🔢 mathematics

Level lines of the Gaussian free field and $c=1$ degenerate conformal blocks

This paper establishes that the crossing probabilities of level lines for both the continuum and metric graph Gaussian free fields with piecewise constant boundary data are determined by specific $c=1$ degenerate conformal blocks and ratios of fused $\mathrm{SLE}_4$ partition functions, which satisfy higher-order BPZ equations and correspond to conformal blocks labeled by generalized Dyck paths subject to monotonicity constraints.

Original authors: Alex Karrila, Eveliina Peltola, Lukas Schoug

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Alex Karrila, Eveliina Peltola, Lukas Schoug

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

The Big Picture: A Random Mountain Range

Imagine a vast, invisible landscape called the Gaussian Free Field (GFF). Think of it not as a solid mountain, but as a "random fog" or a "static-filled TV screen" that fluctuates up and down everywhere at once. Because it fluctuates so wildly, you can't point to a single spot and say, "The height here is exactly 5 feet." It's too chaotic.

However, even though the fog is messy, it has contour lines (like on a topographic map). If you draw a line where the height is exactly 10 feet, that line forms a path. In this paper, the authors study what happens when you draw many of these lines at once on a flat surface (like a piece of paper or a half-plane).

The Setup: The "Jumping" Boundary

Imagine you have a flat sheet of rubber (the domain). You decide to pin down the edges of this sheet at specific heights.

On the left edge, you pin it at height 0.
Then, you jump up to height 2.
Then, you jump up to height 4.
Then, you drop down to height 2 again.
And so on.

These "jumps" in height along the edge are the boundary conditions. The paper looks at what happens when you have a series of these jumps. The "random fog" (the GFF) tries to smooth itself out between these pinned edges, creating a wiggly, random surface.

The Main Discovery: The "Connectivity" Game

The authors are interested in the contour lines (level lines) that appear in this random fog. Specifically, they look at lines that are at "odd" heights (like 1, 3, 5 units high, where the unit is a special number $\lambda$ ).

The Question: If you draw all these lines, how do they connect?

Does the line starting at the first jump connect to the second jump?
Or does it cross over and connect to the third?
Do they form a "rainbow" shape, or do they tangle in a specific way?

This is called a connectivity pattern. It's like a game of "connect the dots," but the dots are connected by random, wiggly lines that obey the rules of the random fog.

The Magic Formula: A Recipe from Physics

The paper's biggest claim is that they found a perfect recipe to calculate the probability of any specific connection pattern.

They discovered that these probabilities are determined by something called Conformal Blocks from a branch of physics called Conformal Field Theory (CFT).

The Analogy: Imagine you have a complex puzzle. Usually, you might have to simulate the puzzle millions of times on a computer to guess how often a specific piece fits.
The Paper's Result: The authors say, "No need to simulate!" There is a specific mathematical formula (a "Conformal Block") that acts like a master key. If you plug in the positions of your jumps, this formula instantly tells you the exact odds of the lines connecting in any specific way.

They also found that these formulas are related to Specht polynomials (a type of math object used in algebra) and that they solve a very specific set of differential equations (the BPZ equations). Think of these equations as the "laws of physics" that these random lines must obey.

The "Resampling" Property: The Shape-Shifting Rule

One of the cool things they proved is a property called Resampling.

The Analogy: Imagine you have a group of hikers (the lines) walking through a forest. If you stop and look at just one hiker, while ignoring where the others are, that single hiker looks like they are walking a standard, random path (called an SLE4 curve).
The Twist: But if you look at the whole group together, they are constrained by each other. They can't cross.
The Result: The paper proves that if you know where all the other hikers are, the remaining hiker will still follow that same standard random path, just confined to the empty space left by the others. This rule is so strict that it uniquely defines how the whole group behaves.

The "Discrete" vs. "Continuous" Connection

The paper also looks at a "pixelated" version of this world (called the Metric Graph GFF). Imagine the smooth rubber sheet is replaced by a grid of tiny squares (like a video game map).

They showed that as you make the squares smaller and smaller (zooming in), the behavior of the lines on the pixelated grid becomes indistinguishable from the smooth, continuous lines on the rubber sheet.
This confirms that their mathematical formulas work for both the "pixelated" world and the "smooth" world.

The "Forbidden" Patterns

Interestingly, the paper notes that not every possible way of connecting the dots is allowed.

The Analogy: Imagine you have a set of ropes tied to the edge of a table. Some knots are physically impossible to tie without the ropes crossing or breaking.
The Result: The authors found that the "jumps" in the boundary conditions act like a filter. They only allow certain "legal" connection patterns (which they call maximally sloped patterns). If you try to force a pattern that violates the rules of the jumps, the probability is zero.

Summary

In short, this paper is a bridge between:

Random Geometry: The messy, wiggly lines of a random field.
Algebra & Physics: The clean, rigid formulas of Conformal Field Theory.

The authors proved that the chaotic behavior of these random lines is actually governed by a precise, elegant mathematical code. They gave us the exact dictionary to translate the "randomness" of the lines into the "order" of algebraic formulas, showing that even in a chaotic system, there is a hidden, perfect structure.

Technical Summary: Level lines of the Gaussian free field and $c=1$ degenerate conformal blocks

Problem Statement
The paper investigates the geometric structure of level lines and first passage sets (FPS) of the Gaussian free field (GFF) on simply connected domains with piecewise constant Dirichlet boundary data. Specifically, it addresses the computation of crossing probabilities for these level lines and their relationship to Conformal Field Theory (CFT). While the coupling between GFF level lines and Schramm-Loewner evolution (SLE) curves is well-established for specific boundary conditions (e.g., alternating data), the behavior under more general, multi-jump boundary data encoded by generalized Dyck paths remains less understood in terms of explicit partition functions and connection probabilities. The authors aim to characterize the topological connectivities of these level lines using $c=1$ degenerate conformal blocks and to establish the existence and uniqueness of a corresponding global multiple SLE $_4$ model.

Methodology
The authors employ a synthesis of probabilistic coupling techniques, combinatorial analysis, and conformal field theory:

GFF and SLE Coupling: The study utilizes the known coupling between the GFF and SLE $_4$ curves. Level lines of the GFF with specific boundary jumps are identified as SLE $_4$ curves (or collections thereof). The authors analyze the "first passage sets" (FPS) of the GFF, showing that their frontiers correspond exactly to the odd-height level lines of the field.
Combinatorial Framework: The topological configurations of the level lines are encoded as valenced link patterns. The authors utilize a bijection between these patterns and generalized Dyck paths. They introduce a binary relation ( $\beta \rightsquigarrow \alpha$ ) derived from the "unfusing" map, which determines which interior connectivity patterns $\alpha$ are realizable for a given boundary condition $\beta$ .
Conformal Blocks and Partition Functions: The core analytical tool involves $c=1$ degenerate conformal block functions, denoted $U_\beta$ , which are ratios of fused Specht polynomials. The authors define "pure partition functions" $Z_\alpha$ as linear combinations of these conformal blocks, utilizing combinatorial quantities such as cover-inclusive Dyck tilings and skew Young diagrams.
Resampling and Uniqueness: To prove the uniqueness of the global multiple SLE $_4$ model, the authors utilize the domain Markov property of the GFF. They demonstrate that the level lines satisfy a "resampling property" (where the conditional law of one curve given the others is a chordal SLE $_4$ in the remaining domain). They argue that any process satisfying this property can be coupled with a GFF, thereby proving uniqueness via the GFF coupling.
Discrete Approximation: To bridge the gap between discrete and continuum models, the authors analyze the metric graph GFF (introduced by Lupu). They rely on established convergence results showing that the FPS of the metric graph GFF converge to those of the continuum GFF in the Hausdorff metric.

Key Contributions and Results

Connection Probabilities (Theorem A): The authors derive an explicit formula for the probability that GFF level lines form a specific topological connectivity pattern $\alpha$ . This probability is expressed as a ratio of a "pure partition function" $Z_\alpha$ to a conformal block function $U_\beta$ , scaled by an incidence matrix entry $M_{\beta, \alpha}$ which is 1 if the pattern is combinatorially possible for the boundary data and 0 otherwise:
$P_\beta[\vartheta_{\text{GFF}} = \alpha] = M_{\beta, \alpha} \frac{Z_\alpha(x_1, \dots, x_p)}{U_\beta(x_1, \dots, x_p)}$
CFT Interpretation (Theorem B): The pure partition functions $Z_\alpha$ are shown to satisfy the Belavin-Polyakov-Zamolodchikov (BPZ) partial differential equations of arbitrary order for a CFT with central charge $c=1$ . They are identified as boundary correlation functions of primary fields with conformal weights $h_{1, s_j+1} = s_j^2/4$ . The authors prove these functions are positive and linearly independent, confirming their status as valid SLE $_4$ partition functions.
Global Multiple SLE $_4$ (Theorems C & D): The paper establishes the existence and uniqueness of a global multiple SLE $_4$ measure for any realizable link pattern $\alpha \in PLP_\varsigma$ . The existence is constructed via GFF level lines, and uniqueness is proven by showing that any such process satisfying the resampling property must coincide with the GFF-coupled level lines.
Metric Graph Convergence (Theorem E): The authors prove that the connection probabilities for the level sets of the metric graph GFF converge to the continuum formulas derived in Theorem A as the mesh size tends to zero. This validates the continuum results as the scaling limit of the discrete metric graph model.
Partition Function Relation (Proposition 3.1): A direct link is established between the GFF partition function with piecewise constant boundary data and the conformal block function $U_\beta$ , showing that the regularized Dirichlet norm of the boundary data yields the logarithm of the conformal block.

Significance and Claims
The paper claims to provide a complete description of the geometry of level-set percolation for the GFF with piecewise constant boundary data in terms of $c=1$ conformal blocks. By identifying the crossing probabilities with ratios of explicit partition functions (fused Specht polynomials), the work connects the probabilistic geometry of the GFF directly to the representation theory of the Virasoro algebra at $c=1$ .

The authors note that while the CFT contains conformal blocks for all generalized Dyck paths, the specific models considered here (GFF level lines) satisfy specific monotonicity constraints, meaning not all combinatorial boundary conditions are realizable. The work extends previous results on alternating boundary data and specific fusion cases to a general setting involving arbitrary valences and boundary jumps. The demonstration that the metric graph GFF converges to these formulas provides a rigorous link between discrete random surface models and their continuum CFT descriptions. The paper does not propose new applications or experimental setups but focuses on the theoretical unification of GFF geometry, SLE, and CFT.