converge to the Lévy semigroup, causing the specific details of the lattice action to wash out in the limit.

The authors note that while distributional-connection approaches (using SPDEs and regularity structures) have established scaling limits for the Brownian class in Banach spaces, this work provides a direct universality statement for holonomy processes and extends it to the full admissible Lévy–Markovian class. They do not claim to realize these non-Brownian fields as random distributional connections, noting this as a potential direction for future investigation.

More like this